[WebAssembly] Regenerate SIMD Arithmetic tests

[RKSimon]

Summary

This is a massive change so wanted to get feedback from the webassembly people if its an acceptable approach or not.

An upcoming SimplifyDemandedBits patch of mine changes the wasm vector shift codegen a huge amount and I was hoping if moving this file to autogeneration is the way to go to make future updates easier.

I've used the update_llc_test_checks script to regenerate the test file, and have added additional check prefixes to reduce duplication.

Because of PR42882, there still needs to be some minor manual tweaking to remove the %BB tags from the checks, if we can remove -asm-verbose=false from the llc command lines then the regeneration would be completely automated.

[tlively]

I have mixed feelings about this. One one hand I understand that this file is huge and cross-cutting changes could cause a lot of manual updating work, but on the other hand each individual test was very readable and succinct. The auto-generated version would certainly be less work to maintain, but its enormous amount of irrelevant information and repetition is unfortunate. @aheejin, what do you think about this?

[quantum]

I have never touched the SIMD code and as a result, am not qualified to review this.

[aheejin]

In D65720#1615544, @tlively wrote:

I have mixed feelings about this. One one hand I understand that this file is huge and cross-cutting changes could cause a lot of manual updating work, but on the other hand each individual test was very readable and succinct. The auto-generated version would certainly be less work to maintain, but its enormous amount of irrelevant information and repetition is unfortunate. @aheejin, what do you think about this?

I have the same concerns. While I understand the convenience of using the auto generation script, the file is now far less readable and contains a lot more contents, and also all different test prefixes, some of which needed only for a few tests, are generated for every function. With the file being huge, it would also be hard to see if subsequent updates to the test are correct.

[RKSimon]

This is what I figured - I'll abandon this, but will probably need you guys to massage the checks for the upcoming SimplifyDemandedBits patch.

[tlively]

Sounds good; I'm happy to help out.

test/CodeGen/WebAssembly/simd-arith.ll

-; RUN: llc < %s -asm-verbose=false -verify-machineinstrs -disable-wasm-fallthrough-return-opt -wasm-disable-explicit-locals -wasm-keep-registers -mattr=+unimplemented-simd128 | FileCheck %s --check-prefixes CHECK,SIMD128,SIMD128-SLOW
-; RUN: llc < %s -asm-verbose=false -verify-machineinstrs -disable-wasm-fallthrough-return-opt -wasm-disable-explicit-locals -wasm-keep-registers -mattr=+unimplemented-simd128 -fast-isel | FileCheck %s --check-prefixes CHECK,SIMD128,SIMD128-FAST
-; RUN: llc < %s -asm-verbose=false -verify-machineinstrs -disable-wasm-fallthrough-return-opt -wasm-disable-explicit-locals -wasm-keep-registers -mattr=+simd128 | FileCheck %s --check-prefixes CHECK,SIMD128-VM
-; RUN: llc < %s -asm-verbose=false -verify-machineinstrs -disable-wasm-fallthrough-return-opt -wasm-disable-explicit-locals -wasm-keep-registers -mattr=+simd128 -fast-isel | FileCheck %s --check-prefixes CHECK,SIMD128-VM
-; RUN: llc < %s -asm-verbose=false -verify-machineinstrs -disable-wasm-fallthrough-return-opt -wasm-disable-explicit-locals -wasm-keep-registers | FileCheck %s --check-prefixes CHECK,NO-SIMD128
-; RUN: llc < %s -asm-verbose=false -verify-machineinstrs -disable-wasm-fallthrough-return-opt -wasm-disable-explicit-locals -wasm-keep-registers -fast-isel | FileCheck %s --check-prefixes CHECK,NO-SIMD128
+; RUN: llc < %s -asm-verbose=false -verify-machineinstrs -disable-wasm-fallthrough-return-opt -wasm-disable-explicit-locals -wasm-keep-registers -mattr=+unimplemented-simd128 | FileCheck %s --check-prefixes=CHECK,SIMD128,SIMD128-SLOW,SIMD128-UNIMPL,SIMD128-UNIMPL-SLOW
+; RUN: llc < %s -asm-verbose=false -verify-machineinstrs -disable-wasm-fallthrough-return-opt -wasm-disable-explicit-locals -wasm-keep-registers -mattr=+unimplemented-simd128 -fast-isel | FileCheck %s --check-prefixes=CHECK,SIMD128,SIMD128-FAST,SIMD128-UNIMPL,SIMD128-UNIMPL-FAST
+; RUN: llc < %s -asm-verbose=false -verify-machineinstrs -disable-wasm-fallthrough-return-opt -wasm-disable-explicit-locals -wasm-keep-registers -mattr=+simd128 | FileCheck %s --check-prefixes=CHECK,SIMD128,SIMD128-SLOW,NO-SIMD64,NO-SIMD64-SLOW,SIMD128-VM,SIMD128-VM-SLOW
+; RUN: llc < %s -asm-verbose=false -verify-machineinstrs -disable-wasm-fallthrough-return-opt -wasm-disable-explicit-locals -wasm-keep-registers -mattr=+simd128 -fast-isel | FileCheck %s --check-prefixes=CHECK,SIMD128,SIMD128-FAST,NO-SIMD64,NO-SIMD64-FAST,SIMD128-VM,SIMD128-VM-FAST
+; RUN: llc < %s -asm-verbose=false -verify-machineinstrs -disable-wasm-fallthrough-return-opt -wasm-disable-explicit-locals -wasm-keep-registers | FileCheck %s --check-prefixes=CHECK,NO-SIMD64,NO-SIMD64-SLOW,NO-SIMD128,NO-SIMD128-SLOW
+; RUN: llc < %s -asm-verbose=false -verify-machineinstrs -disable-wasm-fallthrough-return-opt -wasm-disable-explicit-locals -wasm-keep-registers -fast-isel | FileCheck %s --check-prefixes=CHECK,NO-SIMD64,NO-SIMD64-FAST,NO-SIMD128,NO-SIMD128-FAST
 
 ; check that a non-test run (including explicit locals pass) at least finishes
 ; RUN: llc < %s -O0 -mattr=+unimplemented-simd128
 ; RUN: llc < %s -O2 -mattr=+unimplemented-simd128
 
 ; Test that basic SIMD128 arithmetic operations assemble as expected.
 
 target datalayout = "e-m:e-p:32:32-i64:64-n32:64-S128"
 target triple = "wasm32-unknown-unknown"
 
 ; ==============================================================================
 ; 16 x i8
 ; ==============================================================================
-; CHECK-LABEL: add_v16i8:
-; NO-SIMD128-NOT: i8x16
-; SIMD128-NEXT: .functype add_v16i8 (v128, v128) -> (v128){{$}}
-; SIMD128-NEXT: i8x16.add $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
+
 define <16 x i8> @add_v16i8(<16 x i8> %x, <16 x i8> %y) {
+; SIMD128-LABEL: add_v16i8:
+; SIMD128:         .functype add_v16i8 (v128, v128) -> (v128)
+; SIMD128-NEXT:    i8x16.add $push0=, $0, $1
+; SIMD128-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: add_v16i8:
+; NO-SIMD128-SLOW:         .functype add_v16i8 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push1=, 15
+; NO-SIMD128-SLOW-NEXT:    i32.add $push2=, $0, $pop1
+; NO-SIMD128-SLOW-NEXT:    i32.add $push0=, $16, $32
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop2), $pop0
+; NO-SIMD128-SLOW-NEXT:    i32.const $push4=, 14
+; NO-SIMD128-SLOW-NEXT:    i32.add $push5=, $0, $pop4
+; NO-SIMD128-SLOW-NEXT:    i32.add $push3=, $15, $31
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop5), $pop3
+; NO-SIMD128-SLOW-NEXT:    i32.const $push7=, 13
+; NO-SIMD128-SLOW-NEXT:    i32.add $push8=, $0, $pop7
+; NO-SIMD128-SLOW-NEXT:    i32.add $push6=, $14, $30
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop8), $pop6
+; NO-SIMD128-SLOW-NEXT:    i32.const $push10=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push11=, $0, $pop10
+; NO-SIMD128-SLOW-NEXT:    i32.add $push9=, $13, $29
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop11), $pop9
+; NO-SIMD128-SLOW-NEXT:    i32.const $push13=, 11
+; NO-SIMD128-SLOW-NEXT:    i32.add $push14=, $0, $pop13
+; NO-SIMD128-SLOW-NEXT:    i32.add $push12=, $12, $28
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop14), $pop12
+; NO-SIMD128-SLOW-NEXT:    i32.const $push16=, 10
+; NO-SIMD128-SLOW-NEXT:    i32.add $push17=, $0, $pop16
+; NO-SIMD128-SLOW-NEXT:    i32.add $push15=, $11, $27
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop17), $pop15
+; NO-SIMD128-SLOW-NEXT:    i32.const $push19=, 9
+; NO-SIMD128-SLOW-NEXT:    i32.add $push20=, $0, $pop19
+; NO-SIMD128-SLOW-NEXT:    i32.add $push18=, $10, $26
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop20), $pop18
+; NO-SIMD128-SLOW-NEXT:    i32.add $push21=, $9, $25
+; NO-SIMD128-SLOW-NEXT:    i32.store8 8($0), $pop21
+; NO-SIMD128-SLOW-NEXT:    i32.const $push23=, 7
+; NO-SIMD128-SLOW-NEXT:    i32.add $push24=, $0, $pop23
+; NO-SIMD128-SLOW-NEXT:    i32.add $push22=, $8, $24
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop24), $pop22
+; NO-SIMD128-SLOW-NEXT:    i32.const $push26=, 6
+; NO-SIMD128-SLOW-NEXT:    i32.add $push27=, $0, $pop26
+; NO-SIMD128-SLOW-NEXT:    i32.add $push25=, $7, $23
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop27), $pop25
+; NO-SIMD128-SLOW-NEXT:    i32.const $push29=, 5
+; NO-SIMD128-SLOW-NEXT:    i32.add $push30=, $0, $pop29
+; NO-SIMD128-SLOW-NEXT:    i32.add $push28=, $6, $22
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop30), $pop28
+; NO-SIMD128-SLOW-NEXT:    i32.add $push31=, $5, $21
+; NO-SIMD128-SLOW-NEXT:    i32.store8 4($0), $pop31
+; NO-SIMD128-SLOW-NEXT:    i32.const $push33=, 3
+; NO-SIMD128-SLOW-NEXT:    i32.add $push34=, $0, $pop33
+; NO-SIMD128-SLOW-NEXT:    i32.add $push32=, $4, $20
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop34), $pop32
+; NO-SIMD128-SLOW-NEXT:    i32.add $push35=, $3, $19
+; NO-SIMD128-SLOW-NEXT:    i32.store8 2($0), $pop35
+; NO-SIMD128-SLOW-NEXT:    i32.add $push36=, $2, $18
+; NO-SIMD128-SLOW-NEXT:    i32.store8 1($0), $pop36
+; NO-SIMD128-SLOW-NEXT:    i32.add $push37=, $1, $17
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($0), $pop37
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: add_v16i8:
+; NO-SIMD128-FAST:         .functype add_v16i8 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-FAST-NEXT:    i32.add $push0=, $1, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($0), $pop0
+; NO-SIMD128-FAST-NEXT:    i32.add $push1=, $2, $18
+; NO-SIMD128-FAST-NEXT:    i32.store8 1($0), $pop1
+; NO-SIMD128-FAST-NEXT:    i32.add $push2=, $3, $19
+; NO-SIMD128-FAST-NEXT:    i32.store8 2($0), $pop2
+; NO-SIMD128-FAST-NEXT:    i32.const $push3=, 3
+; NO-SIMD128-FAST-NEXT:    i32.add $push4=, $0, $pop3
+; NO-SIMD128-FAST-NEXT:    i32.add $push5=, $4, $20
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop4), $pop5
+; NO-SIMD128-FAST-NEXT:    i32.add $push6=, $5, $21
+; NO-SIMD128-FAST-NEXT:    i32.store8 4($0), $pop6
+; NO-SIMD128-FAST-NEXT:    i32.const $push7=, 5
+; NO-SIMD128-FAST-NEXT:    i32.add $push8=, $0, $pop7
+; NO-SIMD128-FAST-NEXT:    i32.add $push9=, $6, $22
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop8), $pop9
+; NO-SIMD128-FAST-NEXT:    i32.const $push10=, 6
+; NO-SIMD128-FAST-NEXT:    i32.add $push11=, $0, $pop10
+; NO-SIMD128-FAST-NEXT:    i32.add $push12=, $7, $23
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop11), $pop12
+; NO-SIMD128-FAST-NEXT:    i32.const $push13=, 7
+; NO-SIMD128-FAST-NEXT:    i32.add $push14=, $0, $pop13
+; NO-SIMD128-FAST-NEXT:    i32.add $push15=, $8, $24
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop14), $pop15
+; NO-SIMD128-FAST-NEXT:    i32.add $push16=, $9, $25
+; NO-SIMD128-FAST-NEXT:    i32.store8 8($0), $pop16
+; NO-SIMD128-FAST-NEXT:    i32.const $push17=, 9
+; NO-SIMD128-FAST-NEXT:    i32.add $push18=, $0, $pop17
+; NO-SIMD128-FAST-NEXT:    i32.add $push19=, $10, $26
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop18), $pop19
+; NO-SIMD128-FAST-NEXT:    i32.const $push20=, 10
+; NO-SIMD128-FAST-NEXT:    i32.add $push21=, $0, $pop20
+; NO-SIMD128-FAST-NEXT:    i32.add $push22=, $11, $27
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop21), $pop22
+; NO-SIMD128-FAST-NEXT:    i32.const $push23=, 11
+; NO-SIMD128-FAST-NEXT:    i32.add $push24=, $0, $pop23
+; NO-SIMD128-FAST-NEXT:    i32.add $push25=, $12, $28
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop24), $pop25
+; NO-SIMD128-FAST-NEXT:    i32.const $push26=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push27=, $0, $pop26
+; NO-SIMD128-FAST-NEXT:    i32.add $push28=, $13, $29
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop27), $pop28
+; NO-SIMD128-FAST-NEXT:    i32.const $push29=, 13
+; NO-SIMD128-FAST-NEXT:    i32.add $push30=, $0, $pop29
+; NO-SIMD128-FAST-NEXT:    i32.add $push31=, $14, $30
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop30), $pop31
+; NO-SIMD128-FAST-NEXT:    i32.const $push32=, 14
+; NO-SIMD128-FAST-NEXT:    i32.add $push33=, $0, $pop32
+; NO-SIMD128-FAST-NEXT:    i32.add $push34=, $15, $31
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop33), $pop34
+; NO-SIMD128-FAST-NEXT:    i32.const $push35=, 15
+; NO-SIMD128-FAST-NEXT:    i32.add $push36=, $0, $pop35
+; NO-SIMD128-FAST-NEXT:    i32.add $push37=, $16, $32
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop36), $pop37
+; NO-SIMD128-FAST-NEXT:    return
   %a = add <16 x i8> %x, %y
   ret <16 x i8> %a
 }
 
-; CHECK-LABEL: sub_v16i8:
-; NO-SIMD128-NOT: i8x16
-; SIMD128-NEXT: .functype sub_v16i8 (v128, v128) -> (v128){{$}}
-; SIMD128-NEXT: i8x16.sub $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <16 x i8> @sub_v16i8(<16 x i8> %x, <16 x i8> %y) {
+; SIMD128-LABEL: sub_v16i8:
+; SIMD128:         .functype sub_v16i8 (v128, v128) -> (v128)
+; SIMD128-NEXT:    i8x16.sub $push0=, $0, $1
+; SIMD128-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: sub_v16i8:
+; NO-SIMD128-SLOW:         .functype sub_v16i8 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push1=, 15
+; NO-SIMD128-SLOW-NEXT:    i32.add $push2=, $0, $pop1
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push0=, $16, $32
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop2), $pop0
+; NO-SIMD128-SLOW-NEXT:    i32.const $push4=, 14
+; NO-SIMD128-SLOW-NEXT:    i32.add $push5=, $0, $pop4
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push3=, $15, $31
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop5), $pop3
+; NO-SIMD128-SLOW-NEXT:    i32.const $push7=, 13
+; NO-SIMD128-SLOW-NEXT:    i32.add $push8=, $0, $pop7
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push6=, $14, $30
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop8), $pop6
+; NO-SIMD128-SLOW-NEXT:    i32.const $push10=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push11=, $0, $pop10
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push9=, $13, $29
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop11), $pop9
+; NO-SIMD128-SLOW-NEXT:    i32.const $push13=, 11
+; NO-SIMD128-SLOW-NEXT:    i32.add $push14=, $0, $pop13
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push12=, $12, $28
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop14), $pop12
+; NO-SIMD128-SLOW-NEXT:    i32.const $push16=, 10
+; NO-SIMD128-SLOW-NEXT:    i32.add $push17=, $0, $pop16
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push15=, $11, $27
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop17), $pop15
+; NO-SIMD128-SLOW-NEXT:    i32.const $push19=, 9
+; NO-SIMD128-SLOW-NEXT:    i32.add $push20=, $0, $pop19
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push18=, $10, $26
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop20), $pop18
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push21=, $9, $25
+; NO-SIMD128-SLOW-NEXT:    i32.store8 8($0), $pop21
+; NO-SIMD128-SLOW-NEXT:    i32.const $push23=, 7
+; NO-SIMD128-SLOW-NEXT:    i32.add $push24=, $0, $pop23
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push22=, $8, $24
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop24), $pop22
+; NO-SIMD128-SLOW-NEXT:    i32.const $push26=, 6
+; NO-SIMD128-SLOW-NEXT:    i32.add $push27=, $0, $pop26
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push25=, $7, $23
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop27), $pop25
+; NO-SIMD128-SLOW-NEXT:    i32.const $push29=, 5
+; NO-SIMD128-SLOW-NEXT:    i32.add $push30=, $0, $pop29
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push28=, $6, $22
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop30), $pop28
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push31=, $5, $21
+; NO-SIMD128-SLOW-NEXT:    i32.store8 4($0), $pop31
+; NO-SIMD128-SLOW-NEXT:    i32.const $push33=, 3
+; NO-SIMD128-SLOW-NEXT:    i32.add $push34=, $0, $pop33
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push32=, $4, $20
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop34), $pop32
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push35=, $3, $19
+; NO-SIMD128-SLOW-NEXT:    i32.store8 2($0), $pop35
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push36=, $2, $18
+; NO-SIMD128-SLOW-NEXT:    i32.store8 1($0), $pop36
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push37=, $1, $17
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($0), $pop37
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: sub_v16i8:
+; NO-SIMD128-FAST:         .functype sub_v16i8 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-FAST-NEXT:    i32.sub $push0=, $1, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($0), $pop0
+; NO-SIMD128-FAST-NEXT:    i32.sub $push1=, $2, $18
+; NO-SIMD128-FAST-NEXT:    i32.store8 1($0), $pop1
+; NO-SIMD128-FAST-NEXT:    i32.sub $push2=, $3, $19
+; NO-SIMD128-FAST-NEXT:    i32.store8 2($0), $pop2
+; NO-SIMD128-FAST-NEXT:    i32.const $push3=, 3
+; NO-SIMD128-FAST-NEXT:    i32.add $push4=, $0, $pop3
+; NO-SIMD128-FAST-NEXT:    i32.sub $push5=, $4, $20
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop4), $pop5
+; NO-SIMD128-FAST-NEXT:    i32.sub $push6=, $5, $21
+; NO-SIMD128-FAST-NEXT:    i32.store8 4($0), $pop6
+; NO-SIMD128-FAST-NEXT:    i32.const $push7=, 5
+; NO-SIMD128-FAST-NEXT:    i32.add $push8=, $0, $pop7
+; NO-SIMD128-FAST-NEXT:    i32.sub $push9=, $6, $22
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop8), $pop9
+; NO-SIMD128-FAST-NEXT:    i32.const $push10=, 6
+; NO-SIMD128-FAST-NEXT:    i32.add $push11=, $0, $pop10
+; NO-SIMD128-FAST-NEXT:    i32.sub $push12=, $7, $23
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop11), $pop12
+; NO-SIMD128-FAST-NEXT:    i32.const $push13=, 7
+; NO-SIMD128-FAST-NEXT:    i32.add $push14=, $0, $pop13
+; NO-SIMD128-FAST-NEXT:    i32.sub $push15=, $8, $24
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop14), $pop15
+; NO-SIMD128-FAST-NEXT:    i32.sub $push16=, $9, $25
+; NO-SIMD128-FAST-NEXT:    i32.store8 8($0), $pop16
+; NO-SIMD128-FAST-NEXT:    i32.const $push17=, 9
+; NO-SIMD128-FAST-NEXT:    i32.add $push18=, $0, $pop17
+; NO-SIMD128-FAST-NEXT:    i32.sub $push19=, $10, $26
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop18), $pop19
+; NO-SIMD128-FAST-NEXT:    i32.const $push20=, 10
+; NO-SIMD128-FAST-NEXT:    i32.add $push21=, $0, $pop20
+; NO-SIMD128-FAST-NEXT:    i32.sub $push22=, $11, $27
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop21), $pop22
+; NO-SIMD128-FAST-NEXT:    i32.const $push23=, 11
+; NO-SIMD128-FAST-NEXT:    i32.add $push24=, $0, $pop23
+; NO-SIMD128-FAST-NEXT:    i32.sub $push25=, $12, $28
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop24), $pop25
+; NO-SIMD128-FAST-NEXT:    i32.const $push26=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push27=, $0, $pop26
+; NO-SIMD128-FAST-NEXT:    i32.sub $push28=, $13, $29
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop27), $pop28
+; NO-SIMD128-FAST-NEXT:    i32.const $push29=, 13
+; NO-SIMD128-FAST-NEXT:    i32.add $push30=, $0, $pop29
+; NO-SIMD128-FAST-NEXT:    i32.sub $push31=, $14, $30
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop30), $pop31
+; NO-SIMD128-FAST-NEXT:    i32.const $push32=, 14
+; NO-SIMD128-FAST-NEXT:    i32.add $push33=, $0, $pop32
+; NO-SIMD128-FAST-NEXT:    i32.sub $push34=, $15, $31
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop33), $pop34
+; NO-SIMD128-FAST-NEXT:    i32.const $push35=, 15
+; NO-SIMD128-FAST-NEXT:    i32.add $push36=, $0, $pop35
+; NO-SIMD128-FAST-NEXT:    i32.sub $push37=, $16, $32
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop36), $pop37
+; NO-SIMD128-FAST-NEXT:    return
   %a = sub <16 x i8> %x, %y
   ret <16 x i8> %a
 }
 
-; CHECK-LABEL: mul_v16i8:
-; NO-SIMD128-NOT: i8x16
-; SIMD128-NEXT: .functype mul_v16i8 (v128, v128) -> (v128){{$}}
-; SIMD128-NEXT: i8x16.mul $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <16 x i8> @mul_v16i8(<16 x i8> %x, <16 x i8> %y) {
+; SIMD128-LABEL: mul_v16i8:
+; SIMD128:         .functype mul_v16i8 (v128, v128) -> (v128)
+; SIMD128-NEXT:    i8x16.mul $push0=, $0, $1
+; SIMD128-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: mul_v16i8:
+; NO-SIMD128-SLOW:         .functype mul_v16i8 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push1=, 15
+; NO-SIMD128-SLOW-NEXT:    i32.add $push2=, $0, $pop1
+; NO-SIMD128-SLOW-NEXT:    i32.mul $push0=, $16, $32
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop2), $pop0
+; NO-SIMD128-SLOW-NEXT:    i32.const $push4=, 14
+; NO-SIMD128-SLOW-NEXT:    i32.add $push5=, $0, $pop4
+; NO-SIMD128-SLOW-NEXT:    i32.mul $push3=, $15, $31
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop5), $pop3
+; NO-SIMD128-SLOW-NEXT:    i32.const $push7=, 13
+; NO-SIMD128-SLOW-NEXT:    i32.add $push8=, $0, $pop7
+; NO-SIMD128-SLOW-NEXT:    i32.mul $push6=, $14, $30
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop8), $pop6
+; NO-SIMD128-SLOW-NEXT:    i32.const $push10=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push11=, $0, $pop10
+; NO-SIMD128-SLOW-NEXT:    i32.mul $push9=, $13, $29
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop11), $pop9
+; NO-SIMD128-SLOW-NEXT:    i32.const $push13=, 11
+; NO-SIMD128-SLOW-NEXT:    i32.add $push14=, $0, $pop13
+; NO-SIMD128-SLOW-NEXT:    i32.mul $push12=, $12, $28
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop14), $pop12
+; NO-SIMD128-SLOW-NEXT:    i32.const $push16=, 10
+; NO-SIMD128-SLOW-NEXT:    i32.add $push17=, $0, $pop16
+; NO-SIMD128-SLOW-NEXT:    i32.mul $push15=, $11, $27
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop17), $pop15
+; NO-SIMD128-SLOW-NEXT:    i32.const $push19=, 9
+; NO-SIMD128-SLOW-NEXT:    i32.add $push20=, $0, $pop19
+; NO-SIMD128-SLOW-NEXT:    i32.mul $push18=, $10, $26
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop20), $pop18
+; NO-SIMD128-SLOW-NEXT:    i32.mul $push21=, $9, $25
+; NO-SIMD128-SLOW-NEXT:    i32.store8 8($0), $pop21
+; NO-SIMD128-SLOW-NEXT:    i32.const $push23=, 7
+; NO-SIMD128-SLOW-NEXT:    i32.add $push24=, $0, $pop23
+; NO-SIMD128-SLOW-NEXT:    i32.mul $push22=, $8, $24
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop24), $pop22
+; NO-SIMD128-SLOW-NEXT:    i32.const $push26=, 6
+; NO-SIMD128-SLOW-NEXT:    i32.add $push27=, $0, $pop26
+; NO-SIMD128-SLOW-NEXT:    i32.mul $push25=, $7, $23
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop27), $pop25
+; NO-SIMD128-SLOW-NEXT:    i32.const $push29=, 5
+; NO-SIMD128-SLOW-NEXT:    i32.add $push30=, $0, $pop29
+; NO-SIMD128-SLOW-NEXT:    i32.mul $push28=, $6, $22
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop30), $pop28
+; NO-SIMD128-SLOW-NEXT:    i32.mul $push31=, $5, $21
+; NO-SIMD128-SLOW-NEXT:    i32.store8 4($0), $pop31
+; NO-SIMD128-SLOW-NEXT:    i32.const $push33=, 3
+; NO-SIMD128-SLOW-NEXT:    i32.add $push34=, $0, $pop33
+; NO-SIMD128-SLOW-NEXT:    i32.mul $push32=, $4, $20
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop34), $pop32
+; NO-SIMD128-SLOW-NEXT:    i32.mul $push35=, $3, $19
+; NO-SIMD128-SLOW-NEXT:    i32.store8 2($0), $pop35
+; NO-SIMD128-SLOW-NEXT:    i32.mul $push36=, $2, $18
+; NO-SIMD128-SLOW-NEXT:    i32.store8 1($0), $pop36
+; NO-SIMD128-SLOW-NEXT:    i32.mul $push37=, $1, $17
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($0), $pop37
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: mul_v16i8:
+; NO-SIMD128-FAST:         .functype mul_v16i8 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-FAST-NEXT:    i32.mul $push0=, $1, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($0), $pop0
+; NO-SIMD128-FAST-NEXT:    i32.mul $push1=, $2, $18
+; NO-SIMD128-FAST-NEXT:    i32.store8 1($0), $pop1
+; NO-SIMD128-FAST-NEXT:    i32.mul $push2=, $3, $19
+; NO-SIMD128-FAST-NEXT:    i32.store8 2($0), $pop2
+; NO-SIMD128-FAST-NEXT:    i32.const $push3=, 3
+; NO-SIMD128-FAST-NEXT:    i32.add $push4=, $0, $pop3
+; NO-SIMD128-FAST-NEXT:    i32.mul $push5=, $4, $20
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop4), $pop5
+; NO-SIMD128-FAST-NEXT:    i32.mul $push6=, $5, $21
+; NO-SIMD128-FAST-NEXT:    i32.store8 4($0), $pop6
+; NO-SIMD128-FAST-NEXT:    i32.const $push7=, 5
+; NO-SIMD128-FAST-NEXT:    i32.add $push8=, $0, $pop7
+; NO-SIMD128-FAST-NEXT:    i32.mul $push9=, $6, $22
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop8), $pop9
+; NO-SIMD128-FAST-NEXT:    i32.const $push10=, 6
+; NO-SIMD128-FAST-NEXT:    i32.add $push11=, $0, $pop10
+; NO-SIMD128-FAST-NEXT:    i32.mul $push12=, $7, $23
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop11), $pop12
+; NO-SIMD128-FAST-NEXT:    i32.const $push13=, 7
+; NO-SIMD128-FAST-NEXT:    i32.add $push14=, $0, $pop13
+; NO-SIMD128-FAST-NEXT:    i32.mul $push15=, $8, $24
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop14), $pop15
+; NO-SIMD128-FAST-NEXT:    i32.mul $push16=, $9, $25
+; NO-SIMD128-FAST-NEXT:    i32.store8 8($0), $pop16
+; NO-SIMD128-FAST-NEXT:    i32.const $push17=, 9
+; NO-SIMD128-FAST-NEXT:    i32.add $push18=, $0, $pop17
+; NO-SIMD128-FAST-NEXT:    i32.mul $push19=, $10, $26
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop18), $pop19
+; NO-SIMD128-FAST-NEXT:    i32.const $push20=, 10
+; NO-SIMD128-FAST-NEXT:    i32.add $push21=, $0, $pop20
+; NO-SIMD128-FAST-NEXT:    i32.mul $push22=, $11, $27
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop21), $pop22
+; NO-SIMD128-FAST-NEXT:    i32.const $push23=, 11
+; NO-SIMD128-FAST-NEXT:    i32.add $push24=, $0, $pop23
+; NO-SIMD128-FAST-NEXT:    i32.mul $push25=, $12, $28
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop24), $pop25
+; NO-SIMD128-FAST-NEXT:    i32.const $push26=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push27=, $0, $pop26
+; NO-SIMD128-FAST-NEXT:    i32.mul $push28=, $13, $29
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop27), $pop28
+; NO-SIMD128-FAST-NEXT:    i32.const $push29=, 13
+; NO-SIMD128-FAST-NEXT:    i32.add $push30=, $0, $pop29
+; NO-SIMD128-FAST-NEXT:    i32.mul $push31=, $14, $30
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop30), $pop31
+; NO-SIMD128-FAST-NEXT:    i32.const $push32=, 14
+; NO-SIMD128-FAST-NEXT:    i32.add $push33=, $0, $pop32
+; NO-SIMD128-FAST-NEXT:    i32.mul $push34=, $15, $31
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop33), $pop34
+; NO-SIMD128-FAST-NEXT:    i32.const $push35=, 15
+; NO-SIMD128-FAST-NEXT:    i32.add $push36=, $0, $pop35
+; NO-SIMD128-FAST-NEXT:    i32.mul $push37=, $16, $32
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop36), $pop37
+; NO-SIMD128-FAST-NEXT:    return
   %a = mul <16 x i8> %x, %y
   ret <16 x i8> %a
 }
 
-; CHECK-LABEL: neg_v16i8:
-; NO-SIMD128-NOT: i8x16
-; SIMD128-NEXT: .functype neg_v16i8 (v128) -> (v128){{$}}
-; SIMD128-NEXT: i8x16.neg $push[[R:[0-9]+]]=, $0{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <16 x i8> @neg_v16i8(<16 x i8> %x) {
+; SIMD128-LABEL: neg_v16i8:
+; SIMD128:         .functype neg_v16i8 (v128) -> (v128)
+; SIMD128-NEXT:    i8x16.neg $push0=, $0
+; SIMD128-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: neg_v16i8:
+; NO-SIMD128-SLOW:         .functype neg_v16i8 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push2=, 15
+; NO-SIMD128-SLOW-NEXT:    i32.add $push3=, $0, $pop2
+; NO-SIMD128-SLOW-NEXT:    i32.const $push0=, 0
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push1=, $pop0, $16
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop3), $pop1
+; NO-SIMD128-SLOW-NEXT:    i32.const $push5=, 14
+; NO-SIMD128-SLOW-NEXT:    i32.add $push6=, $0, $pop5
+; NO-SIMD128-SLOW-NEXT:    i32.const $push53=, 0
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push4=, $pop53, $15
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop6), $pop4
+; NO-SIMD128-SLOW-NEXT:    i32.const $push8=, 13
+; NO-SIMD128-SLOW-NEXT:    i32.add $push9=, $0, $pop8
+; NO-SIMD128-SLOW-NEXT:    i32.const $push52=, 0
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push7=, $pop52, $14
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop9), $pop7
+; NO-SIMD128-SLOW-NEXT:    i32.const $push11=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push12=, $0, $pop11
+; NO-SIMD128-SLOW-NEXT:    i32.const $push51=, 0
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push10=, $pop51, $13
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop12), $pop10
+; NO-SIMD128-SLOW-NEXT:    i32.const $push14=, 11
+; NO-SIMD128-SLOW-NEXT:    i32.add $push15=, $0, $pop14
+; NO-SIMD128-SLOW-NEXT:    i32.const $push50=, 0
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push13=, $pop50, $12
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop15), $pop13
+; NO-SIMD128-SLOW-NEXT:    i32.const $push17=, 10
+; NO-SIMD128-SLOW-NEXT:    i32.add $push18=, $0, $pop17
+; NO-SIMD128-SLOW-NEXT:    i32.const $push49=, 0
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push16=, $pop49, $11
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop18), $pop16
+; NO-SIMD128-SLOW-NEXT:    i32.const $push20=, 9
+; NO-SIMD128-SLOW-NEXT:    i32.add $push21=, $0, $pop20
+; NO-SIMD128-SLOW-NEXT:    i32.const $push48=, 0
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push19=, $pop48, $10
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop21), $pop19
+; NO-SIMD128-SLOW-NEXT:    i32.const $push47=, 0
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push22=, $pop47, $9
+; NO-SIMD128-SLOW-NEXT:    i32.store8 8($0), $pop22
+; NO-SIMD128-SLOW-NEXT:    i32.const $push24=, 7
+; NO-SIMD128-SLOW-NEXT:    i32.add $push25=, $0, $pop24
+; NO-SIMD128-SLOW-NEXT:    i32.const $push46=, 0
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push23=, $pop46, $8
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop25), $pop23
+; NO-SIMD128-SLOW-NEXT:    i32.const $push27=, 6
+; NO-SIMD128-SLOW-NEXT:    i32.add $push28=, $0, $pop27
+; NO-SIMD128-SLOW-NEXT:    i32.const $push45=, 0
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push26=, $pop45, $7
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop28), $pop26
+; NO-SIMD128-SLOW-NEXT:    i32.const $push30=, 5
+; NO-SIMD128-SLOW-NEXT:    i32.add $push31=, $0, $pop30
+; NO-SIMD128-SLOW-NEXT:    i32.const $push44=, 0
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push29=, $pop44, $6
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop31), $pop29
+; NO-SIMD128-SLOW-NEXT:    i32.const $push43=, 0
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push32=, $pop43, $5
+; NO-SIMD128-SLOW-NEXT:    i32.store8 4($0), $pop32
+; NO-SIMD128-SLOW-NEXT:    i32.const $push34=, 3
+; NO-SIMD128-SLOW-NEXT:    i32.add $push35=, $0, $pop34
+; NO-SIMD128-SLOW-NEXT:    i32.const $push42=, 0
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push33=, $pop42, $4
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop35), $pop33
+; NO-SIMD128-SLOW-NEXT:    i32.const $push41=, 0
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push36=, $pop41, $3
+; NO-SIMD128-SLOW-NEXT:    i32.store8 2($0), $pop36
+; NO-SIMD128-SLOW-NEXT:    i32.const $push40=, 0
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push37=, $pop40, $2
+; NO-SIMD128-SLOW-NEXT:    i32.store8 1($0), $pop37
+; NO-SIMD128-SLOW-NEXT:    i32.const $push39=, 0
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push38=, $pop39, $1
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($0), $pop38
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: neg_v16i8:
+; NO-SIMD128-FAST:         .functype neg_v16i8 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-FAST-NEXT:    i32.const $push0=, 0
+; NO-SIMD128-FAST-NEXT:    i32.sub $push1=, $pop0, $1
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($0), $pop1
+; NO-SIMD128-FAST-NEXT:    i32.const $push53=, 0
+; NO-SIMD128-FAST-NEXT:    i32.sub $push2=, $pop53, $2
+; NO-SIMD128-FAST-NEXT:    i32.store8 1($0), $pop2
+; NO-SIMD128-FAST-NEXT:    i32.const $push52=, 0
+; NO-SIMD128-FAST-NEXT:    i32.sub $push3=, $pop52, $3
+; NO-SIMD128-FAST-NEXT:    i32.store8 2($0), $pop3
+; NO-SIMD128-FAST-NEXT:    i32.const $push4=, 3
+; NO-SIMD128-FAST-NEXT:    i32.add $push5=, $0, $pop4
+; NO-SIMD128-FAST-NEXT:    i32.const $push51=, 0
+; NO-SIMD128-FAST-NEXT:    i32.sub $push6=, $pop51, $4
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop5), $pop6
+; NO-SIMD128-FAST-NEXT:    i32.const $push50=, 0
+; NO-SIMD128-FAST-NEXT:    i32.sub $push7=, $pop50, $5
+; NO-SIMD128-FAST-NEXT:    i32.store8 4($0), $pop7
+; NO-SIMD128-FAST-NEXT:    i32.const $push8=, 5
+; NO-SIMD128-FAST-NEXT:    i32.add $push9=, $0, $pop8
+; NO-SIMD128-FAST-NEXT:    i32.const $push49=, 0
+; NO-SIMD128-FAST-NEXT:    i32.sub $push10=, $pop49, $6
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop9), $pop10
+; NO-SIMD128-FAST-NEXT:    i32.const $push11=, 6
+; NO-SIMD128-FAST-NEXT:    i32.add $push12=, $0, $pop11
+; NO-SIMD128-FAST-NEXT:    i32.const $push48=, 0
+; NO-SIMD128-FAST-NEXT:    i32.sub $push13=, $pop48, $7
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop12), $pop13
+; NO-SIMD128-FAST-NEXT:    i32.const $push14=, 7
+; NO-SIMD128-FAST-NEXT:    i32.add $push15=, $0, $pop14
+; NO-SIMD128-FAST-NEXT:    i32.const $push47=, 0
+; NO-SIMD128-FAST-NEXT:    i32.sub $push16=, $pop47, $8
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop15), $pop16
+; NO-SIMD128-FAST-NEXT:    i32.const $push46=, 0
+; NO-SIMD128-FAST-NEXT:    i32.sub $push17=, $pop46, $9
+; NO-SIMD128-FAST-NEXT:    i32.store8 8($0), $pop17
+; NO-SIMD128-FAST-NEXT:    i32.const $push18=, 9
+; NO-SIMD128-FAST-NEXT:    i32.add $push19=, $0, $pop18
+; NO-SIMD128-FAST-NEXT:    i32.const $push45=, 0
+; NO-SIMD128-FAST-NEXT:    i32.sub $push20=, $pop45, $10
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop19), $pop20
+; NO-SIMD128-FAST-NEXT:    i32.const $push21=, 10
+; NO-SIMD128-FAST-NEXT:    i32.add $push22=, $0, $pop21
+; NO-SIMD128-FAST-NEXT:    i32.const $push44=, 0
+; NO-SIMD128-FAST-NEXT:    i32.sub $push23=, $pop44, $11
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop22), $pop23
+; NO-SIMD128-FAST-NEXT:    i32.const $push24=, 11
+; NO-SIMD128-FAST-NEXT:    i32.add $push25=, $0, $pop24
+; NO-SIMD128-FAST-NEXT:    i32.const $push43=, 0
+; NO-SIMD128-FAST-NEXT:    i32.sub $push26=, $pop43, $12
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop25), $pop26
+; NO-SIMD128-FAST-NEXT:    i32.const $push27=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push28=, $0, $pop27
+; NO-SIMD128-FAST-NEXT:    i32.const $push42=, 0
+; NO-SIMD128-FAST-NEXT:    i32.sub $push29=, $pop42, $13
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop28), $pop29
+; NO-SIMD128-FAST-NEXT:    i32.const $push30=, 13
+; NO-SIMD128-FAST-NEXT:    i32.add $push31=, $0, $pop30
+; NO-SIMD128-FAST-NEXT:    i32.const $push41=, 0
+; NO-SIMD128-FAST-NEXT:    i32.sub $push32=, $pop41, $14
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop31), $pop32
+; NO-SIMD128-FAST-NEXT:    i32.const $push33=, 14
+; NO-SIMD128-FAST-NEXT:    i32.add $push34=, $0, $pop33
+; NO-SIMD128-FAST-NEXT:    i32.const $push40=, 0
+; NO-SIMD128-FAST-NEXT:    i32.sub $push35=, $pop40, $15
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop34), $pop35
+; NO-SIMD128-FAST-NEXT:    i32.const $push36=, 15
+; NO-SIMD128-FAST-NEXT:    i32.add $push37=, $0, $pop36
+; NO-SIMD128-FAST-NEXT:    i32.const $push39=, 0
+; NO-SIMD128-FAST-NEXT:    i32.sub $push38=, $pop39, $16
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop37), $pop38
+; NO-SIMD128-FAST-NEXT:    return
   %a = sub <16 x i8> <i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0,
                       i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0>,
                      %x
   ret <16 x i8> %a
 }
 
-; CHECK-LABEL: shl_v16i8:
-; NO-SIMD128-NOT: i8x16
-; SIMD128-NEXT: .functype shl_v16i8 (v128, i32) -> (v128){{$}}
-; SIMD128-NEXT: i8x16.shl $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <16 x i8> @shl_v16i8(<16 x i8> %v, i8 %x) {
+; SIMD128-UNIMPL-LABEL: shl_v16i8:
+; SIMD128-UNIMPL:         .functype shl_v16i8 (v128, i32) -> (v128)
+; SIMD128-UNIMPL-NEXT:    i8x16.shl $push0=, $0, $1
+; SIMD128-UNIMPL-NEXT:    return $pop0
+;
+; SIMD128-VM-SLOW-LABEL: shl_v16i8:
+; SIMD128-VM-SLOW:         .functype shl_v16i8 (v128, i32) -> (v128)
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push9=, $0, 0
+; SIMD128-VM-SLOW-NEXT:    i8x16.splat $push2=, $1
+; SIMD128-VM-SLOW-NEXT:    i32.const $push0=, 7
+; SIMD128-VM-SLOW-NEXT:    i8x16.splat $push1=, $pop0
+; SIMD128-VM-SLOW-NEXT:    v128.and $push100=, $pop2, $pop1
+; SIMD128-VM-SLOW-NEXT:    local.tee $push99=, $2=, $pop100
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push7=, $pop99, 0
+; SIMD128-VM-SLOW-NEXT:    i32.const $push98=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push8=, $pop7, $pop98
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push10=, $pop9, $pop8
+; SIMD128-VM-SLOW-NEXT:    i8x16.splat $push11=, $pop10
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push5=, $0, 1
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push3=, $2, 1
+; SIMD128-VM-SLOW-NEXT:    i32.const $push97=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push4=, $pop3, $pop97
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push6=, $pop5, $pop4
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push12=, $pop11, 1, $pop6
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push15=, $0, 2
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push13=, $2, 2
+; SIMD128-VM-SLOW-NEXT:    i32.const $push96=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push14=, $pop13, $pop96
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push16=, $pop15, $pop14
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push17=, $pop12, 2, $pop16
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push20=, $0, 3
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push18=, $2, 3
+; SIMD128-VM-SLOW-NEXT:    i32.const $push95=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push19=, $pop18, $pop95
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push21=, $pop20, $pop19
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push22=, $pop17, 3, $pop21
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push25=, $0, 4
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push23=, $2, 4
+; SIMD128-VM-SLOW-NEXT:    i32.const $push94=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push24=, $pop23, $pop94
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push26=, $pop25, $pop24
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push27=, $pop22, 4, $pop26
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push30=, $0, 5
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push28=, $2, 5
+; SIMD128-VM-SLOW-NEXT:    i32.const $push93=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push29=, $pop28, $pop93
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push31=, $pop30, $pop29
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push32=, $pop27, 5, $pop31
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push35=, $0, 6
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push33=, $2, 6
+; SIMD128-VM-SLOW-NEXT:    i32.const $push92=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push34=, $pop33, $pop92
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push36=, $pop35, $pop34
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push37=, $pop32, 6, $pop36
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push40=, $0, 7
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push38=, $2, 7
+; SIMD128-VM-SLOW-NEXT:    i32.const $push91=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push39=, $pop38, $pop91
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push41=, $pop40, $pop39
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push42=, $pop37, 7, $pop41
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push45=, $0, 8
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push43=, $2, 8
+; SIMD128-VM-SLOW-NEXT:    i32.const $push90=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push44=, $pop43, $pop90
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push46=, $pop45, $pop44
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push47=, $pop42, 8, $pop46
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push50=, $0, 9
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push48=, $2, 9
+; SIMD128-VM-SLOW-NEXT:    i32.const $push89=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push49=, $pop48, $pop89
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push51=, $pop50, $pop49
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push52=, $pop47, 9, $pop51
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push55=, $0, 10
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push53=, $2, 10
+; SIMD128-VM-SLOW-NEXT:    i32.const $push88=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push54=, $pop53, $pop88
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push56=, $pop55, $pop54
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push57=, $pop52, 10, $pop56
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push60=, $0, 11
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push58=, $2, 11
+; SIMD128-VM-SLOW-NEXT:    i32.const $push87=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push59=, $pop58, $pop87
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push61=, $pop60, $pop59
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push62=, $pop57, 11, $pop61
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push65=, $0, 12
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push63=, $2, 12
+; SIMD128-VM-SLOW-NEXT:    i32.const $push86=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push64=, $pop63, $pop86
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push66=, $pop65, $pop64
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push67=, $pop62, 12, $pop66
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push70=, $0, 13
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push68=, $2, 13
+; SIMD128-VM-SLOW-NEXT:    i32.const $push85=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push69=, $pop68, $pop85
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push71=, $pop70, $pop69
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push72=, $pop67, 13, $pop71
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push75=, $0, 14
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push73=, $2, 14
+; SIMD128-VM-SLOW-NEXT:    i32.const $push84=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push74=, $pop73, $pop84
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push76=, $pop75, $pop74
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push77=, $pop72, 14, $pop76
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push80=, $0, 15
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push78=, $2, 15
+; SIMD128-VM-SLOW-NEXT:    i32.const $push83=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push79=, $pop78, $pop83
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push81=, $pop80, $pop79
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push82=, $pop77, 15, $pop81
+; SIMD128-VM-SLOW-NEXT:    return $pop82
+;
+; SIMD128-VM-FAST-LABEL: shl_v16i8:
+; SIMD128-VM-FAST:         .functype shl_v16i8 (v128, i32) -> (v128)
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push10=, $0, 0
+; SIMD128-VM-FAST-NEXT:    i8x16.splat $push1=, $1
+; SIMD128-VM-FAST-NEXT:    i32.const $push2=, 7
+; SIMD128-VM-FAST-NEXT:    i8x16.splat $push3=, $pop2
+; SIMD128-VM-FAST-NEXT:    v128.and $push100=, $pop1, $pop3
+; SIMD128-VM-FAST-NEXT:    local.tee $push99=, $2=, $pop100
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push8=, $pop99, 0
+; SIMD128-VM-FAST-NEXT:    i32.const $push98=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push9=, $pop8, $pop98
+; SIMD128-VM-FAST-NEXT:    i32.shl $push11=, $pop10, $pop9
+; SIMD128-VM-FAST-NEXT:    i8x16.splat $push12=, $pop11
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push6=, $0, 1
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push4=, $2, 1
+; SIMD128-VM-FAST-NEXT:    i32.const $push97=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push5=, $pop4, $pop97
+; SIMD128-VM-FAST-NEXT:    i32.shl $push7=, $pop6, $pop5
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push13=, $pop12, 1, $pop7
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push16=, $0, 2
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push14=, $2, 2
+; SIMD128-VM-FAST-NEXT:    i32.const $push96=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push15=, $pop14, $pop96
+; SIMD128-VM-FAST-NEXT:    i32.shl $push17=, $pop16, $pop15
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push18=, $pop13, 2, $pop17
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push21=, $0, 3
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push19=, $2, 3
+; SIMD128-VM-FAST-NEXT:    i32.const $push95=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push20=, $pop19, $pop95
+; SIMD128-VM-FAST-NEXT:    i32.shl $push22=, $pop21, $pop20
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push23=, $pop18, 3, $pop22
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push26=, $0, 4
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push24=, $2, 4
+; SIMD128-VM-FAST-NEXT:    i32.const $push94=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push25=, $pop24, $pop94
+; SIMD128-VM-FAST-NEXT:    i32.shl $push27=, $pop26, $pop25
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push28=, $pop23, 4, $pop27
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push31=, $0, 5
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push29=, $2, 5
+; SIMD128-VM-FAST-NEXT:    i32.const $push93=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push30=, $pop29, $pop93
+; SIMD128-VM-FAST-NEXT:    i32.shl $push32=, $pop31, $pop30
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push33=, $pop28, 5, $pop32
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push36=, $0, 6
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push34=, $2, 6
+; SIMD128-VM-FAST-NEXT:    i32.const $push92=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push35=, $pop34, $pop92
+; SIMD128-VM-FAST-NEXT:    i32.shl $push37=, $pop36, $pop35
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push38=, $pop33, 6, $pop37
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push41=, $0, 7
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push39=, $2, 7
+; SIMD128-VM-FAST-NEXT:    i32.const $push91=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push40=, $pop39, $pop91
+; SIMD128-VM-FAST-NEXT:    i32.shl $push42=, $pop41, $pop40
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push43=, $pop38, 7, $pop42
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push46=, $0, 8
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push44=, $2, 8
+; SIMD128-VM-FAST-NEXT:    i32.const $push90=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push45=, $pop44, $pop90
+; SIMD128-VM-FAST-NEXT:    i32.shl $push47=, $pop46, $pop45
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push48=, $pop43, 8, $pop47
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push51=, $0, 9
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push49=, $2, 9
+; SIMD128-VM-FAST-NEXT:    i32.const $push89=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push50=, $pop49, $pop89
+; SIMD128-VM-FAST-NEXT:    i32.shl $push52=, $pop51, $pop50
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push53=, $pop48, 9, $pop52
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push56=, $0, 10
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push54=, $2, 10
+; SIMD128-VM-FAST-NEXT:    i32.const $push88=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push55=, $pop54, $pop88
+; SIMD128-VM-FAST-NEXT:    i32.shl $push57=, $pop56, $pop55
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push58=, $pop53, 10, $pop57
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push61=, $0, 11
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push59=, $2, 11
+; SIMD128-VM-FAST-NEXT:    i32.const $push87=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push60=, $pop59, $pop87
+; SIMD128-VM-FAST-NEXT:    i32.shl $push62=, $pop61, $pop60
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push63=, $pop58, 11, $pop62
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push66=, $0, 12
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push64=, $2, 12
+; SIMD128-VM-FAST-NEXT:    i32.const $push86=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push65=, $pop64, $pop86
+; SIMD128-VM-FAST-NEXT:    i32.shl $push67=, $pop66, $pop65
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push68=, $pop63, 12, $pop67
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push71=, $0, 13
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push69=, $2, 13
+; SIMD128-VM-FAST-NEXT:    i32.const $push85=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push70=, $pop69, $pop85
+; SIMD128-VM-FAST-NEXT:    i32.shl $push72=, $pop71, $pop70
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push73=, $pop68, 13, $pop72
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push76=, $0, 14
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push74=, $2, 14
+; SIMD128-VM-FAST-NEXT:    i32.const $push84=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push75=, $pop74, $pop84
+; SIMD128-VM-FAST-NEXT:    i32.shl $push77=, $pop76, $pop75
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push78=, $pop73, 14, $pop77
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push81=, $0, 15
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push79=, $2, 15
+; SIMD128-VM-FAST-NEXT:    i32.const $push83=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push80=, $pop79, $pop83
+; SIMD128-VM-FAST-NEXT:    i32.shl $push82=, $pop81, $pop80
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push0=, $pop78, 15, $pop82
+; SIMD128-VM-FAST-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: shl_v16i8:
+; NO-SIMD128-SLOW:         .functype shl_v16i8 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push0=, 15
+; NO-SIMD128-SLOW-NEXT:    i32.add $push1=, $0, $pop0
+; NO-SIMD128-SLOW-NEXT:    i32.const $push2=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push40=, $17, $pop2
+; NO-SIMD128-SLOW-NEXT:    local.tee $push39=, $17=, $pop40
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push3=, $16, $pop39
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop1), $pop3
+; NO-SIMD128-SLOW-NEXT:    i32.const $push4=, 14
+; NO-SIMD128-SLOW-NEXT:    i32.add $push5=, $0, $pop4
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push6=, $15, $17
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop5), $pop6
+; NO-SIMD128-SLOW-NEXT:    i32.const $push7=, 13
+; NO-SIMD128-SLOW-NEXT:    i32.add $push8=, $0, $pop7
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push9=, $14, $17
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop8), $pop9
+; NO-SIMD128-SLOW-NEXT:    i32.const $push10=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push11=, $0, $pop10
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push12=, $13, $17
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop11), $pop12
+; NO-SIMD128-SLOW-NEXT:    i32.const $push13=, 11
+; NO-SIMD128-SLOW-NEXT:    i32.add $push14=, $0, $pop13
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push15=, $12, $17
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop14), $pop15
+; NO-SIMD128-SLOW-NEXT:    i32.const $push16=, 10
+; NO-SIMD128-SLOW-NEXT:    i32.add $push17=, $0, $pop16
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push18=, $11, $17
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop17), $pop18
+; NO-SIMD128-SLOW-NEXT:    i32.const $push19=, 9
+; NO-SIMD128-SLOW-NEXT:    i32.add $push20=, $0, $pop19
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push21=, $10, $17
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop20), $pop21
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push22=, $9, $17
+; NO-SIMD128-SLOW-NEXT:    i32.store8 8($0), $pop22
+; NO-SIMD128-SLOW-NEXT:    i32.const $push23=, 7
+; NO-SIMD128-SLOW-NEXT:    i32.add $push24=, $0, $pop23
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push25=, $8, $17
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop24), $pop25
+; NO-SIMD128-SLOW-NEXT:    i32.const $push26=, 6
+; NO-SIMD128-SLOW-NEXT:    i32.add $push27=, $0, $pop26
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push28=, $7, $17
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop27), $pop28
+; NO-SIMD128-SLOW-NEXT:    i32.const $push29=, 5
+; NO-SIMD128-SLOW-NEXT:    i32.add $push30=, $0, $pop29
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push31=, $6, $17
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop30), $pop31
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push32=, $5, $17
+; NO-SIMD128-SLOW-NEXT:    i32.store8 4($0), $pop32
+; NO-SIMD128-SLOW-NEXT:    i32.const $push33=, 3
+; NO-SIMD128-SLOW-NEXT:    i32.add $push34=, $0, $pop33
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push35=, $4, $17
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop34), $pop35
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push36=, $3, $17
+; NO-SIMD128-SLOW-NEXT:    i32.store8 2($0), $pop36
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push37=, $2, $17
+; NO-SIMD128-SLOW-NEXT:    i32.store8 1($0), $pop37
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push38=, $1, $17
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($0), $pop38
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: shl_v16i8:
+; NO-SIMD128-FAST:         .functype shl_v16i8 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-FAST-NEXT:    i32.const $push0=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push40=, $17, $pop0
+; NO-SIMD128-FAST-NEXT:    local.tee $push39=, $17=, $pop40
+; NO-SIMD128-FAST-NEXT:    i32.shl $push1=, $2, $pop39
+; NO-SIMD128-FAST-NEXT:    i32.store8 1($0), $pop1
+; NO-SIMD128-FAST-NEXT:    i32.shl $push2=, $1, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($0), $pop2
+; NO-SIMD128-FAST-NEXT:    i32.shl $push3=, $3, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 2($0), $pop3
+; NO-SIMD128-FAST-NEXT:    i32.const $push4=, 3
+; NO-SIMD128-FAST-NEXT:    i32.add $push5=, $0, $pop4
+; NO-SIMD128-FAST-NEXT:    i32.shl $push6=, $4, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop5), $pop6
+; NO-SIMD128-FAST-NEXT:    i32.shl $push7=, $5, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 4($0), $pop7
+; NO-SIMD128-FAST-NEXT:    i32.const $push8=, 5
+; NO-SIMD128-FAST-NEXT:    i32.add $push9=, $0, $pop8
+; NO-SIMD128-FAST-NEXT:    i32.shl $push10=, $6, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop9), $pop10
+; NO-SIMD128-FAST-NEXT:    i32.const $push11=, 6
+; NO-SIMD128-FAST-NEXT:    i32.add $push12=, $0, $pop11
+; NO-SIMD128-FAST-NEXT:    i32.shl $push13=, $7, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop12), $pop13
+; NO-SIMD128-FAST-NEXT:    i32.const $push14=, 7
+; NO-SIMD128-FAST-NEXT:    i32.add $push15=, $0, $pop14
+; NO-SIMD128-FAST-NEXT:    i32.shl $push16=, $8, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop15), $pop16
+; NO-SIMD128-FAST-NEXT:    i32.shl $push17=, $9, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 8($0), $pop17
+; NO-SIMD128-FAST-NEXT:    i32.const $push18=, 9
+; NO-SIMD128-FAST-NEXT:    i32.add $push19=, $0, $pop18
+; NO-SIMD128-FAST-NEXT:    i32.shl $push20=, $10, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop19), $pop20
+; NO-SIMD128-FAST-NEXT:    i32.const $push21=, 10
+; NO-SIMD128-FAST-NEXT:    i32.add $push22=, $0, $pop21
+; NO-SIMD128-FAST-NEXT:    i32.shl $push23=, $11, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop22), $pop23
+; NO-SIMD128-FAST-NEXT:    i32.const $push24=, 11
+; NO-SIMD128-FAST-NEXT:    i32.add $push25=, $0, $pop24
+; NO-SIMD128-FAST-NEXT:    i32.shl $push26=, $12, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop25), $pop26
+; NO-SIMD128-FAST-NEXT:    i32.const $push27=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push28=, $0, $pop27
+; NO-SIMD128-FAST-NEXT:    i32.shl $push29=, $13, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop28), $pop29
+; NO-SIMD128-FAST-NEXT:    i32.const $push30=, 13
+; NO-SIMD128-FAST-NEXT:    i32.add $push31=, $0, $pop30
+; NO-SIMD128-FAST-NEXT:    i32.shl $push32=, $14, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop31), $pop32
+; NO-SIMD128-FAST-NEXT:    i32.const $push33=, 14
+; NO-SIMD128-FAST-NEXT:    i32.add $push34=, $0, $pop33
+; NO-SIMD128-FAST-NEXT:    i32.shl $push35=, $15, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop34), $pop35
+; NO-SIMD128-FAST-NEXT:    i32.const $push36=, 15
+; NO-SIMD128-FAST-NEXT:    i32.add $push37=, $0, $pop36
+; NO-SIMD128-FAST-NEXT:    i32.shl $push38=, $16, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop37), $pop38
+; NO-SIMD128-FAST-NEXT:    return
   %t = insertelement <16 x i8> undef, i8 %x, i32 0
   %s = shufflevector <16 x i8> %t, <16 x i8> undef,
     <16 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0,
                 i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
   %a = shl <16 x i8> %v, %s
   ret <16 x i8> %a
 }
 
-; CHECK-LABEL: shl_const_v16i8:
-; NO-SIMD128-NOT: i8x16
-; SIMD128-NEXT: .functype shl_const_v16i8 (v128) -> (v128){{$}}
-; SIMD128-NEXT: i32.const $push[[L0:[0-9]+]]=, 5
-; SIMD128-NEXT: i8x16.shl $push[[R:[0-9]+]]=, $0, $pop[[L0]]{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <16 x i8> @shl_const_v16i8(<16 x i8> %v) {
+; SIMD128-UNIMPL-SLOW-LABEL: shl_const_v16i8:
+; SIMD128-UNIMPL-SLOW:         .functype shl_const_v16i8 (v128) -> (v128)
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.const $push0=, 5
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.shl $push1=, $0, $pop0
+; SIMD128-UNIMPL-SLOW-NEXT:    return $pop1
+;
+; SIMD128-UNIMPL-FAST-LABEL: shl_const_v16i8:
+; SIMD128-UNIMPL-FAST:         .functype shl_const_v16i8 (v128) -> (v128)
+; SIMD128-UNIMPL-FAST-NEXT:    i32.const $push1=, 5
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.shl $push0=, $0, $pop1
+; SIMD128-UNIMPL-FAST-NEXT:    return $pop0
+;
+; SIMD128-VM-SLOW-LABEL: shl_const_v16i8:
+; SIMD128-VM-SLOW:         .functype shl_const_v16i8 (v128) -> (v128)
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push3=, $0, 0
+; SIMD128-VM-SLOW-NEXT:    i32.const $push1=, 5
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push4=, $pop3, $pop1
+; SIMD128-VM-SLOW-NEXT:    i8x16.splat $push5=, $pop4
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push0=, $0, 1
+; SIMD128-VM-SLOW-NEXT:    i32.const $push63=, 5
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push2=, $pop0, $pop63
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push6=, $pop5, 1, $pop2
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push7=, $0, 2
+; SIMD128-VM-SLOW-NEXT:    i32.const $push62=, 5
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push8=, $pop7, $pop62
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push9=, $pop6, 2, $pop8
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push10=, $0, 3
+; SIMD128-VM-SLOW-NEXT:    i32.const $push61=, 5
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push11=, $pop10, $pop61
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push12=, $pop9, 3, $pop11
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push13=, $0, 4
+; SIMD128-VM-SLOW-NEXT:    i32.const $push60=, 5
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push14=, $pop13, $pop60
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push15=, $pop12, 4, $pop14
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push16=, $0, 5
+; SIMD128-VM-SLOW-NEXT:    i32.const $push59=, 5
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push17=, $pop16, $pop59
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push18=, $pop15, 5, $pop17
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push19=, $0, 6
+; SIMD128-VM-SLOW-NEXT:    i32.const $push58=, 5
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push20=, $pop19, $pop58
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push21=, $pop18, 6, $pop20
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push22=, $0, 7
+; SIMD128-VM-SLOW-NEXT:    i32.const $push57=, 5
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push23=, $pop22, $pop57
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push24=, $pop21, 7, $pop23
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push25=, $0, 8
+; SIMD128-VM-SLOW-NEXT:    i32.const $push56=, 5
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push26=, $pop25, $pop56
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push27=, $pop24, 8, $pop26
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push28=, $0, 9
+; SIMD128-VM-SLOW-NEXT:    i32.const $push55=, 5
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push29=, $pop28, $pop55
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push30=, $pop27, 9, $pop29
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push31=, $0, 10
+; SIMD128-VM-SLOW-NEXT:    i32.const $push54=, 5
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push32=, $pop31, $pop54
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push33=, $pop30, 10, $pop32
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push34=, $0, 11
+; SIMD128-VM-SLOW-NEXT:    i32.const $push53=, 5
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push35=, $pop34, $pop53
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push36=, $pop33, 11, $pop35
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push37=, $0, 12
+; SIMD128-VM-SLOW-NEXT:    i32.const $push52=, 5
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push38=, $pop37, $pop52
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push39=, $pop36, 12, $pop38
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push40=, $0, 13
+; SIMD128-VM-SLOW-NEXT:    i32.const $push51=, 5
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push41=, $pop40, $pop51
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push42=, $pop39, 13, $pop41
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push43=, $0, 14
+; SIMD128-VM-SLOW-NEXT:    i32.const $push50=, 5
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push44=, $pop43, $pop50
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push45=, $pop42, 14, $pop44
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push46=, $0, 15
+; SIMD128-VM-SLOW-NEXT:    i32.const $push49=, 5
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push47=, $pop46, $pop49
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push48=, $pop45, 15, $pop47
+; SIMD128-VM-SLOW-NEXT:    return $pop48
+;
+; SIMD128-VM-FAST-LABEL: shl_const_v16i8:
+; SIMD128-VM-FAST:         .functype shl_const_v16i8 (v128) -> (v128)
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push4=, $0, 0
+; SIMD128-VM-FAST-NEXT:    i32.const $push2=, 5
+; SIMD128-VM-FAST-NEXT:    i32.shl $push5=, $pop4, $pop2
+; SIMD128-VM-FAST-NEXT:    i8x16.splat $push6=, $pop5
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push1=, $0, 1
+; SIMD128-VM-FAST-NEXT:    i32.const $push63=, 5
+; SIMD128-VM-FAST-NEXT:    i32.shl $push3=, $pop1, $pop63
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push7=, $pop6, 1, $pop3
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push8=, $0, 2
+; SIMD128-VM-FAST-NEXT:    i32.const $push62=, 5
+; SIMD128-VM-FAST-NEXT:    i32.shl $push9=, $pop8, $pop62
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push10=, $pop7, 2, $pop9
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push11=, $0, 3
+; SIMD128-VM-FAST-NEXT:    i32.const $push61=, 5
+; SIMD128-VM-FAST-NEXT:    i32.shl $push12=, $pop11, $pop61
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push13=, $pop10, 3, $pop12
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push14=, $0, 4
+; SIMD128-VM-FAST-NEXT:    i32.const $push60=, 5
+; SIMD128-VM-FAST-NEXT:    i32.shl $push15=, $pop14, $pop60
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push16=, $pop13, 4, $pop15
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push17=, $0, 5
+; SIMD128-VM-FAST-NEXT:    i32.const $push59=, 5
+; SIMD128-VM-FAST-NEXT:    i32.shl $push18=, $pop17, $pop59
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push19=, $pop16, 5, $pop18
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push20=, $0, 6
+; SIMD128-VM-FAST-NEXT:    i32.const $push58=, 5
+; SIMD128-VM-FAST-NEXT:    i32.shl $push21=, $pop20, $pop58
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push22=, $pop19, 6, $pop21
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push23=, $0, 7
+; SIMD128-VM-FAST-NEXT:    i32.const $push57=, 5
+; SIMD128-VM-FAST-NEXT:    i32.shl $push24=, $pop23, $pop57
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push25=, $pop22, 7, $pop24
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push26=, $0, 8
+; SIMD128-VM-FAST-NEXT:    i32.const $push56=, 5
+; SIMD128-VM-FAST-NEXT:    i32.shl $push27=, $pop26, $pop56
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push28=, $pop25, 8, $pop27
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push29=, $0, 9
+; SIMD128-VM-FAST-NEXT:    i32.const $push55=, 5
+; SIMD128-VM-FAST-NEXT:    i32.shl $push30=, $pop29, $pop55
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push31=, $pop28, 9, $pop30
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push32=, $0, 10
+; SIMD128-VM-FAST-NEXT:    i32.const $push54=, 5
+; SIMD128-VM-FAST-NEXT:    i32.shl $push33=, $pop32, $pop54
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push34=, $pop31, 10, $pop33
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push35=, $0, 11
+; SIMD128-VM-FAST-NEXT:    i32.const $push53=, 5
+; SIMD128-VM-FAST-NEXT:    i32.shl $push36=, $pop35, $pop53
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push37=, $pop34, 11, $pop36
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push38=, $0, 12
+; SIMD128-VM-FAST-NEXT:    i32.const $push52=, 5
+; SIMD128-VM-FAST-NEXT:    i32.shl $push39=, $pop38, $pop52
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push40=, $pop37, 12, $pop39
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push41=, $0, 13
+; SIMD128-VM-FAST-NEXT:    i32.const $push51=, 5
+; SIMD128-VM-FAST-NEXT:    i32.shl $push42=, $pop41, $pop51
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push43=, $pop40, 13, $pop42
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push44=, $0, 14
+; SIMD128-VM-FAST-NEXT:    i32.const $push50=, 5
+; SIMD128-VM-FAST-NEXT:    i32.shl $push45=, $pop44, $pop50
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push46=, $pop43, 14, $pop45
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push47=, $0, 15
+; SIMD128-VM-FAST-NEXT:    i32.const $push49=, 5
+; SIMD128-VM-FAST-NEXT:    i32.shl $push48=, $pop47, $pop49
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push0=, $pop46, 15, $pop48
+; SIMD128-VM-FAST-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: shl_const_v16i8:
+; NO-SIMD128-SLOW:         .functype shl_const_v16i8 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push2=, 15
+; NO-SIMD128-SLOW-NEXT:    i32.add $push3=, $0, $pop2
+; NO-SIMD128-SLOW-NEXT:    i32.const $push0=, 5
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push1=, $16, $pop0
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop3), $pop1
+; NO-SIMD128-SLOW-NEXT:    i32.const $push5=, 14
+; NO-SIMD128-SLOW-NEXT:    i32.add $push6=, $0, $pop5
+; NO-SIMD128-SLOW-NEXT:    i32.const $push53=, 5
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push4=, $15, $pop53
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop6), $pop4
+; NO-SIMD128-SLOW-NEXT:    i32.const $push8=, 13
+; NO-SIMD128-SLOW-NEXT:    i32.add $push9=, $0, $pop8
+; NO-SIMD128-SLOW-NEXT:    i32.const $push52=, 5
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push7=, $14, $pop52
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop9), $pop7
+; NO-SIMD128-SLOW-NEXT:    i32.const $push11=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push12=, $0, $pop11
+; NO-SIMD128-SLOW-NEXT:    i32.const $push51=, 5
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push10=, $13, $pop51
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop12), $pop10
+; NO-SIMD128-SLOW-NEXT:    i32.const $push14=, 11
+; NO-SIMD128-SLOW-NEXT:    i32.add $push15=, $0, $pop14
+; NO-SIMD128-SLOW-NEXT:    i32.const $push50=, 5
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push13=, $12, $pop50
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop15), $pop13
+; NO-SIMD128-SLOW-NEXT:    i32.const $push17=, 10
+; NO-SIMD128-SLOW-NEXT:    i32.add $push18=, $0, $pop17
+; NO-SIMD128-SLOW-NEXT:    i32.const $push49=, 5
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push16=, $11, $pop49
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop18), $pop16
+; NO-SIMD128-SLOW-NEXT:    i32.const $push20=, 9
+; NO-SIMD128-SLOW-NEXT:    i32.add $push21=, $0, $pop20
+; NO-SIMD128-SLOW-NEXT:    i32.const $push48=, 5
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push19=, $10, $pop48
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop21), $pop19
+; NO-SIMD128-SLOW-NEXT:    i32.const $push47=, 5
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push22=, $9, $pop47
+; NO-SIMD128-SLOW-NEXT:    i32.store8 8($0), $pop22
+; NO-SIMD128-SLOW-NEXT:    i32.const $push24=, 7
+; NO-SIMD128-SLOW-NEXT:    i32.add $push25=, $0, $pop24
+; NO-SIMD128-SLOW-NEXT:    i32.const $push46=, 5
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push23=, $8, $pop46
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop25), $pop23
+; NO-SIMD128-SLOW-NEXT:    i32.const $push27=, 6
+; NO-SIMD128-SLOW-NEXT:    i32.add $push28=, $0, $pop27
+; NO-SIMD128-SLOW-NEXT:    i32.const $push45=, 5
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push26=, $7, $pop45
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop28), $pop26
+; NO-SIMD128-SLOW-NEXT:    i32.const $push44=, 5
+; NO-SIMD128-SLOW-NEXT:    i32.add $push30=, $0, $pop44
+; NO-SIMD128-SLOW-NEXT:    i32.const $push43=, 5
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push29=, $6, $pop43
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop30), $pop29
+; NO-SIMD128-SLOW-NEXT:    i32.const $push42=, 5
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push31=, $5, $pop42
+; NO-SIMD128-SLOW-NEXT:    i32.store8 4($0), $pop31
+; NO-SIMD128-SLOW-NEXT:    i32.const $push33=, 3
+; NO-SIMD128-SLOW-NEXT:    i32.add $push34=, $0, $pop33
+; NO-SIMD128-SLOW-NEXT:    i32.const $push41=, 5
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push32=, $4, $pop41
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop34), $pop32
+; NO-SIMD128-SLOW-NEXT:    i32.const $push40=, 5
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push35=, $3, $pop40
+; NO-SIMD128-SLOW-NEXT:    i32.store8 2($0), $pop35
+; NO-SIMD128-SLOW-NEXT:    i32.const $push39=, 5
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push36=, $2, $pop39
+; NO-SIMD128-SLOW-NEXT:    i32.store8 1($0), $pop36
+; NO-SIMD128-SLOW-NEXT:    i32.const $push38=, 5
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push37=, $1, $pop38
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($0), $pop37
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: shl_const_v16i8:
+; NO-SIMD128-FAST:         .functype shl_const_v16i8 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-FAST-NEXT:    i32.const $push0=, 5
+; NO-SIMD128-FAST-NEXT:    i32.shl $push1=, $1, $pop0
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($0), $pop1
+; NO-SIMD128-FAST-NEXT:    i32.const $push53=, 5
+; NO-SIMD128-FAST-NEXT:    i32.shl $push2=, $2, $pop53
+; NO-SIMD128-FAST-NEXT:    i32.store8 1($0), $pop2
+; NO-SIMD128-FAST-NEXT:    i32.const $push52=, 5
+; NO-SIMD128-FAST-NEXT:    i32.shl $push3=, $3, $pop52
+; NO-SIMD128-FAST-NEXT:    i32.store8 2($0), $pop3
+; NO-SIMD128-FAST-NEXT:    i32.const $push4=, 3
+; NO-SIMD128-FAST-NEXT:    i32.add $push5=, $0, $pop4
+; NO-SIMD128-FAST-NEXT:    i32.const $push51=, 5
+; NO-SIMD128-FAST-NEXT:    i32.shl $push6=, $4, $pop51
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop5), $pop6
+; NO-SIMD128-FAST-NEXT:    i32.const $push50=, 5
+; NO-SIMD128-FAST-NEXT:    i32.shl $push7=, $5, $pop50
+; NO-SIMD128-FAST-NEXT:    i32.store8 4($0), $pop7
+; NO-SIMD128-FAST-NEXT:    i32.const $push49=, 5
+; NO-SIMD128-FAST-NEXT:    i32.add $push8=, $0, $pop49
+; NO-SIMD128-FAST-NEXT:    i32.const $push48=, 5
+; NO-SIMD128-FAST-NEXT:    i32.shl $push9=, $6, $pop48
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop8), $pop9
+; NO-SIMD128-FAST-NEXT:    i32.const $push10=, 6
+; NO-SIMD128-FAST-NEXT:    i32.add $push11=, $0, $pop10
+; NO-SIMD128-FAST-NEXT:    i32.const $push47=, 5
+; NO-SIMD128-FAST-NEXT:    i32.shl $push12=, $7, $pop47
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop11), $pop12
+; NO-SIMD128-FAST-NEXT:    i32.const $push13=, 7
+; NO-SIMD128-FAST-NEXT:    i32.add $push14=, $0, $pop13
+; NO-SIMD128-FAST-NEXT:    i32.const $push46=, 5
+; NO-SIMD128-FAST-NEXT:    i32.shl $push15=, $8, $pop46
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop14), $pop15
+; NO-SIMD128-FAST-NEXT:    i32.const $push45=, 5
+; NO-SIMD128-FAST-NEXT:    i32.shl $push16=, $9, $pop45
+; NO-SIMD128-FAST-NEXT:    i32.store8 8($0), $pop16
+; NO-SIMD128-FAST-NEXT:    i32.const $push17=, 9
+; NO-SIMD128-FAST-NEXT:    i32.add $push18=, $0, $pop17
+; NO-SIMD128-FAST-NEXT:    i32.const $push44=, 5
+; NO-SIMD128-FAST-NEXT:    i32.shl $push19=, $10, $pop44
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop18), $pop19
+; NO-SIMD128-FAST-NEXT:    i32.const $push20=, 10
+; NO-SIMD128-FAST-NEXT:    i32.add $push21=, $0, $pop20
+; NO-SIMD128-FAST-NEXT:    i32.const $push43=, 5
+; NO-SIMD128-FAST-NEXT:    i32.shl $push22=, $11, $pop43
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop21), $pop22
+; NO-SIMD128-FAST-NEXT:    i32.const $push23=, 11
+; NO-SIMD128-FAST-NEXT:    i32.add $push24=, $0, $pop23
+; NO-SIMD128-FAST-NEXT:    i32.const $push42=, 5
+; NO-SIMD128-FAST-NEXT:    i32.shl $push25=, $12, $pop42
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop24), $pop25
+; NO-SIMD128-FAST-NEXT:    i32.const $push26=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push27=, $0, $pop26
+; NO-SIMD128-FAST-NEXT:    i32.const $push41=, 5
+; NO-SIMD128-FAST-NEXT:    i32.shl $push28=, $13, $pop41
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop27), $pop28
+; NO-SIMD128-FAST-NEXT:    i32.const $push29=, 13
+; NO-SIMD128-FAST-NEXT:    i32.add $push30=, $0, $pop29
+; NO-SIMD128-FAST-NEXT:    i32.const $push40=, 5
+; NO-SIMD128-FAST-NEXT:    i32.shl $push31=, $14, $pop40
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop30), $pop31
+; NO-SIMD128-FAST-NEXT:    i32.const $push32=, 14
+; NO-SIMD128-FAST-NEXT:    i32.add $push33=, $0, $pop32
+; NO-SIMD128-FAST-NEXT:    i32.const $push39=, 5
+; NO-SIMD128-FAST-NEXT:    i32.shl $push34=, $15, $pop39
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop33), $pop34
+; NO-SIMD128-FAST-NEXT:    i32.const $push35=, 15
+; NO-SIMD128-FAST-NEXT:    i32.add $push36=, $0, $pop35
+; NO-SIMD128-FAST-NEXT:    i32.const $push38=, 5
+; NO-SIMD128-FAST-NEXT:    i32.shl $push37=, $16, $pop38
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop36), $pop37
+; NO-SIMD128-FAST-NEXT:    return
   %a = shl <16 x i8> %v,
     <i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5,
      i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5>
   ret <16 x i8> %a
 }
 
-; CHECK-LABEL: shl_vec_v16i8:
-; NO-SIMD128-NOT: i8x16
-; SIMD128-NEXT: .functype shl_vec_v16i8 (v128, v128) -> (v128){{$}}
-; SIMD128-NEXT: i8x16.extract_lane_s $push[[L0:[0-9]+]]=, $0, 0{{$}}
-; SIMD128-NEXT: i32.const $push[[M0:[0-9]+]]=, 7{{$}}
-; SIMD128-NEXT: i8x16.splat $push[[M1:[0-9]+]]=, $pop[[M0]]{{$}}
-; SIMD128-NEXT: v128.and $push[[M2:[0-9]+]]=, $1, $pop[[M1]]{{$}}
-; SIMD128-NEXT: local.tee $push[[M:[0-9]+]]=, $1=, $pop[[M2]]{{$}}
-; SIMD128-NEXT: i8x16.extract_lane_u $push[[L1:[0-9]+]]=, $pop[[M]], 0{{$}}
-; SIMD128-NEXT: i32.shl $push[[L2:[0-9]+]]=, $pop[[L0]], $pop[[L1]]{{$}}
-; SIMD128-NEXT: i8x16.splat $push[[L3:[0-9]+]]=, $pop[[L2]]{{$}}
-; Skip 14 lanes
-; SIMD128:      i8x16.extract_lane_s $push[[L4:[0-9]+]]=, $0, 15{{$}}
-; SIMD128-NEXT: i8x16.extract_lane_u $push[[L5:[0-9]+]]=, $1, 15{{$}}
-; SIMD128-NEXT: i32.shl $push[[L6:[0-9]+]]=, $pop[[L4]], $pop[[L5]]{{$}}
-; SIMD128-NEXT: i8x16.replace_lane $push[[R:[0-9]+]]=, $pop[[L7:[0-9]+]], 15, $pop[[L6]]{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <16 x i8> @shl_vec_v16i8(<16 x i8> %v, <16 x i8> %x) {
+; SIMD128-UNIMPL-SLOW-LABEL: shl_vec_v16i8:
+; SIMD128-UNIMPL-SLOW:         .functype shl_vec_v16i8 (v128, v128) -> (v128)
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_s $push6=, $0, 0
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.const $push0=, 7
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.splat $push1=, $pop0
+; SIMD128-UNIMPL-SLOW-NEXT:    v128.and $push67=, $1, $pop1
+; SIMD128-UNIMPL-SLOW-NEXT:    local.tee $push66=, $1=, $pop67
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push5=, $pop66, 0
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shl $push7=, $pop6, $pop5
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.splat $push8=, $pop7
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_s $push3=, $0, 1
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push2=, $1, 1
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shl $push4=, $pop3, $pop2
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.replace_lane $push9=, $pop8, 1, $pop4
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_s $push11=, $0, 2
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push10=, $1, 2
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shl $push12=, $pop11, $pop10
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.replace_lane $push13=, $pop9, 2, $pop12
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_s $push15=, $0, 3
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push14=, $1, 3
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shl $push16=, $pop15, $pop14
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.replace_lane $push17=, $pop13, 3, $pop16
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_s $push19=, $0, 4
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push18=, $1, 4
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shl $push20=, $pop19, $pop18
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.replace_lane $push21=, $pop17, 4, $pop20
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_s $push23=, $0, 5
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push22=, $1, 5
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shl $push24=, $pop23, $pop22
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.replace_lane $push25=, $pop21, 5, $pop24
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_s $push27=, $0, 6
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push26=, $1, 6
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shl $push28=, $pop27, $pop26
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.replace_lane $push29=, $pop25, 6, $pop28
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_s $push31=, $0, 7
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push30=, $1, 7
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shl $push32=, $pop31, $pop30
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.replace_lane $push33=, $pop29, 7, $pop32
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_s $push35=, $0, 8
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push34=, $1, 8
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shl $push36=, $pop35, $pop34
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.replace_lane $push37=, $pop33, 8, $pop36
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_s $push39=, $0, 9
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push38=, $1, 9
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shl $push40=, $pop39, $pop38
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.replace_lane $push41=, $pop37, 9, $pop40
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_s $push43=, $0, 10
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push42=, $1, 10
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shl $push44=, $pop43, $pop42
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.replace_lane $push45=, $pop41, 10, $pop44
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_s $push47=, $0, 11
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push46=, $1, 11
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shl $push48=, $pop47, $pop46
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.replace_lane $push49=, $pop45, 11, $pop48
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_s $push51=, $0, 12
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push50=, $1, 12
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shl $push52=, $pop51, $pop50
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.replace_lane $push53=, $pop49, 12, $pop52
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_s $push55=, $0, 13
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push54=, $1, 13
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shl $push56=, $pop55, $pop54
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.replace_lane $push57=, $pop53, 13, $pop56
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_s $push59=, $0, 14
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push58=, $1, 14
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shl $push60=, $pop59, $pop58
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.replace_lane $push61=, $pop57, 14, $pop60
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_s $push63=, $0, 15
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push62=, $1, 15
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shl $push64=, $pop63, $pop62
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.replace_lane $push65=, $pop61, 15, $pop64
+; SIMD128-UNIMPL-SLOW-NEXT:    return $pop65
+;
+; SIMD128-UNIMPL-FAST-LABEL: shl_vec_v16i8:
+; SIMD128-UNIMPL-FAST:         .functype shl_vec_v16i8 (v128, v128) -> (v128)
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_s $push7=, $0, 0
+; SIMD128-UNIMPL-FAST-NEXT:    i32.const $push1=, 7
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.splat $push2=, $pop1
+; SIMD128-UNIMPL-FAST-NEXT:    v128.and $push67=, $1, $pop2
+; SIMD128-UNIMPL-FAST-NEXT:    local.tee $push66=, $1=, $pop67
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push6=, $pop66, 0
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shl $push8=, $pop7, $pop6
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.splat $push9=, $pop8
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_s $push4=, $0, 1
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push3=, $1, 1
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shl $push5=, $pop4, $pop3
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.replace_lane $push10=, $pop9, 1, $pop5
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_s $push12=, $0, 2
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push11=, $1, 2
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shl $push13=, $pop12, $pop11
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.replace_lane $push14=, $pop10, 2, $pop13
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_s $push16=, $0, 3
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push15=, $1, 3
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shl $push17=, $pop16, $pop15
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.replace_lane $push18=, $pop14, 3, $pop17
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_s $push20=, $0, 4
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push19=, $1, 4
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shl $push21=, $pop20, $pop19
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.replace_lane $push22=, $pop18, 4, $pop21
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_s $push24=, $0, 5
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push23=, $1, 5
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shl $push25=, $pop24, $pop23
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.replace_lane $push26=, $pop22, 5, $pop25
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_s $push28=, $0, 6
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push27=, $1, 6
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shl $push29=, $pop28, $pop27
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.replace_lane $push30=, $pop26, 6, $pop29
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_s $push32=, $0, 7
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push31=, $1, 7
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shl $push33=, $pop32, $pop31
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.replace_lane $push34=, $pop30, 7, $pop33
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_s $push36=, $0, 8
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push35=, $1, 8
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shl $push37=, $pop36, $pop35
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.replace_lane $push38=, $pop34, 8, $pop37
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_s $push40=, $0, 9
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push39=, $1, 9
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shl $push41=, $pop40, $pop39
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.replace_lane $push42=, $pop38, 9, $pop41
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_s $push44=, $0, 10
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push43=, $1, 10
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shl $push45=, $pop44, $pop43
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.replace_lane $push46=, $pop42, 10, $pop45
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_s $push48=, $0, 11
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push47=, $1, 11
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shl $push49=, $pop48, $pop47
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.replace_lane $push50=, $pop46, 11, $pop49
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_s $push52=, $0, 12
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push51=, $1, 12
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shl $push53=, $pop52, $pop51
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.replace_lane $push54=, $pop50, 12, $pop53
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_s $push56=, $0, 13
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push55=, $1, 13
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shl $push57=, $pop56, $pop55
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.replace_lane $push58=, $pop54, 13, $pop57
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_s $push60=, $0, 14
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push59=, $1, 14
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shl $push61=, $pop60, $pop59
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.replace_lane $push62=, $pop58, 14, $pop61
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_s $push64=, $0, 15
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push63=, $1, 15
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shl $push65=, $pop64, $pop63
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.replace_lane $push0=, $pop62, 15, $pop65
+; SIMD128-UNIMPL-FAST-NEXT:    return $pop0
+;
+; SIMD128-VM-SLOW-LABEL: shl_vec_v16i8:
+; SIMD128-VM-SLOW:         .functype shl_vec_v16i8 (v128, v128) -> (v128)
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push8=, $0, 0
+; SIMD128-VM-SLOW-NEXT:    i32.const $push0=, 7
+; SIMD128-VM-SLOW-NEXT:    i8x16.splat $push1=, $pop0
+; SIMD128-VM-SLOW-NEXT:    v128.and $push99=, $1, $pop1
+; SIMD128-VM-SLOW-NEXT:    local.tee $push98=, $1=, $pop99
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push6=, $pop98, 0
+; SIMD128-VM-SLOW-NEXT:    i32.const $push97=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push7=, $pop6, $pop97
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push9=, $pop8, $pop7
+; SIMD128-VM-SLOW-NEXT:    i8x16.splat $push10=, $pop9
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push4=, $0, 1
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push2=, $1, 1
+; SIMD128-VM-SLOW-NEXT:    i32.const $push96=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push3=, $pop2, $pop96
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push5=, $pop4, $pop3
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push11=, $pop10, 1, $pop5
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push14=, $0, 2
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push12=, $1, 2
+; SIMD128-VM-SLOW-NEXT:    i32.const $push95=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push13=, $pop12, $pop95
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push15=, $pop14, $pop13
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push16=, $pop11, 2, $pop15
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push19=, $0, 3
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push17=, $1, 3
+; SIMD128-VM-SLOW-NEXT:    i32.const $push94=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push18=, $pop17, $pop94
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push20=, $pop19, $pop18
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push21=, $pop16, 3, $pop20
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push24=, $0, 4
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push22=, $1, 4
+; SIMD128-VM-SLOW-NEXT:    i32.const $push93=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push23=, $pop22, $pop93
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push25=, $pop24, $pop23
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push26=, $pop21, 4, $pop25
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push29=, $0, 5
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push27=, $1, 5
+; SIMD128-VM-SLOW-NEXT:    i32.const $push92=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push28=, $pop27, $pop92
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push30=, $pop29, $pop28
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push31=, $pop26, 5, $pop30
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push34=, $0, 6
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push32=, $1, 6
+; SIMD128-VM-SLOW-NEXT:    i32.const $push91=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push33=, $pop32, $pop91
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push35=, $pop34, $pop33
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push36=, $pop31, 6, $pop35
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push39=, $0, 7
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push37=, $1, 7
+; SIMD128-VM-SLOW-NEXT:    i32.const $push90=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push38=, $pop37, $pop90
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push40=, $pop39, $pop38
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push41=, $pop36, 7, $pop40
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push44=, $0, 8
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push42=, $1, 8
+; SIMD128-VM-SLOW-NEXT:    i32.const $push89=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push43=, $pop42, $pop89
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push45=, $pop44, $pop43
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push46=, $pop41, 8, $pop45
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push49=, $0, 9
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push47=, $1, 9
+; SIMD128-VM-SLOW-NEXT:    i32.const $push88=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push48=, $pop47, $pop88
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push50=, $pop49, $pop48
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push51=, $pop46, 9, $pop50
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push54=, $0, 10
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push52=, $1, 10
+; SIMD128-VM-SLOW-NEXT:    i32.const $push87=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push53=, $pop52, $pop87
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push55=, $pop54, $pop53
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push56=, $pop51, 10, $pop55
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push59=, $0, 11
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push57=, $1, 11
+; SIMD128-VM-SLOW-NEXT:    i32.const $push86=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push58=, $pop57, $pop86
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push60=, $pop59, $pop58
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push61=, $pop56, 11, $pop60
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push64=, $0, 12
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push62=, $1, 12
+; SIMD128-VM-SLOW-NEXT:    i32.const $push85=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push63=, $pop62, $pop85
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push65=, $pop64, $pop63
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push66=, $pop61, 12, $pop65
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push69=, $0, 13
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push67=, $1, 13
+; SIMD128-VM-SLOW-NEXT:    i32.const $push84=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push68=, $pop67, $pop84
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push70=, $pop69, $pop68
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push71=, $pop66, 13, $pop70
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push74=, $0, 14
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push72=, $1, 14
+; SIMD128-VM-SLOW-NEXT:    i32.const $push83=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push73=, $pop72, $pop83
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push75=, $pop74, $pop73
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push76=, $pop71, 14, $pop75
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push79=, $0, 15
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push77=, $1, 15
+; SIMD128-VM-SLOW-NEXT:    i32.const $push82=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push78=, $pop77, $pop82
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push80=, $pop79, $pop78
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push81=, $pop76, 15, $pop80
+; SIMD128-VM-SLOW-NEXT:    return $pop81
+;
+; SIMD128-VM-FAST-LABEL: shl_vec_v16i8:
+; SIMD128-VM-FAST:         .functype shl_vec_v16i8 (v128, v128) -> (v128)
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push9=, $0, 0
+; SIMD128-VM-FAST-NEXT:    i32.const $push1=, 7
+; SIMD128-VM-FAST-NEXT:    i8x16.splat $push2=, $pop1
+; SIMD128-VM-FAST-NEXT:    v128.and $push99=, $1, $pop2
+; SIMD128-VM-FAST-NEXT:    local.tee $push98=, $1=, $pop99
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push7=, $pop98, 0
+; SIMD128-VM-FAST-NEXT:    i32.const $push97=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push8=, $pop7, $pop97
+; SIMD128-VM-FAST-NEXT:    i32.shl $push10=, $pop9, $pop8
+; SIMD128-VM-FAST-NEXT:    i8x16.splat $push11=, $pop10
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push5=, $0, 1
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push3=, $1, 1
+; SIMD128-VM-FAST-NEXT:    i32.const $push96=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push4=, $pop3, $pop96
+; SIMD128-VM-FAST-NEXT:    i32.shl $push6=, $pop5, $pop4
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push12=, $pop11, 1, $pop6
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push15=, $0, 2
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push13=, $1, 2
+; SIMD128-VM-FAST-NEXT:    i32.const $push95=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push14=, $pop13, $pop95
+; SIMD128-VM-FAST-NEXT:    i32.shl $push16=, $pop15, $pop14
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push17=, $pop12, 2, $pop16
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push20=, $0, 3
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push18=, $1, 3
+; SIMD128-VM-FAST-NEXT:    i32.const $push94=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push19=, $pop18, $pop94
+; SIMD128-VM-FAST-NEXT:    i32.shl $push21=, $pop20, $pop19
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push22=, $pop17, 3, $pop21
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push25=, $0, 4
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push23=, $1, 4
+; SIMD128-VM-FAST-NEXT:    i32.const $push93=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push24=, $pop23, $pop93
+; SIMD128-VM-FAST-NEXT:    i32.shl $push26=, $pop25, $pop24
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push27=, $pop22, 4, $pop26
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push30=, $0, 5
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push28=, $1, 5
+; SIMD128-VM-FAST-NEXT:    i32.const $push92=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push29=, $pop28, $pop92
+; SIMD128-VM-FAST-NEXT:    i32.shl $push31=, $pop30, $pop29
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push32=, $pop27, 5, $pop31
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push35=, $0, 6
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push33=, $1, 6
+; SIMD128-VM-FAST-NEXT:    i32.const $push91=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push34=, $pop33, $pop91
+; SIMD128-VM-FAST-NEXT:    i32.shl $push36=, $pop35, $pop34
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push37=, $pop32, 6, $pop36
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push40=, $0, 7
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push38=, $1, 7
+; SIMD128-VM-FAST-NEXT:    i32.const $push90=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push39=, $pop38, $pop90
+; SIMD128-VM-FAST-NEXT:    i32.shl $push41=, $pop40, $pop39
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push42=, $pop37, 7, $pop41
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push45=, $0, 8
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push43=, $1, 8
+; SIMD128-VM-FAST-NEXT:    i32.const $push89=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push44=, $pop43, $pop89
+; SIMD128-VM-FAST-NEXT:    i32.shl $push46=, $pop45, $pop44
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push47=, $pop42, 8, $pop46
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push50=, $0, 9
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push48=, $1, 9
+; SIMD128-VM-FAST-NEXT:    i32.const $push88=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push49=, $pop48, $pop88
+; SIMD128-VM-FAST-NEXT:    i32.shl $push51=, $pop50, $pop49
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push52=, $pop47, 9, $pop51
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push55=, $0, 10
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push53=, $1, 10
+; SIMD128-VM-FAST-NEXT:    i32.const $push87=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push54=, $pop53, $pop87
+; SIMD128-VM-FAST-NEXT:    i32.shl $push56=, $pop55, $pop54
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push57=, $pop52, 10, $pop56
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push60=, $0, 11
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push58=, $1, 11
+; SIMD128-VM-FAST-NEXT:    i32.const $push86=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push59=, $pop58, $pop86
+; SIMD128-VM-FAST-NEXT:    i32.shl $push61=, $pop60, $pop59
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push62=, $pop57, 11, $pop61
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push65=, $0, 12
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push63=, $1, 12
+; SIMD128-VM-FAST-NEXT:    i32.const $push85=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push64=, $pop63, $pop85
+; SIMD128-VM-FAST-NEXT:    i32.shl $push66=, $pop65, $pop64
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push67=, $pop62, 12, $pop66
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push70=, $0, 13
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push68=, $1, 13
+; SIMD128-VM-FAST-NEXT:    i32.const $push84=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push69=, $pop68, $pop84
+; SIMD128-VM-FAST-NEXT:    i32.shl $push71=, $pop70, $pop69
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push72=, $pop67, 13, $pop71
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push75=, $0, 14
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push73=, $1, 14
+; SIMD128-VM-FAST-NEXT:    i32.const $push83=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push74=, $pop73, $pop83
+; SIMD128-VM-FAST-NEXT:    i32.shl $push76=, $pop75, $pop74
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push77=, $pop72, 14, $pop76
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push80=, $0, 15
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push78=, $1, 15
+; SIMD128-VM-FAST-NEXT:    i32.const $push82=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push79=, $pop78, $pop82
+; SIMD128-VM-FAST-NEXT:    i32.shl $push81=, $pop80, $pop79
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push0=, $pop77, 15, $pop81
+; SIMD128-VM-FAST-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: shl_vec_v16i8:
+; NO-SIMD128-SLOW:         .functype shl_vec_v16i8 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push0=, 15
+; NO-SIMD128-SLOW-NEXT:    i32.add $push1=, $0, $pop0
+; NO-SIMD128-SLOW-NEXT:    i32.const $push2=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push3=, $32, $pop2
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push4=, $16, $pop3
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop1), $pop4
+; NO-SIMD128-SLOW-NEXT:    i32.const $push5=, 14
+; NO-SIMD128-SLOW-NEXT:    i32.add $push6=, $0, $pop5
+; NO-SIMD128-SLOW-NEXT:    i32.const $push69=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push7=, $31, $pop69
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push8=, $15, $pop7
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop6), $pop8
+; NO-SIMD128-SLOW-NEXT:    i32.const $push9=, 13
+; NO-SIMD128-SLOW-NEXT:    i32.add $push10=, $0, $pop9
+; NO-SIMD128-SLOW-NEXT:    i32.const $push68=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push11=, $30, $pop68
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push12=, $14, $pop11
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop10), $pop12
+; NO-SIMD128-SLOW-NEXT:    i32.const $push13=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push14=, $0, $pop13
+; NO-SIMD128-SLOW-NEXT:    i32.const $push67=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push15=, $29, $pop67
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push16=, $13, $pop15
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop14), $pop16
+; NO-SIMD128-SLOW-NEXT:    i32.const $push17=, 11
+; NO-SIMD128-SLOW-NEXT:    i32.add $push18=, $0, $pop17
+; NO-SIMD128-SLOW-NEXT:    i32.const $push66=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push19=, $28, $pop66
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push20=, $12, $pop19
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop18), $pop20
+; NO-SIMD128-SLOW-NEXT:    i32.const $push21=, 10
+; NO-SIMD128-SLOW-NEXT:    i32.add $push22=, $0, $pop21
+; NO-SIMD128-SLOW-NEXT:    i32.const $push65=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push23=, $27, $pop65
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push24=, $11, $pop23
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop22), $pop24
+; NO-SIMD128-SLOW-NEXT:    i32.const $push25=, 9
+; NO-SIMD128-SLOW-NEXT:    i32.add $push26=, $0, $pop25
+; NO-SIMD128-SLOW-NEXT:    i32.const $push64=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push27=, $26, $pop64
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push28=, $10, $pop27
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop26), $pop28
+; NO-SIMD128-SLOW-NEXT:    i32.const $push63=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push29=, $25, $pop63
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push30=, $9, $pop29
+; NO-SIMD128-SLOW-NEXT:    i32.store8 8($0), $pop30
+; NO-SIMD128-SLOW-NEXT:    i32.const $push31=, 7
+; NO-SIMD128-SLOW-NEXT:    i32.add $push32=, $0, $pop31
+; NO-SIMD128-SLOW-NEXT:    i32.const $push62=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push33=, $24, $pop62
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push34=, $8, $pop33
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop32), $pop34
+; NO-SIMD128-SLOW-NEXT:    i32.const $push35=, 6
+; NO-SIMD128-SLOW-NEXT:    i32.add $push36=, $0, $pop35
+; NO-SIMD128-SLOW-NEXT:    i32.const $push61=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push37=, $23, $pop61
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push38=, $7, $pop37
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop36), $pop38
+; NO-SIMD128-SLOW-NEXT:    i32.const $push39=, 5
+; NO-SIMD128-SLOW-NEXT:    i32.add $push40=, $0, $pop39
+; NO-SIMD128-SLOW-NEXT:    i32.const $push60=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push41=, $22, $pop60
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push42=, $6, $pop41
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop40), $pop42
+; NO-SIMD128-SLOW-NEXT:    i32.const $push59=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push43=, $21, $pop59
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push44=, $5, $pop43
+; NO-SIMD128-SLOW-NEXT:    i32.store8 4($0), $pop44
+; NO-SIMD128-SLOW-NEXT:    i32.const $push45=, 3
+; NO-SIMD128-SLOW-NEXT:    i32.add $push46=, $0, $pop45
+; NO-SIMD128-SLOW-NEXT:    i32.const $push58=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push47=, $20, $pop58
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push48=, $4, $pop47
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop46), $pop48
+; NO-SIMD128-SLOW-NEXT:    i32.const $push57=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push49=, $19, $pop57
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push50=, $3, $pop49
+; NO-SIMD128-SLOW-NEXT:    i32.store8 2($0), $pop50
+; NO-SIMD128-SLOW-NEXT:    i32.const $push56=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push51=, $18, $pop56
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push52=, $2, $pop51
+; NO-SIMD128-SLOW-NEXT:    i32.store8 1($0), $pop52
+; NO-SIMD128-SLOW-NEXT:    i32.const $push55=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push53=, $17, $pop55
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push54=, $1, $pop53
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($0), $pop54
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: shl_vec_v16i8:
+; NO-SIMD128-FAST:         .functype shl_vec_v16i8 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-FAST-NEXT:    i32.const $push0=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push1=, $17, $pop0
+; NO-SIMD128-FAST-NEXT:    i32.shl $push2=, $1, $pop1
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($0), $pop2
+; NO-SIMD128-FAST-NEXT:    i32.const $push69=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push3=, $18, $pop69
+; NO-SIMD128-FAST-NEXT:    i32.shl $push4=, $2, $pop3
+; NO-SIMD128-FAST-NEXT:    i32.store8 1($0), $pop4
+; NO-SIMD128-FAST-NEXT:    i32.const $push68=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push5=, $19, $pop68
+; NO-SIMD128-FAST-NEXT:    i32.shl $push6=, $3, $pop5
+; NO-SIMD128-FAST-NEXT:    i32.store8 2($0), $pop6
+; NO-SIMD128-FAST-NEXT:    i32.const $push7=, 3
+; NO-SIMD128-FAST-NEXT:    i32.add $push8=, $0, $pop7
+; NO-SIMD128-FAST-NEXT:    i32.const $push67=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push9=, $20, $pop67
+; NO-SIMD128-FAST-NEXT:    i32.shl $push10=, $4, $pop9
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop8), $pop10
+; NO-SIMD128-FAST-NEXT:    i32.const $push66=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push11=, $21, $pop66
+; NO-SIMD128-FAST-NEXT:    i32.shl $push12=, $5, $pop11
+; NO-SIMD128-FAST-NEXT:    i32.store8 4($0), $pop12
+; NO-SIMD128-FAST-NEXT:    i32.const $push13=, 5
+; NO-SIMD128-FAST-NEXT:    i32.add $push14=, $0, $pop13
+; NO-SIMD128-FAST-NEXT:    i32.const $push65=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push15=, $22, $pop65
+; NO-SIMD128-FAST-NEXT:    i32.shl $push16=, $6, $pop15
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop14), $pop16
+; NO-SIMD128-FAST-NEXT:    i32.const $push17=, 6
+; NO-SIMD128-FAST-NEXT:    i32.add $push18=, $0, $pop17
+; NO-SIMD128-FAST-NEXT:    i32.const $push64=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push19=, $23, $pop64
+; NO-SIMD128-FAST-NEXT:    i32.shl $push20=, $7, $pop19
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop18), $pop20
+; NO-SIMD128-FAST-NEXT:    i32.const $push21=, 7
+; NO-SIMD128-FAST-NEXT:    i32.add $push22=, $0, $pop21
+; NO-SIMD128-FAST-NEXT:    i32.const $push63=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push23=, $24, $pop63
+; NO-SIMD128-FAST-NEXT:    i32.shl $push24=, $8, $pop23
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop22), $pop24
+; NO-SIMD128-FAST-NEXT:    i32.const $push62=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push25=, $25, $pop62
+; NO-SIMD128-FAST-NEXT:    i32.shl $push26=, $9, $pop25
+; NO-SIMD128-FAST-NEXT:    i32.store8 8($0), $pop26
+; NO-SIMD128-FAST-NEXT:    i32.const $push27=, 9
+; NO-SIMD128-FAST-NEXT:    i32.add $push28=, $0, $pop27
+; NO-SIMD128-FAST-NEXT:    i32.const $push61=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push29=, $26, $pop61
+; NO-SIMD128-FAST-NEXT:    i32.shl $push30=, $10, $pop29
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop28), $pop30
+; NO-SIMD128-FAST-NEXT:    i32.const $push31=, 10
+; NO-SIMD128-FAST-NEXT:    i32.add $push32=, $0, $pop31
+; NO-SIMD128-FAST-NEXT:    i32.const $push60=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push33=, $27, $pop60
+; NO-SIMD128-FAST-NEXT:    i32.shl $push34=, $11, $pop33
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop32), $pop34
+; NO-SIMD128-FAST-NEXT:    i32.const $push35=, 11
+; NO-SIMD128-FAST-NEXT:    i32.add $push36=, $0, $pop35
+; NO-SIMD128-FAST-NEXT:    i32.const $push59=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push37=, $28, $pop59
+; NO-SIMD128-FAST-NEXT:    i32.shl $push38=, $12, $pop37
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop36), $pop38
+; NO-SIMD128-FAST-NEXT:    i32.const $push39=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push40=, $0, $pop39
+; NO-SIMD128-FAST-NEXT:    i32.const $push58=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push41=, $29, $pop58
+; NO-SIMD128-FAST-NEXT:    i32.shl $push42=, $13, $pop41
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop40), $pop42
+; NO-SIMD128-FAST-NEXT:    i32.const $push43=, 13
+; NO-SIMD128-FAST-NEXT:    i32.add $push44=, $0, $pop43
+; NO-SIMD128-FAST-NEXT:    i32.const $push57=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push45=, $30, $pop57
+; NO-SIMD128-FAST-NEXT:    i32.shl $push46=, $14, $pop45
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop44), $pop46
+; NO-SIMD128-FAST-NEXT:    i32.const $push47=, 14
+; NO-SIMD128-FAST-NEXT:    i32.add $push48=, $0, $pop47
+; NO-SIMD128-FAST-NEXT:    i32.const $push56=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push49=, $31, $pop56
+; NO-SIMD128-FAST-NEXT:    i32.shl $push50=, $15, $pop49
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop48), $pop50
+; NO-SIMD128-FAST-NEXT:    i32.const $push51=, 15
+; NO-SIMD128-FAST-NEXT:    i32.add $push52=, $0, $pop51
+; NO-SIMD128-FAST-NEXT:    i32.const $push55=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push53=, $32, $pop55
+; NO-SIMD128-FAST-NEXT:    i32.shl $push54=, $16, $pop53
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop52), $pop54
+; NO-SIMD128-FAST-NEXT:    return
   %a = shl <16 x i8> %v, %x
   ret <16 x i8> %a
 }
 
-; CHECK-LABEL: shr_s_v16i8:
-; NO-SIMD128-NOT: i8x16
-; SIMD128-NEXT: .functype shr_s_v16i8 (v128, i32) -> (v128){{$}}
-; SIMD128-NEXT: i8x16.shr_s $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <16 x i8> @shr_s_v16i8(<16 x i8> %v, i8 %x) {
+; SIMD128-UNIMPL-LABEL: shr_s_v16i8:
+; SIMD128-UNIMPL:         .functype shr_s_v16i8 (v128, i32) -> (v128)
+; SIMD128-UNIMPL-NEXT:    i8x16.shr_s $push0=, $0, $1
+; SIMD128-UNIMPL-NEXT:    return $pop0
+;
+; SIMD128-VM-SLOW-LABEL: shr_s_v16i8:
+; SIMD128-VM-SLOW:         .functype shr_s_v16i8 (v128, i32) -> (v128)
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push9=, $0, 0
+; SIMD128-VM-SLOW-NEXT:    i8x16.splat $push2=, $1
+; SIMD128-VM-SLOW-NEXT:    i32.const $push0=, 7
+; SIMD128-VM-SLOW-NEXT:    i8x16.splat $push1=, $pop0
+; SIMD128-VM-SLOW-NEXT:    v128.and $push100=, $pop2, $pop1
+; SIMD128-VM-SLOW-NEXT:    local.tee $push99=, $2=, $pop100
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push7=, $pop99, 0
+; SIMD128-VM-SLOW-NEXT:    i32.const $push98=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push8=, $pop7, $pop98
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push10=, $pop9, $pop8
+; SIMD128-VM-SLOW-NEXT:    i8x16.splat $push11=, $pop10
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push5=, $0, 1
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push3=, $2, 1
+; SIMD128-VM-SLOW-NEXT:    i32.const $push97=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push4=, $pop3, $pop97
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push6=, $pop5, $pop4
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push12=, $pop11, 1, $pop6
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push15=, $0, 2
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push13=, $2, 2
+; SIMD128-VM-SLOW-NEXT:    i32.const $push96=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push14=, $pop13, $pop96
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push16=, $pop15, $pop14
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push17=, $pop12, 2, $pop16
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push20=, $0, 3
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push18=, $2, 3
+; SIMD128-VM-SLOW-NEXT:    i32.const $push95=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push19=, $pop18, $pop95
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push21=, $pop20, $pop19
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push22=, $pop17, 3, $pop21
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push25=, $0, 4
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push23=, $2, 4
+; SIMD128-VM-SLOW-NEXT:    i32.const $push94=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push24=, $pop23, $pop94
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push26=, $pop25, $pop24
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push27=, $pop22, 4, $pop26
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push30=, $0, 5
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push28=, $2, 5
+; SIMD128-VM-SLOW-NEXT:    i32.const $push93=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push29=, $pop28, $pop93
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push31=, $pop30, $pop29
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push32=, $pop27, 5, $pop31
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push35=, $0, 6
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push33=, $2, 6
+; SIMD128-VM-SLOW-NEXT:    i32.const $push92=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push34=, $pop33, $pop92
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push36=, $pop35, $pop34
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push37=, $pop32, 6, $pop36
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push40=, $0, 7
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push38=, $2, 7
+; SIMD128-VM-SLOW-NEXT:    i32.const $push91=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push39=, $pop38, $pop91
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push41=, $pop40, $pop39
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push42=, $pop37, 7, $pop41
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push45=, $0, 8
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push43=, $2, 8
+; SIMD128-VM-SLOW-NEXT:    i32.const $push90=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push44=, $pop43, $pop90
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push46=, $pop45, $pop44
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push47=, $pop42, 8, $pop46
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push50=, $0, 9
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push48=, $2, 9
+; SIMD128-VM-SLOW-NEXT:    i32.const $push89=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push49=, $pop48, $pop89
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push51=, $pop50, $pop49
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push52=, $pop47, 9, $pop51
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push55=, $0, 10
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push53=, $2, 10
+; SIMD128-VM-SLOW-NEXT:    i32.const $push88=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push54=, $pop53, $pop88
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push56=, $pop55, $pop54
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push57=, $pop52, 10, $pop56
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push60=, $0, 11
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push58=, $2, 11
+; SIMD128-VM-SLOW-NEXT:    i32.const $push87=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push59=, $pop58, $pop87
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push61=, $pop60, $pop59
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push62=, $pop57, 11, $pop61
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push65=, $0, 12
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push63=, $2, 12
+; SIMD128-VM-SLOW-NEXT:    i32.const $push86=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push64=, $pop63, $pop86
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push66=, $pop65, $pop64
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push67=, $pop62, 12, $pop66
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push70=, $0, 13
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push68=, $2, 13
+; SIMD128-VM-SLOW-NEXT:    i32.const $push85=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push69=, $pop68, $pop85
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push71=, $pop70, $pop69
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push72=, $pop67, 13, $pop71
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push75=, $0, 14
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push73=, $2, 14
+; SIMD128-VM-SLOW-NEXT:    i32.const $push84=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push74=, $pop73, $pop84
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push76=, $pop75, $pop74
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push77=, $pop72, 14, $pop76
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push80=, $0, 15
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push78=, $2, 15
+; SIMD128-VM-SLOW-NEXT:    i32.const $push83=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push79=, $pop78, $pop83
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push81=, $pop80, $pop79
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push82=, $pop77, 15, $pop81
+; SIMD128-VM-SLOW-NEXT:    return $pop82
+;
+; SIMD128-VM-FAST-LABEL: shr_s_v16i8:
+; SIMD128-VM-FAST:         .functype shr_s_v16i8 (v128, i32) -> (v128)
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push10=, $0, 0
+; SIMD128-VM-FAST-NEXT:    i8x16.splat $push1=, $1
+; SIMD128-VM-FAST-NEXT:    i32.const $push2=, 7
+; SIMD128-VM-FAST-NEXT:    i8x16.splat $push3=, $pop2
+; SIMD128-VM-FAST-NEXT:    v128.and $push100=, $pop1, $pop3
+; SIMD128-VM-FAST-NEXT:    local.tee $push99=, $2=, $pop100
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push8=, $pop99, 0
+; SIMD128-VM-FAST-NEXT:    i32.const $push98=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push9=, $pop8, $pop98
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push11=, $pop10, $pop9
+; SIMD128-VM-FAST-NEXT:    i8x16.splat $push12=, $pop11
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push6=, $0, 1
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push4=, $2, 1
+; SIMD128-VM-FAST-NEXT:    i32.const $push97=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push5=, $pop4, $pop97
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push7=, $pop6, $pop5
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push13=, $pop12, 1, $pop7
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push16=, $0, 2
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push14=, $2, 2
+; SIMD128-VM-FAST-NEXT:    i32.const $push96=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push15=, $pop14, $pop96
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push17=, $pop16, $pop15
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push18=, $pop13, 2, $pop17
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push21=, $0, 3
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push19=, $2, 3
+; SIMD128-VM-FAST-NEXT:    i32.const $push95=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push20=, $pop19, $pop95
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push22=, $pop21, $pop20
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push23=, $pop18, 3, $pop22
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push26=, $0, 4
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push24=, $2, 4
+; SIMD128-VM-FAST-NEXT:    i32.const $push94=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push25=, $pop24, $pop94
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push27=, $pop26, $pop25
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push28=, $pop23, 4, $pop27
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push31=, $0, 5
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push29=, $2, 5
+; SIMD128-VM-FAST-NEXT:    i32.const $push93=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push30=, $pop29, $pop93
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push32=, $pop31, $pop30
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push33=, $pop28, 5, $pop32
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push36=, $0, 6
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push34=, $2, 6
+; SIMD128-VM-FAST-NEXT:    i32.const $push92=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push35=, $pop34, $pop92
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push37=, $pop36, $pop35
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push38=, $pop33, 6, $pop37
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push41=, $0, 7
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push39=, $2, 7
+; SIMD128-VM-FAST-NEXT:    i32.const $push91=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push40=, $pop39, $pop91
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push42=, $pop41, $pop40
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push43=, $pop38, 7, $pop42
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push46=, $0, 8
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push44=, $2, 8
+; SIMD128-VM-FAST-NEXT:    i32.const $push90=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push45=, $pop44, $pop90
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push47=, $pop46, $pop45
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push48=, $pop43, 8, $pop47
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push51=, $0, 9
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push49=, $2, 9
+; SIMD128-VM-FAST-NEXT:    i32.const $push89=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push50=, $pop49, $pop89
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push52=, $pop51, $pop50
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push53=, $pop48, 9, $pop52
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push56=, $0, 10
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push54=, $2, 10
+; SIMD128-VM-FAST-NEXT:    i32.const $push88=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push55=, $pop54, $pop88
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push57=, $pop56, $pop55
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push58=, $pop53, 10, $pop57
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push61=, $0, 11
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push59=, $2, 11
+; SIMD128-VM-FAST-NEXT:    i32.const $push87=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push60=, $pop59, $pop87
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push62=, $pop61, $pop60
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push63=, $pop58, 11, $pop62
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push66=, $0, 12
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push64=, $2, 12
+; SIMD128-VM-FAST-NEXT:    i32.const $push86=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push65=, $pop64, $pop86
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push67=, $pop66, $pop65
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push68=, $pop63, 12, $pop67
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push71=, $0, 13
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push69=, $2, 13
+; SIMD128-VM-FAST-NEXT:    i32.const $push85=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push70=, $pop69, $pop85
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push72=, $pop71, $pop70
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push73=, $pop68, 13, $pop72
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push76=, $0, 14
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push74=, $2, 14
+; SIMD128-VM-FAST-NEXT:    i32.const $push84=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push75=, $pop74, $pop84
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push77=, $pop76, $pop75
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push78=, $pop73, 14, $pop77
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push81=, $0, 15
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push79=, $2, 15
+; SIMD128-VM-FAST-NEXT:    i32.const $push83=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push80=, $pop79, $pop83
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push82=, $pop81, $pop80
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push0=, $pop78, 15, $pop82
+; SIMD128-VM-FAST-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: shr_s_v16i8:
+; NO-SIMD128-SLOW:         .functype shr_s_v16i8 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push5=, 15
+; NO-SIMD128-SLOW-NEXT:    i32.add $push6=, $0, $pop5
+; NO-SIMD128-SLOW-NEXT:    i32.const $push1=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push2=, $16, $pop1
+; NO-SIMD128-SLOW-NEXT:    i32.const $push104=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push3=, $pop2, $pop104
+; NO-SIMD128-SLOW-NEXT:    i32.const $push0=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push103=, $17, $pop0
+; NO-SIMD128-SLOW-NEXT:    local.tee $push102=, $16=, $pop103
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push4=, $pop3, $pop102
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop6), $pop4
+; NO-SIMD128-SLOW-NEXT:    i32.const $push10=, 14
+; NO-SIMD128-SLOW-NEXT:    i32.add $push11=, $0, $pop10
+; NO-SIMD128-SLOW-NEXT:    i32.const $push101=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push7=, $15, $pop101
+; NO-SIMD128-SLOW-NEXT:    i32.const $push100=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push8=, $pop7, $pop100
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push9=, $pop8, $16
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop11), $pop9
+; NO-SIMD128-SLOW-NEXT:    i32.const $push15=, 13
+; NO-SIMD128-SLOW-NEXT:    i32.add $push16=, $0, $pop15
+; NO-SIMD128-SLOW-NEXT:    i32.const $push99=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push12=, $14, $pop99
+; NO-SIMD128-SLOW-NEXT:    i32.const $push98=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push13=, $pop12, $pop98
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push14=, $pop13, $16
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop16), $pop14
+; NO-SIMD128-SLOW-NEXT:    i32.const $push20=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push21=, $0, $pop20
+; NO-SIMD128-SLOW-NEXT:    i32.const $push97=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push17=, $13, $pop97
+; NO-SIMD128-SLOW-NEXT:    i32.const $push96=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push18=, $pop17, $pop96
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push19=, $pop18, $16
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop21), $pop19
+; NO-SIMD128-SLOW-NEXT:    i32.const $push25=, 11
+; NO-SIMD128-SLOW-NEXT:    i32.add $push26=, $0, $pop25
+; NO-SIMD128-SLOW-NEXT:    i32.const $push95=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push22=, $12, $pop95
+; NO-SIMD128-SLOW-NEXT:    i32.const $push94=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push23=, $pop22, $pop94
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push24=, $pop23, $16
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop26), $pop24
+; NO-SIMD128-SLOW-NEXT:    i32.const $push30=, 10
+; NO-SIMD128-SLOW-NEXT:    i32.add $push31=, $0, $pop30
+; NO-SIMD128-SLOW-NEXT:    i32.const $push93=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push27=, $11, $pop93
+; NO-SIMD128-SLOW-NEXT:    i32.const $push92=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push28=, $pop27, $pop92
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push29=, $pop28, $16
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop31), $pop29
+; NO-SIMD128-SLOW-NEXT:    i32.const $push35=, 9
+; NO-SIMD128-SLOW-NEXT:    i32.add $push36=, $0, $pop35
+; NO-SIMD128-SLOW-NEXT:    i32.const $push91=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push32=, $10, $pop91
+; NO-SIMD128-SLOW-NEXT:    i32.const $push90=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push33=, $pop32, $pop90
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push34=, $pop33, $16
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop36), $pop34
+; NO-SIMD128-SLOW-NEXT:    i32.const $push89=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push37=, $9, $pop89
+; NO-SIMD128-SLOW-NEXT:    i32.const $push88=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push38=, $pop37, $pop88
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push39=, $pop38, $16
+; NO-SIMD128-SLOW-NEXT:    i32.store8 8($0), $pop39
+; NO-SIMD128-SLOW-NEXT:    i32.const $push43=, 7
+; NO-SIMD128-SLOW-NEXT:    i32.add $push44=, $0, $pop43
+; NO-SIMD128-SLOW-NEXT:    i32.const $push87=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push40=, $8, $pop87
+; NO-SIMD128-SLOW-NEXT:    i32.const $push86=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push41=, $pop40, $pop86
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push42=, $pop41, $16
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop44), $pop42
+; NO-SIMD128-SLOW-NEXT:    i32.const $push48=, 6
+; NO-SIMD128-SLOW-NEXT:    i32.add $push49=, $0, $pop48
+; NO-SIMD128-SLOW-NEXT:    i32.const $push85=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push45=, $7, $pop85
+; NO-SIMD128-SLOW-NEXT:    i32.const $push84=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push46=, $pop45, $pop84
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push47=, $pop46, $16
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop49), $pop47
+; NO-SIMD128-SLOW-NEXT:    i32.const $push53=, 5
+; NO-SIMD128-SLOW-NEXT:    i32.add $push54=, $0, $pop53
+; NO-SIMD128-SLOW-NEXT:    i32.const $push83=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push50=, $6, $pop83
+; NO-SIMD128-SLOW-NEXT:    i32.const $push82=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push51=, $pop50, $pop82
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push52=, $pop51, $16
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop54), $pop52
+; NO-SIMD128-SLOW-NEXT:    i32.const $push81=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push55=, $5, $pop81
+; NO-SIMD128-SLOW-NEXT:    i32.const $push80=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push56=, $pop55, $pop80
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push57=, $pop56, $16
+; NO-SIMD128-SLOW-NEXT:    i32.store8 4($0), $pop57
+; NO-SIMD128-SLOW-NEXT:    i32.const $push61=, 3
+; NO-SIMD128-SLOW-NEXT:    i32.add $push62=, $0, $pop61
+; NO-SIMD128-SLOW-NEXT:    i32.const $push79=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push58=, $4, $pop79
+; NO-SIMD128-SLOW-NEXT:    i32.const $push78=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push59=, $pop58, $pop78
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push60=, $pop59, $16
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop62), $pop60
+; NO-SIMD128-SLOW-NEXT:    i32.const $push77=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push63=, $3, $pop77
+; NO-SIMD128-SLOW-NEXT:    i32.const $push76=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push64=, $pop63, $pop76
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push65=, $pop64, $16
+; NO-SIMD128-SLOW-NEXT:    i32.store8 2($0), $pop65
+; NO-SIMD128-SLOW-NEXT:    i32.const $push75=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push66=, $2, $pop75
+; NO-SIMD128-SLOW-NEXT:    i32.const $push74=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push67=, $pop66, $pop74
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push68=, $pop67, $16
+; NO-SIMD128-SLOW-NEXT:    i32.store8 1($0), $pop68
+; NO-SIMD128-SLOW-NEXT:    i32.const $push73=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push69=, $1, $pop73
+; NO-SIMD128-SLOW-NEXT:    i32.const $push72=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push70=, $pop69, $pop72
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push71=, $pop70, $16
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($0), $pop71
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: shr_s_v16i8:
+; NO-SIMD128-FAST:         .functype shr_s_v16i8 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-FAST-NEXT:    i32.const $push1=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shl $push2=, $1, $pop1
+; NO-SIMD128-FAST-NEXT:    i32.const $push104=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push3=, $pop2, $pop104
+; NO-SIMD128-FAST-NEXT:    i32.const $push0=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push103=, $17, $pop0
+; NO-SIMD128-FAST-NEXT:    local.tee $push102=, $17=, $pop103
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push4=, $pop3, $pop102
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($0), $pop4
+; NO-SIMD128-FAST-NEXT:    i32.const $push101=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shl $push5=, $2, $pop101
+; NO-SIMD128-FAST-NEXT:    i32.const $push100=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push6=, $pop5, $pop100
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push7=, $pop6, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 1($0), $pop7
+; NO-SIMD128-FAST-NEXT:    i32.const $push99=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shl $push8=, $3, $pop99
+; NO-SIMD128-FAST-NEXT:    i32.const $push98=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push9=, $pop8, $pop98
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push10=, $pop9, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 2($0), $pop10
+; NO-SIMD128-FAST-NEXT:    i32.const $push14=, 3
+; NO-SIMD128-FAST-NEXT:    i32.add $push15=, $0, $pop14
+; NO-SIMD128-FAST-NEXT:    i32.const $push97=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shl $push11=, $4, $pop97
+; NO-SIMD128-FAST-NEXT:    i32.const $push96=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push12=, $pop11, $pop96
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push13=, $pop12, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop15), $pop13
+; NO-SIMD128-FAST-NEXT:    i32.const $push95=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shl $push16=, $5, $pop95
+; NO-SIMD128-FAST-NEXT:    i32.const $push94=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push17=, $pop16, $pop94
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push18=, $pop17, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 4($0), $pop18
+; NO-SIMD128-FAST-NEXT:    i32.const $push22=, 5
+; NO-SIMD128-FAST-NEXT:    i32.add $push23=, $0, $pop22
+; NO-SIMD128-FAST-NEXT:    i32.const $push93=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shl $push19=, $6, $pop93
+; NO-SIMD128-FAST-NEXT:    i32.const $push92=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push20=, $pop19, $pop92
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push21=, $pop20, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop23), $pop21
+; NO-SIMD128-FAST-NEXT:    i32.const $push27=, 6
+; NO-SIMD128-FAST-NEXT:    i32.add $push28=, $0, $pop27
+; NO-SIMD128-FAST-NEXT:    i32.const $push91=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shl $push24=, $7, $pop91
+; NO-SIMD128-FAST-NEXT:    i32.const $push90=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push25=, $pop24, $pop90
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push26=, $pop25, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop28), $pop26
+; NO-SIMD128-FAST-NEXT:    i32.const $push32=, 7
+; NO-SIMD128-FAST-NEXT:    i32.add $push33=, $0, $pop32
+; NO-SIMD128-FAST-NEXT:    i32.const $push89=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shl $push29=, $8, $pop89
+; NO-SIMD128-FAST-NEXT:    i32.const $push88=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push30=, $pop29, $pop88
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push31=, $pop30, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop33), $pop31
+; NO-SIMD128-FAST-NEXT:    i32.const $push87=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shl $push34=, $9, $pop87
+; NO-SIMD128-FAST-NEXT:    i32.const $push86=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push35=, $pop34, $pop86
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push36=, $pop35, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 8($0), $pop36
+; NO-SIMD128-FAST-NEXT:    i32.const $push40=, 9
+; NO-SIMD128-FAST-NEXT:    i32.add $push41=, $0, $pop40
+; NO-SIMD128-FAST-NEXT:    i32.const $push85=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shl $push37=, $10, $pop85
+; NO-SIMD128-FAST-NEXT:    i32.const $push84=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push38=, $pop37, $pop84
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push39=, $pop38, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop41), $pop39
+; NO-SIMD128-FAST-NEXT:    i32.const $push45=, 10
+; NO-SIMD128-FAST-NEXT:    i32.add $push46=, $0, $pop45
+; NO-SIMD128-FAST-NEXT:    i32.const $push83=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shl $push42=, $11, $pop83
+; NO-SIMD128-FAST-NEXT:    i32.const $push82=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push43=, $pop42, $pop82
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push44=, $pop43, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop46), $pop44
+; NO-SIMD128-FAST-NEXT:    i32.const $push50=, 11
+; NO-SIMD128-FAST-NEXT:    i32.add $push51=, $0, $pop50
+; NO-SIMD128-FAST-NEXT:    i32.const $push81=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shl $push47=, $12, $pop81
+; NO-SIMD128-FAST-NEXT:    i32.const $push80=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push48=, $pop47, $pop80
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push49=, $pop48, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop51), $pop49
+; NO-SIMD128-FAST-NEXT:    i32.const $push55=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push56=, $0, $pop55
+; NO-SIMD128-FAST-NEXT:    i32.const $push79=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shl $push52=, $13, $pop79
+; NO-SIMD128-FAST-NEXT:    i32.const $push78=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push53=, $pop52, $pop78
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push54=, $pop53, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop56), $pop54
+; NO-SIMD128-FAST-NEXT:    i32.const $push60=, 13
+; NO-SIMD128-FAST-NEXT:    i32.add $push61=, $0, $pop60
+; NO-SIMD128-FAST-NEXT:    i32.const $push77=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shl $push57=, $14, $pop77
+; NO-SIMD128-FAST-NEXT:    i32.const $push76=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push58=, $pop57, $pop76
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push59=, $pop58, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop61), $pop59
+; NO-SIMD128-FAST-NEXT:    i32.const $push65=, 14
+; NO-SIMD128-FAST-NEXT:    i32.add $push66=, $0, $pop65
+; NO-SIMD128-FAST-NEXT:    i32.const $push75=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shl $push62=, $15, $pop75
+; NO-SIMD128-FAST-NEXT:    i32.const $push74=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push63=, $pop62, $pop74
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push64=, $pop63, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop66), $pop64
+; NO-SIMD128-FAST-NEXT:    i32.const $push70=, 15
+; NO-SIMD128-FAST-NEXT:    i32.add $push71=, $0, $pop70
+; NO-SIMD128-FAST-NEXT:    i32.const $push73=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shl $push67=, $16, $pop73
+; NO-SIMD128-FAST-NEXT:    i32.const $push72=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push68=, $pop67, $pop72
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push69=, $pop68, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop71), $pop69
+; NO-SIMD128-FAST-NEXT:    return
   %t = insertelement <16 x i8> undef, i8 %x, i32 0
   %s = shufflevector <16 x i8> %t, <16 x i8> undef,
     <16 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0,
                 i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
   %a = ashr <16 x i8> %v, %s
   ret <16 x i8> %a
 }
 
-; CHECK-LABEL: shr_s_vec_v16i8:
-; NO-SIMD128-NOT: i8x16
-; SIMD128-NEXT: .functype shr_s_vec_v16i8 (v128, v128) -> (v128){{$}}
-; SIMD128-NEXT: i8x16.extract_lane_s $push[[L0:[0-9]+]]=, $0, 0{{$}}
-; SIMD128-NEXT: i32.const $push[[M0:[0-9]+]]=, 7{{$}}
-; SIMD128-NEXT: i8x16.splat $push[[M1:[0-9]+]]=, $pop[[M0]]{{$}}
-; SIMD128-NEXT: v128.and $push[[M2:[0-9]+]]=, $1, $pop[[M1]]{{$}}
-; SIMD128-NEXT: local.tee $push[[M:[0-9]+]]=, $1=, $pop[[M2]]{{$}}
-; SIMD128-NEXT: i8x16.extract_lane_u $push[[L1:[0-9]+]]=, $pop[[M]], 0{{$}}
-; SIMD128-NEXT: i32.shr_s $push[[L2:[0-9]+]]=, $pop[[L0]], $pop[[L1]]{{$}}
-; SIMD128-NEXT: i8x16.splat $push[[L3:[0-9]+]]=, $pop[[L2]]{{$}}
-; Skip 14 lanes
-; SIMD128:      i8x16.extract_lane_s $push[[L0:[0-9]+]]=, $0, 15{{$}}
-; SIMD128-NEXT: i8x16.extract_lane_u $push[[L1:[0-9]+]]=, $1, 15{{$}}
-; SIMD128-NEXT: i32.shr_s $push[[L2:[0-9]+]]=, $pop[[L0]], $pop[[L1]]{{$}}
-; SIMD128-NEXT: i8x16.replace_lane $push[[R:[0-9]+]]=, $pop{{[0-9]+}}, 15, $pop[[L2]]{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <16 x i8> @shr_s_vec_v16i8(<16 x i8> %v, <16 x i8> %x) {
+; SIMD128-UNIMPL-SLOW-LABEL: shr_s_vec_v16i8:
+; SIMD128-UNIMPL-SLOW:         .functype shr_s_vec_v16i8 (v128, v128) -> (v128)
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_s $push6=, $0, 0
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.const $push0=, 7
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.splat $push1=, $pop0
+; SIMD128-UNIMPL-SLOW-NEXT:    v128.and $push67=, $1, $pop1
+; SIMD128-UNIMPL-SLOW-NEXT:    local.tee $push66=, $1=, $pop67
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push5=, $pop66, 0
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shr_s $push7=, $pop6, $pop5
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.splat $push8=, $pop7
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_s $push3=, $0, 1
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push2=, $1, 1
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shr_s $push4=, $pop3, $pop2
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.replace_lane $push9=, $pop8, 1, $pop4
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_s $push11=, $0, 2
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push10=, $1, 2
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shr_s $push12=, $pop11, $pop10
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.replace_lane $push13=, $pop9, 2, $pop12
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_s $push15=, $0, 3
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push14=, $1, 3
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shr_s $push16=, $pop15, $pop14
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.replace_lane $push17=, $pop13, 3, $pop16
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_s $push19=, $0, 4
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push18=, $1, 4
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shr_s $push20=, $pop19, $pop18
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.replace_lane $push21=, $pop17, 4, $pop20
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_s $push23=, $0, 5
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push22=, $1, 5
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shr_s $push24=, $pop23, $pop22
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.replace_lane $push25=, $pop21, 5, $pop24
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_s $push27=, $0, 6
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push26=, $1, 6
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shr_s $push28=, $pop27, $pop26
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.replace_lane $push29=, $pop25, 6, $pop28
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_s $push31=, $0, 7
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push30=, $1, 7
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shr_s $push32=, $pop31, $pop30
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.replace_lane $push33=, $pop29, 7, $pop32
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_s $push35=, $0, 8
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push34=, $1, 8
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shr_s $push36=, $pop35, $pop34
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.replace_lane $push37=, $pop33, 8, $pop36
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_s $push39=, $0, 9
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push38=, $1, 9
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shr_s $push40=, $pop39, $pop38
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.replace_lane $push41=, $pop37, 9, $pop40
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_s $push43=, $0, 10
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push42=, $1, 10
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shr_s $push44=, $pop43, $pop42
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.replace_lane $push45=, $pop41, 10, $pop44
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_s $push47=, $0, 11
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push46=, $1, 11
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shr_s $push48=, $pop47, $pop46
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.replace_lane $push49=, $pop45, 11, $pop48
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_s $push51=, $0, 12
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push50=, $1, 12
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shr_s $push52=, $pop51, $pop50
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.replace_lane $push53=, $pop49, 12, $pop52
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_s $push55=, $0, 13
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push54=, $1, 13
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shr_s $push56=, $pop55, $pop54
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.replace_lane $push57=, $pop53, 13, $pop56
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_s $push59=, $0, 14
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push58=, $1, 14
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shr_s $push60=, $pop59, $pop58
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.replace_lane $push61=, $pop57, 14, $pop60
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_s $push63=, $0, 15
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push62=, $1, 15
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shr_s $push64=, $pop63, $pop62
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.replace_lane $push65=, $pop61, 15, $pop64
+; SIMD128-UNIMPL-SLOW-NEXT:    return $pop65
+;
+; SIMD128-UNIMPL-FAST-LABEL: shr_s_vec_v16i8:
+; SIMD128-UNIMPL-FAST:         .functype shr_s_vec_v16i8 (v128, v128) -> (v128)
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_s $push7=, $0, 0
+; SIMD128-UNIMPL-FAST-NEXT:    i32.const $push1=, 7
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.splat $push2=, $pop1
+; SIMD128-UNIMPL-FAST-NEXT:    v128.and $push67=, $1, $pop2
+; SIMD128-UNIMPL-FAST-NEXT:    local.tee $push66=, $1=, $pop67
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push6=, $pop66, 0
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shr_s $push8=, $pop7, $pop6
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.splat $push9=, $pop8
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_s $push4=, $0, 1
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push3=, $1, 1
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shr_s $push5=, $pop4, $pop3
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.replace_lane $push10=, $pop9, 1, $pop5
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_s $push12=, $0, 2
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push11=, $1, 2
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shr_s $push13=, $pop12, $pop11
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.replace_lane $push14=, $pop10, 2, $pop13
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_s $push16=, $0, 3
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push15=, $1, 3
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shr_s $push17=, $pop16, $pop15
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.replace_lane $push18=, $pop14, 3, $pop17
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_s $push20=, $0, 4
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push19=, $1, 4
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shr_s $push21=, $pop20, $pop19
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.replace_lane $push22=, $pop18, 4, $pop21
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_s $push24=, $0, 5
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push23=, $1, 5
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shr_s $push25=, $pop24, $pop23
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.replace_lane $push26=, $pop22, 5, $pop25
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_s $push28=, $0, 6
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push27=, $1, 6
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shr_s $push29=, $pop28, $pop27
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.replace_lane $push30=, $pop26, 6, $pop29
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_s $push32=, $0, 7
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push31=, $1, 7
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shr_s $push33=, $pop32, $pop31
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.replace_lane $push34=, $pop30, 7, $pop33
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_s $push36=, $0, 8
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push35=, $1, 8
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shr_s $push37=, $pop36, $pop35
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.replace_lane $push38=, $pop34, 8, $pop37
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_s $push40=, $0, 9
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push39=, $1, 9
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shr_s $push41=, $pop40, $pop39
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.replace_lane $push42=, $pop38, 9, $pop41
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_s $push44=, $0, 10
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push43=, $1, 10
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shr_s $push45=, $pop44, $pop43
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.replace_lane $push46=, $pop42, 10, $pop45
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_s $push48=, $0, 11
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push47=, $1, 11
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shr_s $push49=, $pop48, $pop47
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.replace_lane $push50=, $pop46, 11, $pop49
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_s $push52=, $0, 12
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push51=, $1, 12
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shr_s $push53=, $pop52, $pop51
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.replace_lane $push54=, $pop50, 12, $pop53
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_s $push56=, $0, 13
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push55=, $1, 13
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shr_s $push57=, $pop56, $pop55
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.replace_lane $push58=, $pop54, 13, $pop57
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_s $push60=, $0, 14
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push59=, $1, 14
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shr_s $push61=, $pop60, $pop59
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.replace_lane $push62=, $pop58, 14, $pop61
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_s $push64=, $0, 15
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push63=, $1, 15
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shr_s $push65=, $pop64, $pop63
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.replace_lane $push0=, $pop62, 15, $pop65
+; SIMD128-UNIMPL-FAST-NEXT:    return $pop0
+;
+; SIMD128-VM-SLOW-LABEL: shr_s_vec_v16i8:
+; SIMD128-VM-SLOW:         .functype shr_s_vec_v16i8 (v128, v128) -> (v128)
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push8=, $0, 0
+; SIMD128-VM-SLOW-NEXT:    i32.const $push0=, 7
+; SIMD128-VM-SLOW-NEXT:    i8x16.splat $push1=, $pop0
+; SIMD128-VM-SLOW-NEXT:    v128.and $push99=, $1, $pop1
+; SIMD128-VM-SLOW-NEXT:    local.tee $push98=, $1=, $pop99
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push6=, $pop98, 0
+; SIMD128-VM-SLOW-NEXT:    i32.const $push97=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push7=, $pop6, $pop97
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push9=, $pop8, $pop7
+; SIMD128-VM-SLOW-NEXT:    i8x16.splat $push10=, $pop9
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push4=, $0, 1
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push2=, $1, 1
+; SIMD128-VM-SLOW-NEXT:    i32.const $push96=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push3=, $pop2, $pop96
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push5=, $pop4, $pop3
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push11=, $pop10, 1, $pop5
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push14=, $0, 2
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push12=, $1, 2
+; SIMD128-VM-SLOW-NEXT:    i32.const $push95=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push13=, $pop12, $pop95
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push15=, $pop14, $pop13
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push16=, $pop11, 2, $pop15
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push19=, $0, 3
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push17=, $1, 3
+; SIMD128-VM-SLOW-NEXT:    i32.const $push94=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push18=, $pop17, $pop94
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push20=, $pop19, $pop18
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push21=, $pop16, 3, $pop20
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push24=, $0, 4
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push22=, $1, 4
+; SIMD128-VM-SLOW-NEXT:    i32.const $push93=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push23=, $pop22, $pop93
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push25=, $pop24, $pop23
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push26=, $pop21, 4, $pop25
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push29=, $0, 5
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push27=, $1, 5
+; SIMD128-VM-SLOW-NEXT:    i32.const $push92=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push28=, $pop27, $pop92
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push30=, $pop29, $pop28
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push31=, $pop26, 5, $pop30
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push34=, $0, 6
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push32=, $1, 6
+; SIMD128-VM-SLOW-NEXT:    i32.const $push91=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push33=, $pop32, $pop91
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push35=, $pop34, $pop33
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push36=, $pop31, 6, $pop35
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push39=, $0, 7
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push37=, $1, 7
+; SIMD128-VM-SLOW-NEXT:    i32.const $push90=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push38=, $pop37, $pop90
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push40=, $pop39, $pop38
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push41=, $pop36, 7, $pop40
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push44=, $0, 8
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push42=, $1, 8
+; SIMD128-VM-SLOW-NEXT:    i32.const $push89=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push43=, $pop42, $pop89
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push45=, $pop44, $pop43
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push46=, $pop41, 8, $pop45
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push49=, $0, 9
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push47=, $1, 9
+; SIMD128-VM-SLOW-NEXT:    i32.const $push88=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push48=, $pop47, $pop88
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push50=, $pop49, $pop48
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push51=, $pop46, 9, $pop50
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push54=, $0, 10
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push52=, $1, 10
+; SIMD128-VM-SLOW-NEXT:    i32.const $push87=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push53=, $pop52, $pop87
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push55=, $pop54, $pop53
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push56=, $pop51, 10, $pop55
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push59=, $0, 11
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push57=, $1, 11
+; SIMD128-VM-SLOW-NEXT:    i32.const $push86=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push58=, $pop57, $pop86
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push60=, $pop59, $pop58
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push61=, $pop56, 11, $pop60
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push64=, $0, 12
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push62=, $1, 12
+; SIMD128-VM-SLOW-NEXT:    i32.const $push85=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push63=, $pop62, $pop85
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push65=, $pop64, $pop63
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push66=, $pop61, 12, $pop65
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push69=, $0, 13
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push67=, $1, 13
+; SIMD128-VM-SLOW-NEXT:    i32.const $push84=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push68=, $pop67, $pop84
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push70=, $pop69, $pop68
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push71=, $pop66, 13, $pop70
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push74=, $0, 14
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push72=, $1, 14
+; SIMD128-VM-SLOW-NEXT:    i32.const $push83=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push73=, $pop72, $pop83
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push75=, $pop74, $pop73
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push76=, $pop71, 14, $pop75
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push79=, $0, 15
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push77=, $1, 15
+; SIMD128-VM-SLOW-NEXT:    i32.const $push82=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push78=, $pop77, $pop82
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push80=, $pop79, $pop78
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push81=, $pop76, 15, $pop80
+; SIMD128-VM-SLOW-NEXT:    return $pop81
+;
+; SIMD128-VM-FAST-LABEL: shr_s_vec_v16i8:
+; SIMD128-VM-FAST:         .functype shr_s_vec_v16i8 (v128, v128) -> (v128)
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push9=, $0, 0
+; SIMD128-VM-FAST-NEXT:    i32.const $push1=, 7
+; SIMD128-VM-FAST-NEXT:    i8x16.splat $push2=, $pop1
+; SIMD128-VM-FAST-NEXT:    v128.and $push99=, $1, $pop2
+; SIMD128-VM-FAST-NEXT:    local.tee $push98=, $1=, $pop99
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push7=, $pop98, 0
+; SIMD128-VM-FAST-NEXT:    i32.const $push97=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push8=, $pop7, $pop97
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push10=, $pop9, $pop8
+; SIMD128-VM-FAST-NEXT:    i8x16.splat $push11=, $pop10
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push5=, $0, 1
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push3=, $1, 1
+; SIMD128-VM-FAST-NEXT:    i32.const $push96=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push4=, $pop3, $pop96
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push6=, $pop5, $pop4
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push12=, $pop11, 1, $pop6
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push15=, $0, 2
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push13=, $1, 2
+; SIMD128-VM-FAST-NEXT:    i32.const $push95=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push14=, $pop13, $pop95
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push16=, $pop15, $pop14
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push17=, $pop12, 2, $pop16
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push20=, $0, 3
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push18=, $1, 3
+; SIMD128-VM-FAST-NEXT:    i32.const $push94=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push19=, $pop18, $pop94
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push21=, $pop20, $pop19
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push22=, $pop17, 3, $pop21
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push25=, $0, 4
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push23=, $1, 4
+; SIMD128-VM-FAST-NEXT:    i32.const $push93=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push24=, $pop23, $pop93
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push26=, $pop25, $pop24
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push27=, $pop22, 4, $pop26
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push30=, $0, 5
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push28=, $1, 5
+; SIMD128-VM-FAST-NEXT:    i32.const $push92=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push29=, $pop28, $pop92
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push31=, $pop30, $pop29
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push32=, $pop27, 5, $pop31
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push35=, $0, 6
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push33=, $1, 6
+; SIMD128-VM-FAST-NEXT:    i32.const $push91=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push34=, $pop33, $pop91
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push36=, $pop35, $pop34
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push37=, $pop32, 6, $pop36
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push40=, $0, 7
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push38=, $1, 7
+; SIMD128-VM-FAST-NEXT:    i32.const $push90=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push39=, $pop38, $pop90
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push41=, $pop40, $pop39
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push42=, $pop37, 7, $pop41
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push45=, $0, 8
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push43=, $1, 8
+; SIMD128-VM-FAST-NEXT:    i32.const $push89=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push44=, $pop43, $pop89
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push46=, $pop45, $pop44
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push47=, $pop42, 8, $pop46
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push50=, $0, 9
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push48=, $1, 9
+; SIMD128-VM-FAST-NEXT:    i32.const $push88=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push49=, $pop48, $pop88
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push51=, $pop50, $pop49
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push52=, $pop47, 9, $pop51
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push55=, $0, 10
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push53=, $1, 10
+; SIMD128-VM-FAST-NEXT:    i32.const $push87=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push54=, $pop53, $pop87
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push56=, $pop55, $pop54
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push57=, $pop52, 10, $pop56
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push60=, $0, 11
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push58=, $1, 11
+; SIMD128-VM-FAST-NEXT:    i32.const $push86=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push59=, $pop58, $pop86
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push61=, $pop60, $pop59
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push62=, $pop57, 11, $pop61
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push65=, $0, 12
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push63=, $1, 12
+; SIMD128-VM-FAST-NEXT:    i32.const $push85=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push64=, $pop63, $pop85
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push66=, $pop65, $pop64
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push67=, $pop62, 12, $pop66
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push70=, $0, 13
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push68=, $1, 13
+; SIMD128-VM-FAST-NEXT:    i32.const $push84=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push69=, $pop68, $pop84
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push71=, $pop70, $pop69
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push72=, $pop67, 13, $pop71
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push75=, $0, 14
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push73=, $1, 14
+; SIMD128-VM-FAST-NEXT:    i32.const $push83=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push74=, $pop73, $pop83
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push76=, $pop75, $pop74
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push77=, $pop72, 14, $pop76
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push80=, $0, 15
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push78=, $1, 15
+; SIMD128-VM-FAST-NEXT:    i32.const $push82=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push79=, $pop78, $pop82
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push81=, $pop80, $pop79
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push0=, $pop77, 15, $pop81
+; SIMD128-VM-FAST-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: shr_s_vec_v16i8:
+; NO-SIMD128-SLOW:         .functype shr_s_vec_v16i8 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push6=, 15
+; NO-SIMD128-SLOW-NEXT:    i32.add $push7=, $0, $pop6
+; NO-SIMD128-SLOW-NEXT:    i32.const $push2=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push3=, $16, $pop2
+; NO-SIMD128-SLOW-NEXT:    i32.const $push133=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push4=, $pop3, $pop133
+; NO-SIMD128-SLOW-NEXT:    i32.const $push0=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push1=, $32, $pop0
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push5=, $pop4, $pop1
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop7), $pop5
+; NO-SIMD128-SLOW-NEXT:    i32.const $push12=, 14
+; NO-SIMD128-SLOW-NEXT:    i32.add $push13=, $0, $pop12
+; NO-SIMD128-SLOW-NEXT:    i32.const $push132=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push9=, $15, $pop132
+; NO-SIMD128-SLOW-NEXT:    i32.const $push131=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push10=, $pop9, $pop131
+; NO-SIMD128-SLOW-NEXT:    i32.const $push130=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push8=, $31, $pop130
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push11=, $pop10, $pop8
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop13), $pop11
+; NO-SIMD128-SLOW-NEXT:    i32.const $push18=, 13
+; NO-SIMD128-SLOW-NEXT:    i32.add $push19=, $0, $pop18
+; NO-SIMD128-SLOW-NEXT:    i32.const $push129=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push15=, $14, $pop129
+; NO-SIMD128-SLOW-NEXT:    i32.const $push128=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push16=, $pop15, $pop128
+; NO-SIMD128-SLOW-NEXT:    i32.const $push127=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push14=, $30, $pop127
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push17=, $pop16, $pop14
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop19), $pop17
+; NO-SIMD128-SLOW-NEXT:    i32.const $push24=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push25=, $0, $pop24
+; NO-SIMD128-SLOW-NEXT:    i32.const $push126=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push21=, $13, $pop126
+; NO-SIMD128-SLOW-NEXT:    i32.const $push125=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push22=, $pop21, $pop125
+; NO-SIMD128-SLOW-NEXT:    i32.const $push124=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push20=, $29, $pop124
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push23=, $pop22, $pop20
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop25), $pop23
+; NO-SIMD128-SLOW-NEXT:    i32.const $push30=, 11
+; NO-SIMD128-SLOW-NEXT:    i32.add $push31=, $0, $pop30
+; NO-SIMD128-SLOW-NEXT:    i32.const $push123=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push27=, $12, $pop123
+; NO-SIMD128-SLOW-NEXT:    i32.const $push122=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push28=, $pop27, $pop122
+; NO-SIMD128-SLOW-NEXT:    i32.const $push121=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push26=, $28, $pop121
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push29=, $pop28, $pop26
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop31), $pop29
+; NO-SIMD128-SLOW-NEXT:    i32.const $push36=, 10
+; NO-SIMD128-SLOW-NEXT:    i32.add $push37=, $0, $pop36
+; NO-SIMD128-SLOW-NEXT:    i32.const $push120=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push33=, $11, $pop120
+; NO-SIMD128-SLOW-NEXT:    i32.const $push119=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push34=, $pop33, $pop119
+; NO-SIMD128-SLOW-NEXT:    i32.const $push118=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push32=, $27, $pop118
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push35=, $pop34, $pop32
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop37), $pop35
+; NO-SIMD128-SLOW-NEXT:    i32.const $push42=, 9
+; NO-SIMD128-SLOW-NEXT:    i32.add $push43=, $0, $pop42
+; NO-SIMD128-SLOW-NEXT:    i32.const $push117=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push39=, $10, $pop117
+; NO-SIMD128-SLOW-NEXT:    i32.const $push116=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push40=, $pop39, $pop116
+; NO-SIMD128-SLOW-NEXT:    i32.const $push115=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push38=, $26, $pop115
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push41=, $pop40, $pop38
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop43), $pop41
+; NO-SIMD128-SLOW-NEXT:    i32.const $push114=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push45=, $9, $pop114
+; NO-SIMD128-SLOW-NEXT:    i32.const $push113=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push46=, $pop45, $pop113
+; NO-SIMD128-SLOW-NEXT:    i32.const $push112=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push44=, $25, $pop112
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push47=, $pop46, $pop44
+; NO-SIMD128-SLOW-NEXT:    i32.store8 8($0), $pop47
+; NO-SIMD128-SLOW-NEXT:    i32.const $push52=, 7
+; NO-SIMD128-SLOW-NEXT:    i32.add $push53=, $0, $pop52
+; NO-SIMD128-SLOW-NEXT:    i32.const $push111=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push49=, $8, $pop111
+; NO-SIMD128-SLOW-NEXT:    i32.const $push110=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push50=, $pop49, $pop110
+; NO-SIMD128-SLOW-NEXT:    i32.const $push109=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push48=, $24, $pop109
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push51=, $pop50, $pop48
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop53), $pop51
+; NO-SIMD128-SLOW-NEXT:    i32.const $push58=, 6
+; NO-SIMD128-SLOW-NEXT:    i32.add $push59=, $0, $pop58
+; NO-SIMD128-SLOW-NEXT:    i32.const $push108=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push55=, $7, $pop108
+; NO-SIMD128-SLOW-NEXT:    i32.const $push107=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push56=, $pop55, $pop107
+; NO-SIMD128-SLOW-NEXT:    i32.const $push106=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push54=, $23, $pop106
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push57=, $pop56, $pop54
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop59), $pop57
+; NO-SIMD128-SLOW-NEXT:    i32.const $push64=, 5
+; NO-SIMD128-SLOW-NEXT:    i32.add $push65=, $0, $pop64
+; NO-SIMD128-SLOW-NEXT:    i32.const $push105=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push61=, $6, $pop105
+; NO-SIMD128-SLOW-NEXT:    i32.const $push104=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push62=, $pop61, $pop104
+; NO-SIMD128-SLOW-NEXT:    i32.const $push103=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push60=, $22, $pop103
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push63=, $pop62, $pop60
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop65), $pop63
+; NO-SIMD128-SLOW-NEXT:    i32.const $push102=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push67=, $5, $pop102
+; NO-SIMD128-SLOW-NEXT:    i32.const $push101=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push68=, $pop67, $pop101
+; NO-SIMD128-SLOW-NEXT:    i32.const $push100=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push66=, $21, $pop100
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push69=, $pop68, $pop66
+; NO-SIMD128-SLOW-NEXT:    i32.store8 4($0), $pop69
+; NO-SIMD128-SLOW-NEXT:    i32.const $push74=, 3
+; NO-SIMD128-SLOW-NEXT:    i32.add $push75=, $0, $pop74
+; NO-SIMD128-SLOW-NEXT:    i32.const $push99=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push71=, $4, $pop99
+; NO-SIMD128-SLOW-NEXT:    i32.const $push98=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push72=, $pop71, $pop98
+; NO-SIMD128-SLOW-NEXT:    i32.const $push97=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push70=, $20, $pop97
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push73=, $pop72, $pop70
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop75), $pop73
+; NO-SIMD128-SLOW-NEXT:    i32.const $push96=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push77=, $3, $pop96
+; NO-SIMD128-SLOW-NEXT:    i32.const $push95=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push78=, $pop77, $pop95
+; NO-SIMD128-SLOW-NEXT:    i32.const $push94=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push76=, $19, $pop94
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push79=, $pop78, $pop76
+; NO-SIMD128-SLOW-NEXT:    i32.store8 2($0), $pop79
+; NO-SIMD128-SLOW-NEXT:    i32.const $push93=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push81=, $2, $pop93
+; NO-SIMD128-SLOW-NEXT:    i32.const $push92=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push82=, $pop81, $pop92
+; NO-SIMD128-SLOW-NEXT:    i32.const $push91=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push80=, $18, $pop91
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push83=, $pop82, $pop80
+; NO-SIMD128-SLOW-NEXT:    i32.store8 1($0), $pop83
+; NO-SIMD128-SLOW-NEXT:    i32.const $push90=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push85=, $1, $pop90
+; NO-SIMD128-SLOW-NEXT:    i32.const $push89=, 24
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push86=, $pop85, $pop89
+; NO-SIMD128-SLOW-NEXT:    i32.const $push88=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push84=, $17, $pop88
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push87=, $pop86, $pop84
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($0), $pop87
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: shr_s_vec_v16i8:
+; NO-SIMD128-FAST:         .functype shr_s_vec_v16i8 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-FAST-NEXT:    i32.const $push2=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shl $push3=, $1, $pop2
+; NO-SIMD128-FAST-NEXT:    i32.const $push133=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push4=, $pop3, $pop133
+; NO-SIMD128-FAST-NEXT:    i32.const $push0=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push1=, $17, $pop0
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push5=, $pop4, $pop1
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($0), $pop5
+; NO-SIMD128-FAST-NEXT:    i32.const $push132=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shl $push7=, $2, $pop132
+; NO-SIMD128-FAST-NEXT:    i32.const $push131=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push8=, $pop7, $pop131
+; NO-SIMD128-FAST-NEXT:    i32.const $push130=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push6=, $18, $pop130
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push9=, $pop8, $pop6
+; NO-SIMD128-FAST-NEXT:    i32.store8 1($0), $pop9
+; NO-SIMD128-FAST-NEXT:    i32.const $push129=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shl $push11=, $3, $pop129
+; NO-SIMD128-FAST-NEXT:    i32.const $push128=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push12=, $pop11, $pop128
+; NO-SIMD128-FAST-NEXT:    i32.const $push127=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push10=, $19, $pop127
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push13=, $pop12, $pop10
+; NO-SIMD128-FAST-NEXT:    i32.store8 2($0), $pop13
+; NO-SIMD128-FAST-NEXT:    i32.const $push18=, 3
+; NO-SIMD128-FAST-NEXT:    i32.add $push19=, $0, $pop18
+; NO-SIMD128-FAST-NEXT:    i32.const $push126=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shl $push15=, $4, $pop126
+; NO-SIMD128-FAST-NEXT:    i32.const $push125=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push16=, $pop15, $pop125
+; NO-SIMD128-FAST-NEXT:    i32.const $push124=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push14=, $20, $pop124
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push17=, $pop16, $pop14
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop19), $pop17
+; NO-SIMD128-FAST-NEXT:    i32.const $push123=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shl $push21=, $5, $pop123
+; NO-SIMD128-FAST-NEXT:    i32.const $push122=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push22=, $pop21, $pop122
+; NO-SIMD128-FAST-NEXT:    i32.const $push121=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push20=, $21, $pop121
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push23=, $pop22, $pop20
+; NO-SIMD128-FAST-NEXT:    i32.store8 4($0), $pop23
+; NO-SIMD128-FAST-NEXT:    i32.const $push28=, 5
+; NO-SIMD128-FAST-NEXT:    i32.add $push29=, $0, $pop28
+; NO-SIMD128-FAST-NEXT:    i32.const $push120=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shl $push25=, $6, $pop120
+; NO-SIMD128-FAST-NEXT:    i32.const $push119=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push26=, $pop25, $pop119
+; NO-SIMD128-FAST-NEXT:    i32.const $push118=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push24=, $22, $pop118
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push27=, $pop26, $pop24
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop29), $pop27
+; NO-SIMD128-FAST-NEXT:    i32.const $push34=, 6
+; NO-SIMD128-FAST-NEXT:    i32.add $push35=, $0, $pop34
+; NO-SIMD128-FAST-NEXT:    i32.const $push117=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shl $push31=, $7, $pop117
+; NO-SIMD128-FAST-NEXT:    i32.const $push116=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push32=, $pop31, $pop116
+; NO-SIMD128-FAST-NEXT:    i32.const $push115=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push30=, $23, $pop115
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push33=, $pop32, $pop30
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop35), $pop33
+; NO-SIMD128-FAST-NEXT:    i32.const $push40=, 7
+; NO-SIMD128-FAST-NEXT:    i32.add $push41=, $0, $pop40
+; NO-SIMD128-FAST-NEXT:    i32.const $push114=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shl $push37=, $8, $pop114
+; NO-SIMD128-FAST-NEXT:    i32.const $push113=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push38=, $pop37, $pop113
+; NO-SIMD128-FAST-NEXT:    i32.const $push112=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push36=, $24, $pop112
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push39=, $pop38, $pop36
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop41), $pop39
+; NO-SIMD128-FAST-NEXT:    i32.const $push111=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shl $push43=, $9, $pop111
+; NO-SIMD128-FAST-NEXT:    i32.const $push110=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push44=, $pop43, $pop110
+; NO-SIMD128-FAST-NEXT:    i32.const $push109=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push42=, $25, $pop109
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push45=, $pop44, $pop42
+; NO-SIMD128-FAST-NEXT:    i32.store8 8($0), $pop45
+; NO-SIMD128-FAST-NEXT:    i32.const $push50=, 9
+; NO-SIMD128-FAST-NEXT:    i32.add $push51=, $0, $pop50
+; NO-SIMD128-FAST-NEXT:    i32.const $push108=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shl $push47=, $10, $pop108
+; NO-SIMD128-FAST-NEXT:    i32.const $push107=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push48=, $pop47, $pop107
+; NO-SIMD128-FAST-NEXT:    i32.const $push106=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push46=, $26, $pop106
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push49=, $pop48, $pop46
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop51), $pop49
+; NO-SIMD128-FAST-NEXT:    i32.const $push56=, 10
+; NO-SIMD128-FAST-NEXT:    i32.add $push57=, $0, $pop56
+; NO-SIMD128-FAST-NEXT:    i32.const $push105=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shl $push53=, $11, $pop105
+; NO-SIMD128-FAST-NEXT:    i32.const $push104=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push54=, $pop53, $pop104
+; NO-SIMD128-FAST-NEXT:    i32.const $push103=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push52=, $27, $pop103
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push55=, $pop54, $pop52
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop57), $pop55
+; NO-SIMD128-FAST-NEXT:    i32.const $push62=, 11
+; NO-SIMD128-FAST-NEXT:    i32.add $push63=, $0, $pop62
+; NO-SIMD128-FAST-NEXT:    i32.const $push102=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shl $push59=, $12, $pop102
+; NO-SIMD128-FAST-NEXT:    i32.const $push101=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push60=, $pop59, $pop101
+; NO-SIMD128-FAST-NEXT:    i32.const $push100=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push58=, $28, $pop100
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push61=, $pop60, $pop58
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop63), $pop61
+; NO-SIMD128-FAST-NEXT:    i32.const $push68=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push69=, $0, $pop68
+; NO-SIMD128-FAST-NEXT:    i32.const $push99=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shl $push65=, $13, $pop99
+; NO-SIMD128-FAST-NEXT:    i32.const $push98=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push66=, $pop65, $pop98
+; NO-SIMD128-FAST-NEXT:    i32.const $push97=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push64=, $29, $pop97
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push67=, $pop66, $pop64
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop69), $pop67
+; NO-SIMD128-FAST-NEXT:    i32.const $push74=, 13
+; NO-SIMD128-FAST-NEXT:    i32.add $push75=, $0, $pop74
+; NO-SIMD128-FAST-NEXT:    i32.const $push96=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shl $push71=, $14, $pop96
+; NO-SIMD128-FAST-NEXT:    i32.const $push95=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push72=, $pop71, $pop95
+; NO-SIMD128-FAST-NEXT:    i32.const $push94=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push70=, $30, $pop94
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push73=, $pop72, $pop70
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop75), $pop73
+; NO-SIMD128-FAST-NEXT:    i32.const $push80=, 14
+; NO-SIMD128-FAST-NEXT:    i32.add $push81=, $0, $pop80
+; NO-SIMD128-FAST-NEXT:    i32.const $push93=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shl $push77=, $15, $pop93
+; NO-SIMD128-FAST-NEXT:    i32.const $push92=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push78=, $pop77, $pop92
+; NO-SIMD128-FAST-NEXT:    i32.const $push91=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push76=, $31, $pop91
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push79=, $pop78, $pop76
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop81), $pop79
+; NO-SIMD128-FAST-NEXT:    i32.const $push86=, 15
+; NO-SIMD128-FAST-NEXT:    i32.add $push87=, $0, $pop86
+; NO-SIMD128-FAST-NEXT:    i32.const $push90=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shl $push83=, $16, $pop90
+; NO-SIMD128-FAST-NEXT:    i32.const $push89=, 24
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push84=, $pop83, $pop89
+; NO-SIMD128-FAST-NEXT:    i32.const $push88=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push82=, $32, $pop88
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push85=, $pop84, $pop82
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop87), $pop85
+; NO-SIMD128-FAST-NEXT:    return
   %a = ashr <16 x i8> %v, %x
   ret <16 x i8> %a
 }
 
-; CHECK-LABEL: shr_u_v16i8:
-; NO-SIMD128-NOT: i8x16
-; SIMD128-NEXT: .functype shr_u_v16i8 (v128, i32) -> (v128){{$}}
-; SIMD128-NEXT: i8x16.shr_u $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <16 x i8> @shr_u_v16i8(<16 x i8> %v, i8 %x) {
+; SIMD128-UNIMPL-LABEL: shr_u_v16i8:
+; SIMD128-UNIMPL:         .functype shr_u_v16i8 (v128, i32) -> (v128)
+; SIMD128-UNIMPL-NEXT:    i8x16.shr_u $push0=, $0, $1
+; SIMD128-UNIMPL-NEXT:    return $pop0
+;
+; SIMD128-VM-SLOW-LABEL: shr_u_v16i8:
+; SIMD128-VM-SLOW:         .functype shr_u_v16i8 (v128, i32) -> (v128)
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push9=, $0, 0
+; SIMD128-VM-SLOW-NEXT:    i32.const $push1=, 255
+; SIMD128-VM-SLOW-NEXT:    i32.and $push10=, $pop9, $pop1
+; SIMD128-VM-SLOW-NEXT:    i8x16.splat $push5=, $1
+; SIMD128-VM-SLOW-NEXT:    i32.const $push3=, 7
+; SIMD128-VM-SLOW-NEXT:    i8x16.splat $push4=, $pop3
+; SIMD128-VM-SLOW-NEXT:    v128.and $push132=, $pop5, $pop4
+; SIMD128-VM-SLOW-NEXT:    local.tee $push131=, $2=, $pop132
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push11=, $pop131, 0
+; SIMD128-VM-SLOW-NEXT:    i32.const $push130=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push12=, $pop11, $pop130
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push13=, $pop10, $pop12
+; SIMD128-VM-SLOW-NEXT:    i8x16.splat $push14=, $pop13
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push0=, $0, 1
+; SIMD128-VM-SLOW-NEXT:    i32.const $push129=, 255
+; SIMD128-VM-SLOW-NEXT:    i32.and $push2=, $pop0, $pop129
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push6=, $2, 1
+; SIMD128-VM-SLOW-NEXT:    i32.const $push128=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push7=, $pop6, $pop128
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push8=, $pop2, $pop7
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push15=, $pop14, 1, $pop8
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push16=, $0, 2
+; SIMD128-VM-SLOW-NEXT:    i32.const $push127=, 255
+; SIMD128-VM-SLOW-NEXT:    i32.and $push17=, $pop16, $pop127
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push18=, $2, 2
+; SIMD128-VM-SLOW-NEXT:    i32.const $push126=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push19=, $pop18, $pop126
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push20=, $pop17, $pop19
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push21=, $pop15, 2, $pop20
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push22=, $0, 3
+; SIMD128-VM-SLOW-NEXT:    i32.const $push125=, 255
+; SIMD128-VM-SLOW-NEXT:    i32.and $push23=, $pop22, $pop125
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push24=, $2, 3
+; SIMD128-VM-SLOW-NEXT:    i32.const $push124=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push25=, $pop24, $pop124
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push26=, $pop23, $pop25
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push27=, $pop21, 3, $pop26
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push28=, $0, 4
+; SIMD128-VM-SLOW-NEXT:    i32.const $push123=, 255
+; SIMD128-VM-SLOW-NEXT:    i32.and $push29=, $pop28, $pop123
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push30=, $2, 4
+; SIMD128-VM-SLOW-NEXT:    i32.const $push122=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push31=, $pop30, $pop122
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push32=, $pop29, $pop31
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push33=, $pop27, 4, $pop32
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push34=, $0, 5
+; SIMD128-VM-SLOW-NEXT:    i32.const $push121=, 255
+; SIMD128-VM-SLOW-NEXT:    i32.and $push35=, $pop34, $pop121
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push36=, $2, 5
+; SIMD128-VM-SLOW-NEXT:    i32.const $push120=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push37=, $pop36, $pop120
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push38=, $pop35, $pop37
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push39=, $pop33, 5, $pop38
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push40=, $0, 6
+; SIMD128-VM-SLOW-NEXT:    i32.const $push119=, 255
+; SIMD128-VM-SLOW-NEXT:    i32.and $push41=, $pop40, $pop119
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push42=, $2, 6
+; SIMD128-VM-SLOW-NEXT:    i32.const $push118=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push43=, $pop42, $pop118
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push44=, $pop41, $pop43
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push45=, $pop39, 6, $pop44
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push46=, $0, 7
+; SIMD128-VM-SLOW-NEXT:    i32.const $push117=, 255
+; SIMD128-VM-SLOW-NEXT:    i32.and $push47=, $pop46, $pop117
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push48=, $2, 7
+; SIMD128-VM-SLOW-NEXT:    i32.const $push116=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push49=, $pop48, $pop116
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push50=, $pop47, $pop49
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push51=, $pop45, 7, $pop50
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push52=, $0, 8
+; SIMD128-VM-SLOW-NEXT:    i32.const $push115=, 255
+; SIMD128-VM-SLOW-NEXT:    i32.and $push53=, $pop52, $pop115
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push54=, $2, 8
+; SIMD128-VM-SLOW-NEXT:    i32.const $push114=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push55=, $pop54, $pop114
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push56=, $pop53, $pop55
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push57=, $pop51, 8, $pop56
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push58=, $0, 9
+; SIMD128-VM-SLOW-NEXT:    i32.const $push113=, 255
+; SIMD128-VM-SLOW-NEXT:    i32.and $push59=, $pop58, $pop113
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push60=, $2, 9
+; SIMD128-VM-SLOW-NEXT:    i32.const $push112=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push61=, $pop60, $pop112
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push62=, $pop59, $pop61
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push63=, $pop57, 9, $pop62
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push64=, $0, 10
+; SIMD128-VM-SLOW-NEXT:    i32.const $push111=, 255
+; SIMD128-VM-SLOW-NEXT:    i32.and $push65=, $pop64, $pop111
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push66=, $2, 10
+; SIMD128-VM-SLOW-NEXT:    i32.const $push110=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push67=, $pop66, $pop110
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push68=, $pop65, $pop67
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push69=, $pop63, 10, $pop68
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push70=, $0, 11
+; SIMD128-VM-SLOW-NEXT:    i32.const $push109=, 255
+; SIMD128-VM-SLOW-NEXT:    i32.and $push71=, $pop70, $pop109
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push72=, $2, 11
+; SIMD128-VM-SLOW-NEXT:    i32.const $push108=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push73=, $pop72, $pop108
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push74=, $pop71, $pop73
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push75=, $pop69, 11, $pop74
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push76=, $0, 12
+; SIMD128-VM-SLOW-NEXT:    i32.const $push107=, 255
+; SIMD128-VM-SLOW-NEXT:    i32.and $push77=, $pop76, $pop107
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push78=, $2, 12
+; SIMD128-VM-SLOW-NEXT:    i32.const $push106=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push79=, $pop78, $pop106
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push80=, $pop77, $pop79
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push81=, $pop75, 12, $pop80
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push82=, $0, 13
+; SIMD128-VM-SLOW-NEXT:    i32.const $push105=, 255
+; SIMD128-VM-SLOW-NEXT:    i32.and $push83=, $pop82, $pop105
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push84=, $2, 13
+; SIMD128-VM-SLOW-NEXT:    i32.const $push104=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push85=, $pop84, $pop104
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push86=, $pop83, $pop85
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push87=, $pop81, 13, $pop86
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push88=, $0, 14
+; SIMD128-VM-SLOW-NEXT:    i32.const $push103=, 255
+; SIMD128-VM-SLOW-NEXT:    i32.and $push89=, $pop88, $pop103
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push90=, $2, 14
+; SIMD128-VM-SLOW-NEXT:    i32.const $push102=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push91=, $pop90, $pop102
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push92=, $pop89, $pop91
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push93=, $pop87, 14, $pop92
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push94=, $0, 15
+; SIMD128-VM-SLOW-NEXT:    i32.const $push101=, 255
+; SIMD128-VM-SLOW-NEXT:    i32.and $push95=, $pop94, $pop101
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push96=, $2, 15
+; SIMD128-VM-SLOW-NEXT:    i32.const $push100=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push97=, $pop96, $pop100
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push98=, $pop95, $pop97
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push99=, $pop93, 15, $pop98
+; SIMD128-VM-SLOW-NEXT:    return $pop99
+;
+; SIMD128-VM-FAST-LABEL: shr_u_v16i8:
+; SIMD128-VM-FAST:         .functype shr_u_v16i8 (v128, i32) -> (v128)
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push10=, $0, 0
+; SIMD128-VM-FAST-NEXT:    i32.const $push3=, 255
+; SIMD128-VM-FAST-NEXT:    i32.and $push11=, $pop10, $pop3
+; SIMD128-VM-FAST-NEXT:    i8x16.splat $push1=, $1
+; SIMD128-VM-FAST-NEXT:    i32.const $push5=, 7
+; SIMD128-VM-FAST-NEXT:    i8x16.splat $push6=, $pop5
+; SIMD128-VM-FAST-NEXT:    v128.and $push132=, $pop1, $pop6
+; SIMD128-VM-FAST-NEXT:    local.tee $push131=, $2=, $pop132
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push12=, $pop131, 0
+; SIMD128-VM-FAST-NEXT:    i32.const $push130=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push13=, $pop12, $pop130
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push14=, $pop11, $pop13
+; SIMD128-VM-FAST-NEXT:    i8x16.splat $push15=, $pop14
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push2=, $0, 1
+; SIMD128-VM-FAST-NEXT:    i32.const $push129=, 255
+; SIMD128-VM-FAST-NEXT:    i32.and $push4=, $pop2, $pop129
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push7=, $2, 1
+; SIMD128-VM-FAST-NEXT:    i32.const $push128=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push8=, $pop7, $pop128
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push9=, $pop4, $pop8
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push16=, $pop15, 1, $pop9
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push17=, $0, 2
+; SIMD128-VM-FAST-NEXT:    i32.const $push127=, 255
+; SIMD128-VM-FAST-NEXT:    i32.and $push18=, $pop17, $pop127
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push19=, $2, 2
+; SIMD128-VM-FAST-NEXT:    i32.const $push126=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push20=, $pop19, $pop126
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push21=, $pop18, $pop20
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push22=, $pop16, 2, $pop21
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push23=, $0, 3
+; SIMD128-VM-FAST-NEXT:    i32.const $push125=, 255
+; SIMD128-VM-FAST-NEXT:    i32.and $push24=, $pop23, $pop125
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push25=, $2, 3
+; SIMD128-VM-FAST-NEXT:    i32.const $push124=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push26=, $pop25, $pop124
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push27=, $pop24, $pop26
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push28=, $pop22, 3, $pop27
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push29=, $0, 4
+; SIMD128-VM-FAST-NEXT:    i32.const $push123=, 255
+; SIMD128-VM-FAST-NEXT:    i32.and $push30=, $pop29, $pop123
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push31=, $2, 4
+; SIMD128-VM-FAST-NEXT:    i32.const $push122=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push32=, $pop31, $pop122
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push33=, $pop30, $pop32
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push34=, $pop28, 4, $pop33
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push35=, $0, 5
+; SIMD128-VM-FAST-NEXT:    i32.const $push121=, 255
+; SIMD128-VM-FAST-NEXT:    i32.and $push36=, $pop35, $pop121
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push37=, $2, 5
+; SIMD128-VM-FAST-NEXT:    i32.const $push120=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push38=, $pop37, $pop120
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push39=, $pop36, $pop38
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push40=, $pop34, 5, $pop39
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push41=, $0, 6
+; SIMD128-VM-FAST-NEXT:    i32.const $push119=, 255
+; SIMD128-VM-FAST-NEXT:    i32.and $push42=, $pop41, $pop119
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push43=, $2, 6
+; SIMD128-VM-FAST-NEXT:    i32.const $push118=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push44=, $pop43, $pop118
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push45=, $pop42, $pop44
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push46=, $pop40, 6, $pop45
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push47=, $0, 7
+; SIMD128-VM-FAST-NEXT:    i32.const $push117=, 255
+; SIMD128-VM-FAST-NEXT:    i32.and $push48=, $pop47, $pop117
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push49=, $2, 7
+; SIMD128-VM-FAST-NEXT:    i32.const $push116=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push50=, $pop49, $pop116
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push51=, $pop48, $pop50
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push52=, $pop46, 7, $pop51
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push53=, $0, 8
+; SIMD128-VM-FAST-NEXT:    i32.const $push115=, 255
+; SIMD128-VM-FAST-NEXT:    i32.and $push54=, $pop53, $pop115
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push55=, $2, 8
+; SIMD128-VM-FAST-NEXT:    i32.const $push114=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push56=, $pop55, $pop114
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push57=, $pop54, $pop56
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push58=, $pop52, 8, $pop57
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push59=, $0, 9
+; SIMD128-VM-FAST-NEXT:    i32.const $push113=, 255
+; SIMD128-VM-FAST-NEXT:    i32.and $push60=, $pop59, $pop113
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push61=, $2, 9
+; SIMD128-VM-FAST-NEXT:    i32.const $push112=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push62=, $pop61, $pop112
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push63=, $pop60, $pop62
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push64=, $pop58, 9, $pop63
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push65=, $0, 10
+; SIMD128-VM-FAST-NEXT:    i32.const $push111=, 255
+; SIMD128-VM-FAST-NEXT:    i32.and $push66=, $pop65, $pop111
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push67=, $2, 10
+; SIMD128-VM-FAST-NEXT:    i32.const $push110=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push68=, $pop67, $pop110
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push69=, $pop66, $pop68
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push70=, $pop64, 10, $pop69
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push71=, $0, 11
+; SIMD128-VM-FAST-NEXT:    i32.const $push109=, 255
+; SIMD128-VM-FAST-NEXT:    i32.and $push72=, $pop71, $pop109
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push73=, $2, 11
+; SIMD128-VM-FAST-NEXT:    i32.const $push108=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push74=, $pop73, $pop108
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push75=, $pop72, $pop74
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push76=, $pop70, 11, $pop75
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push77=, $0, 12
+; SIMD128-VM-FAST-NEXT:    i32.const $push107=, 255
+; SIMD128-VM-FAST-NEXT:    i32.and $push78=, $pop77, $pop107
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push79=, $2, 12
+; SIMD128-VM-FAST-NEXT:    i32.const $push106=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push80=, $pop79, $pop106
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push81=, $pop78, $pop80
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push82=, $pop76, 12, $pop81
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push83=, $0, 13
+; SIMD128-VM-FAST-NEXT:    i32.const $push105=, 255
+; SIMD128-VM-FAST-NEXT:    i32.and $push84=, $pop83, $pop105
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push85=, $2, 13
+; SIMD128-VM-FAST-NEXT:    i32.const $push104=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push86=, $pop85, $pop104
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push87=, $pop84, $pop86
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push88=, $pop82, 13, $pop87
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push89=, $0, 14
+; SIMD128-VM-FAST-NEXT:    i32.const $push103=, 255
+; SIMD128-VM-FAST-NEXT:    i32.and $push90=, $pop89, $pop103
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push91=, $2, 14
+; SIMD128-VM-FAST-NEXT:    i32.const $push102=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push92=, $pop91, $pop102
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push93=, $pop90, $pop92
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push94=, $pop88, 14, $pop93
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push95=, $0, 15
+; SIMD128-VM-FAST-NEXT:    i32.const $push101=, 255
+; SIMD128-VM-FAST-NEXT:    i32.and $push96=, $pop95, $pop101
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push97=, $2, 15
+; SIMD128-VM-FAST-NEXT:    i32.const $push100=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push98=, $pop97, $pop100
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push99=, $pop96, $pop98
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push0=, $pop94, 15, $pop99
+; SIMD128-VM-FAST-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: shr_u_v16i8:
+; NO-SIMD128-SLOW:         .functype shr_u_v16i8 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push3=, 15
+; NO-SIMD128-SLOW-NEXT:    i32.add $push4=, $0, $pop3
+; NO-SIMD128-SLOW-NEXT:    i32.const $push0=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push1=, $16, $pop0
+; NO-SIMD128-SLOW-NEXT:    i32.const $push72=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push71=, $17, $pop72
+; NO-SIMD128-SLOW-NEXT:    local.tee $push70=, $16=, $pop71
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push2=, $pop1, $pop70
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop4), $pop2
+; NO-SIMD128-SLOW-NEXT:    i32.const $push7=, 14
+; NO-SIMD128-SLOW-NEXT:    i32.add $push8=, $0, $pop7
+; NO-SIMD128-SLOW-NEXT:    i32.const $push69=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push5=, $15, $pop69
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push6=, $pop5, $16
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop8), $pop6
+; NO-SIMD128-SLOW-NEXT:    i32.const $push11=, 13
+; NO-SIMD128-SLOW-NEXT:    i32.add $push12=, $0, $pop11
+; NO-SIMD128-SLOW-NEXT:    i32.const $push68=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push9=, $14, $pop68
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push10=, $pop9, $16
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop12), $pop10
+; NO-SIMD128-SLOW-NEXT:    i32.const $push15=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push16=, $0, $pop15
+; NO-SIMD128-SLOW-NEXT:    i32.const $push67=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push13=, $13, $pop67
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push14=, $pop13, $16
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop16), $pop14
+; NO-SIMD128-SLOW-NEXT:    i32.const $push19=, 11
+; NO-SIMD128-SLOW-NEXT:    i32.add $push20=, $0, $pop19
+; NO-SIMD128-SLOW-NEXT:    i32.const $push66=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push17=, $12, $pop66
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push18=, $pop17, $16
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop20), $pop18
+; NO-SIMD128-SLOW-NEXT:    i32.const $push23=, 10
+; NO-SIMD128-SLOW-NEXT:    i32.add $push24=, $0, $pop23
+; NO-SIMD128-SLOW-NEXT:    i32.const $push65=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push21=, $11, $pop65
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push22=, $pop21, $16
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop24), $pop22
+; NO-SIMD128-SLOW-NEXT:    i32.const $push27=, 9
+; NO-SIMD128-SLOW-NEXT:    i32.add $push28=, $0, $pop27
+; NO-SIMD128-SLOW-NEXT:    i32.const $push64=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push25=, $10, $pop64
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push26=, $pop25, $16
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop28), $pop26
+; NO-SIMD128-SLOW-NEXT:    i32.const $push63=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push29=, $9, $pop63
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push30=, $pop29, $16
+; NO-SIMD128-SLOW-NEXT:    i32.store8 8($0), $pop30
+; NO-SIMD128-SLOW-NEXT:    i32.const $push33=, 7
+; NO-SIMD128-SLOW-NEXT:    i32.add $push34=, $0, $pop33
+; NO-SIMD128-SLOW-NEXT:    i32.const $push62=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push31=, $8, $pop62
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push32=, $pop31, $16
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop34), $pop32
+; NO-SIMD128-SLOW-NEXT:    i32.const $push37=, 6
+; NO-SIMD128-SLOW-NEXT:    i32.add $push38=, $0, $pop37
+; NO-SIMD128-SLOW-NEXT:    i32.const $push61=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push35=, $7, $pop61
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push36=, $pop35, $16
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop38), $pop36
+; NO-SIMD128-SLOW-NEXT:    i32.const $push41=, 5
+; NO-SIMD128-SLOW-NEXT:    i32.add $push42=, $0, $pop41
+; NO-SIMD128-SLOW-NEXT:    i32.const $push60=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push39=, $6, $pop60
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push40=, $pop39, $16
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop42), $pop40
+; NO-SIMD128-SLOW-NEXT:    i32.const $push59=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push43=, $5, $pop59
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push44=, $pop43, $16
+; NO-SIMD128-SLOW-NEXT:    i32.store8 4($0), $pop44
+; NO-SIMD128-SLOW-NEXT:    i32.const $push47=, 3
+; NO-SIMD128-SLOW-NEXT:    i32.add $push48=, $0, $pop47
+; NO-SIMD128-SLOW-NEXT:    i32.const $push58=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push45=, $4, $pop58
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push46=, $pop45, $16
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop48), $pop46
+; NO-SIMD128-SLOW-NEXT:    i32.const $push57=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push49=, $3, $pop57
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push50=, $pop49, $16
+; NO-SIMD128-SLOW-NEXT:    i32.store8 2($0), $pop50
+; NO-SIMD128-SLOW-NEXT:    i32.const $push56=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push51=, $2, $pop56
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push52=, $pop51, $16
+; NO-SIMD128-SLOW-NEXT:    i32.store8 1($0), $pop52
+; NO-SIMD128-SLOW-NEXT:    i32.const $push55=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push53=, $1, $pop55
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push54=, $pop53, $16
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($0), $pop54
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: shr_u_v16i8:
+; NO-SIMD128-FAST:         .functype shr_u_v16i8 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-FAST-NEXT:    i32.const $push0=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push1=, $1, $pop0
+; NO-SIMD128-FAST-NEXT:    i32.const $push72=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push71=, $17, $pop72
+; NO-SIMD128-FAST-NEXT:    local.tee $push70=, $17=, $pop71
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push2=, $pop1, $pop70
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($0), $pop2
+; NO-SIMD128-FAST-NEXT:    i32.const $push69=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push3=, $2, $pop69
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push4=, $pop3, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 1($0), $pop4
+; NO-SIMD128-FAST-NEXT:    i32.const $push68=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push5=, $3, $pop68
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push6=, $pop5, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 2($0), $pop6
+; NO-SIMD128-FAST-NEXT:    i32.const $push9=, 3
+; NO-SIMD128-FAST-NEXT:    i32.add $push10=, $0, $pop9
+; NO-SIMD128-FAST-NEXT:    i32.const $push67=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push7=, $4, $pop67
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push8=, $pop7, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop10), $pop8
+; NO-SIMD128-FAST-NEXT:    i32.const $push66=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push11=, $5, $pop66
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push12=, $pop11, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 4($0), $pop12
+; NO-SIMD128-FAST-NEXT:    i32.const $push15=, 5
+; NO-SIMD128-FAST-NEXT:    i32.add $push16=, $0, $pop15
+; NO-SIMD128-FAST-NEXT:    i32.const $push65=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push13=, $6, $pop65
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push14=, $pop13, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop16), $pop14
+; NO-SIMD128-FAST-NEXT:    i32.const $push19=, 6
+; NO-SIMD128-FAST-NEXT:    i32.add $push20=, $0, $pop19
+; NO-SIMD128-FAST-NEXT:    i32.const $push64=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push17=, $7, $pop64
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push18=, $pop17, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop20), $pop18
+; NO-SIMD128-FAST-NEXT:    i32.const $push23=, 7
+; NO-SIMD128-FAST-NEXT:    i32.add $push24=, $0, $pop23
+; NO-SIMD128-FAST-NEXT:    i32.const $push63=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push21=, $8, $pop63
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push22=, $pop21, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop24), $pop22
+; NO-SIMD128-FAST-NEXT:    i32.const $push62=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push25=, $9, $pop62
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push26=, $pop25, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 8($0), $pop26
+; NO-SIMD128-FAST-NEXT:    i32.const $push29=, 9
+; NO-SIMD128-FAST-NEXT:    i32.add $push30=, $0, $pop29
+; NO-SIMD128-FAST-NEXT:    i32.const $push61=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push27=, $10, $pop61
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push28=, $pop27, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop30), $pop28
+; NO-SIMD128-FAST-NEXT:    i32.const $push33=, 10
+; NO-SIMD128-FAST-NEXT:    i32.add $push34=, $0, $pop33
+; NO-SIMD128-FAST-NEXT:    i32.const $push60=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push31=, $11, $pop60
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push32=, $pop31, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop34), $pop32
+; NO-SIMD128-FAST-NEXT:    i32.const $push37=, 11
+; NO-SIMD128-FAST-NEXT:    i32.add $push38=, $0, $pop37
+; NO-SIMD128-FAST-NEXT:    i32.const $push59=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push35=, $12, $pop59
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push36=, $pop35, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop38), $pop36
+; NO-SIMD128-FAST-NEXT:    i32.const $push41=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push42=, $0, $pop41
+; NO-SIMD128-FAST-NEXT:    i32.const $push58=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push39=, $13, $pop58
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push40=, $pop39, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop42), $pop40
+; NO-SIMD128-FAST-NEXT:    i32.const $push45=, 13
+; NO-SIMD128-FAST-NEXT:    i32.add $push46=, $0, $pop45
+; NO-SIMD128-FAST-NEXT:    i32.const $push57=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push43=, $14, $pop57
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push44=, $pop43, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop46), $pop44
+; NO-SIMD128-FAST-NEXT:    i32.const $push49=, 14
+; NO-SIMD128-FAST-NEXT:    i32.add $push50=, $0, $pop49
+; NO-SIMD128-FAST-NEXT:    i32.const $push56=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push47=, $15, $pop56
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push48=, $pop47, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop50), $pop48
+; NO-SIMD128-FAST-NEXT:    i32.const $push53=, 15
+; NO-SIMD128-FAST-NEXT:    i32.add $push54=, $0, $pop53
+; NO-SIMD128-FAST-NEXT:    i32.const $push55=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push51=, $16, $pop55
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push52=, $pop51, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop54), $pop52
+; NO-SIMD128-FAST-NEXT:    return
   %t = insertelement <16 x i8> undef, i8 %x, i32 0
   %s = shufflevector <16 x i8> %t, <16 x i8> undef,
     <16 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0,
                 i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
   %a = lshr <16 x i8> %v, %s
   ret <16 x i8> %a
 }
 
-; CHECK-LABEL: shr_u_vec_v16i8:
-; NO-SIMD128-NOT: i8x16
-; SIMD128-NEXT: .functype shr_u_vec_v16i8 (v128, v128) -> (v128){{$}}
-; SIMD128-NEXT: i8x16.extract_lane_u $push[[L0:[0-9]+]]=, $0, 0{{$}}
-; SIMD128-NEXT: i32.const $push[[M0:[0-9]+]]=, 7{{$}}
-; SIMD128-NEXT: i8x16.splat $push[[M1:[0-9]+]]=, $pop[[M0]]{{$}}
-; SIMD128-NEXT: v128.and $push[[M2:[0-9]+]]=, $1, $pop[[M1]]{{$}}
-; SIMD128-NEXT: local.tee $push[[M:[0-9]+]]=, $1=, $pop[[M2]]{{$}}
-; SIMD128-NEXT: i8x16.extract_lane_u $push[[L1:[0-9]+]]=, $pop[[M]], 0{{$}}
-; SIMD128-NEXT: i32.shr_u $push[[L2:[0-9]+]]=, $pop[[L0]], $pop[[L1]]{{$}}
-; SIMD128-NEXT: i8x16.splat $push[[L3:[0-9]+]]=, $pop[[L2]]{{$}}
-; Skip 14 lanes
-; SIMD128:      i8x16.extract_lane_u $push[[L4:[0-9]+]]=, $0, 15{{$}}
-; SIMD128-NEXT: i8x16.extract_lane_u $push[[L5:[0-9]+]]=, $1, 15{{$}}
-; SIMD128-NEXT: i32.shr_u $push[[L6:[0-9]+]]=, $pop[[L4]], $pop[[L5]]{{$}}
-; SIMD128-NEXT: i8x16.replace_lane $push[[R:[0-9]+]]=, $pop[[L7:[0-9]+]], 15, $pop[[L6]]{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <16 x i8> @shr_u_vec_v16i8(<16 x i8> %v, <16 x i8> %x) {
+; SIMD128-UNIMPL-SLOW-LABEL: shr_u_vec_v16i8:
+; SIMD128-UNIMPL-SLOW:         .functype shr_u_vec_v16i8 (v128, v128) -> (v128)
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push6=, $0, 0
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.const $push0=, 7
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.splat $push1=, $pop0
+; SIMD128-UNIMPL-SLOW-NEXT:    v128.and $push67=, $1, $pop1
+; SIMD128-UNIMPL-SLOW-NEXT:    local.tee $push66=, $1=, $pop67
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push5=, $pop66, 0
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shr_u $push7=, $pop6, $pop5
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.splat $push8=, $pop7
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push3=, $0, 1
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push2=, $1, 1
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shr_u $push4=, $pop3, $pop2
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.replace_lane $push9=, $pop8, 1, $pop4
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push11=, $0, 2
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push10=, $1, 2
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shr_u $push12=, $pop11, $pop10
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.replace_lane $push13=, $pop9, 2, $pop12
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push15=, $0, 3
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push14=, $1, 3
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shr_u $push16=, $pop15, $pop14
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.replace_lane $push17=, $pop13, 3, $pop16
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push19=, $0, 4
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push18=, $1, 4
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shr_u $push20=, $pop19, $pop18
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.replace_lane $push21=, $pop17, 4, $pop20
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push23=, $0, 5
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push22=, $1, 5
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shr_u $push24=, $pop23, $pop22
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.replace_lane $push25=, $pop21, 5, $pop24
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push27=, $0, 6
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push26=, $1, 6
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shr_u $push28=, $pop27, $pop26
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.replace_lane $push29=, $pop25, 6, $pop28
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push31=, $0, 7
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push30=, $1, 7
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shr_u $push32=, $pop31, $pop30
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.replace_lane $push33=, $pop29, 7, $pop32
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push35=, $0, 8
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push34=, $1, 8
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shr_u $push36=, $pop35, $pop34
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.replace_lane $push37=, $pop33, 8, $pop36
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push39=, $0, 9
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push38=, $1, 9
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shr_u $push40=, $pop39, $pop38
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.replace_lane $push41=, $pop37, 9, $pop40
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push43=, $0, 10
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push42=, $1, 10
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shr_u $push44=, $pop43, $pop42
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.replace_lane $push45=, $pop41, 10, $pop44
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push47=, $0, 11
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push46=, $1, 11
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shr_u $push48=, $pop47, $pop46
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.replace_lane $push49=, $pop45, 11, $pop48
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push51=, $0, 12
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push50=, $1, 12
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shr_u $push52=, $pop51, $pop50
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.replace_lane $push53=, $pop49, 12, $pop52
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push55=, $0, 13
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push54=, $1, 13
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shr_u $push56=, $pop55, $pop54
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.replace_lane $push57=, $pop53, 13, $pop56
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push59=, $0, 14
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push58=, $1, 14
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shr_u $push60=, $pop59, $pop58
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.replace_lane $push61=, $pop57, 14, $pop60
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push63=, $0, 15
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.extract_lane_u $push62=, $1, 15
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shr_u $push64=, $pop63, $pop62
+; SIMD128-UNIMPL-SLOW-NEXT:    i8x16.replace_lane $push65=, $pop61, 15, $pop64
+; SIMD128-UNIMPL-SLOW-NEXT:    return $pop65
+;
+; SIMD128-UNIMPL-FAST-LABEL: shr_u_vec_v16i8:
+; SIMD128-UNIMPL-FAST:         .functype shr_u_vec_v16i8 (v128, v128) -> (v128)
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push7=, $0, 0
+; SIMD128-UNIMPL-FAST-NEXT:    i32.const $push1=, 7
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.splat $push2=, $pop1
+; SIMD128-UNIMPL-FAST-NEXT:    v128.and $push67=, $1, $pop2
+; SIMD128-UNIMPL-FAST-NEXT:    local.tee $push66=, $1=, $pop67
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push6=, $pop66, 0
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shr_u $push8=, $pop7, $pop6
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.splat $push9=, $pop8
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push4=, $0, 1
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push3=, $1, 1
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shr_u $push5=, $pop4, $pop3
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.replace_lane $push10=, $pop9, 1, $pop5
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push12=, $0, 2
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push11=, $1, 2
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shr_u $push13=, $pop12, $pop11
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.replace_lane $push14=, $pop10, 2, $pop13
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push16=, $0, 3
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push15=, $1, 3
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shr_u $push17=, $pop16, $pop15
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.replace_lane $push18=, $pop14, 3, $pop17
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push20=, $0, 4
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push19=, $1, 4
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shr_u $push21=, $pop20, $pop19
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.replace_lane $push22=, $pop18, 4, $pop21
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push24=, $0, 5
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push23=, $1, 5
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shr_u $push25=, $pop24, $pop23
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.replace_lane $push26=, $pop22, 5, $pop25
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push28=, $0, 6
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push27=, $1, 6
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shr_u $push29=, $pop28, $pop27
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.replace_lane $push30=, $pop26, 6, $pop29
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push32=, $0, 7
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push31=, $1, 7
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shr_u $push33=, $pop32, $pop31
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.replace_lane $push34=, $pop30, 7, $pop33
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push36=, $0, 8
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push35=, $1, 8
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shr_u $push37=, $pop36, $pop35
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.replace_lane $push38=, $pop34, 8, $pop37
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push40=, $0, 9
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push39=, $1, 9
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shr_u $push41=, $pop40, $pop39
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.replace_lane $push42=, $pop38, 9, $pop41
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push44=, $0, 10
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push43=, $1, 10
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shr_u $push45=, $pop44, $pop43
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.replace_lane $push46=, $pop42, 10, $pop45
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push48=, $0, 11
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push47=, $1, 11
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shr_u $push49=, $pop48, $pop47
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.replace_lane $push50=, $pop46, 11, $pop49
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push52=, $0, 12
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push51=, $1, 12
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shr_u $push53=, $pop52, $pop51
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.replace_lane $push54=, $pop50, 12, $pop53
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push56=, $0, 13
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push55=, $1, 13
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shr_u $push57=, $pop56, $pop55
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.replace_lane $push58=, $pop54, 13, $pop57
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push60=, $0, 14
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push59=, $1, 14
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shr_u $push61=, $pop60, $pop59
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.replace_lane $push62=, $pop58, 14, $pop61
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push64=, $0, 15
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.extract_lane_u $push63=, $1, 15
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shr_u $push65=, $pop64, $pop63
+; SIMD128-UNIMPL-FAST-NEXT:    i8x16.replace_lane $push0=, $pop62, 15, $pop65
+; SIMD128-UNIMPL-FAST-NEXT:    return $pop0
+;
+; SIMD128-VM-SLOW-LABEL: shr_u_vec_v16i8:
+; SIMD128-VM-SLOW:         .functype shr_u_vec_v16i8 (v128, v128) -> (v128)
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push8=, $0, 0
+; SIMD128-VM-SLOW-NEXT:    i32.const $push1=, 255
+; SIMD128-VM-SLOW-NEXT:    i32.and $push9=, $pop8, $pop1
+; SIMD128-VM-SLOW-NEXT:    i32.const $push3=, 7
+; SIMD128-VM-SLOW-NEXT:    i8x16.splat $push4=, $pop3
+; SIMD128-VM-SLOW-NEXT:    v128.and $push131=, $1, $pop4
+; SIMD128-VM-SLOW-NEXT:    local.tee $push130=, $1=, $pop131
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push10=, $pop130, 0
+; SIMD128-VM-SLOW-NEXT:    i32.const $push129=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push11=, $pop10, $pop129
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push12=, $pop9, $pop11
+; SIMD128-VM-SLOW-NEXT:    i8x16.splat $push13=, $pop12
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push0=, $0, 1
+; SIMD128-VM-SLOW-NEXT:    i32.const $push128=, 255
+; SIMD128-VM-SLOW-NEXT:    i32.and $push2=, $pop0, $pop128
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push5=, $1, 1
+; SIMD128-VM-SLOW-NEXT:    i32.const $push127=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push6=, $pop5, $pop127
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push7=, $pop2, $pop6
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push14=, $pop13, 1, $pop7
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push15=, $0, 2
+; SIMD128-VM-SLOW-NEXT:    i32.const $push126=, 255
+; SIMD128-VM-SLOW-NEXT:    i32.and $push16=, $pop15, $pop126
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push17=, $1, 2
+; SIMD128-VM-SLOW-NEXT:    i32.const $push125=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push18=, $pop17, $pop125
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push19=, $pop16, $pop18
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push20=, $pop14, 2, $pop19
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push21=, $0, 3
+; SIMD128-VM-SLOW-NEXT:    i32.const $push124=, 255
+; SIMD128-VM-SLOW-NEXT:    i32.and $push22=, $pop21, $pop124
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push23=, $1, 3
+; SIMD128-VM-SLOW-NEXT:    i32.const $push123=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push24=, $pop23, $pop123
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push25=, $pop22, $pop24
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push26=, $pop20, 3, $pop25
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push27=, $0, 4
+; SIMD128-VM-SLOW-NEXT:    i32.const $push122=, 255
+; SIMD128-VM-SLOW-NEXT:    i32.and $push28=, $pop27, $pop122
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push29=, $1, 4
+; SIMD128-VM-SLOW-NEXT:    i32.const $push121=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push30=, $pop29, $pop121
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push31=, $pop28, $pop30
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push32=, $pop26, 4, $pop31
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push33=, $0, 5
+; SIMD128-VM-SLOW-NEXT:    i32.const $push120=, 255
+; SIMD128-VM-SLOW-NEXT:    i32.and $push34=, $pop33, $pop120
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push35=, $1, 5
+; SIMD128-VM-SLOW-NEXT:    i32.const $push119=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push36=, $pop35, $pop119
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push37=, $pop34, $pop36
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push38=, $pop32, 5, $pop37
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push39=, $0, 6
+; SIMD128-VM-SLOW-NEXT:    i32.const $push118=, 255
+; SIMD128-VM-SLOW-NEXT:    i32.and $push40=, $pop39, $pop118
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push41=, $1, 6
+; SIMD128-VM-SLOW-NEXT:    i32.const $push117=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push42=, $pop41, $pop117
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push43=, $pop40, $pop42
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push44=, $pop38, 6, $pop43
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push45=, $0, 7
+; SIMD128-VM-SLOW-NEXT:    i32.const $push116=, 255
+; SIMD128-VM-SLOW-NEXT:    i32.and $push46=, $pop45, $pop116
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push47=, $1, 7
+; SIMD128-VM-SLOW-NEXT:    i32.const $push115=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push48=, $pop47, $pop115
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push49=, $pop46, $pop48
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push50=, $pop44, 7, $pop49
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push51=, $0, 8
+; SIMD128-VM-SLOW-NEXT:    i32.const $push114=, 255
+; SIMD128-VM-SLOW-NEXT:    i32.and $push52=, $pop51, $pop114
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push53=, $1, 8
+; SIMD128-VM-SLOW-NEXT:    i32.const $push113=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push54=, $pop53, $pop113
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push55=, $pop52, $pop54
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push56=, $pop50, 8, $pop55
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push57=, $0, 9
+; SIMD128-VM-SLOW-NEXT:    i32.const $push112=, 255
+; SIMD128-VM-SLOW-NEXT:    i32.and $push58=, $pop57, $pop112
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push59=, $1, 9
+; SIMD128-VM-SLOW-NEXT:    i32.const $push111=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push60=, $pop59, $pop111
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push61=, $pop58, $pop60
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push62=, $pop56, 9, $pop61
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push63=, $0, 10
+; SIMD128-VM-SLOW-NEXT:    i32.const $push110=, 255
+; SIMD128-VM-SLOW-NEXT:    i32.and $push64=, $pop63, $pop110
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push65=, $1, 10
+; SIMD128-VM-SLOW-NEXT:    i32.const $push109=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push66=, $pop65, $pop109
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push67=, $pop64, $pop66
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push68=, $pop62, 10, $pop67
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push69=, $0, 11
+; SIMD128-VM-SLOW-NEXT:    i32.const $push108=, 255
+; SIMD128-VM-SLOW-NEXT:    i32.and $push70=, $pop69, $pop108
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push71=, $1, 11
+; SIMD128-VM-SLOW-NEXT:    i32.const $push107=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push72=, $pop71, $pop107
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push73=, $pop70, $pop72
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push74=, $pop68, 11, $pop73
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push75=, $0, 12
+; SIMD128-VM-SLOW-NEXT:    i32.const $push106=, 255
+; SIMD128-VM-SLOW-NEXT:    i32.and $push76=, $pop75, $pop106
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push77=, $1, 12
+; SIMD128-VM-SLOW-NEXT:    i32.const $push105=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push78=, $pop77, $pop105
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push79=, $pop76, $pop78
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push80=, $pop74, 12, $pop79
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push81=, $0, 13
+; SIMD128-VM-SLOW-NEXT:    i32.const $push104=, 255
+; SIMD128-VM-SLOW-NEXT:    i32.and $push82=, $pop81, $pop104
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push83=, $1, 13
+; SIMD128-VM-SLOW-NEXT:    i32.const $push103=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push84=, $pop83, $pop103
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push85=, $pop82, $pop84
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push86=, $pop80, 13, $pop85
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push87=, $0, 14
+; SIMD128-VM-SLOW-NEXT:    i32.const $push102=, 255
+; SIMD128-VM-SLOW-NEXT:    i32.and $push88=, $pop87, $pop102
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push89=, $1, 14
+; SIMD128-VM-SLOW-NEXT:    i32.const $push101=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push90=, $pop89, $pop101
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push91=, $pop88, $pop90
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push92=, $pop86, 14, $pop91
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push93=, $0, 15
+; SIMD128-VM-SLOW-NEXT:    i32.const $push100=, 255
+; SIMD128-VM-SLOW-NEXT:    i32.and $push94=, $pop93, $pop100
+; SIMD128-VM-SLOW-NEXT:    i8x16.extract_lane_s $push95=, $1, 15
+; SIMD128-VM-SLOW-NEXT:    i32.const $push99=, 7
+; SIMD128-VM-SLOW-NEXT:    i32.and $push96=, $pop95, $pop99
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push97=, $pop94, $pop96
+; SIMD128-VM-SLOW-NEXT:    i8x16.replace_lane $push98=, $pop92, 15, $pop97
+; SIMD128-VM-SLOW-NEXT:    return $pop98
+;
+; SIMD128-VM-FAST-LABEL: shr_u_vec_v16i8:
+; SIMD128-VM-FAST:         .functype shr_u_vec_v16i8 (v128, v128) -> (v128)
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push9=, $0, 0
+; SIMD128-VM-FAST-NEXT:    i32.const $push2=, 255
+; SIMD128-VM-FAST-NEXT:    i32.and $push10=, $pop9, $pop2
+; SIMD128-VM-FAST-NEXT:    i32.const $push4=, 7
+; SIMD128-VM-FAST-NEXT:    i8x16.splat $push5=, $pop4
+; SIMD128-VM-FAST-NEXT:    v128.and $push131=, $1, $pop5
+; SIMD128-VM-FAST-NEXT:    local.tee $push130=, $1=, $pop131
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push11=, $pop130, 0
+; SIMD128-VM-FAST-NEXT:    i32.const $push129=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push12=, $pop11, $pop129
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push13=, $pop10, $pop12
+; SIMD128-VM-FAST-NEXT:    i8x16.splat $push14=, $pop13
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push1=, $0, 1
+; SIMD128-VM-FAST-NEXT:    i32.const $push128=, 255
+; SIMD128-VM-FAST-NEXT:    i32.and $push3=, $pop1, $pop128
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push6=, $1, 1
+; SIMD128-VM-FAST-NEXT:    i32.const $push127=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push7=, $pop6, $pop127
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push8=, $pop3, $pop7
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push15=, $pop14, 1, $pop8
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push16=, $0, 2
+; SIMD128-VM-FAST-NEXT:    i32.const $push126=, 255
+; SIMD128-VM-FAST-NEXT:    i32.and $push17=, $pop16, $pop126
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push18=, $1, 2
+; SIMD128-VM-FAST-NEXT:    i32.const $push125=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push19=, $pop18, $pop125
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push20=, $pop17, $pop19
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push21=, $pop15, 2, $pop20
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push22=, $0, 3
+; SIMD128-VM-FAST-NEXT:    i32.const $push124=, 255
+; SIMD128-VM-FAST-NEXT:    i32.and $push23=, $pop22, $pop124
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push24=, $1, 3
+; SIMD128-VM-FAST-NEXT:    i32.const $push123=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push25=, $pop24, $pop123
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push26=, $pop23, $pop25
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push27=, $pop21, 3, $pop26
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push28=, $0, 4
+; SIMD128-VM-FAST-NEXT:    i32.const $push122=, 255
+; SIMD128-VM-FAST-NEXT:    i32.and $push29=, $pop28, $pop122
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push30=, $1, 4
+; SIMD128-VM-FAST-NEXT:    i32.const $push121=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push31=, $pop30, $pop121
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push32=, $pop29, $pop31
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push33=, $pop27, 4, $pop32
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push34=, $0, 5
+; SIMD128-VM-FAST-NEXT:    i32.const $push120=, 255
+; SIMD128-VM-FAST-NEXT:    i32.and $push35=, $pop34, $pop120
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push36=, $1, 5
+; SIMD128-VM-FAST-NEXT:    i32.const $push119=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push37=, $pop36, $pop119
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push38=, $pop35, $pop37
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push39=, $pop33, 5, $pop38
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push40=, $0, 6
+; SIMD128-VM-FAST-NEXT:    i32.const $push118=, 255
+; SIMD128-VM-FAST-NEXT:    i32.and $push41=, $pop40, $pop118
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push42=, $1, 6
+; SIMD128-VM-FAST-NEXT:    i32.const $push117=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push43=, $pop42, $pop117
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push44=, $pop41, $pop43
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push45=, $pop39, 6, $pop44
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push46=, $0, 7
+; SIMD128-VM-FAST-NEXT:    i32.const $push116=, 255
+; SIMD128-VM-FAST-NEXT:    i32.and $push47=, $pop46, $pop116
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push48=, $1, 7
+; SIMD128-VM-FAST-NEXT:    i32.const $push115=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push49=, $pop48, $pop115
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push50=, $pop47, $pop49
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push51=, $pop45, 7, $pop50
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push52=, $0, 8
+; SIMD128-VM-FAST-NEXT:    i32.const $push114=, 255
+; SIMD128-VM-FAST-NEXT:    i32.and $push53=, $pop52, $pop114
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push54=, $1, 8
+; SIMD128-VM-FAST-NEXT:    i32.const $push113=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push55=, $pop54, $pop113
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push56=, $pop53, $pop55
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push57=, $pop51, 8, $pop56
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push58=, $0, 9
+; SIMD128-VM-FAST-NEXT:    i32.const $push112=, 255
+; SIMD128-VM-FAST-NEXT:    i32.and $push59=, $pop58, $pop112
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push60=, $1, 9
+; SIMD128-VM-FAST-NEXT:    i32.const $push111=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push61=, $pop60, $pop111
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push62=, $pop59, $pop61
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push63=, $pop57, 9, $pop62
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push64=, $0, 10
+; SIMD128-VM-FAST-NEXT:    i32.const $push110=, 255
+; SIMD128-VM-FAST-NEXT:    i32.and $push65=, $pop64, $pop110
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push66=, $1, 10
+; SIMD128-VM-FAST-NEXT:    i32.const $push109=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push67=, $pop66, $pop109
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push68=, $pop65, $pop67
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push69=, $pop63, 10, $pop68
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push70=, $0, 11
+; SIMD128-VM-FAST-NEXT:    i32.const $push108=, 255
+; SIMD128-VM-FAST-NEXT:    i32.and $push71=, $pop70, $pop108
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push72=, $1, 11
+; SIMD128-VM-FAST-NEXT:    i32.const $push107=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push73=, $pop72, $pop107
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push74=, $pop71, $pop73
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push75=, $pop69, 11, $pop74
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push76=, $0, 12
+; SIMD128-VM-FAST-NEXT:    i32.const $push106=, 255
+; SIMD128-VM-FAST-NEXT:    i32.and $push77=, $pop76, $pop106
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push78=, $1, 12
+; SIMD128-VM-FAST-NEXT:    i32.const $push105=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push79=, $pop78, $pop105
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push80=, $pop77, $pop79
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push81=, $pop75, 12, $pop80
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push82=, $0, 13
+; SIMD128-VM-FAST-NEXT:    i32.const $push104=, 255
+; SIMD128-VM-FAST-NEXT:    i32.and $push83=, $pop82, $pop104
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push84=, $1, 13
+; SIMD128-VM-FAST-NEXT:    i32.const $push103=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push85=, $pop84, $pop103
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push86=, $pop83, $pop85
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push87=, $pop81, 13, $pop86
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push88=, $0, 14
+; SIMD128-VM-FAST-NEXT:    i32.const $push102=, 255
+; SIMD128-VM-FAST-NEXT:    i32.and $push89=, $pop88, $pop102
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push90=, $1, 14
+; SIMD128-VM-FAST-NEXT:    i32.const $push101=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push91=, $pop90, $pop101
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push92=, $pop89, $pop91
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push93=, $pop87, 14, $pop92
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push94=, $0, 15
+; SIMD128-VM-FAST-NEXT:    i32.const $push100=, 255
+; SIMD128-VM-FAST-NEXT:    i32.and $push95=, $pop94, $pop100
+; SIMD128-VM-FAST-NEXT:    i8x16.extract_lane_s $push96=, $1, 15
+; SIMD128-VM-FAST-NEXT:    i32.const $push99=, 7
+; SIMD128-VM-FAST-NEXT:    i32.and $push97=, $pop96, $pop99
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push98=, $pop95, $pop97
+; SIMD128-VM-FAST-NEXT:    i8x16.replace_lane $push0=, $pop93, 15, $pop98
+; SIMD128-VM-FAST-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: shr_u_vec_v16i8:
+; NO-SIMD128-SLOW:         .functype shr_u_vec_v16i8 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push4=, 15
+; NO-SIMD128-SLOW-NEXT:    i32.add $push5=, $0, $pop4
+; NO-SIMD128-SLOW-NEXT:    i32.const $push0=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push2=, $16, $pop0
+; NO-SIMD128-SLOW-NEXT:    i32.const $push101=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push1=, $32, $pop101
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push3=, $pop2, $pop1
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop5), $pop3
+; NO-SIMD128-SLOW-NEXT:    i32.const $push9=, 14
+; NO-SIMD128-SLOW-NEXT:    i32.add $push10=, $0, $pop9
+; NO-SIMD128-SLOW-NEXT:    i32.const $push100=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push7=, $15, $pop100
+; NO-SIMD128-SLOW-NEXT:    i32.const $push99=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push6=, $31, $pop99
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push8=, $pop7, $pop6
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop10), $pop8
+; NO-SIMD128-SLOW-NEXT:    i32.const $push14=, 13
+; NO-SIMD128-SLOW-NEXT:    i32.add $push15=, $0, $pop14
+; NO-SIMD128-SLOW-NEXT:    i32.const $push98=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push12=, $14, $pop98
+; NO-SIMD128-SLOW-NEXT:    i32.const $push97=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push11=, $30, $pop97
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push13=, $pop12, $pop11
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop15), $pop13
+; NO-SIMD128-SLOW-NEXT:    i32.const $push19=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push20=, $0, $pop19
+; NO-SIMD128-SLOW-NEXT:    i32.const $push96=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push17=, $13, $pop96
+; NO-SIMD128-SLOW-NEXT:    i32.const $push95=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push16=, $29, $pop95
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push18=, $pop17, $pop16
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop20), $pop18
+; NO-SIMD128-SLOW-NEXT:    i32.const $push24=, 11
+; NO-SIMD128-SLOW-NEXT:    i32.add $push25=, $0, $pop24
+; NO-SIMD128-SLOW-NEXT:    i32.const $push94=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push22=, $12, $pop94
+; NO-SIMD128-SLOW-NEXT:    i32.const $push93=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push21=, $28, $pop93
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push23=, $pop22, $pop21
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop25), $pop23
+; NO-SIMD128-SLOW-NEXT:    i32.const $push29=, 10
+; NO-SIMD128-SLOW-NEXT:    i32.add $push30=, $0, $pop29
+; NO-SIMD128-SLOW-NEXT:    i32.const $push92=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push27=, $11, $pop92
+; NO-SIMD128-SLOW-NEXT:    i32.const $push91=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push26=, $27, $pop91
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push28=, $pop27, $pop26
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop30), $pop28
+; NO-SIMD128-SLOW-NEXT:    i32.const $push34=, 9
+; NO-SIMD128-SLOW-NEXT:    i32.add $push35=, $0, $pop34
+; NO-SIMD128-SLOW-NEXT:    i32.const $push90=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push32=, $10, $pop90
+; NO-SIMD128-SLOW-NEXT:    i32.const $push89=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push31=, $26, $pop89
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push33=, $pop32, $pop31
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop35), $pop33
+; NO-SIMD128-SLOW-NEXT:    i32.const $push88=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push37=, $9, $pop88
+; NO-SIMD128-SLOW-NEXT:    i32.const $push87=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push36=, $25, $pop87
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push38=, $pop37, $pop36
+; NO-SIMD128-SLOW-NEXT:    i32.store8 8($0), $pop38
+; NO-SIMD128-SLOW-NEXT:    i32.const $push42=, 7
+; NO-SIMD128-SLOW-NEXT:    i32.add $push43=, $0, $pop42
+; NO-SIMD128-SLOW-NEXT:    i32.const $push86=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push40=, $8, $pop86
+; NO-SIMD128-SLOW-NEXT:    i32.const $push85=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push39=, $24, $pop85
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push41=, $pop40, $pop39
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop43), $pop41
+; NO-SIMD128-SLOW-NEXT:    i32.const $push47=, 6
+; NO-SIMD128-SLOW-NEXT:    i32.add $push48=, $0, $pop47
+; NO-SIMD128-SLOW-NEXT:    i32.const $push84=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push45=, $7, $pop84
+; NO-SIMD128-SLOW-NEXT:    i32.const $push83=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push44=, $23, $pop83
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push46=, $pop45, $pop44
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop48), $pop46
+; NO-SIMD128-SLOW-NEXT:    i32.const $push52=, 5
+; NO-SIMD128-SLOW-NEXT:    i32.add $push53=, $0, $pop52
+; NO-SIMD128-SLOW-NEXT:    i32.const $push82=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push50=, $6, $pop82
+; NO-SIMD128-SLOW-NEXT:    i32.const $push81=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push49=, $22, $pop81
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push51=, $pop50, $pop49
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop53), $pop51
+; NO-SIMD128-SLOW-NEXT:    i32.const $push80=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push55=, $5, $pop80
+; NO-SIMD128-SLOW-NEXT:    i32.const $push79=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push54=, $21, $pop79
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push56=, $pop55, $pop54
+; NO-SIMD128-SLOW-NEXT:    i32.store8 4($0), $pop56
+; NO-SIMD128-SLOW-NEXT:    i32.const $push60=, 3
+; NO-SIMD128-SLOW-NEXT:    i32.add $push61=, $0, $pop60
+; NO-SIMD128-SLOW-NEXT:    i32.const $push78=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push58=, $4, $pop78
+; NO-SIMD128-SLOW-NEXT:    i32.const $push77=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push57=, $20, $pop77
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push59=, $pop58, $pop57
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop61), $pop59
+; NO-SIMD128-SLOW-NEXT:    i32.const $push76=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push63=, $3, $pop76
+; NO-SIMD128-SLOW-NEXT:    i32.const $push75=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push62=, $19, $pop75
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push64=, $pop63, $pop62
+; NO-SIMD128-SLOW-NEXT:    i32.store8 2($0), $pop64
+; NO-SIMD128-SLOW-NEXT:    i32.const $push74=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push66=, $2, $pop74
+; NO-SIMD128-SLOW-NEXT:    i32.const $push73=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push65=, $18, $pop73
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push67=, $pop66, $pop65
+; NO-SIMD128-SLOW-NEXT:    i32.store8 1($0), $pop67
+; NO-SIMD128-SLOW-NEXT:    i32.const $push72=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push69=, $1, $pop72
+; NO-SIMD128-SLOW-NEXT:    i32.const $push71=, 255
+; NO-SIMD128-SLOW-NEXT:    i32.and $push68=, $17, $pop71
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push70=, $pop69, $pop68
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($0), $pop70
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: shr_u_vec_v16i8:
+; NO-SIMD128-FAST:         .functype shr_u_vec_v16i8 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-FAST-NEXT:    i32.const $push0=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push2=, $1, $pop0
+; NO-SIMD128-FAST-NEXT:    i32.const $push101=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push1=, $17, $pop101
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push3=, $pop2, $pop1
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($0), $pop3
+; NO-SIMD128-FAST-NEXT:    i32.const $push100=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push5=, $2, $pop100
+; NO-SIMD128-FAST-NEXT:    i32.const $push99=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push4=, $18, $pop99
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push6=, $pop5, $pop4
+; NO-SIMD128-FAST-NEXT:    i32.store8 1($0), $pop6
+; NO-SIMD128-FAST-NEXT:    i32.const $push98=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push8=, $3, $pop98
+; NO-SIMD128-FAST-NEXT:    i32.const $push97=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push7=, $19, $pop97
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push9=, $pop8, $pop7
+; NO-SIMD128-FAST-NEXT:    i32.store8 2($0), $pop9
+; NO-SIMD128-FAST-NEXT:    i32.const $push13=, 3
+; NO-SIMD128-FAST-NEXT:    i32.add $push14=, $0, $pop13
+; NO-SIMD128-FAST-NEXT:    i32.const $push96=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push11=, $4, $pop96
+; NO-SIMD128-FAST-NEXT:    i32.const $push95=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push10=, $20, $pop95
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push12=, $pop11, $pop10
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop14), $pop12
+; NO-SIMD128-FAST-NEXT:    i32.const $push94=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push16=, $5, $pop94
+; NO-SIMD128-FAST-NEXT:    i32.const $push93=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push15=, $21, $pop93
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push17=, $pop16, $pop15
+; NO-SIMD128-FAST-NEXT:    i32.store8 4($0), $pop17
+; NO-SIMD128-FAST-NEXT:    i32.const $push21=, 5
+; NO-SIMD128-FAST-NEXT:    i32.add $push22=, $0, $pop21
+; NO-SIMD128-FAST-NEXT:    i32.const $push92=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push19=, $6, $pop92
+; NO-SIMD128-FAST-NEXT:    i32.const $push91=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push18=, $22, $pop91
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push20=, $pop19, $pop18
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop22), $pop20
+; NO-SIMD128-FAST-NEXT:    i32.const $push26=, 6
+; NO-SIMD128-FAST-NEXT:    i32.add $push27=, $0, $pop26
+; NO-SIMD128-FAST-NEXT:    i32.const $push90=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push24=, $7, $pop90
+; NO-SIMD128-FAST-NEXT:    i32.const $push89=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push23=, $23, $pop89
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push25=, $pop24, $pop23
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop27), $pop25
+; NO-SIMD128-FAST-NEXT:    i32.const $push31=, 7
+; NO-SIMD128-FAST-NEXT:    i32.add $push32=, $0, $pop31
+; NO-SIMD128-FAST-NEXT:    i32.const $push88=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push29=, $8, $pop88
+; NO-SIMD128-FAST-NEXT:    i32.const $push87=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push28=, $24, $pop87
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push30=, $pop29, $pop28
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop32), $pop30
+; NO-SIMD128-FAST-NEXT:    i32.const $push86=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push34=, $9, $pop86
+; NO-SIMD128-FAST-NEXT:    i32.const $push85=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push33=, $25, $pop85
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push35=, $pop34, $pop33
+; NO-SIMD128-FAST-NEXT:    i32.store8 8($0), $pop35
+; NO-SIMD128-FAST-NEXT:    i32.const $push39=, 9
+; NO-SIMD128-FAST-NEXT:    i32.add $push40=, $0, $pop39
+; NO-SIMD128-FAST-NEXT:    i32.const $push84=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push37=, $10, $pop84
+; NO-SIMD128-FAST-NEXT:    i32.const $push83=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push36=, $26, $pop83
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push38=, $pop37, $pop36
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop40), $pop38
+; NO-SIMD128-FAST-NEXT:    i32.const $push44=, 10
+; NO-SIMD128-FAST-NEXT:    i32.add $push45=, $0, $pop44
+; NO-SIMD128-FAST-NEXT:    i32.const $push82=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push42=, $11, $pop82
+; NO-SIMD128-FAST-NEXT:    i32.const $push81=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push41=, $27, $pop81
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push43=, $pop42, $pop41
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop45), $pop43
+; NO-SIMD128-FAST-NEXT:    i32.const $push49=, 11
+; NO-SIMD128-FAST-NEXT:    i32.add $push50=, $0, $pop49
+; NO-SIMD128-FAST-NEXT:    i32.const $push80=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push47=, $12, $pop80
+; NO-SIMD128-FAST-NEXT:    i32.const $push79=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push46=, $28, $pop79
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push48=, $pop47, $pop46
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop50), $pop48
+; NO-SIMD128-FAST-NEXT:    i32.const $push54=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push55=, $0, $pop54
+; NO-SIMD128-FAST-NEXT:    i32.const $push78=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push52=, $13, $pop78
+; NO-SIMD128-FAST-NEXT:    i32.const $push77=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push51=, $29, $pop77
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push53=, $pop52, $pop51
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop55), $pop53
+; NO-SIMD128-FAST-NEXT:    i32.const $push59=, 13
+; NO-SIMD128-FAST-NEXT:    i32.add $push60=, $0, $pop59
+; NO-SIMD128-FAST-NEXT:    i32.const $push76=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push57=, $14, $pop76
+; NO-SIMD128-FAST-NEXT:    i32.const $push75=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push56=, $30, $pop75
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push58=, $pop57, $pop56
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop60), $pop58
+; NO-SIMD128-FAST-NEXT:    i32.const $push64=, 14
+; NO-SIMD128-FAST-NEXT:    i32.add $push65=, $0, $pop64
+; NO-SIMD128-FAST-NEXT:    i32.const $push74=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push62=, $15, $pop74
+; NO-SIMD128-FAST-NEXT:    i32.const $push73=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push61=, $31, $pop73
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push63=, $pop62, $pop61
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop65), $pop63
+; NO-SIMD128-FAST-NEXT:    i32.const $push69=, 15
+; NO-SIMD128-FAST-NEXT:    i32.add $push70=, $0, $pop69
+; NO-SIMD128-FAST-NEXT:    i32.const $push72=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push67=, $16, $pop72
+; NO-SIMD128-FAST-NEXT:    i32.const $push71=, 255
+; NO-SIMD128-FAST-NEXT:    i32.and $push66=, $32, $pop71
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push68=, $pop67, $pop66
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop70), $pop68
+; NO-SIMD128-FAST-NEXT:    return
   %a = lshr <16 x i8> %v, %x
   ret <16 x i8> %a
 }
 
-; CHECK-LABEL: and_v16i8:
-; NO-SIMD128-NOT: v128
-; SIMD128-NEXT: .functype and_v16i8 (v128, v128) -> (v128){{$}}
-; SIMD128-NEXT: v128.and $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <16 x i8> @and_v16i8(<16 x i8> %x, <16 x i8> %y) {
+; SIMD128-LABEL: and_v16i8:
+; SIMD128:         .functype and_v16i8 (v128, v128) -> (v128)
+; SIMD128-NEXT:    v128.and $push0=, $0, $1
+; SIMD128-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: and_v16i8:
+; NO-SIMD128-SLOW:         .functype and_v16i8 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push1=, 15
+; NO-SIMD128-SLOW-NEXT:    i32.add $push2=, $0, $pop1
+; NO-SIMD128-SLOW-NEXT:    i32.and $push0=, $16, $32
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop2), $pop0
+; NO-SIMD128-SLOW-NEXT:    i32.const $push4=, 14
+; NO-SIMD128-SLOW-NEXT:    i32.add $push5=, $0, $pop4
+; NO-SIMD128-SLOW-NEXT:    i32.and $push3=, $15, $31
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop5), $pop3
+; NO-SIMD128-SLOW-NEXT:    i32.const $push7=, 13
+; NO-SIMD128-SLOW-NEXT:    i32.add $push8=, $0, $pop7
+; NO-SIMD128-SLOW-NEXT:    i32.and $push6=, $14, $30
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop8), $pop6
+; NO-SIMD128-SLOW-NEXT:    i32.const $push10=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push11=, $0, $pop10
+; NO-SIMD128-SLOW-NEXT:    i32.and $push9=, $13, $29
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop11), $pop9
+; NO-SIMD128-SLOW-NEXT:    i32.const $push13=, 11
+; NO-SIMD128-SLOW-NEXT:    i32.add $push14=, $0, $pop13
+; NO-SIMD128-SLOW-NEXT:    i32.and $push12=, $12, $28
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop14), $pop12
+; NO-SIMD128-SLOW-NEXT:    i32.const $push16=, 10
+; NO-SIMD128-SLOW-NEXT:    i32.add $push17=, $0, $pop16
+; NO-SIMD128-SLOW-NEXT:    i32.and $push15=, $11, $27
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop17), $pop15
+; NO-SIMD128-SLOW-NEXT:    i32.const $push19=, 9
+; NO-SIMD128-SLOW-NEXT:    i32.add $push20=, $0, $pop19
+; NO-SIMD128-SLOW-NEXT:    i32.and $push18=, $10, $26
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop20), $pop18
+; NO-SIMD128-SLOW-NEXT:    i32.and $push21=, $9, $25
+; NO-SIMD128-SLOW-NEXT:    i32.store8 8($0), $pop21
+; NO-SIMD128-SLOW-NEXT:    i32.const $push23=, 7
+; NO-SIMD128-SLOW-NEXT:    i32.add $push24=, $0, $pop23
+; NO-SIMD128-SLOW-NEXT:    i32.and $push22=, $8, $24
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop24), $pop22
+; NO-SIMD128-SLOW-NEXT:    i32.const $push26=, 6
+; NO-SIMD128-SLOW-NEXT:    i32.add $push27=, $0, $pop26
+; NO-SIMD128-SLOW-NEXT:    i32.and $push25=, $7, $23
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop27), $pop25
+; NO-SIMD128-SLOW-NEXT:    i32.const $push29=, 5
+; NO-SIMD128-SLOW-NEXT:    i32.add $push30=, $0, $pop29
+; NO-SIMD128-SLOW-NEXT:    i32.and $push28=, $6, $22
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop30), $pop28
+; NO-SIMD128-SLOW-NEXT:    i32.and $push31=, $5, $21
+; NO-SIMD128-SLOW-NEXT:    i32.store8 4($0), $pop31
+; NO-SIMD128-SLOW-NEXT:    i32.const $push33=, 3
+; NO-SIMD128-SLOW-NEXT:    i32.add $push34=, $0, $pop33
+; NO-SIMD128-SLOW-NEXT:    i32.and $push32=, $4, $20
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop34), $pop32
+; NO-SIMD128-SLOW-NEXT:    i32.and $push35=, $3, $19
+; NO-SIMD128-SLOW-NEXT:    i32.store8 2($0), $pop35
+; NO-SIMD128-SLOW-NEXT:    i32.and $push36=, $2, $18
+; NO-SIMD128-SLOW-NEXT:    i32.store8 1($0), $pop36
+; NO-SIMD128-SLOW-NEXT:    i32.and $push37=, $1, $17
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($0), $pop37
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: and_v16i8:
+; NO-SIMD128-FAST:         .functype and_v16i8 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-FAST-NEXT:    i32.and $push0=, $1, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($0), $pop0
+; NO-SIMD128-FAST-NEXT:    i32.and $push1=, $2, $18
+; NO-SIMD128-FAST-NEXT:    i32.store8 1($0), $pop1
+; NO-SIMD128-FAST-NEXT:    i32.and $push2=, $3, $19
+; NO-SIMD128-FAST-NEXT:    i32.store8 2($0), $pop2
+; NO-SIMD128-FAST-NEXT:    i32.const $push3=, 3
+; NO-SIMD128-FAST-NEXT:    i32.add $push4=, $0, $pop3
+; NO-SIMD128-FAST-NEXT:    i32.and $push5=, $4, $20
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop4), $pop5
+; NO-SIMD128-FAST-NEXT:    i32.and $push6=, $5, $21
+; NO-SIMD128-FAST-NEXT:    i32.store8 4($0), $pop6
+; NO-SIMD128-FAST-NEXT:    i32.const $push7=, 5
+; NO-SIMD128-FAST-NEXT:    i32.add $push8=, $0, $pop7
+; NO-SIMD128-FAST-NEXT:    i32.and $push9=, $6, $22
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop8), $pop9
+; NO-SIMD128-FAST-NEXT:    i32.const $push10=, 6
+; NO-SIMD128-FAST-NEXT:    i32.add $push11=, $0, $pop10
+; NO-SIMD128-FAST-NEXT:    i32.and $push12=, $7, $23
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop11), $pop12
+; NO-SIMD128-FAST-NEXT:    i32.const $push13=, 7
+; NO-SIMD128-FAST-NEXT:    i32.add $push14=, $0, $pop13
+; NO-SIMD128-FAST-NEXT:    i32.and $push15=, $8, $24
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop14), $pop15
+; NO-SIMD128-FAST-NEXT:    i32.and $push16=, $9, $25
+; NO-SIMD128-FAST-NEXT:    i32.store8 8($0), $pop16
+; NO-SIMD128-FAST-NEXT:    i32.const $push17=, 9
+; NO-SIMD128-FAST-NEXT:    i32.add $push18=, $0, $pop17
+; NO-SIMD128-FAST-NEXT:    i32.and $push19=, $10, $26
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop18), $pop19
+; NO-SIMD128-FAST-NEXT:    i32.const $push20=, 10
+; NO-SIMD128-FAST-NEXT:    i32.add $push21=, $0, $pop20
+; NO-SIMD128-FAST-NEXT:    i32.and $push22=, $11, $27
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop21), $pop22
+; NO-SIMD128-FAST-NEXT:    i32.const $push23=, 11
+; NO-SIMD128-FAST-NEXT:    i32.add $push24=, $0, $pop23
+; NO-SIMD128-FAST-NEXT:    i32.and $push25=, $12, $28
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop24), $pop25
+; NO-SIMD128-FAST-NEXT:    i32.const $push26=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push27=, $0, $pop26
+; NO-SIMD128-FAST-NEXT:    i32.and $push28=, $13, $29
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop27), $pop28
+; NO-SIMD128-FAST-NEXT:    i32.const $push29=, 13
+; NO-SIMD128-FAST-NEXT:    i32.add $push30=, $0, $pop29
+; NO-SIMD128-FAST-NEXT:    i32.and $push31=, $14, $30
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop30), $pop31
+; NO-SIMD128-FAST-NEXT:    i32.const $push32=, 14
+; NO-SIMD128-FAST-NEXT:    i32.add $push33=, $0, $pop32
+; NO-SIMD128-FAST-NEXT:    i32.and $push34=, $15, $31
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop33), $pop34
+; NO-SIMD128-FAST-NEXT:    i32.const $push35=, 15
+; NO-SIMD128-FAST-NEXT:    i32.add $push36=, $0, $pop35
+; NO-SIMD128-FAST-NEXT:    i32.and $push37=, $16, $32
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop36), $pop37
+; NO-SIMD128-FAST-NEXT:    return
   %a = and <16 x i8> %x, %y
   ret <16 x i8> %a
 }
 
-; CHECK-LABEL: or_v16i8:
-; NO-SIMD128-NOT: v128
-; SIMD128-NEXT: .functype or_v16i8 (v128, v128) -> (v128){{$}}
-; SIMD128-NEXT: v128.or $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <16 x i8> @or_v16i8(<16 x i8> %x, <16 x i8> %y) {
+; SIMD128-LABEL: or_v16i8:
+; SIMD128:         .functype or_v16i8 (v128, v128) -> (v128)
+; SIMD128-NEXT:    v128.or $push0=, $0, $1
+; SIMD128-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: or_v16i8:
+; NO-SIMD128-SLOW:         .functype or_v16i8 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push1=, 15
+; NO-SIMD128-SLOW-NEXT:    i32.add $push2=, $0, $pop1
+; NO-SIMD128-SLOW-NEXT:    i32.or $push0=, $16, $32
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop2), $pop0
+; NO-SIMD128-SLOW-NEXT:    i32.const $push4=, 14
+; NO-SIMD128-SLOW-NEXT:    i32.add $push5=, $0, $pop4
+; NO-SIMD128-SLOW-NEXT:    i32.or $push3=, $15, $31
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop5), $pop3
+; NO-SIMD128-SLOW-NEXT:    i32.const $push7=, 13
+; NO-SIMD128-SLOW-NEXT:    i32.add $push8=, $0, $pop7
+; NO-SIMD128-SLOW-NEXT:    i32.or $push6=, $14, $30
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop8), $pop6
+; NO-SIMD128-SLOW-NEXT:    i32.const $push10=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push11=, $0, $pop10
+; NO-SIMD128-SLOW-NEXT:    i32.or $push9=, $13, $29
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop11), $pop9
+; NO-SIMD128-SLOW-NEXT:    i32.const $push13=, 11
+; NO-SIMD128-SLOW-NEXT:    i32.add $push14=, $0, $pop13
+; NO-SIMD128-SLOW-NEXT:    i32.or $push12=, $12, $28
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop14), $pop12
+; NO-SIMD128-SLOW-NEXT:    i32.const $push16=, 10
+; NO-SIMD128-SLOW-NEXT:    i32.add $push17=, $0, $pop16
+; NO-SIMD128-SLOW-NEXT:    i32.or $push15=, $11, $27
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop17), $pop15
+; NO-SIMD128-SLOW-NEXT:    i32.const $push19=, 9
+; NO-SIMD128-SLOW-NEXT:    i32.add $push20=, $0, $pop19
+; NO-SIMD128-SLOW-NEXT:    i32.or $push18=, $10, $26
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop20), $pop18
+; NO-SIMD128-SLOW-NEXT:    i32.or $push21=, $9, $25
+; NO-SIMD128-SLOW-NEXT:    i32.store8 8($0), $pop21
+; NO-SIMD128-SLOW-NEXT:    i32.const $push23=, 7
+; NO-SIMD128-SLOW-NEXT:    i32.add $push24=, $0, $pop23
+; NO-SIMD128-SLOW-NEXT:    i32.or $push22=, $8, $24
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop24), $pop22
+; NO-SIMD128-SLOW-NEXT:    i32.const $push26=, 6
+; NO-SIMD128-SLOW-NEXT:    i32.add $push27=, $0, $pop26
+; NO-SIMD128-SLOW-NEXT:    i32.or $push25=, $7, $23
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop27), $pop25
+; NO-SIMD128-SLOW-NEXT:    i32.const $push29=, 5
+; NO-SIMD128-SLOW-NEXT:    i32.add $push30=, $0, $pop29
+; NO-SIMD128-SLOW-NEXT:    i32.or $push28=, $6, $22
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop30), $pop28
+; NO-SIMD128-SLOW-NEXT:    i32.or $push31=, $5, $21
+; NO-SIMD128-SLOW-NEXT:    i32.store8 4($0), $pop31
+; NO-SIMD128-SLOW-NEXT:    i32.const $push33=, 3
+; NO-SIMD128-SLOW-NEXT:    i32.add $push34=, $0, $pop33
+; NO-SIMD128-SLOW-NEXT:    i32.or $push32=, $4, $20
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop34), $pop32
+; NO-SIMD128-SLOW-NEXT:    i32.or $push35=, $3, $19
+; NO-SIMD128-SLOW-NEXT:    i32.store8 2($0), $pop35
+; NO-SIMD128-SLOW-NEXT:    i32.or $push36=, $2, $18
+; NO-SIMD128-SLOW-NEXT:    i32.store8 1($0), $pop36
+; NO-SIMD128-SLOW-NEXT:    i32.or $push37=, $1, $17
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($0), $pop37
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: or_v16i8:
+; NO-SIMD128-FAST:         .functype or_v16i8 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-FAST-NEXT:    i32.or $push0=, $1, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($0), $pop0
+; NO-SIMD128-FAST-NEXT:    i32.or $push1=, $2, $18
+; NO-SIMD128-FAST-NEXT:    i32.store8 1($0), $pop1
+; NO-SIMD128-FAST-NEXT:    i32.or $push2=, $3, $19
+; NO-SIMD128-FAST-NEXT:    i32.store8 2($0), $pop2
+; NO-SIMD128-FAST-NEXT:    i32.const $push3=, 3
+; NO-SIMD128-FAST-NEXT:    i32.add $push4=, $0, $pop3
+; NO-SIMD128-FAST-NEXT:    i32.or $push5=, $4, $20
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop4), $pop5
+; NO-SIMD128-FAST-NEXT:    i32.or $push6=, $5, $21
+; NO-SIMD128-FAST-NEXT:    i32.store8 4($0), $pop6
+; NO-SIMD128-FAST-NEXT:    i32.const $push7=, 5
+; NO-SIMD128-FAST-NEXT:    i32.add $push8=, $0, $pop7
+; NO-SIMD128-FAST-NEXT:    i32.or $push9=, $6, $22
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop8), $pop9
+; NO-SIMD128-FAST-NEXT:    i32.const $push10=, 6
+; NO-SIMD128-FAST-NEXT:    i32.add $push11=, $0, $pop10
+; NO-SIMD128-FAST-NEXT:    i32.or $push12=, $7, $23
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop11), $pop12
+; NO-SIMD128-FAST-NEXT:    i32.const $push13=, 7
+; NO-SIMD128-FAST-NEXT:    i32.add $push14=, $0, $pop13
+; NO-SIMD128-FAST-NEXT:    i32.or $push15=, $8, $24
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop14), $pop15
+; NO-SIMD128-FAST-NEXT:    i32.or $push16=, $9, $25
+; NO-SIMD128-FAST-NEXT:    i32.store8 8($0), $pop16
+; NO-SIMD128-FAST-NEXT:    i32.const $push17=, 9
+; NO-SIMD128-FAST-NEXT:    i32.add $push18=, $0, $pop17
+; NO-SIMD128-FAST-NEXT:    i32.or $push19=, $10, $26
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop18), $pop19
+; NO-SIMD128-FAST-NEXT:    i32.const $push20=, 10
+; NO-SIMD128-FAST-NEXT:    i32.add $push21=, $0, $pop20
+; NO-SIMD128-FAST-NEXT:    i32.or $push22=, $11, $27
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop21), $pop22
+; NO-SIMD128-FAST-NEXT:    i32.const $push23=, 11
+; NO-SIMD128-FAST-NEXT:    i32.add $push24=, $0, $pop23
+; NO-SIMD128-FAST-NEXT:    i32.or $push25=, $12, $28
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop24), $pop25
+; NO-SIMD128-FAST-NEXT:    i32.const $push26=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push27=, $0, $pop26
+; NO-SIMD128-FAST-NEXT:    i32.or $push28=, $13, $29
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop27), $pop28
+; NO-SIMD128-FAST-NEXT:    i32.const $push29=, 13
+; NO-SIMD128-FAST-NEXT:    i32.add $push30=, $0, $pop29
+; NO-SIMD128-FAST-NEXT:    i32.or $push31=, $14, $30
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop30), $pop31
+; NO-SIMD128-FAST-NEXT:    i32.const $push32=, 14
+; NO-SIMD128-FAST-NEXT:    i32.add $push33=, $0, $pop32
+; NO-SIMD128-FAST-NEXT:    i32.or $push34=, $15, $31
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop33), $pop34
+; NO-SIMD128-FAST-NEXT:    i32.const $push35=, 15
+; NO-SIMD128-FAST-NEXT:    i32.add $push36=, $0, $pop35
+; NO-SIMD128-FAST-NEXT:    i32.or $push37=, $16, $32
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop36), $pop37
+; NO-SIMD128-FAST-NEXT:    return
   %a = or <16 x i8> %x, %y
   ret <16 x i8> %a
 }
 
-; CHECK-LABEL: xor_v16i8:
-; NO-SIMD128-NOT: v128
-; SIMD128-NEXT: .functype xor_v16i8 (v128, v128) -> (v128){{$}}
-; SIMD128-NEXT: v128.xor $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <16 x i8> @xor_v16i8(<16 x i8> %x, <16 x i8> %y) {
+; SIMD128-LABEL: xor_v16i8:
+; SIMD128:         .functype xor_v16i8 (v128, v128) -> (v128)
+; SIMD128-NEXT:    v128.xor $push0=, $0, $1
+; SIMD128-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: xor_v16i8:
+; NO-SIMD128-SLOW:         .functype xor_v16i8 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push1=, 15
+; NO-SIMD128-SLOW-NEXT:    i32.add $push2=, $0, $pop1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push0=, $16, $32
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop2), $pop0
+; NO-SIMD128-SLOW-NEXT:    i32.const $push4=, 14
+; NO-SIMD128-SLOW-NEXT:    i32.add $push5=, $0, $pop4
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push3=, $15, $31
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop5), $pop3
+; NO-SIMD128-SLOW-NEXT:    i32.const $push7=, 13
+; NO-SIMD128-SLOW-NEXT:    i32.add $push8=, $0, $pop7
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push6=, $14, $30
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop8), $pop6
+; NO-SIMD128-SLOW-NEXT:    i32.const $push10=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push11=, $0, $pop10
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push9=, $13, $29
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop11), $pop9
+; NO-SIMD128-SLOW-NEXT:    i32.const $push13=, 11
+; NO-SIMD128-SLOW-NEXT:    i32.add $push14=, $0, $pop13
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push12=, $12, $28
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop14), $pop12
+; NO-SIMD128-SLOW-NEXT:    i32.const $push16=, 10
+; NO-SIMD128-SLOW-NEXT:    i32.add $push17=, $0, $pop16
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push15=, $11, $27
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop17), $pop15
+; NO-SIMD128-SLOW-NEXT:    i32.const $push19=, 9
+; NO-SIMD128-SLOW-NEXT:    i32.add $push20=, $0, $pop19
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push18=, $10, $26
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop20), $pop18
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push21=, $9, $25
+; NO-SIMD128-SLOW-NEXT:    i32.store8 8($0), $pop21
+; NO-SIMD128-SLOW-NEXT:    i32.const $push23=, 7
+; NO-SIMD128-SLOW-NEXT:    i32.add $push24=, $0, $pop23
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push22=, $8, $24
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop24), $pop22
+; NO-SIMD128-SLOW-NEXT:    i32.const $push26=, 6
+; NO-SIMD128-SLOW-NEXT:    i32.add $push27=, $0, $pop26
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push25=, $7, $23
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop27), $pop25
+; NO-SIMD128-SLOW-NEXT:    i32.const $push29=, 5
+; NO-SIMD128-SLOW-NEXT:    i32.add $push30=, $0, $pop29
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push28=, $6, $22
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop30), $pop28
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push31=, $5, $21
+; NO-SIMD128-SLOW-NEXT:    i32.store8 4($0), $pop31
+; NO-SIMD128-SLOW-NEXT:    i32.const $push33=, 3
+; NO-SIMD128-SLOW-NEXT:    i32.add $push34=, $0, $pop33
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push32=, $4, $20
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop34), $pop32
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push35=, $3, $19
+; NO-SIMD128-SLOW-NEXT:    i32.store8 2($0), $pop35
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push36=, $2, $18
+; NO-SIMD128-SLOW-NEXT:    i32.store8 1($0), $pop36
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push37=, $1, $17
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($0), $pop37
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: xor_v16i8:
+; NO-SIMD128-FAST:         .functype xor_v16i8 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-FAST-NEXT:    i32.xor $push0=, $1, $17
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($0), $pop0
+; NO-SIMD128-FAST-NEXT:    i32.xor $push1=, $2, $18
+; NO-SIMD128-FAST-NEXT:    i32.store8 1($0), $pop1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push2=, $3, $19
+; NO-SIMD128-FAST-NEXT:    i32.store8 2($0), $pop2
+; NO-SIMD128-FAST-NEXT:    i32.const $push3=, 3
+; NO-SIMD128-FAST-NEXT:    i32.add $push4=, $0, $pop3
+; NO-SIMD128-FAST-NEXT:    i32.xor $push5=, $4, $20
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop4), $pop5
+; NO-SIMD128-FAST-NEXT:    i32.xor $push6=, $5, $21
+; NO-SIMD128-FAST-NEXT:    i32.store8 4($0), $pop6
+; NO-SIMD128-FAST-NEXT:    i32.const $push7=, 5
+; NO-SIMD128-FAST-NEXT:    i32.add $push8=, $0, $pop7
+; NO-SIMD128-FAST-NEXT:    i32.xor $push9=, $6, $22
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop8), $pop9
+; NO-SIMD128-FAST-NEXT:    i32.const $push10=, 6
+; NO-SIMD128-FAST-NEXT:    i32.add $push11=, $0, $pop10
+; NO-SIMD128-FAST-NEXT:    i32.xor $push12=, $7, $23
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop11), $pop12
+; NO-SIMD128-FAST-NEXT:    i32.const $push13=, 7
+; NO-SIMD128-FAST-NEXT:    i32.add $push14=, $0, $pop13
+; NO-SIMD128-FAST-NEXT:    i32.xor $push15=, $8, $24
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop14), $pop15
+; NO-SIMD128-FAST-NEXT:    i32.xor $push16=, $9, $25
+; NO-SIMD128-FAST-NEXT:    i32.store8 8($0), $pop16
+; NO-SIMD128-FAST-NEXT:    i32.const $push17=, 9
+; NO-SIMD128-FAST-NEXT:    i32.add $push18=, $0, $pop17
+; NO-SIMD128-FAST-NEXT:    i32.xor $push19=, $10, $26
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop18), $pop19
+; NO-SIMD128-FAST-NEXT:    i32.const $push20=, 10
+; NO-SIMD128-FAST-NEXT:    i32.add $push21=, $0, $pop20
+; NO-SIMD128-FAST-NEXT:    i32.xor $push22=, $11, $27
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop21), $pop22
+; NO-SIMD128-FAST-NEXT:    i32.const $push23=, 11
+; NO-SIMD128-FAST-NEXT:    i32.add $push24=, $0, $pop23
+; NO-SIMD128-FAST-NEXT:    i32.xor $push25=, $12, $28
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop24), $pop25
+; NO-SIMD128-FAST-NEXT:    i32.const $push26=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push27=, $0, $pop26
+; NO-SIMD128-FAST-NEXT:    i32.xor $push28=, $13, $29
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop27), $pop28
+; NO-SIMD128-FAST-NEXT:    i32.const $push29=, 13
+; NO-SIMD128-FAST-NEXT:    i32.add $push30=, $0, $pop29
+; NO-SIMD128-FAST-NEXT:    i32.xor $push31=, $14, $30
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop30), $pop31
+; NO-SIMD128-FAST-NEXT:    i32.const $push32=, 14
+; NO-SIMD128-FAST-NEXT:    i32.add $push33=, $0, $pop32
+; NO-SIMD128-FAST-NEXT:    i32.xor $push34=, $15, $31
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop33), $pop34
+; NO-SIMD128-FAST-NEXT:    i32.const $push35=, 15
+; NO-SIMD128-FAST-NEXT:    i32.add $push36=, $0, $pop35
+; NO-SIMD128-FAST-NEXT:    i32.xor $push37=, $16, $32
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop36), $pop37
+; NO-SIMD128-FAST-NEXT:    return
   %a = xor <16 x i8> %x, %y
   ret <16 x i8> %a
 }
 
-; CHECK-LABEL: not_v16i8:
-; NO-SIMD128-NOT: v128
-; SIMD128-NEXT: .functype not_v16i8 (v128) -> (v128){{$}}
-; SIMD128-NEXT: v128.not $push[[R:[0-9]+]]=, $0{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <16 x i8> @not_v16i8(<16 x i8> %x) {
+; SIMD128-LABEL: not_v16i8:
+; SIMD128:         .functype not_v16i8 (v128) -> (v128)
+; SIMD128-NEXT:    v128.not $push0=, $0
+; SIMD128-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: not_v16i8:
+; NO-SIMD128-SLOW:         .functype not_v16i8 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push2=, 15
+; NO-SIMD128-SLOW-NEXT:    i32.add $push3=, $0, $pop2
+; NO-SIMD128-SLOW-NEXT:    i32.const $push0=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push1=, $16, $pop0
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop3), $pop1
+; NO-SIMD128-SLOW-NEXT:    i32.const $push5=, 14
+; NO-SIMD128-SLOW-NEXT:    i32.add $push6=, $0, $pop5
+; NO-SIMD128-SLOW-NEXT:    i32.const $push53=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push4=, $15, $pop53
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop6), $pop4
+; NO-SIMD128-SLOW-NEXT:    i32.const $push8=, 13
+; NO-SIMD128-SLOW-NEXT:    i32.add $push9=, $0, $pop8
+; NO-SIMD128-SLOW-NEXT:    i32.const $push52=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push7=, $14, $pop52
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop9), $pop7
+; NO-SIMD128-SLOW-NEXT:    i32.const $push11=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push12=, $0, $pop11
+; NO-SIMD128-SLOW-NEXT:    i32.const $push51=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push10=, $13, $pop51
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop12), $pop10
+; NO-SIMD128-SLOW-NEXT:    i32.const $push14=, 11
+; NO-SIMD128-SLOW-NEXT:    i32.add $push15=, $0, $pop14
+; NO-SIMD128-SLOW-NEXT:    i32.const $push50=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push13=, $12, $pop50
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop15), $pop13
+; NO-SIMD128-SLOW-NEXT:    i32.const $push17=, 10
+; NO-SIMD128-SLOW-NEXT:    i32.add $push18=, $0, $pop17
+; NO-SIMD128-SLOW-NEXT:    i32.const $push49=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push16=, $11, $pop49
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop18), $pop16
+; NO-SIMD128-SLOW-NEXT:    i32.const $push20=, 9
+; NO-SIMD128-SLOW-NEXT:    i32.add $push21=, $0, $pop20
+; NO-SIMD128-SLOW-NEXT:    i32.const $push48=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push19=, $10, $pop48
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop21), $pop19
+; NO-SIMD128-SLOW-NEXT:    i32.const $push47=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push22=, $9, $pop47
+; NO-SIMD128-SLOW-NEXT:    i32.store8 8($0), $pop22
+; NO-SIMD128-SLOW-NEXT:    i32.const $push24=, 7
+; NO-SIMD128-SLOW-NEXT:    i32.add $push25=, $0, $pop24
+; NO-SIMD128-SLOW-NEXT:    i32.const $push46=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push23=, $8, $pop46
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop25), $pop23
+; NO-SIMD128-SLOW-NEXT:    i32.const $push27=, 6
+; NO-SIMD128-SLOW-NEXT:    i32.add $push28=, $0, $pop27
+; NO-SIMD128-SLOW-NEXT:    i32.const $push45=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push26=, $7, $pop45
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop28), $pop26
+; NO-SIMD128-SLOW-NEXT:    i32.const $push30=, 5
+; NO-SIMD128-SLOW-NEXT:    i32.add $push31=, $0, $pop30
+; NO-SIMD128-SLOW-NEXT:    i32.const $push44=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push29=, $6, $pop44
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop31), $pop29
+; NO-SIMD128-SLOW-NEXT:    i32.const $push43=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push32=, $5, $pop43
+; NO-SIMD128-SLOW-NEXT:    i32.store8 4($0), $pop32
+; NO-SIMD128-SLOW-NEXT:    i32.const $push34=, 3
+; NO-SIMD128-SLOW-NEXT:    i32.add $push35=, $0, $pop34
+; NO-SIMD128-SLOW-NEXT:    i32.const $push42=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push33=, $4, $pop42
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop35), $pop33
+; NO-SIMD128-SLOW-NEXT:    i32.const $push41=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push36=, $3, $pop41
+; NO-SIMD128-SLOW-NEXT:    i32.store8 2($0), $pop36
+; NO-SIMD128-SLOW-NEXT:    i32.const $push40=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push37=, $2, $pop40
+; NO-SIMD128-SLOW-NEXT:    i32.store8 1($0), $pop37
+; NO-SIMD128-SLOW-NEXT:    i32.const $push39=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push38=, $1, $pop39
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($0), $pop38
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: not_v16i8:
+; NO-SIMD128-FAST:         .functype not_v16i8 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-FAST-NEXT:    i32.const $push0=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push1=, $1, $pop0
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($0), $pop1
+; NO-SIMD128-FAST-NEXT:    i32.const $push53=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push2=, $2, $pop53
+; NO-SIMD128-FAST-NEXT:    i32.store8 1($0), $pop2
+; NO-SIMD128-FAST-NEXT:    i32.const $push52=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push3=, $3, $pop52
+; NO-SIMD128-FAST-NEXT:    i32.store8 2($0), $pop3
+; NO-SIMD128-FAST-NEXT:    i32.const $push4=, 3
+; NO-SIMD128-FAST-NEXT:    i32.add $push5=, $0, $pop4
+; NO-SIMD128-FAST-NEXT:    i32.const $push51=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push6=, $4, $pop51
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop5), $pop6
+; NO-SIMD128-FAST-NEXT:    i32.const $push50=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push7=, $5, $pop50
+; NO-SIMD128-FAST-NEXT:    i32.store8 4($0), $pop7
+; NO-SIMD128-FAST-NEXT:    i32.const $push8=, 5
+; NO-SIMD128-FAST-NEXT:    i32.add $push9=, $0, $pop8
+; NO-SIMD128-FAST-NEXT:    i32.const $push49=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push10=, $6, $pop49
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop9), $pop10
+; NO-SIMD128-FAST-NEXT:    i32.const $push11=, 6
+; NO-SIMD128-FAST-NEXT:    i32.add $push12=, $0, $pop11
+; NO-SIMD128-FAST-NEXT:    i32.const $push48=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push13=, $7, $pop48
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop12), $pop13
+; NO-SIMD128-FAST-NEXT:    i32.const $push14=, 7
+; NO-SIMD128-FAST-NEXT:    i32.add $push15=, $0, $pop14
+; NO-SIMD128-FAST-NEXT:    i32.const $push47=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push16=, $8, $pop47
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop15), $pop16
+; NO-SIMD128-FAST-NEXT:    i32.const $push46=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push17=, $9, $pop46
+; NO-SIMD128-FAST-NEXT:    i32.store8 8($0), $pop17
+; NO-SIMD128-FAST-NEXT:    i32.const $push18=, 9
+; NO-SIMD128-FAST-NEXT:    i32.add $push19=, $0, $pop18
+; NO-SIMD128-FAST-NEXT:    i32.const $push45=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push20=, $10, $pop45
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop19), $pop20
+; NO-SIMD128-FAST-NEXT:    i32.const $push21=, 10
+; NO-SIMD128-FAST-NEXT:    i32.add $push22=, $0, $pop21
+; NO-SIMD128-FAST-NEXT:    i32.const $push44=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push23=, $11, $pop44
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop22), $pop23
+; NO-SIMD128-FAST-NEXT:    i32.const $push24=, 11
+; NO-SIMD128-FAST-NEXT:    i32.add $push25=, $0, $pop24
+; NO-SIMD128-FAST-NEXT:    i32.const $push43=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push26=, $12, $pop43
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop25), $pop26
+; NO-SIMD128-FAST-NEXT:    i32.const $push27=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push28=, $0, $pop27
+; NO-SIMD128-FAST-NEXT:    i32.const $push42=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push29=, $13, $pop42
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop28), $pop29
+; NO-SIMD128-FAST-NEXT:    i32.const $push30=, 13
+; NO-SIMD128-FAST-NEXT:    i32.add $push31=, $0, $pop30
+; NO-SIMD128-FAST-NEXT:    i32.const $push41=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push32=, $14, $pop41
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop31), $pop32
+; NO-SIMD128-FAST-NEXT:    i32.const $push33=, 14
+; NO-SIMD128-FAST-NEXT:    i32.add $push34=, $0, $pop33
+; NO-SIMD128-FAST-NEXT:    i32.const $push40=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push35=, $15, $pop40
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop34), $pop35
+; NO-SIMD128-FAST-NEXT:    i32.const $push36=, 15
+; NO-SIMD128-FAST-NEXT:    i32.add $push37=, $0, $pop36
+; NO-SIMD128-FAST-NEXT:    i32.const $push39=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push38=, $16, $pop39
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop37), $pop38
+; NO-SIMD128-FAST-NEXT:    return
   %a = xor <16 x i8> %x, <i8 -1, i8 -1, i8 -1, i8 -1,
                           i8 -1, i8 -1, i8 -1, i8 -1,
                           i8 -1, i8 -1, i8 -1, i8 -1,
                           i8 -1, i8 -1, i8 -1, i8 -1>
   ret <16 x i8> %a
 }
 
-; CHECK-LABEL: bitselect_v16i8:
-; NO-SIMD128-NOT: v128
-; SIMD128-NEXT: .functype bitselect_v16i8 (v128, v128, v128) -> (v128){{$}}
-; SIMD128-SLOW-NEXT: v128.bitselect $push[[R:[0-9]+]]=, $1, $2, $0{{$}}
-; SIMD128-SLOW-NEXT: return $pop[[R]]{{$}}
-; SIMD128-FAST-NEXT: v128.and
-; SIMD128-FAST-NEXT: v128.not
-; SIMD128-FAST-NEXT: v128.and
-; SIMD128-FAST-NEXT: v128.or
-; SIMD128-FAST-NEXT: return
 define <16 x i8> @bitselect_v16i8(<16 x i8> %c, <16 x i8> %v1, <16 x i8> %v2) {
+; SIMD128-SLOW-LABEL: bitselect_v16i8:
+; SIMD128-SLOW:         .functype bitselect_v16i8 (v128, v128, v128) -> (v128)
+; SIMD128-SLOW-NEXT:    v128.bitselect $push0=, $1, $2, $0
+; SIMD128-SLOW-NEXT:    return $pop0
+;
+; SIMD128-FAST-LABEL: bitselect_v16i8:
+; SIMD128-FAST:         .functype bitselect_v16i8 (v128, v128, v128) -> (v128)
+; SIMD128-FAST-NEXT:    v128.and $push0=, $0, $1
+; SIMD128-FAST-NEXT:    v128.not $push2=, $0
+; SIMD128-FAST-NEXT:    v128.and $push3=, $pop2, $2
+; SIMD128-FAST-NEXT:    v128.or $push1=, $pop0, $pop3
+; SIMD128-FAST-NEXT:    return $pop1
+;
+; NO-SIMD128-SLOW-LABEL: bitselect_v16i8:
+; NO-SIMD128-SLOW:         .functype bitselect_v16i8 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push5=, 15
+; NO-SIMD128-SLOW-NEXT:    i32.add $push6=, $0, $pop5
+; NO-SIMD128-SLOW-NEXT:    i32.and $push0=, $16, $32
+; NO-SIMD128-SLOW-NEXT:    i32.const $push1=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push2=, $16, $pop1
+; NO-SIMD128-SLOW-NEXT:    i32.and $push3=, $pop2, $48
+; NO-SIMD128-SLOW-NEXT:    i32.or $push4=, $pop0, $pop3
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop6), $pop4
+; NO-SIMD128-SLOW-NEXT:    i32.const $push11=, 14
+; NO-SIMD128-SLOW-NEXT:    i32.add $push12=, $0, $pop11
+; NO-SIMD128-SLOW-NEXT:    i32.and $push7=, $15, $31
+; NO-SIMD128-SLOW-NEXT:    i32.const $push101=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push8=, $15, $pop101
+; NO-SIMD128-SLOW-NEXT:    i32.and $push9=, $pop8, $47
+; NO-SIMD128-SLOW-NEXT:    i32.or $push10=, $pop7, $pop9
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop12), $pop10
+; NO-SIMD128-SLOW-NEXT:    i32.const $push17=, 13
+; NO-SIMD128-SLOW-NEXT:    i32.add $push18=, $0, $pop17
+; NO-SIMD128-SLOW-NEXT:    i32.and $push13=, $14, $30
+; NO-SIMD128-SLOW-NEXT:    i32.const $push100=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push14=, $14, $pop100
+; NO-SIMD128-SLOW-NEXT:    i32.and $push15=, $pop14, $46
+; NO-SIMD128-SLOW-NEXT:    i32.or $push16=, $pop13, $pop15
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop18), $pop16
+; NO-SIMD128-SLOW-NEXT:    i32.const $push23=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push24=, $0, $pop23
+; NO-SIMD128-SLOW-NEXT:    i32.and $push19=, $13, $29
+; NO-SIMD128-SLOW-NEXT:    i32.const $push99=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push20=, $13, $pop99
+; NO-SIMD128-SLOW-NEXT:    i32.and $push21=, $pop20, $45
+; NO-SIMD128-SLOW-NEXT:    i32.or $push22=, $pop19, $pop21
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop24), $pop22
+; NO-SIMD128-SLOW-NEXT:    i32.const $push29=, 11
+; NO-SIMD128-SLOW-NEXT:    i32.add $push30=, $0, $pop29
+; NO-SIMD128-SLOW-NEXT:    i32.and $push25=, $12, $28
+; NO-SIMD128-SLOW-NEXT:    i32.const $push98=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push26=, $12, $pop98
+; NO-SIMD128-SLOW-NEXT:    i32.and $push27=, $pop26, $44
+; NO-SIMD128-SLOW-NEXT:    i32.or $push28=, $pop25, $pop27
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop30), $pop28
+; NO-SIMD128-SLOW-NEXT:    i32.const $push35=, 10
+; NO-SIMD128-SLOW-NEXT:    i32.add $push36=, $0, $pop35
+; NO-SIMD128-SLOW-NEXT:    i32.and $push31=, $11, $27
+; NO-SIMD128-SLOW-NEXT:    i32.const $push97=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push32=, $11, $pop97
+; NO-SIMD128-SLOW-NEXT:    i32.and $push33=, $pop32, $43
+; NO-SIMD128-SLOW-NEXT:    i32.or $push34=, $pop31, $pop33
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop36), $pop34
+; NO-SIMD128-SLOW-NEXT:    i32.const $push41=, 9
+; NO-SIMD128-SLOW-NEXT:    i32.add $push42=, $0, $pop41
+; NO-SIMD128-SLOW-NEXT:    i32.and $push37=, $10, $26
+; NO-SIMD128-SLOW-NEXT:    i32.const $push96=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push38=, $10, $pop96
+; NO-SIMD128-SLOW-NEXT:    i32.and $push39=, $pop38, $42
+; NO-SIMD128-SLOW-NEXT:    i32.or $push40=, $pop37, $pop39
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop42), $pop40
+; NO-SIMD128-SLOW-NEXT:    i32.and $push43=, $9, $25
+; NO-SIMD128-SLOW-NEXT:    i32.const $push95=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push44=, $9, $pop95
+; NO-SIMD128-SLOW-NEXT:    i32.and $push45=, $pop44, $41
+; NO-SIMD128-SLOW-NEXT:    i32.or $push46=, $pop43, $pop45
+; NO-SIMD128-SLOW-NEXT:    i32.store8 8($0), $pop46
+; NO-SIMD128-SLOW-NEXT:    i32.const $push51=, 7
+; NO-SIMD128-SLOW-NEXT:    i32.add $push52=, $0, $pop51
+; NO-SIMD128-SLOW-NEXT:    i32.and $push47=, $8, $24
+; NO-SIMD128-SLOW-NEXT:    i32.const $push94=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push48=, $8, $pop94
+; NO-SIMD128-SLOW-NEXT:    i32.and $push49=, $pop48, $40
+; NO-SIMD128-SLOW-NEXT:    i32.or $push50=, $pop47, $pop49
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop52), $pop50
+; NO-SIMD128-SLOW-NEXT:    i32.const $push57=, 6
+; NO-SIMD128-SLOW-NEXT:    i32.add $push58=, $0, $pop57
+; NO-SIMD128-SLOW-NEXT:    i32.and $push53=, $7, $23
+; NO-SIMD128-SLOW-NEXT:    i32.const $push93=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push54=, $7, $pop93
+; NO-SIMD128-SLOW-NEXT:    i32.and $push55=, $pop54, $39
+; NO-SIMD128-SLOW-NEXT:    i32.or $push56=, $pop53, $pop55
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop58), $pop56
+; NO-SIMD128-SLOW-NEXT:    i32.const $push63=, 5
+; NO-SIMD128-SLOW-NEXT:    i32.add $push64=, $0, $pop63
+; NO-SIMD128-SLOW-NEXT:    i32.and $push59=, $6, $22
+; NO-SIMD128-SLOW-NEXT:    i32.const $push92=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push60=, $6, $pop92
+; NO-SIMD128-SLOW-NEXT:    i32.and $push61=, $pop60, $38
+; NO-SIMD128-SLOW-NEXT:    i32.or $push62=, $pop59, $pop61
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop64), $pop62
+; NO-SIMD128-SLOW-NEXT:    i32.and $push65=, $5, $21
+; NO-SIMD128-SLOW-NEXT:    i32.const $push91=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push66=, $5, $pop91
+; NO-SIMD128-SLOW-NEXT:    i32.and $push67=, $pop66, $37
+; NO-SIMD128-SLOW-NEXT:    i32.or $push68=, $pop65, $pop67
+; NO-SIMD128-SLOW-NEXT:    i32.store8 4($0), $pop68
+; NO-SIMD128-SLOW-NEXT:    i32.const $push73=, 3
+; NO-SIMD128-SLOW-NEXT:    i32.add $push74=, $0, $pop73
+; NO-SIMD128-SLOW-NEXT:    i32.and $push69=, $4, $20
+; NO-SIMD128-SLOW-NEXT:    i32.const $push90=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push70=, $4, $pop90
+; NO-SIMD128-SLOW-NEXT:    i32.and $push71=, $pop70, $36
+; NO-SIMD128-SLOW-NEXT:    i32.or $push72=, $pop69, $pop71
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($pop74), $pop72
+; NO-SIMD128-SLOW-NEXT:    i32.and $push75=, $3, $19
+; NO-SIMD128-SLOW-NEXT:    i32.const $push89=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push76=, $3, $pop89
+; NO-SIMD128-SLOW-NEXT:    i32.and $push77=, $pop76, $35
+; NO-SIMD128-SLOW-NEXT:    i32.or $push78=, $pop75, $pop77
+; NO-SIMD128-SLOW-NEXT:    i32.store8 2($0), $pop78
+; NO-SIMD128-SLOW-NEXT:    i32.and $push79=, $2, $18
+; NO-SIMD128-SLOW-NEXT:    i32.const $push88=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push80=, $2, $pop88
+; NO-SIMD128-SLOW-NEXT:    i32.and $push81=, $pop80, $34
+; NO-SIMD128-SLOW-NEXT:    i32.or $push82=, $pop79, $pop81
+; NO-SIMD128-SLOW-NEXT:    i32.store8 1($0), $pop82
+; NO-SIMD128-SLOW-NEXT:    i32.and $push83=, $1, $17
+; NO-SIMD128-SLOW-NEXT:    i32.const $push87=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push84=, $1, $pop87
+; NO-SIMD128-SLOW-NEXT:    i32.and $push85=, $pop84, $33
+; NO-SIMD128-SLOW-NEXT:    i32.or $push86=, $pop83, $pop85
+; NO-SIMD128-SLOW-NEXT:    i32.store8 0($0), $pop86
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: bitselect_v16i8:
+; NO-SIMD128-FAST:         .functype bitselect_v16i8 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-FAST-NEXT:    i32.and $push0=, $1, $17
+; NO-SIMD128-FAST-NEXT:    i32.const $push1=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push2=, $1, $pop1
+; NO-SIMD128-FAST-NEXT:    i32.and $push3=, $pop2, $33
+; NO-SIMD128-FAST-NEXT:    i32.or $push4=, $pop0, $pop3
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($0), $pop4
+; NO-SIMD128-FAST-NEXT:    i32.and $push5=, $2, $18
+; NO-SIMD128-FAST-NEXT:    i32.const $push101=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push6=, $2, $pop101
+; NO-SIMD128-FAST-NEXT:    i32.and $push7=, $pop6, $34
+; NO-SIMD128-FAST-NEXT:    i32.or $push8=, $pop5, $pop7
+; NO-SIMD128-FAST-NEXT:    i32.store8 1($0), $pop8
+; NO-SIMD128-FAST-NEXT:    i32.and $push9=, $3, $19
+; NO-SIMD128-FAST-NEXT:    i32.const $push100=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push10=, $3, $pop100
+; NO-SIMD128-FAST-NEXT:    i32.and $push11=, $pop10, $35
+; NO-SIMD128-FAST-NEXT:    i32.or $push12=, $pop9, $pop11
+; NO-SIMD128-FAST-NEXT:    i32.store8 2($0), $pop12
+; NO-SIMD128-FAST-NEXT:    i32.const $push17=, 3
+; NO-SIMD128-FAST-NEXT:    i32.add $push18=, $0, $pop17
+; NO-SIMD128-FAST-NEXT:    i32.and $push13=, $4, $20
+; NO-SIMD128-FAST-NEXT:    i32.const $push99=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push14=, $4, $pop99
+; NO-SIMD128-FAST-NEXT:    i32.and $push15=, $pop14, $36
+; NO-SIMD128-FAST-NEXT:    i32.or $push16=, $pop13, $pop15
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop18), $pop16
+; NO-SIMD128-FAST-NEXT:    i32.and $push19=, $5, $21
+; NO-SIMD128-FAST-NEXT:    i32.const $push98=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push20=, $5, $pop98
+; NO-SIMD128-FAST-NEXT:    i32.and $push21=, $pop20, $37
+; NO-SIMD128-FAST-NEXT:    i32.or $push22=, $pop19, $pop21
+; NO-SIMD128-FAST-NEXT:    i32.store8 4($0), $pop22
+; NO-SIMD128-FAST-NEXT:    i32.const $push27=, 5
+; NO-SIMD128-FAST-NEXT:    i32.add $push28=, $0, $pop27
+; NO-SIMD128-FAST-NEXT:    i32.and $push23=, $6, $22
+; NO-SIMD128-FAST-NEXT:    i32.const $push97=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push24=, $6, $pop97
+; NO-SIMD128-FAST-NEXT:    i32.and $push25=, $pop24, $38
+; NO-SIMD128-FAST-NEXT:    i32.or $push26=, $pop23, $pop25
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop28), $pop26
+; NO-SIMD128-FAST-NEXT:    i32.const $push33=, 6
+; NO-SIMD128-FAST-NEXT:    i32.add $push34=, $0, $pop33
+; NO-SIMD128-FAST-NEXT:    i32.and $push29=, $7, $23
+; NO-SIMD128-FAST-NEXT:    i32.const $push96=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push30=, $7, $pop96
+; NO-SIMD128-FAST-NEXT:    i32.and $push31=, $pop30, $39
+; NO-SIMD128-FAST-NEXT:    i32.or $push32=, $pop29, $pop31
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop34), $pop32
+; NO-SIMD128-FAST-NEXT:    i32.const $push39=, 7
+; NO-SIMD128-FAST-NEXT:    i32.add $push40=, $0, $pop39
+; NO-SIMD128-FAST-NEXT:    i32.and $push35=, $8, $24
+; NO-SIMD128-FAST-NEXT:    i32.const $push95=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push36=, $8, $pop95
+; NO-SIMD128-FAST-NEXT:    i32.and $push37=, $pop36, $40
+; NO-SIMD128-FAST-NEXT:    i32.or $push38=, $pop35, $pop37
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop40), $pop38
+; NO-SIMD128-FAST-NEXT:    i32.and $push41=, $9, $25
+; NO-SIMD128-FAST-NEXT:    i32.const $push94=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push42=, $9, $pop94
+; NO-SIMD128-FAST-NEXT:    i32.and $push43=, $pop42, $41
+; NO-SIMD128-FAST-NEXT:    i32.or $push44=, $pop41, $pop43
+; NO-SIMD128-FAST-NEXT:    i32.store8 8($0), $pop44
+; NO-SIMD128-FAST-NEXT:    i32.const $push49=, 9
+; NO-SIMD128-FAST-NEXT:    i32.add $push50=, $0, $pop49
+; NO-SIMD128-FAST-NEXT:    i32.and $push45=, $10, $26
+; NO-SIMD128-FAST-NEXT:    i32.const $push93=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push46=, $10, $pop93
+; NO-SIMD128-FAST-NEXT:    i32.and $push47=, $pop46, $42
+; NO-SIMD128-FAST-NEXT:    i32.or $push48=, $pop45, $pop47
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop50), $pop48
+; NO-SIMD128-FAST-NEXT:    i32.const $push55=, 10
+; NO-SIMD128-FAST-NEXT:    i32.add $push56=, $0, $pop55
+; NO-SIMD128-FAST-NEXT:    i32.and $push51=, $11, $27
+; NO-SIMD128-FAST-NEXT:    i32.const $push92=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push52=, $11, $pop92
+; NO-SIMD128-FAST-NEXT:    i32.and $push53=, $pop52, $43
+; NO-SIMD128-FAST-NEXT:    i32.or $push54=, $pop51, $pop53
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop56), $pop54
+; NO-SIMD128-FAST-NEXT:    i32.const $push61=, 11
+; NO-SIMD128-FAST-NEXT:    i32.add $push62=, $0, $pop61
+; NO-SIMD128-FAST-NEXT:    i32.and $push57=, $12, $28
+; NO-SIMD128-FAST-NEXT:    i32.const $push91=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push58=, $12, $pop91
+; NO-SIMD128-FAST-NEXT:    i32.and $push59=, $pop58, $44
+; NO-SIMD128-FAST-NEXT:    i32.or $push60=, $pop57, $pop59
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop62), $pop60
+; NO-SIMD128-FAST-NEXT:    i32.const $push67=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push68=, $0, $pop67
+; NO-SIMD128-FAST-NEXT:    i32.and $push63=, $13, $29
+; NO-SIMD128-FAST-NEXT:    i32.const $push90=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push64=, $13, $pop90
+; NO-SIMD128-FAST-NEXT:    i32.and $push65=, $pop64, $45
+; NO-SIMD128-FAST-NEXT:    i32.or $push66=, $pop63, $pop65
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop68), $pop66
+; NO-SIMD128-FAST-NEXT:    i32.const $push73=, 13
+; NO-SIMD128-FAST-NEXT:    i32.add $push74=, $0, $pop73
+; NO-SIMD128-FAST-NEXT:    i32.and $push69=, $14, $30
+; NO-SIMD128-FAST-NEXT:    i32.const $push89=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push70=, $14, $pop89
+; NO-SIMD128-FAST-NEXT:    i32.and $push71=, $pop70, $46
+; NO-SIMD128-FAST-NEXT:    i32.or $push72=, $pop69, $pop71
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop74), $pop72
+; NO-SIMD128-FAST-NEXT:    i32.const $push79=, 14
+; NO-SIMD128-FAST-NEXT:    i32.add $push80=, $0, $pop79
+; NO-SIMD128-FAST-NEXT:    i32.and $push75=, $15, $31
+; NO-SIMD128-FAST-NEXT:    i32.const $push88=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push76=, $15, $pop88
+; NO-SIMD128-FAST-NEXT:    i32.and $push77=, $pop76, $47
+; NO-SIMD128-FAST-NEXT:    i32.or $push78=, $pop75, $pop77
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop80), $pop78
+; NO-SIMD128-FAST-NEXT:    i32.const $push85=, 15
+; NO-SIMD128-FAST-NEXT:    i32.add $push86=, $0, $pop85
+; NO-SIMD128-FAST-NEXT:    i32.and $push81=, $16, $32
+; NO-SIMD128-FAST-NEXT:    i32.const $push87=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push82=, $16, $pop87
+; NO-SIMD128-FAST-NEXT:    i32.and $push83=, $pop82, $48
+; NO-SIMD128-FAST-NEXT:    i32.or $push84=, $pop81, $pop83
+; NO-SIMD128-FAST-NEXT:    i32.store8 0($pop86), $pop84
+; NO-SIMD128-FAST-NEXT:    return
   %masked_v1 = and <16 x i8> %c, %v1
   %inv_mask = xor <16 x i8> %c,
     <i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1,
      i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1>
   %masked_v2 = and <16 x i8> %inv_mask, %v2
   %a = or <16 x i8> %masked_v1, %masked_v2
   ret <16 x i8> %a
 }
 
 ; ==============================================================================
 ; 8 x i16
 ; ==============================================================================
-; CHECK-LABEL: add_v8i16:
-; NO-SIMD128-NOT: i16x8
-; SIMD128-NEXT: .functype add_v8i16 (v128, v128) -> (v128){{$}}
-; SIMD128-NEXT: i16x8.add $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
+
 define <8 x i16> @add_v8i16(<8 x i16> %x, <8 x i16> %y) {
+; SIMD128-LABEL: add_v8i16:
+; SIMD128:         .functype add_v8i16 (v128, v128) -> (v128)
+; SIMD128-NEXT:    i16x8.add $push0=, $0, $1
+; SIMD128-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: add_v8i16:
+; NO-SIMD128-SLOW:         .functype add_v8i16 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push1=, 14
+; NO-SIMD128-SLOW-NEXT:    i32.add $push2=, $0, $pop1
+; NO-SIMD128-SLOW-NEXT:    i32.add $push0=, $8, $16
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop2), $pop0
+; NO-SIMD128-SLOW-NEXT:    i32.const $push4=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push5=, $0, $pop4
+; NO-SIMD128-SLOW-NEXT:    i32.add $push3=, $7, $15
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop5), $pop3
+; NO-SIMD128-SLOW-NEXT:    i32.const $push7=, 10
+; NO-SIMD128-SLOW-NEXT:    i32.add $push8=, $0, $pop7
+; NO-SIMD128-SLOW-NEXT:    i32.add $push6=, $6, $14
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop8), $pop6
+; NO-SIMD128-SLOW-NEXT:    i32.add $push9=, $5, $13
+; NO-SIMD128-SLOW-NEXT:    i32.store16 8($0), $pop9
+; NO-SIMD128-SLOW-NEXT:    i32.const $push11=, 6
+; NO-SIMD128-SLOW-NEXT:    i32.add $push12=, $0, $pop11
+; NO-SIMD128-SLOW-NEXT:    i32.add $push10=, $4, $12
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop12), $pop10
+; NO-SIMD128-SLOW-NEXT:    i32.add $push13=, $3, $11
+; NO-SIMD128-SLOW-NEXT:    i32.store16 4($0), $pop13
+; NO-SIMD128-SLOW-NEXT:    i32.add $push14=, $2, $10
+; NO-SIMD128-SLOW-NEXT:    i32.store16 2($0), $pop14
+; NO-SIMD128-SLOW-NEXT:    i32.add $push15=, $1, $9
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($0), $pop15
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: add_v8i16:
+; NO-SIMD128-FAST:         .functype add_v8i16 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-FAST-NEXT:    i32.add $push0=, $1, $9
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($0), $pop0
+; NO-SIMD128-FAST-NEXT:    i32.add $push1=, $2, $10
+; NO-SIMD128-FAST-NEXT:    i32.store16 2($0), $pop1
+; NO-SIMD128-FAST-NEXT:    i32.add $push2=, $3, $11
+; NO-SIMD128-FAST-NEXT:    i32.store16 4($0), $pop2
+; NO-SIMD128-FAST-NEXT:    i32.const $push3=, 6
+; NO-SIMD128-FAST-NEXT:    i32.add $push4=, $0, $pop3
+; NO-SIMD128-FAST-NEXT:    i32.add $push5=, $4, $12
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop4), $pop5
+; NO-SIMD128-FAST-NEXT:    i32.add $push6=, $5, $13
+; NO-SIMD128-FAST-NEXT:    i32.store16 8($0), $pop6
+; NO-SIMD128-FAST-NEXT:    i32.const $push7=, 10
+; NO-SIMD128-FAST-NEXT:    i32.add $push8=, $0, $pop7
+; NO-SIMD128-FAST-NEXT:    i32.add $push9=, $6, $14
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop8), $pop9
+; NO-SIMD128-FAST-NEXT:    i32.const $push10=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push11=, $0, $pop10
+; NO-SIMD128-FAST-NEXT:    i32.add $push12=, $7, $15
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop11), $pop12
+; NO-SIMD128-FAST-NEXT:    i32.const $push13=, 14
+; NO-SIMD128-FAST-NEXT:    i32.add $push14=, $0, $pop13
+; NO-SIMD128-FAST-NEXT:    i32.add $push15=, $8, $16
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop14), $pop15
+; NO-SIMD128-FAST-NEXT:    return
   %a = add <8 x i16> %x, %y
   ret <8 x i16> %a
 }
 
-; CHECK-LABEL: sub_v8i16:
-; NO-SIMD128-NOT: i16x8
-; SIMD128-NEXT: .functype sub_v8i16 (v128, v128) -> (v128){{$}}
-; SIMD128-NEXT: i16x8.sub $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <8 x i16> @sub_v8i16(<8 x i16> %x, <8 x i16> %y) {
+; SIMD128-LABEL: sub_v8i16:
+; SIMD128:         .functype sub_v8i16 (v128, v128) -> (v128)
+; SIMD128-NEXT:    i16x8.sub $push0=, $0, $1
+; SIMD128-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: sub_v8i16:
+; NO-SIMD128-SLOW:         .functype sub_v8i16 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push1=, 14
+; NO-SIMD128-SLOW-NEXT:    i32.add $push2=, $0, $pop1
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push0=, $8, $16
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop2), $pop0
+; NO-SIMD128-SLOW-NEXT:    i32.const $push4=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push5=, $0, $pop4
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push3=, $7, $15
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop5), $pop3
+; NO-SIMD128-SLOW-NEXT:    i32.const $push7=, 10
+; NO-SIMD128-SLOW-NEXT:    i32.add $push8=, $0, $pop7
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push6=, $6, $14
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop8), $pop6
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push9=, $5, $13
+; NO-SIMD128-SLOW-NEXT:    i32.store16 8($0), $pop9
+; NO-SIMD128-SLOW-NEXT:    i32.const $push11=, 6
+; NO-SIMD128-SLOW-NEXT:    i32.add $push12=, $0, $pop11
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push10=, $4, $12
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop12), $pop10
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push13=, $3, $11
+; NO-SIMD128-SLOW-NEXT:    i32.store16 4($0), $pop13
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push14=, $2, $10
+; NO-SIMD128-SLOW-NEXT:    i32.store16 2($0), $pop14
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push15=, $1, $9
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($0), $pop15
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: sub_v8i16:
+; NO-SIMD128-FAST:         .functype sub_v8i16 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-FAST-NEXT:    i32.sub $push0=, $1, $9
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($0), $pop0
+; NO-SIMD128-FAST-NEXT:    i32.sub $push1=, $2, $10
+; NO-SIMD128-FAST-NEXT:    i32.store16 2($0), $pop1
+; NO-SIMD128-FAST-NEXT:    i32.sub $push2=, $3, $11
+; NO-SIMD128-FAST-NEXT:    i32.store16 4($0), $pop2
+; NO-SIMD128-FAST-NEXT:    i32.const $push3=, 6
+; NO-SIMD128-FAST-NEXT:    i32.add $push4=, $0, $pop3
+; NO-SIMD128-FAST-NEXT:    i32.sub $push5=, $4, $12
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop4), $pop5
+; NO-SIMD128-FAST-NEXT:    i32.sub $push6=, $5, $13
+; NO-SIMD128-FAST-NEXT:    i32.store16 8($0), $pop6
+; NO-SIMD128-FAST-NEXT:    i32.const $push7=, 10
+; NO-SIMD128-FAST-NEXT:    i32.add $push8=, $0, $pop7
+; NO-SIMD128-FAST-NEXT:    i32.sub $push9=, $6, $14
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop8), $pop9
+; NO-SIMD128-FAST-NEXT:    i32.const $push10=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push11=, $0, $pop10
+; NO-SIMD128-FAST-NEXT:    i32.sub $push12=, $7, $15
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop11), $pop12
+; NO-SIMD128-FAST-NEXT:    i32.const $push13=, 14
+; NO-SIMD128-FAST-NEXT:    i32.add $push14=, $0, $pop13
+; NO-SIMD128-FAST-NEXT:    i32.sub $push15=, $8, $16
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop14), $pop15
+; NO-SIMD128-FAST-NEXT:    return
   %a = sub <8 x i16> %x, %y
   ret <8 x i16> %a
 }
 
-; CHECK-LABEL: mul_v8i16:
-; NO-SIMD128-NOT: i16x8
-; SIMD128-NEXT: .functype mul_v8i16 (v128, v128) -> (v128){{$}}
-; SIMD128-NEXT: i16x8.mul $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <8 x i16> @mul_v8i16(<8 x i16> %x, <8 x i16> %y) {
+; SIMD128-LABEL: mul_v8i16:
+; SIMD128:         .functype mul_v8i16 (v128, v128) -> (v128)
+; SIMD128-NEXT:    i16x8.mul $push0=, $0, $1
+; SIMD128-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: mul_v8i16:
+; NO-SIMD128-SLOW:         .functype mul_v8i16 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push1=, 14
+; NO-SIMD128-SLOW-NEXT:    i32.add $push2=, $0, $pop1
+; NO-SIMD128-SLOW-NEXT:    i32.mul $push0=, $8, $16
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop2), $pop0
+; NO-SIMD128-SLOW-NEXT:    i32.const $push4=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push5=, $0, $pop4
+; NO-SIMD128-SLOW-NEXT:    i32.mul $push3=, $7, $15
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop5), $pop3
+; NO-SIMD128-SLOW-NEXT:    i32.const $push7=, 10
+; NO-SIMD128-SLOW-NEXT:    i32.add $push8=, $0, $pop7
+; NO-SIMD128-SLOW-NEXT:    i32.mul $push6=, $6, $14
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop8), $pop6
+; NO-SIMD128-SLOW-NEXT:    i32.mul $push9=, $5, $13
+; NO-SIMD128-SLOW-NEXT:    i32.store16 8($0), $pop9
+; NO-SIMD128-SLOW-NEXT:    i32.const $push11=, 6
+; NO-SIMD128-SLOW-NEXT:    i32.add $push12=, $0, $pop11
+; NO-SIMD128-SLOW-NEXT:    i32.mul $push10=, $4, $12
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop12), $pop10
+; NO-SIMD128-SLOW-NEXT:    i32.mul $push13=, $3, $11
+; NO-SIMD128-SLOW-NEXT:    i32.store16 4($0), $pop13
+; NO-SIMD128-SLOW-NEXT:    i32.mul $push14=, $2, $10
+; NO-SIMD128-SLOW-NEXT:    i32.store16 2($0), $pop14
+; NO-SIMD128-SLOW-NEXT:    i32.mul $push15=, $1, $9
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($0), $pop15
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: mul_v8i16:
+; NO-SIMD128-FAST:         .functype mul_v8i16 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-FAST-NEXT:    i32.mul $push0=, $1, $9
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($0), $pop0
+; NO-SIMD128-FAST-NEXT:    i32.mul $push1=, $2, $10
+; NO-SIMD128-FAST-NEXT:    i32.store16 2($0), $pop1
+; NO-SIMD128-FAST-NEXT:    i32.mul $push2=, $3, $11
+; NO-SIMD128-FAST-NEXT:    i32.store16 4($0), $pop2
+; NO-SIMD128-FAST-NEXT:    i32.const $push3=, 6
+; NO-SIMD128-FAST-NEXT:    i32.add $push4=, $0, $pop3
+; NO-SIMD128-FAST-NEXT:    i32.mul $push5=, $4, $12
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop4), $pop5
+; NO-SIMD128-FAST-NEXT:    i32.mul $push6=, $5, $13
+; NO-SIMD128-FAST-NEXT:    i32.store16 8($0), $pop6
+; NO-SIMD128-FAST-NEXT:    i32.const $push7=, 10
+; NO-SIMD128-FAST-NEXT:    i32.add $push8=, $0, $pop7
+; NO-SIMD128-FAST-NEXT:    i32.mul $push9=, $6, $14
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop8), $pop9
+; NO-SIMD128-FAST-NEXT:    i32.const $push10=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push11=, $0, $pop10
+; NO-SIMD128-FAST-NEXT:    i32.mul $push12=, $7, $15
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop11), $pop12
+; NO-SIMD128-FAST-NEXT:    i32.const $push13=, 14
+; NO-SIMD128-FAST-NEXT:    i32.add $push14=, $0, $pop13
+; NO-SIMD128-FAST-NEXT:    i32.mul $push15=, $8, $16
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop14), $pop15
+; NO-SIMD128-FAST-NEXT:    return
   %a = mul <8 x i16> %x, %y
   ret <8 x i16> %a
 }
 
-; CHECK-LABEL: neg_v8i16:
-; NO-SIMD128-NOT: i16x8
-; SIMD128-NEXT: .functype neg_v8i16 (v128) -> (v128){{$}}
-; SIMD128-NEXT: i16x8.neg $push[[R:[0-9]+]]=, $0{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <8 x i16> @neg_v8i16(<8 x i16> %x) {
+; SIMD128-LABEL: neg_v8i16:
+; SIMD128:         .functype neg_v8i16 (v128) -> (v128)
+; SIMD128-NEXT:    i16x8.neg $push0=, $0
+; SIMD128-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: neg_v8i16:
+; NO-SIMD128-SLOW:         .functype neg_v8i16 (i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push2=, 14
+; NO-SIMD128-SLOW-NEXT:    i32.add $push3=, $0, $pop2
+; NO-SIMD128-SLOW-NEXT:    i32.const $push0=, 0
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push1=, $pop0, $8
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop3), $pop1
+; NO-SIMD128-SLOW-NEXT:    i32.const $push5=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push6=, $0, $pop5
+; NO-SIMD128-SLOW-NEXT:    i32.const $push23=, 0
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push4=, $pop23, $7
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop6), $pop4
+; NO-SIMD128-SLOW-NEXT:    i32.const $push8=, 10
+; NO-SIMD128-SLOW-NEXT:    i32.add $push9=, $0, $pop8
+; NO-SIMD128-SLOW-NEXT:    i32.const $push22=, 0
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push7=, $pop22, $6
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop9), $pop7
+; NO-SIMD128-SLOW-NEXT:    i32.const $push21=, 0
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push10=, $pop21, $5
+; NO-SIMD128-SLOW-NEXT:    i32.store16 8($0), $pop10
+; NO-SIMD128-SLOW-NEXT:    i32.const $push12=, 6
+; NO-SIMD128-SLOW-NEXT:    i32.add $push13=, $0, $pop12
+; NO-SIMD128-SLOW-NEXT:    i32.const $push20=, 0
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push11=, $pop20, $4
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop13), $pop11
+; NO-SIMD128-SLOW-NEXT:    i32.const $push19=, 0
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push14=, $pop19, $3
+; NO-SIMD128-SLOW-NEXT:    i32.store16 4($0), $pop14
+; NO-SIMD128-SLOW-NEXT:    i32.const $push18=, 0
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push15=, $pop18, $2
+; NO-SIMD128-SLOW-NEXT:    i32.store16 2($0), $pop15
+; NO-SIMD128-SLOW-NEXT:    i32.const $push17=, 0
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push16=, $pop17, $1
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($0), $pop16
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: neg_v8i16:
+; NO-SIMD128-FAST:         .functype neg_v8i16 (i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-FAST-NEXT:    i32.const $push0=, 0
+; NO-SIMD128-FAST-NEXT:    i32.sub $push1=, $pop0, $1
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($0), $pop1
+; NO-SIMD128-FAST-NEXT:    i32.const $push23=, 0
+; NO-SIMD128-FAST-NEXT:    i32.sub $push2=, $pop23, $2
+; NO-SIMD128-FAST-NEXT:    i32.store16 2($0), $pop2
+; NO-SIMD128-FAST-NEXT:    i32.const $push22=, 0
+; NO-SIMD128-FAST-NEXT:    i32.sub $push3=, $pop22, $3
+; NO-SIMD128-FAST-NEXT:    i32.store16 4($0), $pop3
+; NO-SIMD128-FAST-NEXT:    i32.const $push4=, 6
+; NO-SIMD128-FAST-NEXT:    i32.add $push5=, $0, $pop4
+; NO-SIMD128-FAST-NEXT:    i32.const $push21=, 0
+; NO-SIMD128-FAST-NEXT:    i32.sub $push6=, $pop21, $4
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop5), $pop6
+; NO-SIMD128-FAST-NEXT:    i32.const $push20=, 0
+; NO-SIMD128-FAST-NEXT:    i32.sub $push7=, $pop20, $5
+; NO-SIMD128-FAST-NEXT:    i32.store16 8($0), $pop7
+; NO-SIMD128-FAST-NEXT:    i32.const $push8=, 10
+; NO-SIMD128-FAST-NEXT:    i32.add $push9=, $0, $pop8
+; NO-SIMD128-FAST-NEXT:    i32.const $push19=, 0
+; NO-SIMD128-FAST-NEXT:    i32.sub $push10=, $pop19, $6
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop9), $pop10
+; NO-SIMD128-FAST-NEXT:    i32.const $push11=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push12=, $0, $pop11
+; NO-SIMD128-FAST-NEXT:    i32.const $push18=, 0
+; NO-SIMD128-FAST-NEXT:    i32.sub $push13=, $pop18, $7
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop12), $pop13
+; NO-SIMD128-FAST-NEXT:    i32.const $push14=, 14
+; NO-SIMD128-FAST-NEXT:    i32.add $push15=, $0, $pop14
+; NO-SIMD128-FAST-NEXT:    i32.const $push17=, 0
+; NO-SIMD128-FAST-NEXT:    i32.sub $push16=, $pop17, $8
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop15), $pop16
+; NO-SIMD128-FAST-NEXT:    return
   %a = sub <8 x i16> <i16 0, i16 0, i16 0, i16 0, i16 0, i16 0, i16 0, i16 0>,
                      %x
   ret <8 x i16> %a
 }
 
-; CHECK-LABEL: shl_v8i16:
-; NO-SIMD128-NOT: i16x8
-; SIMD128-NEXT: .functype shl_v8i16 (v128, i32) -> (v128){{$}}
-; SIMD128-NEXT: i16x8.shl $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <8 x i16> @shl_v8i16(<8 x i16> %v, i16 %x) {
+; SIMD128-UNIMPL-LABEL: shl_v8i16:
+; SIMD128-UNIMPL:         .functype shl_v8i16 (v128, i32) -> (v128)
+; SIMD128-UNIMPL-NEXT:    i16x8.shl $push0=, $0, $1
+; SIMD128-UNIMPL-NEXT:    return $pop0
+;
+; SIMD128-VM-SLOW-LABEL: shl_v8i16:
+; SIMD128-VM-SLOW:         .functype shl_v8i16 (v128, i32) -> (v128)
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push9=, $0, 0
+; SIMD128-VM-SLOW-NEXT:    i16x8.splat $push2=, $1
+; SIMD128-VM-SLOW-NEXT:    i32.const $push0=, 15
+; SIMD128-VM-SLOW-NEXT:    i16x8.splat $push1=, $pop0
+; SIMD128-VM-SLOW-NEXT:    v128.and $push52=, $pop2, $pop1
+; SIMD128-VM-SLOW-NEXT:    local.tee $push51=, $2=, $pop52
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push7=, $pop51, 0
+; SIMD128-VM-SLOW-NEXT:    i32.const $push50=, 15
+; SIMD128-VM-SLOW-NEXT:    i32.and $push8=, $pop7, $pop50
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push10=, $pop9, $pop8
+; SIMD128-VM-SLOW-NEXT:    i16x8.splat $push11=, $pop10
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push5=, $0, 1
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push3=, $2, 1
+; SIMD128-VM-SLOW-NEXT:    i32.const $push49=, 15
+; SIMD128-VM-SLOW-NEXT:    i32.and $push4=, $pop3, $pop49
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push6=, $pop5, $pop4
+; SIMD128-VM-SLOW-NEXT:    i16x8.replace_lane $push12=, $pop11, 1, $pop6
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push15=, $0, 2
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push13=, $2, 2
+; SIMD128-VM-SLOW-NEXT:    i32.const $push48=, 15
+; SIMD128-VM-SLOW-NEXT:    i32.and $push14=, $pop13, $pop48
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push16=, $pop15, $pop14
+; SIMD128-VM-SLOW-NEXT:    i16x8.replace_lane $push17=, $pop12, 2, $pop16
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push20=, $0, 3
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push18=, $2, 3
+; SIMD128-VM-SLOW-NEXT:    i32.const $push47=, 15
+; SIMD128-VM-SLOW-NEXT:    i32.and $push19=, $pop18, $pop47
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push21=, $pop20, $pop19
+; SIMD128-VM-SLOW-NEXT:    i16x8.replace_lane $push22=, $pop17, 3, $pop21
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push25=, $0, 4
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push23=, $2, 4
+; SIMD128-VM-SLOW-NEXT:    i32.const $push46=, 15
+; SIMD128-VM-SLOW-NEXT:    i32.and $push24=, $pop23, $pop46
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push26=, $pop25, $pop24
+; SIMD128-VM-SLOW-NEXT:    i16x8.replace_lane $push27=, $pop22, 4, $pop26
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push30=, $0, 5
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push28=, $2, 5
+; SIMD128-VM-SLOW-NEXT:    i32.const $push45=, 15
+; SIMD128-VM-SLOW-NEXT:    i32.and $push29=, $pop28, $pop45
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push31=, $pop30, $pop29
+; SIMD128-VM-SLOW-NEXT:    i16x8.replace_lane $push32=, $pop27, 5, $pop31
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push35=, $0, 6
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push33=, $2, 6
+; SIMD128-VM-SLOW-NEXT:    i32.const $push44=, 15
+; SIMD128-VM-SLOW-NEXT:    i32.and $push34=, $pop33, $pop44
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push36=, $pop35, $pop34
+; SIMD128-VM-SLOW-NEXT:    i16x8.replace_lane $push37=, $pop32, 6, $pop36
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push40=, $0, 7
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push38=, $2, 7
+; SIMD128-VM-SLOW-NEXT:    i32.const $push43=, 15
+; SIMD128-VM-SLOW-NEXT:    i32.and $push39=, $pop38, $pop43
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push41=, $pop40, $pop39
+; SIMD128-VM-SLOW-NEXT:    i16x8.replace_lane $push42=, $pop37, 7, $pop41
+; SIMD128-VM-SLOW-NEXT:    return $pop42
+;
+; SIMD128-VM-FAST-LABEL: shl_v8i16:
+; SIMD128-VM-FAST:         .functype shl_v8i16 (v128, i32) -> (v128)
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push10=, $0, 0
+; SIMD128-VM-FAST-NEXT:    i16x8.splat $push1=, $1
+; SIMD128-VM-FAST-NEXT:    i32.const $push2=, 15
+; SIMD128-VM-FAST-NEXT:    i16x8.splat $push3=, $pop2
+; SIMD128-VM-FAST-NEXT:    v128.and $push52=, $pop1, $pop3
+; SIMD128-VM-FAST-NEXT:    local.tee $push51=, $2=, $pop52
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push8=, $pop51, 0
+; SIMD128-VM-FAST-NEXT:    i32.const $push50=, 15
+; SIMD128-VM-FAST-NEXT:    i32.and $push9=, $pop8, $pop50
+; SIMD128-VM-FAST-NEXT:    i32.shl $push11=, $pop10, $pop9
+; SIMD128-VM-FAST-NEXT:    i16x8.splat $push12=, $pop11
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push6=, $0, 1
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push4=, $2, 1
+; SIMD128-VM-FAST-NEXT:    i32.const $push49=, 15
+; SIMD128-VM-FAST-NEXT:    i32.and $push5=, $pop4, $pop49
+; SIMD128-VM-FAST-NEXT:    i32.shl $push7=, $pop6, $pop5
+; SIMD128-VM-FAST-NEXT:    i16x8.replace_lane $push13=, $pop12, 1, $pop7
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push16=, $0, 2
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push14=, $2, 2
+; SIMD128-VM-FAST-NEXT:    i32.const $push48=, 15
+; SIMD128-VM-FAST-NEXT:    i32.and $push15=, $pop14, $pop48
+; SIMD128-VM-FAST-NEXT:    i32.shl $push17=, $pop16, $pop15
+; SIMD128-VM-FAST-NEXT:    i16x8.replace_lane $push18=, $pop13, 2, $pop17
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push21=, $0, 3
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push19=, $2, 3
+; SIMD128-VM-FAST-NEXT:    i32.const $push47=, 15
+; SIMD128-VM-FAST-NEXT:    i32.and $push20=, $pop19, $pop47
+; SIMD128-VM-FAST-NEXT:    i32.shl $push22=, $pop21, $pop20
+; SIMD128-VM-FAST-NEXT:    i16x8.replace_lane $push23=, $pop18, 3, $pop22
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push26=, $0, 4
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push24=, $2, 4
+; SIMD128-VM-FAST-NEXT:    i32.const $push46=, 15
+; SIMD128-VM-FAST-NEXT:    i32.and $push25=, $pop24, $pop46
+; SIMD128-VM-FAST-NEXT:    i32.shl $push27=, $pop26, $pop25
+; SIMD128-VM-FAST-NEXT:    i16x8.replace_lane $push28=, $pop23, 4, $pop27
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push31=, $0, 5
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push29=, $2, 5
+; SIMD128-VM-FAST-NEXT:    i32.const $push45=, 15
+; SIMD128-VM-FAST-NEXT:    i32.and $push30=, $pop29, $pop45
+; SIMD128-VM-FAST-NEXT:    i32.shl $push32=, $pop31, $pop30
+; SIMD128-VM-FAST-NEXT:    i16x8.replace_lane $push33=, $pop28, 5, $pop32
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push36=, $0, 6
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push34=, $2, 6
+; SIMD128-VM-FAST-NEXT:    i32.const $push44=, 15
+; SIMD128-VM-FAST-NEXT:    i32.and $push35=, $pop34, $pop44
+; SIMD128-VM-FAST-NEXT:    i32.shl $push37=, $pop36, $pop35
+; SIMD128-VM-FAST-NEXT:    i16x8.replace_lane $push38=, $pop33, 6, $pop37
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push41=, $0, 7
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push39=, $2, 7
+; SIMD128-VM-FAST-NEXT:    i32.const $push43=, 15
+; SIMD128-VM-FAST-NEXT:    i32.and $push40=, $pop39, $pop43
+; SIMD128-VM-FAST-NEXT:    i32.shl $push42=, $pop41, $pop40
+; SIMD128-VM-FAST-NEXT:    i16x8.replace_lane $push0=, $pop38, 7, $pop42
+; SIMD128-VM-FAST-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: shl_v8i16:
+; NO-SIMD128-SLOW:         .functype shl_v8i16 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push0=, 14
+; NO-SIMD128-SLOW-NEXT:    i32.add $push1=, $0, $pop0
+; NO-SIMD128-SLOW-NEXT:    i32.const $push2=, 65535
+; NO-SIMD128-SLOW-NEXT:    i32.and $push18=, $9, $pop2
+; NO-SIMD128-SLOW-NEXT:    local.tee $push17=, $9=, $pop18
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push3=, $8, $pop17
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop1), $pop3
+; NO-SIMD128-SLOW-NEXT:    i32.const $push4=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push5=, $0, $pop4
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push6=, $7, $9
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop5), $pop6
+; NO-SIMD128-SLOW-NEXT:    i32.const $push7=, 10
+; NO-SIMD128-SLOW-NEXT:    i32.add $push8=, $0, $pop7
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push9=, $6, $9
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop8), $pop9
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push10=, $5, $9
+; NO-SIMD128-SLOW-NEXT:    i32.store16 8($0), $pop10
+; NO-SIMD128-SLOW-NEXT:    i32.const $push11=, 6
+; NO-SIMD128-SLOW-NEXT:    i32.add $push12=, $0, $pop11
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push13=, $4, $9
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop12), $pop13
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push14=, $3, $9
+; NO-SIMD128-SLOW-NEXT:    i32.store16 4($0), $pop14
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push15=, $2, $9
+; NO-SIMD128-SLOW-NEXT:    i32.store16 2($0), $pop15
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push16=, $1, $9
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($0), $pop16
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: shl_v8i16:
+; NO-SIMD128-FAST:         .functype shl_v8i16 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-FAST-NEXT:    i32.const $push0=, 65535
+; NO-SIMD128-FAST-NEXT:    i32.and $push18=, $9, $pop0
+; NO-SIMD128-FAST-NEXT:    local.tee $push17=, $9=, $pop18
+; NO-SIMD128-FAST-NEXT:    i32.shl $push1=, $2, $pop17
+; NO-SIMD128-FAST-NEXT:    i32.store16 2($0), $pop1
+; NO-SIMD128-FAST-NEXT:    i32.shl $push2=, $1, $9
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($0), $pop2
+; NO-SIMD128-FAST-NEXT:    i32.shl $push3=, $3, $9
+; NO-SIMD128-FAST-NEXT:    i32.store16 4($0), $pop3
+; NO-SIMD128-FAST-NEXT:    i32.const $push4=, 6
+; NO-SIMD128-FAST-NEXT:    i32.add $push5=, $0, $pop4
+; NO-SIMD128-FAST-NEXT:    i32.shl $push6=, $4, $9
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop5), $pop6
+; NO-SIMD128-FAST-NEXT:    i32.shl $push7=, $5, $9
+; NO-SIMD128-FAST-NEXT:    i32.store16 8($0), $pop7
+; NO-SIMD128-FAST-NEXT:    i32.const $push8=, 10
+; NO-SIMD128-FAST-NEXT:    i32.add $push9=, $0, $pop8
+; NO-SIMD128-FAST-NEXT:    i32.shl $push10=, $6, $9
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop9), $pop10
+; NO-SIMD128-FAST-NEXT:    i32.const $push11=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push12=, $0, $pop11
+; NO-SIMD128-FAST-NEXT:    i32.shl $push13=, $7, $9
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop12), $pop13
+; NO-SIMD128-FAST-NEXT:    i32.const $push14=, 14
+; NO-SIMD128-FAST-NEXT:    i32.add $push15=, $0, $pop14
+; NO-SIMD128-FAST-NEXT:    i32.shl $push16=, $8, $9
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop15), $pop16
+; NO-SIMD128-FAST-NEXT:    return
   %t = insertelement <8 x i16> undef, i16 %x, i32 0
   %s = shufflevector <8 x i16> %t, <8 x i16> undef,
     <8 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
   %a = shl <8 x i16> %v, %s
   ret <8 x i16> %a
 }
 
-; CHECK-LABEL: shl_const_v8i16:
-; NO-SIMD128-NOT: i16x8
-; SIMD128-NEXT: .functype shl_const_v8i16 (v128) -> (v128){{$}}
-; SIMD128-NEXT: i32.const $push[[L0:[0-9]+]]=, 5
-; SIMD128-NEXT: i16x8.shl $push[[R:[0-9]+]]=, $0, $pop[[L0]]{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <8 x i16> @shl_const_v8i16(<8 x i16> %v) {
+; SIMD128-UNIMPL-SLOW-LABEL: shl_const_v8i16:
+; SIMD128-UNIMPL-SLOW:         .functype shl_const_v8i16 (v128) -> (v128)
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.const $push0=, 5
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.shl $push1=, $0, $pop0
+; SIMD128-UNIMPL-SLOW-NEXT:    return $pop1
+;
+; SIMD128-UNIMPL-FAST-LABEL: shl_const_v8i16:
+; SIMD128-UNIMPL-FAST:         .functype shl_const_v8i16 (v128) -> (v128)
+; SIMD128-UNIMPL-FAST-NEXT:    i32.const $push1=, 5
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.shl $push0=, $0, $pop1
+; SIMD128-UNIMPL-FAST-NEXT:    return $pop0
+;
+; SIMD128-VM-SLOW-LABEL: shl_const_v8i16:
+; SIMD128-VM-SLOW:         .functype shl_const_v8i16 (v128) -> (v128)
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push3=, $0, 0
+; SIMD128-VM-SLOW-NEXT:    i32.const $push1=, 5
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push4=, $pop3, $pop1
+; SIMD128-VM-SLOW-NEXT:    i16x8.splat $push5=, $pop4
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push0=, $0, 1
+; SIMD128-VM-SLOW-NEXT:    i32.const $push31=, 5
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push2=, $pop0, $pop31
+; SIMD128-VM-SLOW-NEXT:    i16x8.replace_lane $push6=, $pop5, 1, $pop2
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push7=, $0, 2
+; SIMD128-VM-SLOW-NEXT:    i32.const $push30=, 5
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push8=, $pop7, $pop30
+; SIMD128-VM-SLOW-NEXT:    i16x8.replace_lane $push9=, $pop6, 2, $pop8
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push10=, $0, 3
+; SIMD128-VM-SLOW-NEXT:    i32.const $push29=, 5
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push11=, $pop10, $pop29
+; SIMD128-VM-SLOW-NEXT:    i16x8.replace_lane $push12=, $pop9, 3, $pop11
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push13=, $0, 4
+; SIMD128-VM-SLOW-NEXT:    i32.const $push28=, 5
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push14=, $pop13, $pop28
+; SIMD128-VM-SLOW-NEXT:    i16x8.replace_lane $push15=, $pop12, 4, $pop14
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push16=, $0, 5
+; SIMD128-VM-SLOW-NEXT:    i32.const $push27=, 5
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push17=, $pop16, $pop27
+; SIMD128-VM-SLOW-NEXT:    i16x8.replace_lane $push18=, $pop15, 5, $pop17
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push19=, $0, 6
+; SIMD128-VM-SLOW-NEXT:    i32.const $push26=, 5
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push20=, $pop19, $pop26
+; SIMD128-VM-SLOW-NEXT:    i16x8.replace_lane $push21=, $pop18, 6, $pop20
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push22=, $0, 7
+; SIMD128-VM-SLOW-NEXT:    i32.const $push25=, 5
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push23=, $pop22, $pop25
+; SIMD128-VM-SLOW-NEXT:    i16x8.replace_lane $push24=, $pop21, 7, $pop23
+; SIMD128-VM-SLOW-NEXT:    return $pop24
+;
+; SIMD128-VM-FAST-LABEL: shl_const_v8i16:
+; SIMD128-VM-FAST:         .functype shl_const_v8i16 (v128) -> (v128)
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push4=, $0, 0
+; SIMD128-VM-FAST-NEXT:    i32.const $push2=, 5
+; SIMD128-VM-FAST-NEXT:    i32.shl $push5=, $pop4, $pop2
+; SIMD128-VM-FAST-NEXT:    i16x8.splat $push6=, $pop5
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push1=, $0, 1
+; SIMD128-VM-FAST-NEXT:    i32.const $push31=, 5
+; SIMD128-VM-FAST-NEXT:    i32.shl $push3=, $pop1, $pop31
+; SIMD128-VM-FAST-NEXT:    i16x8.replace_lane $push7=, $pop6, 1, $pop3
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push8=, $0, 2
+; SIMD128-VM-FAST-NEXT:    i32.const $push30=, 5
+; SIMD128-VM-FAST-NEXT:    i32.shl $push9=, $pop8, $pop30
+; SIMD128-VM-FAST-NEXT:    i16x8.replace_lane $push10=, $pop7, 2, $pop9
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push11=, $0, 3
+; SIMD128-VM-FAST-NEXT:    i32.const $push29=, 5
+; SIMD128-VM-FAST-NEXT:    i32.shl $push12=, $pop11, $pop29
+; SIMD128-VM-FAST-NEXT:    i16x8.replace_lane $push13=, $pop10, 3, $pop12
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push14=, $0, 4
+; SIMD128-VM-FAST-NEXT:    i32.const $push28=, 5
+; SIMD128-VM-FAST-NEXT:    i32.shl $push15=, $pop14, $pop28
+; SIMD128-VM-FAST-NEXT:    i16x8.replace_lane $push16=, $pop13, 4, $pop15
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push17=, $0, 5
+; SIMD128-VM-FAST-NEXT:    i32.const $push27=, 5
+; SIMD128-VM-FAST-NEXT:    i32.shl $push18=, $pop17, $pop27
+; SIMD128-VM-FAST-NEXT:    i16x8.replace_lane $push19=, $pop16, 5, $pop18
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push20=, $0, 6
+; SIMD128-VM-FAST-NEXT:    i32.const $push26=, 5
+; SIMD128-VM-FAST-NEXT:    i32.shl $push21=, $pop20, $pop26
+; SIMD128-VM-FAST-NEXT:    i16x8.replace_lane $push22=, $pop19, 6, $pop21
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push23=, $0, 7
+; SIMD128-VM-FAST-NEXT:    i32.const $push25=, 5
+; SIMD128-VM-FAST-NEXT:    i32.shl $push24=, $pop23, $pop25
+; SIMD128-VM-FAST-NEXT:    i16x8.replace_lane $push0=, $pop22, 7, $pop24
+; SIMD128-VM-FAST-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: shl_const_v8i16:
+; NO-SIMD128-SLOW:         .functype shl_const_v8i16 (i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push2=, 14
+; NO-SIMD128-SLOW-NEXT:    i32.add $push3=, $0, $pop2
+; NO-SIMD128-SLOW-NEXT:    i32.const $push0=, 5
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push1=, $8, $pop0
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop3), $pop1
+; NO-SIMD128-SLOW-NEXT:    i32.const $push5=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push6=, $0, $pop5
+; NO-SIMD128-SLOW-NEXT:    i32.const $push23=, 5
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push4=, $7, $pop23
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop6), $pop4
+; NO-SIMD128-SLOW-NEXT:    i32.const $push8=, 10
+; NO-SIMD128-SLOW-NEXT:    i32.add $push9=, $0, $pop8
+; NO-SIMD128-SLOW-NEXT:    i32.const $push22=, 5
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push7=, $6, $pop22
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop9), $pop7
+; NO-SIMD128-SLOW-NEXT:    i32.const $push21=, 5
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push10=, $5, $pop21
+; NO-SIMD128-SLOW-NEXT:    i32.store16 8($0), $pop10
+; NO-SIMD128-SLOW-NEXT:    i32.const $push12=, 6
+; NO-SIMD128-SLOW-NEXT:    i32.add $push13=, $0, $pop12
+; NO-SIMD128-SLOW-NEXT:    i32.const $push20=, 5
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push11=, $4, $pop20
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop13), $pop11
+; NO-SIMD128-SLOW-NEXT:    i32.const $push19=, 5
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push14=, $3, $pop19
+; NO-SIMD128-SLOW-NEXT:    i32.store16 4($0), $pop14
+; NO-SIMD128-SLOW-NEXT:    i32.const $push18=, 5
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push15=, $2, $pop18
+; NO-SIMD128-SLOW-NEXT:    i32.store16 2($0), $pop15
+; NO-SIMD128-SLOW-NEXT:    i32.const $push17=, 5
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push16=, $1, $pop17
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($0), $pop16
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: shl_const_v8i16:
+; NO-SIMD128-FAST:         .functype shl_const_v8i16 (i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-FAST-NEXT:    i32.const $push0=, 5
+; NO-SIMD128-FAST-NEXT:    i32.shl $push1=, $1, $pop0
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($0), $pop1
+; NO-SIMD128-FAST-NEXT:    i32.const $push23=, 5
+; NO-SIMD128-FAST-NEXT:    i32.shl $push2=, $2, $pop23
+; NO-SIMD128-FAST-NEXT:    i32.store16 2($0), $pop2
+; NO-SIMD128-FAST-NEXT:    i32.const $push22=, 5
+; NO-SIMD128-FAST-NEXT:    i32.shl $push3=, $3, $pop22
+; NO-SIMD128-FAST-NEXT:    i32.store16 4($0), $pop3
+; NO-SIMD128-FAST-NEXT:    i32.const $push4=, 6
+; NO-SIMD128-FAST-NEXT:    i32.add $push5=, $0, $pop4
+; NO-SIMD128-FAST-NEXT:    i32.const $push21=, 5
+; NO-SIMD128-FAST-NEXT:    i32.shl $push6=, $4, $pop21
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop5), $pop6
+; NO-SIMD128-FAST-NEXT:    i32.const $push20=, 5
+; NO-SIMD128-FAST-NEXT:    i32.shl $push7=, $5, $pop20
+; NO-SIMD128-FAST-NEXT:    i32.store16 8($0), $pop7
+; NO-SIMD128-FAST-NEXT:    i32.const $push8=, 10
+; NO-SIMD128-FAST-NEXT:    i32.add $push9=, $0, $pop8
+; NO-SIMD128-FAST-NEXT:    i32.const $push19=, 5
+; NO-SIMD128-FAST-NEXT:    i32.shl $push10=, $6, $pop19
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop9), $pop10
+; NO-SIMD128-FAST-NEXT:    i32.const $push11=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push12=, $0, $pop11
+; NO-SIMD128-FAST-NEXT:    i32.const $push18=, 5
+; NO-SIMD128-FAST-NEXT:    i32.shl $push13=, $7, $pop18
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop12), $pop13
+; NO-SIMD128-FAST-NEXT:    i32.const $push14=, 14
+; NO-SIMD128-FAST-NEXT:    i32.add $push15=, $0, $pop14
+; NO-SIMD128-FAST-NEXT:    i32.const $push17=, 5
+; NO-SIMD128-FAST-NEXT:    i32.shl $push16=, $8, $pop17
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop15), $pop16
+; NO-SIMD128-FAST-NEXT:    return
   %a = shl <8 x i16> %v,
     <i16 5, i16 5, i16 5, i16 5, i16 5, i16 5, i16 5, i16 5>
   ret <8 x i16> %a
 }
 
-; CHECK-LABEL: shl_vec_v8i16:
-; NO-SIMD128-NOT: i16x8
-; SIMD128-NEXT: .functype shl_vec_v8i16 (v128, v128) -> (v128){{$}}
-; SIMD128-NEXT: i16x8.extract_lane_s $push[[L0:[0-9]+]]=, $0, 0{{$}}
-; SIMD128-NEXT: i32.const $push[[M0:[0-9]+]]=, 15{{$}}
-; SIMD128-NEXT: i16x8.splat $push[[M1:[0-9]+]]=, $pop[[M0]]{{$}}
-; SIMD128-NEXT: v128.and $push[[M2:[0-9]+]]=, $1, $pop[[M1]]{{$}}
-; SIMD128-NEXT: local.tee $push[[M:[0-9]+]]=, $1=, $pop[[M2]]{{$}}
-; SIMD128-NEXT: i16x8.extract_lane_u $push[[L1:[0-9]+]]=, $pop[[M]], 0{{$}}
-; SIMD128-NEXT: i32.shl $push[[L2:[0-9]+]]=, $pop[[L0]], $pop[[L1]]{{$}}
-; SIMD128-NEXT: i16x8.splat $push[[L3:[0-9]+]]=, $pop[[L2]]{{$}}
-; Skip 6 lanes
-; SIMD128:      i16x8.extract_lane_s $push[[L4:[0-9]+]]=, $0, 7{{$}}
-; SIMD128-NEXT: i16x8.extract_lane_u $push[[L5:[0-9]+]]=, $1, 7{{$}}
-; SIMD128-NEXT: i32.shl $push[[L6:[0-9]+]]=, $pop[[L4]], $pop[[L5]]{{$}}
-; SIMD128-NEXT: i16x8.replace_lane $push[[R:[0-9]+]]=, $pop[[L7:[0-9]+]], 7, $pop[[L6]]{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <8 x i16> @shl_vec_v8i16(<8 x i16> %v, <8 x i16> %x) {
+; SIMD128-UNIMPL-SLOW-LABEL: shl_vec_v8i16:
+; SIMD128-UNIMPL-SLOW:         .functype shl_vec_v8i16 (v128, v128) -> (v128)
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.extract_lane_s $push6=, $0, 0
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.const $push0=, 15
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.splat $push1=, $pop0
+; SIMD128-UNIMPL-SLOW-NEXT:    v128.and $push35=, $1, $pop1
+; SIMD128-UNIMPL-SLOW-NEXT:    local.tee $push34=, $1=, $pop35
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.extract_lane_u $push5=, $pop34, 0
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shl $push7=, $pop6, $pop5
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.splat $push8=, $pop7
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.extract_lane_s $push3=, $0, 1
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.extract_lane_u $push2=, $1, 1
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shl $push4=, $pop3, $pop2
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.replace_lane $push9=, $pop8, 1, $pop4
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.extract_lane_s $push11=, $0, 2
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.extract_lane_u $push10=, $1, 2
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shl $push12=, $pop11, $pop10
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.replace_lane $push13=, $pop9, 2, $pop12
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.extract_lane_s $push15=, $0, 3
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.extract_lane_u $push14=, $1, 3
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shl $push16=, $pop15, $pop14
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.replace_lane $push17=, $pop13, 3, $pop16
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.extract_lane_s $push19=, $0, 4
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.extract_lane_u $push18=, $1, 4
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shl $push20=, $pop19, $pop18
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.replace_lane $push21=, $pop17, 4, $pop20
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.extract_lane_s $push23=, $0, 5
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.extract_lane_u $push22=, $1, 5
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shl $push24=, $pop23, $pop22
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.replace_lane $push25=, $pop21, 5, $pop24
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.extract_lane_s $push27=, $0, 6
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.extract_lane_u $push26=, $1, 6
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shl $push28=, $pop27, $pop26
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.replace_lane $push29=, $pop25, 6, $pop28
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.extract_lane_s $push31=, $0, 7
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.extract_lane_u $push30=, $1, 7
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shl $push32=, $pop31, $pop30
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.replace_lane $push33=, $pop29, 7, $pop32
+; SIMD128-UNIMPL-SLOW-NEXT:    return $pop33
+;
+; SIMD128-UNIMPL-FAST-LABEL: shl_vec_v8i16:
+; SIMD128-UNIMPL-FAST:         .functype shl_vec_v8i16 (v128, v128) -> (v128)
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.extract_lane_s $push7=, $0, 0
+; SIMD128-UNIMPL-FAST-NEXT:    i32.const $push1=, 15
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.splat $push2=, $pop1
+; SIMD128-UNIMPL-FAST-NEXT:    v128.and $push35=, $1, $pop2
+; SIMD128-UNIMPL-FAST-NEXT:    local.tee $push34=, $1=, $pop35
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.extract_lane_u $push6=, $pop34, 0
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shl $push8=, $pop7, $pop6
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.splat $push9=, $pop8
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.extract_lane_s $push4=, $0, 1
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.extract_lane_u $push3=, $1, 1
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shl $push5=, $pop4, $pop3
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.replace_lane $push10=, $pop9, 1, $pop5
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.extract_lane_s $push12=, $0, 2
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.extract_lane_u $push11=, $1, 2
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shl $push13=, $pop12, $pop11
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.replace_lane $push14=, $pop10, 2, $pop13
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.extract_lane_s $push16=, $0, 3
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.extract_lane_u $push15=, $1, 3
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shl $push17=, $pop16, $pop15
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.replace_lane $push18=, $pop14, 3, $pop17
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.extract_lane_s $push20=, $0, 4
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.extract_lane_u $push19=, $1, 4
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shl $push21=, $pop20, $pop19
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.replace_lane $push22=, $pop18, 4, $pop21
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.extract_lane_s $push24=, $0, 5
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.extract_lane_u $push23=, $1, 5
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shl $push25=, $pop24, $pop23
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.replace_lane $push26=, $pop22, 5, $pop25
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.extract_lane_s $push28=, $0, 6
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.extract_lane_u $push27=, $1, 6
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shl $push29=, $pop28, $pop27
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.replace_lane $push30=, $pop26, 6, $pop29
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.extract_lane_s $push32=, $0, 7
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.extract_lane_u $push31=, $1, 7
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shl $push33=, $pop32, $pop31
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.replace_lane $push0=, $pop30, 7, $pop33
+; SIMD128-UNIMPL-FAST-NEXT:    return $pop0
+;
+; SIMD128-VM-SLOW-LABEL: shl_vec_v8i16:
+; SIMD128-VM-SLOW:         .functype shl_vec_v8i16 (v128, v128) -> (v128)
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push8=, $0, 0
+; SIMD128-VM-SLOW-NEXT:    i32.const $push0=, 15
+; SIMD128-VM-SLOW-NEXT:    i16x8.splat $push1=, $pop0
+; SIMD128-VM-SLOW-NEXT:    v128.and $push51=, $1, $pop1
+; SIMD128-VM-SLOW-NEXT:    local.tee $push50=, $1=, $pop51
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push6=, $pop50, 0
+; SIMD128-VM-SLOW-NEXT:    i32.const $push49=, 15
+; SIMD128-VM-SLOW-NEXT:    i32.and $push7=, $pop6, $pop49
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push9=, $pop8, $pop7
+; SIMD128-VM-SLOW-NEXT:    i16x8.splat $push10=, $pop9
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push4=, $0, 1
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push2=, $1, 1
+; SIMD128-VM-SLOW-NEXT:    i32.const $push48=, 15
+; SIMD128-VM-SLOW-NEXT:    i32.and $push3=, $pop2, $pop48
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push5=, $pop4, $pop3
+; SIMD128-VM-SLOW-NEXT:    i16x8.replace_lane $push11=, $pop10, 1, $pop5
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push14=, $0, 2
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push12=, $1, 2
+; SIMD128-VM-SLOW-NEXT:    i32.const $push47=, 15
+; SIMD128-VM-SLOW-NEXT:    i32.and $push13=, $pop12, $pop47
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push15=, $pop14, $pop13
+; SIMD128-VM-SLOW-NEXT:    i16x8.replace_lane $push16=, $pop11, 2, $pop15
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push19=, $0, 3
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push17=, $1, 3
+; SIMD128-VM-SLOW-NEXT:    i32.const $push46=, 15
+; SIMD128-VM-SLOW-NEXT:    i32.and $push18=, $pop17, $pop46
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push20=, $pop19, $pop18
+; SIMD128-VM-SLOW-NEXT:    i16x8.replace_lane $push21=, $pop16, 3, $pop20
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push24=, $0, 4
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push22=, $1, 4
+; SIMD128-VM-SLOW-NEXT:    i32.const $push45=, 15
+; SIMD128-VM-SLOW-NEXT:    i32.and $push23=, $pop22, $pop45
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push25=, $pop24, $pop23
+; SIMD128-VM-SLOW-NEXT:    i16x8.replace_lane $push26=, $pop21, 4, $pop25
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push29=, $0, 5
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push27=, $1, 5
+; SIMD128-VM-SLOW-NEXT:    i32.const $push44=, 15
+; SIMD128-VM-SLOW-NEXT:    i32.and $push28=, $pop27, $pop44
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push30=, $pop29, $pop28
+; SIMD128-VM-SLOW-NEXT:    i16x8.replace_lane $push31=, $pop26, 5, $pop30
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push34=, $0, 6
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push32=, $1, 6
+; SIMD128-VM-SLOW-NEXT:    i32.const $push43=, 15
+; SIMD128-VM-SLOW-NEXT:    i32.and $push33=, $pop32, $pop43
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push35=, $pop34, $pop33
+; SIMD128-VM-SLOW-NEXT:    i16x8.replace_lane $push36=, $pop31, 6, $pop35
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push39=, $0, 7
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push37=, $1, 7
+; SIMD128-VM-SLOW-NEXT:    i32.const $push42=, 15
+; SIMD128-VM-SLOW-NEXT:    i32.and $push38=, $pop37, $pop42
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push40=, $pop39, $pop38
+; SIMD128-VM-SLOW-NEXT:    i16x8.replace_lane $push41=, $pop36, 7, $pop40
+; SIMD128-VM-SLOW-NEXT:    return $pop41
+;
+; SIMD128-VM-FAST-LABEL: shl_vec_v8i16:
+; SIMD128-VM-FAST:         .functype shl_vec_v8i16 (v128, v128) -> (v128)
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push9=, $0, 0
+; SIMD128-VM-FAST-NEXT:    i32.const $push1=, 15
+; SIMD128-VM-FAST-NEXT:    i16x8.splat $push2=, $pop1
+; SIMD128-VM-FAST-NEXT:    v128.and $push51=, $1, $pop2
+; SIMD128-VM-FAST-NEXT:    local.tee $push50=, $1=, $pop51
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push7=, $pop50, 0
+; SIMD128-VM-FAST-NEXT:    i32.const $push49=, 15
+; SIMD128-VM-FAST-NEXT:    i32.and $push8=, $pop7, $pop49
+; SIMD128-VM-FAST-NEXT:    i32.shl $push10=, $pop9, $pop8
+; SIMD128-VM-FAST-NEXT:    i16x8.splat $push11=, $pop10
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push5=, $0, 1
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push3=, $1, 1
+; SIMD128-VM-FAST-NEXT:    i32.const $push48=, 15
+; SIMD128-VM-FAST-NEXT:    i32.and $push4=, $pop3, $pop48
+; SIMD128-VM-FAST-NEXT:    i32.shl $push6=, $pop5, $pop4
+; SIMD128-VM-FAST-NEXT:    i16x8.replace_lane $push12=, $pop11, 1, $pop6
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push15=, $0, 2
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push13=, $1, 2
+; SIMD128-VM-FAST-NEXT:    i32.const $push47=, 15
+; SIMD128-VM-FAST-NEXT:    i32.and $push14=, $pop13, $pop47
+; SIMD128-VM-FAST-NEXT:    i32.shl $push16=, $pop15, $pop14
+; SIMD128-VM-FAST-NEXT:    i16x8.replace_lane $push17=, $pop12, 2, $pop16
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push20=, $0, 3
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push18=, $1, 3
+; SIMD128-VM-FAST-NEXT:    i32.const $push46=, 15
+; SIMD128-VM-FAST-NEXT:    i32.and $push19=, $pop18, $pop46
+; SIMD128-VM-FAST-NEXT:    i32.shl $push21=, $pop20, $pop19
+; SIMD128-VM-FAST-NEXT:    i16x8.replace_lane $push22=, $pop17, 3, $pop21
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push25=, $0, 4
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push23=, $1, 4
+; SIMD128-VM-FAST-NEXT:    i32.const $push45=, 15
+; SIMD128-VM-FAST-NEXT:    i32.and $push24=, $pop23, $pop45
+; SIMD128-VM-FAST-NEXT:    i32.shl $push26=, $pop25, $pop24
+; SIMD128-VM-FAST-NEXT:    i16x8.replace_lane $push27=, $pop22, 4, $pop26
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push30=, $0, 5
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push28=, $1, 5
+; SIMD128-VM-FAST-NEXT:    i32.const $push44=, 15
+; SIMD128-VM-FAST-NEXT:    i32.and $push29=, $pop28, $pop44
+; SIMD128-VM-FAST-NEXT:    i32.shl $push31=, $pop30, $pop29
+; SIMD128-VM-FAST-NEXT:    i16x8.replace_lane $push32=, $pop27, 5, $pop31
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push35=, $0, 6
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push33=, $1, 6
+; SIMD128-VM-FAST-NEXT:    i32.const $push43=, 15
+; SIMD128-VM-FAST-NEXT:    i32.and $push34=, $pop33, $pop43
+; SIMD128-VM-FAST-NEXT:    i32.shl $push36=, $pop35, $pop34
+; SIMD128-VM-FAST-NEXT:    i16x8.replace_lane $push37=, $pop32, 6, $pop36
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push40=, $0, 7
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push38=, $1, 7
+; SIMD128-VM-FAST-NEXT:    i32.const $push42=, 15
+; SIMD128-VM-FAST-NEXT:    i32.and $push39=, $pop38, $pop42
+; SIMD128-VM-FAST-NEXT:    i32.shl $push41=, $pop40, $pop39
+; SIMD128-VM-FAST-NEXT:    i16x8.replace_lane $push0=, $pop37, 7, $pop41
+; SIMD128-VM-FAST-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: shl_vec_v8i16:
+; NO-SIMD128-SLOW:         .functype shl_vec_v8i16 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push0=, 14
+; NO-SIMD128-SLOW-NEXT:    i32.add $push1=, $0, $pop0
+; NO-SIMD128-SLOW-NEXT:    i32.const $push2=, 65535
+; NO-SIMD128-SLOW-NEXT:    i32.and $push3=, $16, $pop2
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push4=, $8, $pop3
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop1), $pop4
+; NO-SIMD128-SLOW-NEXT:    i32.const $push5=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push6=, $0, $pop5
+; NO-SIMD128-SLOW-NEXT:    i32.const $push31=, 65535
+; NO-SIMD128-SLOW-NEXT:    i32.and $push7=, $15, $pop31
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push8=, $7, $pop7
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop6), $pop8
+; NO-SIMD128-SLOW-NEXT:    i32.const $push9=, 10
+; NO-SIMD128-SLOW-NEXT:    i32.add $push10=, $0, $pop9
+; NO-SIMD128-SLOW-NEXT:    i32.const $push30=, 65535
+; NO-SIMD128-SLOW-NEXT:    i32.and $push11=, $14, $pop30
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push12=, $6, $pop11
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop10), $pop12
+; NO-SIMD128-SLOW-NEXT:    i32.const $push29=, 65535
+; NO-SIMD128-SLOW-NEXT:    i32.and $push13=, $13, $pop29
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push14=, $5, $pop13
+; NO-SIMD128-SLOW-NEXT:    i32.store16 8($0), $pop14
+; NO-SIMD128-SLOW-NEXT:    i32.const $push15=, 6
+; NO-SIMD128-SLOW-NEXT:    i32.add $push16=, $0, $pop15
+; NO-SIMD128-SLOW-NEXT:    i32.const $push28=, 65535
+; NO-SIMD128-SLOW-NEXT:    i32.and $push17=, $12, $pop28
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push18=, $4, $pop17
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop16), $pop18
+; NO-SIMD128-SLOW-NEXT:    i32.const $push27=, 65535
+; NO-SIMD128-SLOW-NEXT:    i32.and $push19=, $11, $pop27
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push20=, $3, $pop19
+; NO-SIMD128-SLOW-NEXT:    i32.store16 4($0), $pop20
+; NO-SIMD128-SLOW-NEXT:    i32.const $push26=, 65535
+; NO-SIMD128-SLOW-NEXT:    i32.and $push21=, $10, $pop26
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push22=, $2, $pop21
+; NO-SIMD128-SLOW-NEXT:    i32.store16 2($0), $pop22
+; NO-SIMD128-SLOW-NEXT:    i32.const $push25=, 65535
+; NO-SIMD128-SLOW-NEXT:    i32.and $push23=, $9, $pop25
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push24=, $1, $pop23
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($0), $pop24
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: shl_vec_v8i16:
+; NO-SIMD128-FAST:         .functype shl_vec_v8i16 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-FAST-NEXT:    i32.const $push0=, 65535
+; NO-SIMD128-FAST-NEXT:    i32.and $push1=, $9, $pop0
+; NO-SIMD128-FAST-NEXT:    i32.shl $push2=, $1, $pop1
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($0), $pop2
+; NO-SIMD128-FAST-NEXT:    i32.const $push31=, 65535
+; NO-SIMD128-FAST-NEXT:    i32.and $push3=, $10, $pop31
+; NO-SIMD128-FAST-NEXT:    i32.shl $push4=, $2, $pop3
+; NO-SIMD128-FAST-NEXT:    i32.store16 2($0), $pop4
+; NO-SIMD128-FAST-NEXT:    i32.const $push30=, 65535
+; NO-SIMD128-FAST-NEXT:    i32.and $push5=, $11, $pop30
+; NO-SIMD128-FAST-NEXT:    i32.shl $push6=, $3, $pop5
+; NO-SIMD128-FAST-NEXT:    i32.store16 4($0), $pop6
+; NO-SIMD128-FAST-NEXT:    i32.const $push7=, 6
+; NO-SIMD128-FAST-NEXT:    i32.add $push8=, $0, $pop7
+; NO-SIMD128-FAST-NEXT:    i32.const $push29=, 65535
+; NO-SIMD128-FAST-NEXT:    i32.and $push9=, $12, $pop29
+; NO-SIMD128-FAST-NEXT:    i32.shl $push10=, $4, $pop9
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop8), $pop10
+; NO-SIMD128-FAST-NEXT:    i32.const $push28=, 65535
+; NO-SIMD128-FAST-NEXT:    i32.and $push11=, $13, $pop28
+; NO-SIMD128-FAST-NEXT:    i32.shl $push12=, $5, $pop11
+; NO-SIMD128-FAST-NEXT:    i32.store16 8($0), $pop12
+; NO-SIMD128-FAST-NEXT:    i32.const $push13=, 10
+; NO-SIMD128-FAST-NEXT:    i32.add $push14=, $0, $pop13
+; NO-SIMD128-FAST-NEXT:    i32.const $push27=, 65535
+; NO-SIMD128-FAST-NEXT:    i32.and $push15=, $14, $pop27
+; NO-SIMD128-FAST-NEXT:    i32.shl $push16=, $6, $pop15
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop14), $pop16
+; NO-SIMD128-FAST-NEXT:    i32.const $push17=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push18=, $0, $pop17
+; NO-SIMD128-FAST-NEXT:    i32.const $push26=, 65535
+; NO-SIMD128-FAST-NEXT:    i32.and $push19=, $15, $pop26
+; NO-SIMD128-FAST-NEXT:    i32.shl $push20=, $7, $pop19
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop18), $pop20
+; NO-SIMD128-FAST-NEXT:    i32.const $push21=, 14
+; NO-SIMD128-FAST-NEXT:    i32.add $push22=, $0, $pop21
+; NO-SIMD128-FAST-NEXT:    i32.const $push25=, 65535
+; NO-SIMD128-FAST-NEXT:    i32.and $push23=, $16, $pop25
+; NO-SIMD128-FAST-NEXT:    i32.shl $push24=, $8, $pop23
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop22), $pop24
+; NO-SIMD128-FAST-NEXT:    return
   %a = shl <8 x i16> %v, %x
   ret <8 x i16> %a
 }
 
-; CHECK-LABEL: shr_s_v8i16:
-; NO-SIMD128-NOT: i16x8
-; SIMD128-NEXT: .functype shr_s_v8i16 (v128, i32) -> (v128){{$}}
-; SIMD128-NEXT: i16x8.shr_s $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <8 x i16> @shr_s_v8i16(<8 x i16> %v, i16 %x) {
+; SIMD128-UNIMPL-LABEL: shr_s_v8i16:
+; SIMD128-UNIMPL:         .functype shr_s_v8i16 (v128, i32) -> (v128)
+; SIMD128-UNIMPL-NEXT:    i16x8.shr_s $push0=, $0, $1
+; SIMD128-UNIMPL-NEXT:    return $pop0
+;
+; SIMD128-VM-SLOW-LABEL: shr_s_v8i16:
+; SIMD128-VM-SLOW:         .functype shr_s_v8i16 (v128, i32) -> (v128)
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push9=, $0, 0
+; SIMD128-VM-SLOW-NEXT:    i16x8.splat $push2=, $1
+; SIMD128-VM-SLOW-NEXT:    i32.const $push0=, 15
+; SIMD128-VM-SLOW-NEXT:    i16x8.splat $push1=, $pop0
+; SIMD128-VM-SLOW-NEXT:    v128.and $push52=, $pop2, $pop1
+; SIMD128-VM-SLOW-NEXT:    local.tee $push51=, $2=, $pop52
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push7=, $pop51, 0
+; SIMD128-VM-SLOW-NEXT:    i32.const $push50=, 15
+; SIMD128-VM-SLOW-NEXT:    i32.and $push8=, $pop7, $pop50
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push10=, $pop9, $pop8
+; SIMD128-VM-SLOW-NEXT:    i16x8.splat $push11=, $pop10
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push5=, $0, 1
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push3=, $2, 1
+; SIMD128-VM-SLOW-NEXT:    i32.const $push49=, 15
+; SIMD128-VM-SLOW-NEXT:    i32.and $push4=, $pop3, $pop49
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push6=, $pop5, $pop4
+; SIMD128-VM-SLOW-NEXT:    i16x8.replace_lane $push12=, $pop11, 1, $pop6
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push15=, $0, 2
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push13=, $2, 2
+; SIMD128-VM-SLOW-NEXT:    i32.const $push48=, 15
+; SIMD128-VM-SLOW-NEXT:    i32.and $push14=, $pop13, $pop48
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push16=, $pop15, $pop14
+; SIMD128-VM-SLOW-NEXT:    i16x8.replace_lane $push17=, $pop12, 2, $pop16
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push20=, $0, 3
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push18=, $2, 3
+; SIMD128-VM-SLOW-NEXT:    i32.const $push47=, 15
+; SIMD128-VM-SLOW-NEXT:    i32.and $push19=, $pop18, $pop47
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push21=, $pop20, $pop19
+; SIMD128-VM-SLOW-NEXT:    i16x8.replace_lane $push22=, $pop17, 3, $pop21
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push25=, $0, 4
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push23=, $2, 4
+; SIMD128-VM-SLOW-NEXT:    i32.const $push46=, 15
+; SIMD128-VM-SLOW-NEXT:    i32.and $push24=, $pop23, $pop46
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push26=, $pop25, $pop24
+; SIMD128-VM-SLOW-NEXT:    i16x8.replace_lane $push27=, $pop22, 4, $pop26
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push30=, $0, 5
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push28=, $2, 5
+; SIMD128-VM-SLOW-NEXT:    i32.const $push45=, 15
+; SIMD128-VM-SLOW-NEXT:    i32.and $push29=, $pop28, $pop45
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push31=, $pop30, $pop29
+; SIMD128-VM-SLOW-NEXT:    i16x8.replace_lane $push32=, $pop27, 5, $pop31
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push35=, $0, 6
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push33=, $2, 6
+; SIMD128-VM-SLOW-NEXT:    i32.const $push44=, 15
+; SIMD128-VM-SLOW-NEXT:    i32.and $push34=, $pop33, $pop44
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push36=, $pop35, $pop34
+; SIMD128-VM-SLOW-NEXT:    i16x8.replace_lane $push37=, $pop32, 6, $pop36
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push40=, $0, 7
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push38=, $2, 7
+; SIMD128-VM-SLOW-NEXT:    i32.const $push43=, 15
+; SIMD128-VM-SLOW-NEXT:    i32.and $push39=, $pop38, $pop43
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push41=, $pop40, $pop39
+; SIMD128-VM-SLOW-NEXT:    i16x8.replace_lane $push42=, $pop37, 7, $pop41
+; SIMD128-VM-SLOW-NEXT:    return $pop42
+;
+; SIMD128-VM-FAST-LABEL: shr_s_v8i16:
+; SIMD128-VM-FAST:         .functype shr_s_v8i16 (v128, i32) -> (v128)
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push10=, $0, 0
+; SIMD128-VM-FAST-NEXT:    i16x8.splat $push1=, $1
+; SIMD128-VM-FAST-NEXT:    i32.const $push2=, 15
+; SIMD128-VM-FAST-NEXT:    i16x8.splat $push3=, $pop2
+; SIMD128-VM-FAST-NEXT:    v128.and $push52=, $pop1, $pop3
+; SIMD128-VM-FAST-NEXT:    local.tee $push51=, $2=, $pop52
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push8=, $pop51, 0
+; SIMD128-VM-FAST-NEXT:    i32.const $push50=, 15
+; SIMD128-VM-FAST-NEXT:    i32.and $push9=, $pop8, $pop50
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push11=, $pop10, $pop9
+; SIMD128-VM-FAST-NEXT:    i16x8.splat $push12=, $pop11
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push6=, $0, 1
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push4=, $2, 1
+; SIMD128-VM-FAST-NEXT:    i32.const $push49=, 15
+; SIMD128-VM-FAST-NEXT:    i32.and $push5=, $pop4, $pop49
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push7=, $pop6, $pop5
+; SIMD128-VM-FAST-NEXT:    i16x8.replace_lane $push13=, $pop12, 1, $pop7
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push16=, $0, 2
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push14=, $2, 2
+; SIMD128-VM-FAST-NEXT:    i32.const $push48=, 15
+; SIMD128-VM-FAST-NEXT:    i32.and $push15=, $pop14, $pop48
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push17=, $pop16, $pop15
+; SIMD128-VM-FAST-NEXT:    i16x8.replace_lane $push18=, $pop13, 2, $pop17
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push21=, $0, 3
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push19=, $2, 3
+; SIMD128-VM-FAST-NEXT:    i32.const $push47=, 15
+; SIMD128-VM-FAST-NEXT:    i32.and $push20=, $pop19, $pop47
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push22=, $pop21, $pop20
+; SIMD128-VM-FAST-NEXT:    i16x8.replace_lane $push23=, $pop18, 3, $pop22
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push26=, $0, 4
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push24=, $2, 4
+; SIMD128-VM-FAST-NEXT:    i32.const $push46=, 15
+; SIMD128-VM-FAST-NEXT:    i32.and $push25=, $pop24, $pop46
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push27=, $pop26, $pop25
+; SIMD128-VM-FAST-NEXT:    i16x8.replace_lane $push28=, $pop23, 4, $pop27
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push31=, $0, 5
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push29=, $2, 5
+; SIMD128-VM-FAST-NEXT:    i32.const $push45=, 15
+; SIMD128-VM-FAST-NEXT:    i32.and $push30=, $pop29, $pop45
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push32=, $pop31, $pop30
+; SIMD128-VM-FAST-NEXT:    i16x8.replace_lane $push33=, $pop28, 5, $pop32
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push36=, $0, 6
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push34=, $2, 6
+; SIMD128-VM-FAST-NEXT:    i32.const $push44=, 15
+; SIMD128-VM-FAST-NEXT:    i32.and $push35=, $pop34, $pop44
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push37=, $pop36, $pop35
+; SIMD128-VM-FAST-NEXT:    i16x8.replace_lane $push38=, $pop33, 6, $pop37
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push41=, $0, 7
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push39=, $2, 7
+; SIMD128-VM-FAST-NEXT:    i32.const $push43=, 15
+; SIMD128-VM-FAST-NEXT:    i32.and $push40=, $pop39, $pop43
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push42=, $pop41, $pop40
+; SIMD128-VM-FAST-NEXT:    i16x8.replace_lane $push0=, $pop38, 7, $pop42
+; SIMD128-VM-FAST-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: shr_s_v8i16:
+; NO-SIMD128-SLOW:         .functype shr_s_v8i16 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push5=, 14
+; NO-SIMD128-SLOW-NEXT:    i32.add $push6=, $0, $pop5
+; NO-SIMD128-SLOW-NEXT:    i32.const $push1=, 16
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push2=, $8, $pop1
+; NO-SIMD128-SLOW-NEXT:    i32.const $push50=, 16
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push3=, $pop2, $pop50
+; NO-SIMD128-SLOW-NEXT:    i32.const $push0=, 65535
+; NO-SIMD128-SLOW-NEXT:    i32.and $push49=, $9, $pop0
+; NO-SIMD128-SLOW-NEXT:    local.tee $push48=, $8=, $pop49
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push4=, $pop3, $pop48
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop6), $pop4
+; NO-SIMD128-SLOW-NEXT:    i32.const $push10=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push11=, $0, $pop10
+; NO-SIMD128-SLOW-NEXT:    i32.const $push47=, 16
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push7=, $7, $pop47
+; NO-SIMD128-SLOW-NEXT:    i32.const $push46=, 16
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push8=, $pop7, $pop46
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push9=, $pop8, $8
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop11), $pop9
+; NO-SIMD128-SLOW-NEXT:    i32.const $push15=, 10
+; NO-SIMD128-SLOW-NEXT:    i32.add $push16=, $0, $pop15
+; NO-SIMD128-SLOW-NEXT:    i32.const $push45=, 16
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push12=, $6, $pop45
+; NO-SIMD128-SLOW-NEXT:    i32.const $push44=, 16
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push13=, $pop12, $pop44
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push14=, $pop13, $8
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop16), $pop14
+; NO-SIMD128-SLOW-NEXT:    i32.const $push43=, 16
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push17=, $5, $pop43
+; NO-SIMD128-SLOW-NEXT:    i32.const $push42=, 16
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push18=, $pop17, $pop42
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push19=, $pop18, $8
+; NO-SIMD128-SLOW-NEXT:    i32.store16 8($0), $pop19
+; NO-SIMD128-SLOW-NEXT:    i32.const $push23=, 6
+; NO-SIMD128-SLOW-NEXT:    i32.add $push24=, $0, $pop23
+; NO-SIMD128-SLOW-NEXT:    i32.const $push41=, 16
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push20=, $4, $pop41
+; NO-SIMD128-SLOW-NEXT:    i32.const $push40=, 16
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push21=, $pop20, $pop40
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push22=, $pop21, $8
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop24), $pop22
+; NO-SIMD128-SLOW-NEXT:    i32.const $push39=, 16
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push25=, $3, $pop39
+; NO-SIMD128-SLOW-NEXT:    i32.const $push38=, 16
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push26=, $pop25, $pop38
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push27=, $pop26, $8
+; NO-SIMD128-SLOW-NEXT:    i32.store16 4($0), $pop27
+; NO-SIMD128-SLOW-NEXT:    i32.const $push37=, 16
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push28=, $2, $pop37
+; NO-SIMD128-SLOW-NEXT:    i32.const $push36=, 16
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push29=, $pop28, $pop36
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push30=, $pop29, $8
+; NO-SIMD128-SLOW-NEXT:    i32.store16 2($0), $pop30
+; NO-SIMD128-SLOW-NEXT:    i32.const $push35=, 16
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push31=, $1, $pop35
+; NO-SIMD128-SLOW-NEXT:    i32.const $push34=, 16
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push32=, $pop31, $pop34
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push33=, $pop32, $8
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($0), $pop33
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: shr_s_v8i16:
+; NO-SIMD128-FAST:         .functype shr_s_v8i16 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-FAST-NEXT:    i32.const $push1=, 16
+; NO-SIMD128-FAST-NEXT:    i32.shl $push2=, $1, $pop1
+; NO-SIMD128-FAST-NEXT:    i32.const $push50=, 16
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push3=, $pop2, $pop50
+; NO-SIMD128-FAST-NEXT:    i32.const $push0=, 65535
+; NO-SIMD128-FAST-NEXT:    i32.and $push49=, $9, $pop0
+; NO-SIMD128-FAST-NEXT:    local.tee $push48=, $9=, $pop49
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push4=, $pop3, $pop48
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($0), $pop4
+; NO-SIMD128-FAST-NEXT:    i32.const $push47=, 16
+; NO-SIMD128-FAST-NEXT:    i32.shl $push5=, $2, $pop47
+; NO-SIMD128-FAST-NEXT:    i32.const $push46=, 16
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push6=, $pop5, $pop46
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push7=, $pop6, $9
+; NO-SIMD128-FAST-NEXT:    i32.store16 2($0), $pop7
+; NO-SIMD128-FAST-NEXT:    i32.const $push45=, 16
+; NO-SIMD128-FAST-NEXT:    i32.shl $push8=, $3, $pop45
+; NO-SIMD128-FAST-NEXT:    i32.const $push44=, 16
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push9=, $pop8, $pop44
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push10=, $pop9, $9
+; NO-SIMD128-FAST-NEXT:    i32.store16 4($0), $pop10
+; NO-SIMD128-FAST-NEXT:    i32.const $push14=, 6
+; NO-SIMD128-FAST-NEXT:    i32.add $push15=, $0, $pop14
+; NO-SIMD128-FAST-NEXT:    i32.const $push43=, 16
+; NO-SIMD128-FAST-NEXT:    i32.shl $push11=, $4, $pop43
+; NO-SIMD128-FAST-NEXT:    i32.const $push42=, 16
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push12=, $pop11, $pop42
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push13=, $pop12, $9
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop15), $pop13
+; NO-SIMD128-FAST-NEXT:    i32.const $push41=, 16
+; NO-SIMD128-FAST-NEXT:    i32.shl $push16=, $5, $pop41
+; NO-SIMD128-FAST-NEXT:    i32.const $push40=, 16
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push17=, $pop16, $pop40
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push18=, $pop17, $9
+; NO-SIMD128-FAST-NEXT:    i32.store16 8($0), $pop18
+; NO-SIMD128-FAST-NEXT:    i32.const $push22=, 10
+; NO-SIMD128-FAST-NEXT:    i32.add $push23=, $0, $pop22
+; NO-SIMD128-FAST-NEXT:    i32.const $push39=, 16
+; NO-SIMD128-FAST-NEXT:    i32.shl $push19=, $6, $pop39
+; NO-SIMD128-FAST-NEXT:    i32.const $push38=, 16
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push20=, $pop19, $pop38
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push21=, $pop20, $9
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop23), $pop21
+; NO-SIMD128-FAST-NEXT:    i32.const $push27=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push28=, $0, $pop27
+; NO-SIMD128-FAST-NEXT:    i32.const $push37=, 16
+; NO-SIMD128-FAST-NEXT:    i32.shl $push24=, $7, $pop37
+; NO-SIMD128-FAST-NEXT:    i32.const $push36=, 16
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push25=, $pop24, $pop36
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push26=, $pop25, $9
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop28), $pop26
+; NO-SIMD128-FAST-NEXT:    i32.const $push32=, 14
+; NO-SIMD128-FAST-NEXT:    i32.add $push33=, $0, $pop32
+; NO-SIMD128-FAST-NEXT:    i32.const $push35=, 16
+; NO-SIMD128-FAST-NEXT:    i32.shl $push29=, $8, $pop35
+; NO-SIMD128-FAST-NEXT:    i32.const $push34=, 16
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push30=, $pop29, $pop34
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push31=, $pop30, $9
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop33), $pop31
+; NO-SIMD128-FAST-NEXT:    return
   %t = insertelement <8 x i16> undef, i16 %x, i32 0
   %s = shufflevector <8 x i16> %t, <8 x i16> undef,
     <8 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
   %a = ashr <8 x i16> %v, %s
   ret <8 x i16> %a
 }
 
-; CHECK-LABEL: shr_s_vec_v8i16:
-; NO-SIMD128-NOT: i16x8
-; SIMD128-NEXT: .functype shr_s_vec_v8i16 (v128, v128) -> (v128){{$}}
-; SIMD128-NEXT: i16x8.extract_lane_s $push[[L0:[0-9]+]]=, $0, 0{{$}}
-; SIMD128-NEXT: i32.const $push[[M0:[0-9]+]]=, 15{{$}}
-; SIMD128-NEXT: i16x8.splat $push[[M1:[0-9]+]]=, $pop[[M0]]{{$}}
-; SIMD128-NEXT: v128.and $push[[M2:[0-9]+]]=, $1, $pop[[M1]]{{$}}
-; SIMD128-NEXT: local.tee $push[[M:[0-9]+]]=, $1=, $pop[[M2]]{{$}}
-; SIMD128-NEXT: i16x8.extract_lane_u $push[[L1:[0-9]+]]=, $pop[[M]], 0{{$}}
-; SIMD128-NEXT: i32.shr_s $push[[L2:[0-9]+]]=, $pop[[L0]], $pop[[L1]]{{$}}
-; SIMD128-NEXT: i16x8.splat $push[[L3:[0-9]+]]=, $pop[[L2]]{{$}}
-; Skip 6 lanes
-; SIMD128:      i16x8.extract_lane_s $push[[L0:[0-9]+]]=, $0, 7{{$}}
-; SIMD128-NEXT: i16x8.extract_lane_u $push[[L1:[0-9]+]]=, $1, 7{{$}}
-; SIMD128-NEXT: i32.shr_s $push[[L2:[0-9]+]]=, $pop[[L0]], $pop[[L1]]{{$}}
-; SIMD128-NEXT: i16x8.replace_lane $push[[R:[0-9]+]]=, $pop{{[0-9]+}}, 7, $pop[[L2]]{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <8 x i16> @shr_s_vec_v8i16(<8 x i16> %v, <8 x i16> %x) {
+; SIMD128-UNIMPL-SLOW-LABEL: shr_s_vec_v8i16:
+; SIMD128-UNIMPL-SLOW:         .functype shr_s_vec_v8i16 (v128, v128) -> (v128)
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.extract_lane_s $push6=, $0, 0
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.const $push0=, 15
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.splat $push1=, $pop0
+; SIMD128-UNIMPL-SLOW-NEXT:    v128.and $push35=, $1, $pop1
+; SIMD128-UNIMPL-SLOW-NEXT:    local.tee $push34=, $1=, $pop35
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.extract_lane_u $push5=, $pop34, 0
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shr_s $push7=, $pop6, $pop5
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.splat $push8=, $pop7
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.extract_lane_s $push3=, $0, 1
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.extract_lane_u $push2=, $1, 1
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shr_s $push4=, $pop3, $pop2
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.replace_lane $push9=, $pop8, 1, $pop4
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.extract_lane_s $push11=, $0, 2
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.extract_lane_u $push10=, $1, 2
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shr_s $push12=, $pop11, $pop10
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.replace_lane $push13=, $pop9, 2, $pop12
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.extract_lane_s $push15=, $0, 3
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.extract_lane_u $push14=, $1, 3
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shr_s $push16=, $pop15, $pop14
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.replace_lane $push17=, $pop13, 3, $pop16
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.extract_lane_s $push19=, $0, 4
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.extract_lane_u $push18=, $1, 4
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shr_s $push20=, $pop19, $pop18
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.replace_lane $push21=, $pop17, 4, $pop20
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.extract_lane_s $push23=, $0, 5
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.extract_lane_u $push22=, $1, 5
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shr_s $push24=, $pop23, $pop22
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.replace_lane $push25=, $pop21, 5, $pop24
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.extract_lane_s $push27=, $0, 6
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.extract_lane_u $push26=, $1, 6
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shr_s $push28=, $pop27, $pop26
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.replace_lane $push29=, $pop25, 6, $pop28
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.extract_lane_s $push31=, $0, 7
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.extract_lane_u $push30=, $1, 7
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shr_s $push32=, $pop31, $pop30
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.replace_lane $push33=, $pop29, 7, $pop32
+; SIMD128-UNIMPL-SLOW-NEXT:    return $pop33
+;
+; SIMD128-UNIMPL-FAST-LABEL: shr_s_vec_v8i16:
+; SIMD128-UNIMPL-FAST:         .functype shr_s_vec_v8i16 (v128, v128) -> (v128)
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.extract_lane_s $push7=, $0, 0
+; SIMD128-UNIMPL-FAST-NEXT:    i32.const $push1=, 15
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.splat $push2=, $pop1
+; SIMD128-UNIMPL-FAST-NEXT:    v128.and $push35=, $1, $pop2
+; SIMD128-UNIMPL-FAST-NEXT:    local.tee $push34=, $1=, $pop35
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.extract_lane_u $push6=, $pop34, 0
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shr_s $push8=, $pop7, $pop6
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.splat $push9=, $pop8
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.extract_lane_s $push4=, $0, 1
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.extract_lane_u $push3=, $1, 1
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shr_s $push5=, $pop4, $pop3
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.replace_lane $push10=, $pop9, 1, $pop5
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.extract_lane_s $push12=, $0, 2
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.extract_lane_u $push11=, $1, 2
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shr_s $push13=, $pop12, $pop11
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.replace_lane $push14=, $pop10, 2, $pop13
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.extract_lane_s $push16=, $0, 3
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.extract_lane_u $push15=, $1, 3
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shr_s $push17=, $pop16, $pop15
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.replace_lane $push18=, $pop14, 3, $pop17
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.extract_lane_s $push20=, $0, 4
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.extract_lane_u $push19=, $1, 4
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shr_s $push21=, $pop20, $pop19
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.replace_lane $push22=, $pop18, 4, $pop21
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.extract_lane_s $push24=, $0, 5
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.extract_lane_u $push23=, $1, 5
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shr_s $push25=, $pop24, $pop23
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.replace_lane $push26=, $pop22, 5, $pop25
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.extract_lane_s $push28=, $0, 6
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.extract_lane_u $push27=, $1, 6
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shr_s $push29=, $pop28, $pop27
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.replace_lane $push30=, $pop26, 6, $pop29
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.extract_lane_s $push32=, $0, 7
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.extract_lane_u $push31=, $1, 7
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shr_s $push33=, $pop32, $pop31
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.replace_lane $push0=, $pop30, 7, $pop33
+; SIMD128-UNIMPL-FAST-NEXT:    return $pop0
+;
+; SIMD128-VM-SLOW-LABEL: shr_s_vec_v8i16:
+; SIMD128-VM-SLOW:         .functype shr_s_vec_v8i16 (v128, v128) -> (v128)
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push8=, $0, 0
+; SIMD128-VM-SLOW-NEXT:    i32.const $push0=, 15
+; SIMD128-VM-SLOW-NEXT:    i16x8.splat $push1=, $pop0
+; SIMD128-VM-SLOW-NEXT:    v128.and $push51=, $1, $pop1
+; SIMD128-VM-SLOW-NEXT:    local.tee $push50=, $1=, $pop51
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push6=, $pop50, 0
+; SIMD128-VM-SLOW-NEXT:    i32.const $push49=, 15
+; SIMD128-VM-SLOW-NEXT:    i32.and $push7=, $pop6, $pop49
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push9=, $pop8, $pop7
+; SIMD128-VM-SLOW-NEXT:    i16x8.splat $push10=, $pop9
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push4=, $0, 1
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push2=, $1, 1
+; SIMD128-VM-SLOW-NEXT:    i32.const $push48=, 15
+; SIMD128-VM-SLOW-NEXT:    i32.and $push3=, $pop2, $pop48
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push5=, $pop4, $pop3
+; SIMD128-VM-SLOW-NEXT:    i16x8.replace_lane $push11=, $pop10, 1, $pop5
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push14=, $0, 2
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push12=, $1, 2
+; SIMD128-VM-SLOW-NEXT:    i32.const $push47=, 15
+; SIMD128-VM-SLOW-NEXT:    i32.and $push13=, $pop12, $pop47
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push15=, $pop14, $pop13
+; SIMD128-VM-SLOW-NEXT:    i16x8.replace_lane $push16=, $pop11, 2, $pop15
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push19=, $0, 3
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push17=, $1, 3
+; SIMD128-VM-SLOW-NEXT:    i32.const $push46=, 15
+; SIMD128-VM-SLOW-NEXT:    i32.and $push18=, $pop17, $pop46
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push20=, $pop19, $pop18
+; SIMD128-VM-SLOW-NEXT:    i16x8.replace_lane $push21=, $pop16, 3, $pop20
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push24=, $0, 4
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push22=, $1, 4
+; SIMD128-VM-SLOW-NEXT:    i32.const $push45=, 15
+; SIMD128-VM-SLOW-NEXT:    i32.and $push23=, $pop22, $pop45
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push25=, $pop24, $pop23
+; SIMD128-VM-SLOW-NEXT:    i16x8.replace_lane $push26=, $pop21, 4, $pop25
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push29=, $0, 5
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push27=, $1, 5
+; SIMD128-VM-SLOW-NEXT:    i32.const $push44=, 15
+; SIMD128-VM-SLOW-NEXT:    i32.and $push28=, $pop27, $pop44
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push30=, $pop29, $pop28
+; SIMD128-VM-SLOW-NEXT:    i16x8.replace_lane $push31=, $pop26, 5, $pop30
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push34=, $0, 6
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push32=, $1, 6
+; SIMD128-VM-SLOW-NEXT:    i32.const $push43=, 15
+; SIMD128-VM-SLOW-NEXT:    i32.and $push33=, $pop32, $pop43
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push35=, $pop34, $pop33
+; SIMD128-VM-SLOW-NEXT:    i16x8.replace_lane $push36=, $pop31, 6, $pop35
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push39=, $0, 7
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push37=, $1, 7
+; SIMD128-VM-SLOW-NEXT:    i32.const $push42=, 15
+; SIMD128-VM-SLOW-NEXT:    i32.and $push38=, $pop37, $pop42
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push40=, $pop39, $pop38
+; SIMD128-VM-SLOW-NEXT:    i16x8.replace_lane $push41=, $pop36, 7, $pop40
+; SIMD128-VM-SLOW-NEXT:    return $pop41
+;
+; SIMD128-VM-FAST-LABEL: shr_s_vec_v8i16:
+; SIMD128-VM-FAST:         .functype shr_s_vec_v8i16 (v128, v128) -> (v128)
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push9=, $0, 0
+; SIMD128-VM-FAST-NEXT:    i32.const $push1=, 15
+; SIMD128-VM-FAST-NEXT:    i16x8.splat $push2=, $pop1
+; SIMD128-VM-FAST-NEXT:    v128.and $push51=, $1, $pop2
+; SIMD128-VM-FAST-NEXT:    local.tee $push50=, $1=, $pop51
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push7=, $pop50, 0
+; SIMD128-VM-FAST-NEXT:    i32.const $push49=, 15
+; SIMD128-VM-FAST-NEXT:    i32.and $push8=, $pop7, $pop49
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push10=, $pop9, $pop8
+; SIMD128-VM-FAST-NEXT:    i16x8.splat $push11=, $pop10
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push5=, $0, 1
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push3=, $1, 1
+; SIMD128-VM-FAST-NEXT:    i32.const $push48=, 15
+; SIMD128-VM-FAST-NEXT:    i32.and $push4=, $pop3, $pop48
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push6=, $pop5, $pop4
+; SIMD128-VM-FAST-NEXT:    i16x8.replace_lane $push12=, $pop11, 1, $pop6
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push15=, $0, 2
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push13=, $1, 2
+; SIMD128-VM-FAST-NEXT:    i32.const $push47=, 15
+; SIMD128-VM-FAST-NEXT:    i32.and $push14=, $pop13, $pop47
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push16=, $pop15, $pop14
+; SIMD128-VM-FAST-NEXT:    i16x8.replace_lane $push17=, $pop12, 2, $pop16
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push20=, $0, 3
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push18=, $1, 3
+; SIMD128-VM-FAST-NEXT:    i32.const $push46=, 15
+; SIMD128-VM-FAST-NEXT:    i32.and $push19=, $pop18, $pop46
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push21=, $pop20, $pop19
+; SIMD128-VM-FAST-NEXT:    i16x8.replace_lane $push22=, $pop17, 3, $pop21
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push25=, $0, 4
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push23=, $1, 4
+; SIMD128-VM-FAST-NEXT:    i32.const $push45=, 15
+; SIMD128-VM-FAST-NEXT:    i32.and $push24=, $pop23, $pop45
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push26=, $pop25, $pop24
+; SIMD128-VM-FAST-NEXT:    i16x8.replace_lane $push27=, $pop22, 4, $pop26
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push30=, $0, 5
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push28=, $1, 5
+; SIMD128-VM-FAST-NEXT:    i32.const $push44=, 15
+; SIMD128-VM-FAST-NEXT:    i32.and $push29=, $pop28, $pop44
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push31=, $pop30, $pop29
+; SIMD128-VM-FAST-NEXT:    i16x8.replace_lane $push32=, $pop27, 5, $pop31
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push35=, $0, 6
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push33=, $1, 6
+; SIMD128-VM-FAST-NEXT:    i32.const $push43=, 15
+; SIMD128-VM-FAST-NEXT:    i32.and $push34=, $pop33, $pop43
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push36=, $pop35, $pop34
+; SIMD128-VM-FAST-NEXT:    i16x8.replace_lane $push37=, $pop32, 6, $pop36
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push40=, $0, 7
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push38=, $1, 7
+; SIMD128-VM-FAST-NEXT:    i32.const $push42=, 15
+; SIMD128-VM-FAST-NEXT:    i32.and $push39=, $pop38, $pop42
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push41=, $pop40, $pop39
+; SIMD128-VM-FAST-NEXT:    i16x8.replace_lane $push0=, $pop37, 7, $pop41
+; SIMD128-VM-FAST-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: shr_s_vec_v8i16:
+; NO-SIMD128-SLOW:         .functype shr_s_vec_v8i16 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push6=, 14
+; NO-SIMD128-SLOW-NEXT:    i32.add $push7=, $0, $pop6
+; NO-SIMD128-SLOW-NEXT:    i32.const $push2=, 16
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push3=, $8, $pop2
+; NO-SIMD128-SLOW-NEXT:    i32.const $push63=, 16
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push4=, $pop3, $pop63
+; NO-SIMD128-SLOW-NEXT:    i32.const $push0=, 65535
+; NO-SIMD128-SLOW-NEXT:    i32.and $push1=, $16, $pop0
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push5=, $pop4, $pop1
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop7), $pop5
+; NO-SIMD128-SLOW-NEXT:    i32.const $push12=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push13=, $0, $pop12
+; NO-SIMD128-SLOW-NEXT:    i32.const $push62=, 16
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push9=, $7, $pop62
+; NO-SIMD128-SLOW-NEXT:    i32.const $push61=, 16
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push10=, $pop9, $pop61
+; NO-SIMD128-SLOW-NEXT:    i32.const $push60=, 65535
+; NO-SIMD128-SLOW-NEXT:    i32.and $push8=, $15, $pop60
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push11=, $pop10, $pop8
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop13), $pop11
+; NO-SIMD128-SLOW-NEXT:    i32.const $push18=, 10
+; NO-SIMD128-SLOW-NEXT:    i32.add $push19=, $0, $pop18
+; NO-SIMD128-SLOW-NEXT:    i32.const $push59=, 16
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push15=, $6, $pop59
+; NO-SIMD128-SLOW-NEXT:    i32.const $push58=, 16
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push16=, $pop15, $pop58
+; NO-SIMD128-SLOW-NEXT:    i32.const $push57=, 65535
+; NO-SIMD128-SLOW-NEXT:    i32.and $push14=, $14, $pop57
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push17=, $pop16, $pop14
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop19), $pop17
+; NO-SIMD128-SLOW-NEXT:    i32.const $push56=, 16
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push21=, $5, $pop56
+; NO-SIMD128-SLOW-NEXT:    i32.const $push55=, 16
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push22=, $pop21, $pop55
+; NO-SIMD128-SLOW-NEXT:    i32.const $push54=, 65535
+; NO-SIMD128-SLOW-NEXT:    i32.and $push20=, $13, $pop54
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push23=, $pop22, $pop20
+; NO-SIMD128-SLOW-NEXT:    i32.store16 8($0), $pop23
+; NO-SIMD128-SLOW-NEXT:    i32.const $push28=, 6
+; NO-SIMD128-SLOW-NEXT:    i32.add $push29=, $0, $pop28
+; NO-SIMD128-SLOW-NEXT:    i32.const $push53=, 16
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push25=, $4, $pop53
+; NO-SIMD128-SLOW-NEXT:    i32.const $push52=, 16
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push26=, $pop25, $pop52
+; NO-SIMD128-SLOW-NEXT:    i32.const $push51=, 65535
+; NO-SIMD128-SLOW-NEXT:    i32.and $push24=, $12, $pop51
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push27=, $pop26, $pop24
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop29), $pop27
+; NO-SIMD128-SLOW-NEXT:    i32.const $push50=, 16
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push31=, $3, $pop50
+; NO-SIMD128-SLOW-NEXT:    i32.const $push49=, 16
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push32=, $pop31, $pop49
+; NO-SIMD128-SLOW-NEXT:    i32.const $push48=, 65535
+; NO-SIMD128-SLOW-NEXT:    i32.and $push30=, $11, $pop48
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push33=, $pop32, $pop30
+; NO-SIMD128-SLOW-NEXT:    i32.store16 4($0), $pop33
+; NO-SIMD128-SLOW-NEXT:    i32.const $push47=, 16
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push35=, $2, $pop47
+; NO-SIMD128-SLOW-NEXT:    i32.const $push46=, 16
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push36=, $pop35, $pop46
+; NO-SIMD128-SLOW-NEXT:    i32.const $push45=, 65535
+; NO-SIMD128-SLOW-NEXT:    i32.and $push34=, $10, $pop45
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push37=, $pop36, $pop34
+; NO-SIMD128-SLOW-NEXT:    i32.store16 2($0), $pop37
+; NO-SIMD128-SLOW-NEXT:    i32.const $push44=, 16
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push39=, $1, $pop44
+; NO-SIMD128-SLOW-NEXT:    i32.const $push43=, 16
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push40=, $pop39, $pop43
+; NO-SIMD128-SLOW-NEXT:    i32.const $push42=, 65535
+; NO-SIMD128-SLOW-NEXT:    i32.and $push38=, $9, $pop42
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push41=, $pop40, $pop38
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($0), $pop41
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: shr_s_vec_v8i16:
+; NO-SIMD128-FAST:         .functype shr_s_vec_v8i16 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-FAST-NEXT:    i32.const $push2=, 16
+; NO-SIMD128-FAST-NEXT:    i32.shl $push3=, $1, $pop2
+; NO-SIMD128-FAST-NEXT:    i32.const $push63=, 16
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push4=, $pop3, $pop63
+; NO-SIMD128-FAST-NEXT:    i32.const $push0=, 65535
+; NO-SIMD128-FAST-NEXT:    i32.and $push1=, $9, $pop0
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push5=, $pop4, $pop1
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($0), $pop5
+; NO-SIMD128-FAST-NEXT:    i32.const $push62=, 16
+; NO-SIMD128-FAST-NEXT:    i32.shl $push7=, $2, $pop62
+; NO-SIMD128-FAST-NEXT:    i32.const $push61=, 16
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push8=, $pop7, $pop61
+; NO-SIMD128-FAST-NEXT:    i32.const $push60=, 65535
+; NO-SIMD128-FAST-NEXT:    i32.and $push6=, $10, $pop60
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push9=, $pop8, $pop6
+; NO-SIMD128-FAST-NEXT:    i32.store16 2($0), $pop9
+; NO-SIMD128-FAST-NEXT:    i32.const $push59=, 16
+; NO-SIMD128-FAST-NEXT:    i32.shl $push11=, $3, $pop59
+; NO-SIMD128-FAST-NEXT:    i32.const $push58=, 16
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push12=, $pop11, $pop58
+; NO-SIMD128-FAST-NEXT:    i32.const $push57=, 65535
+; NO-SIMD128-FAST-NEXT:    i32.and $push10=, $11, $pop57
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push13=, $pop12, $pop10
+; NO-SIMD128-FAST-NEXT:    i32.store16 4($0), $pop13
+; NO-SIMD128-FAST-NEXT:    i32.const $push18=, 6
+; NO-SIMD128-FAST-NEXT:    i32.add $push19=, $0, $pop18
+; NO-SIMD128-FAST-NEXT:    i32.const $push56=, 16
+; NO-SIMD128-FAST-NEXT:    i32.shl $push15=, $4, $pop56
+; NO-SIMD128-FAST-NEXT:    i32.const $push55=, 16
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push16=, $pop15, $pop55
+; NO-SIMD128-FAST-NEXT:    i32.const $push54=, 65535
+; NO-SIMD128-FAST-NEXT:    i32.and $push14=, $12, $pop54
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push17=, $pop16, $pop14
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop19), $pop17
+; NO-SIMD128-FAST-NEXT:    i32.const $push53=, 16
+; NO-SIMD128-FAST-NEXT:    i32.shl $push21=, $5, $pop53
+; NO-SIMD128-FAST-NEXT:    i32.const $push52=, 16
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push22=, $pop21, $pop52
+; NO-SIMD128-FAST-NEXT:    i32.const $push51=, 65535
+; NO-SIMD128-FAST-NEXT:    i32.and $push20=, $13, $pop51
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push23=, $pop22, $pop20
+; NO-SIMD128-FAST-NEXT:    i32.store16 8($0), $pop23
+; NO-SIMD128-FAST-NEXT:    i32.const $push28=, 10
+; NO-SIMD128-FAST-NEXT:    i32.add $push29=, $0, $pop28
+; NO-SIMD128-FAST-NEXT:    i32.const $push50=, 16
+; NO-SIMD128-FAST-NEXT:    i32.shl $push25=, $6, $pop50
+; NO-SIMD128-FAST-NEXT:    i32.const $push49=, 16
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push26=, $pop25, $pop49
+; NO-SIMD128-FAST-NEXT:    i32.const $push48=, 65535
+; NO-SIMD128-FAST-NEXT:    i32.and $push24=, $14, $pop48
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push27=, $pop26, $pop24
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop29), $pop27
+; NO-SIMD128-FAST-NEXT:    i32.const $push34=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push35=, $0, $pop34
+; NO-SIMD128-FAST-NEXT:    i32.const $push47=, 16
+; NO-SIMD128-FAST-NEXT:    i32.shl $push31=, $7, $pop47
+; NO-SIMD128-FAST-NEXT:    i32.const $push46=, 16
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push32=, $pop31, $pop46
+; NO-SIMD128-FAST-NEXT:    i32.const $push45=, 65535
+; NO-SIMD128-FAST-NEXT:    i32.and $push30=, $15, $pop45
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push33=, $pop32, $pop30
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop35), $pop33
+; NO-SIMD128-FAST-NEXT:    i32.const $push40=, 14
+; NO-SIMD128-FAST-NEXT:    i32.add $push41=, $0, $pop40
+; NO-SIMD128-FAST-NEXT:    i32.const $push44=, 16
+; NO-SIMD128-FAST-NEXT:    i32.shl $push37=, $8, $pop44
+; NO-SIMD128-FAST-NEXT:    i32.const $push43=, 16
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push38=, $pop37, $pop43
+; NO-SIMD128-FAST-NEXT:    i32.const $push42=, 65535
+; NO-SIMD128-FAST-NEXT:    i32.and $push36=, $16, $pop42
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push39=, $pop38, $pop36
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop41), $pop39
+; NO-SIMD128-FAST-NEXT:    return
   %a = ashr <8 x i16> %v, %x
   ret <8 x i16> %a
 }
 
-; CHECK-LABEL: shr_u_v8i16:
-; NO-SIMD128-NOT: i16x8
-; SIMD128-NEXT: .functype shr_u_v8i16 (v128, i32) -> (v128){{$}}
-; SIMD128-NEXT: i16x8.shr_u $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <8 x i16> @shr_u_v8i16(<8 x i16> %v, i16 %x) {
+; SIMD128-UNIMPL-LABEL: shr_u_v8i16:
+; SIMD128-UNIMPL:         .functype shr_u_v8i16 (v128, i32) -> (v128)
+; SIMD128-UNIMPL-NEXT:    i16x8.shr_u $push0=, $0, $1
+; SIMD128-UNIMPL-NEXT:    return $pop0
+;
+; SIMD128-VM-SLOW-LABEL: shr_u_v8i16:
+; SIMD128-VM-SLOW:         .functype shr_u_v8i16 (v128, i32) -> (v128)
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push9=, $0, 0
+; SIMD128-VM-SLOW-NEXT:    i32.const $push1=, 65535
+; SIMD128-VM-SLOW-NEXT:    i32.and $push10=, $pop9, $pop1
+; SIMD128-VM-SLOW-NEXT:    i16x8.splat $push5=, $1
+; SIMD128-VM-SLOW-NEXT:    i32.const $push3=, 15
+; SIMD128-VM-SLOW-NEXT:    i16x8.splat $push4=, $pop3
+; SIMD128-VM-SLOW-NEXT:    v128.and $push68=, $pop5, $pop4
+; SIMD128-VM-SLOW-NEXT:    local.tee $push67=, $2=, $pop68
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push11=, $pop67, 0
+; SIMD128-VM-SLOW-NEXT:    i32.const $push66=, 15
+; SIMD128-VM-SLOW-NEXT:    i32.and $push12=, $pop11, $pop66
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push13=, $pop10, $pop12
+; SIMD128-VM-SLOW-NEXT:    i16x8.splat $push14=, $pop13
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push0=, $0, 1
+; SIMD128-VM-SLOW-NEXT:    i32.const $push65=, 65535
+; SIMD128-VM-SLOW-NEXT:    i32.and $push2=, $pop0, $pop65
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push6=, $2, 1
+; SIMD128-VM-SLOW-NEXT:    i32.const $push64=, 15
+; SIMD128-VM-SLOW-NEXT:    i32.and $push7=, $pop6, $pop64
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push8=, $pop2, $pop7
+; SIMD128-VM-SLOW-NEXT:    i16x8.replace_lane $push15=, $pop14, 1, $pop8
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push16=, $0, 2
+; SIMD128-VM-SLOW-NEXT:    i32.const $push63=, 65535
+; SIMD128-VM-SLOW-NEXT:    i32.and $push17=, $pop16, $pop63
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push18=, $2, 2
+; SIMD128-VM-SLOW-NEXT:    i32.const $push62=, 15
+; SIMD128-VM-SLOW-NEXT:    i32.and $push19=, $pop18, $pop62
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push20=, $pop17, $pop19
+; SIMD128-VM-SLOW-NEXT:    i16x8.replace_lane $push21=, $pop15, 2, $pop20
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push22=, $0, 3
+; SIMD128-VM-SLOW-NEXT:    i32.const $push61=, 65535
+; SIMD128-VM-SLOW-NEXT:    i32.and $push23=, $pop22, $pop61
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push24=, $2, 3
+; SIMD128-VM-SLOW-NEXT:    i32.const $push60=, 15
+; SIMD128-VM-SLOW-NEXT:    i32.and $push25=, $pop24, $pop60
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push26=, $pop23, $pop25
+; SIMD128-VM-SLOW-NEXT:    i16x8.replace_lane $push27=, $pop21, 3, $pop26
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push28=, $0, 4
+; SIMD128-VM-SLOW-NEXT:    i32.const $push59=, 65535
+; SIMD128-VM-SLOW-NEXT:    i32.and $push29=, $pop28, $pop59
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push30=, $2, 4
+; SIMD128-VM-SLOW-NEXT:    i32.const $push58=, 15
+; SIMD128-VM-SLOW-NEXT:    i32.and $push31=, $pop30, $pop58
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push32=, $pop29, $pop31
+; SIMD128-VM-SLOW-NEXT:    i16x8.replace_lane $push33=, $pop27, 4, $pop32
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push34=, $0, 5
+; SIMD128-VM-SLOW-NEXT:    i32.const $push57=, 65535
+; SIMD128-VM-SLOW-NEXT:    i32.and $push35=, $pop34, $pop57
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push36=, $2, 5
+; SIMD128-VM-SLOW-NEXT:    i32.const $push56=, 15
+; SIMD128-VM-SLOW-NEXT:    i32.and $push37=, $pop36, $pop56
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push38=, $pop35, $pop37
+; SIMD128-VM-SLOW-NEXT:    i16x8.replace_lane $push39=, $pop33, 5, $pop38
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push40=, $0, 6
+; SIMD128-VM-SLOW-NEXT:    i32.const $push55=, 65535
+; SIMD128-VM-SLOW-NEXT:    i32.and $push41=, $pop40, $pop55
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push42=, $2, 6
+; SIMD128-VM-SLOW-NEXT:    i32.const $push54=, 15
+; SIMD128-VM-SLOW-NEXT:    i32.and $push43=, $pop42, $pop54
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push44=, $pop41, $pop43
+; SIMD128-VM-SLOW-NEXT:    i16x8.replace_lane $push45=, $pop39, 6, $pop44
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push46=, $0, 7
+; SIMD128-VM-SLOW-NEXT:    i32.const $push53=, 65535
+; SIMD128-VM-SLOW-NEXT:    i32.and $push47=, $pop46, $pop53
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push48=, $2, 7
+; SIMD128-VM-SLOW-NEXT:    i32.const $push52=, 15
+; SIMD128-VM-SLOW-NEXT:    i32.and $push49=, $pop48, $pop52
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push50=, $pop47, $pop49
+; SIMD128-VM-SLOW-NEXT:    i16x8.replace_lane $push51=, $pop45, 7, $pop50
+; SIMD128-VM-SLOW-NEXT:    return $pop51
+;
+; SIMD128-VM-FAST-LABEL: shr_u_v8i16:
+; SIMD128-VM-FAST:         .functype shr_u_v8i16 (v128, i32) -> (v128)
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push10=, $0, 0
+; SIMD128-VM-FAST-NEXT:    i32.const $push3=, 65535
+; SIMD128-VM-FAST-NEXT:    i32.and $push11=, $pop10, $pop3
+; SIMD128-VM-FAST-NEXT:    i16x8.splat $push1=, $1
+; SIMD128-VM-FAST-NEXT:    i32.const $push5=, 15
+; SIMD128-VM-FAST-NEXT:    i16x8.splat $push6=, $pop5
+; SIMD128-VM-FAST-NEXT:    v128.and $push68=, $pop1, $pop6
+; SIMD128-VM-FAST-NEXT:    local.tee $push67=, $2=, $pop68
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push12=, $pop67, 0
+; SIMD128-VM-FAST-NEXT:    i32.const $push66=, 15
+; SIMD128-VM-FAST-NEXT:    i32.and $push13=, $pop12, $pop66
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push14=, $pop11, $pop13
+; SIMD128-VM-FAST-NEXT:    i16x8.splat $push15=, $pop14
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push2=, $0, 1
+; SIMD128-VM-FAST-NEXT:    i32.const $push65=, 65535
+; SIMD128-VM-FAST-NEXT:    i32.and $push4=, $pop2, $pop65
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push7=, $2, 1
+; SIMD128-VM-FAST-NEXT:    i32.const $push64=, 15
+; SIMD128-VM-FAST-NEXT:    i32.and $push8=, $pop7, $pop64
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push9=, $pop4, $pop8
+; SIMD128-VM-FAST-NEXT:    i16x8.replace_lane $push16=, $pop15, 1, $pop9
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push17=, $0, 2
+; SIMD128-VM-FAST-NEXT:    i32.const $push63=, 65535
+; SIMD128-VM-FAST-NEXT:    i32.and $push18=, $pop17, $pop63
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push19=, $2, 2
+; SIMD128-VM-FAST-NEXT:    i32.const $push62=, 15
+; SIMD128-VM-FAST-NEXT:    i32.and $push20=, $pop19, $pop62
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push21=, $pop18, $pop20
+; SIMD128-VM-FAST-NEXT:    i16x8.replace_lane $push22=, $pop16, 2, $pop21
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push23=, $0, 3
+; SIMD128-VM-FAST-NEXT:    i32.const $push61=, 65535
+; SIMD128-VM-FAST-NEXT:    i32.and $push24=, $pop23, $pop61
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push25=, $2, 3
+; SIMD128-VM-FAST-NEXT:    i32.const $push60=, 15
+; SIMD128-VM-FAST-NEXT:    i32.and $push26=, $pop25, $pop60
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push27=, $pop24, $pop26
+; SIMD128-VM-FAST-NEXT:    i16x8.replace_lane $push28=, $pop22, 3, $pop27
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push29=, $0, 4
+; SIMD128-VM-FAST-NEXT:    i32.const $push59=, 65535
+; SIMD128-VM-FAST-NEXT:    i32.and $push30=, $pop29, $pop59
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push31=, $2, 4
+; SIMD128-VM-FAST-NEXT:    i32.const $push58=, 15
+; SIMD128-VM-FAST-NEXT:    i32.and $push32=, $pop31, $pop58
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push33=, $pop30, $pop32
+; SIMD128-VM-FAST-NEXT:    i16x8.replace_lane $push34=, $pop28, 4, $pop33
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push35=, $0, 5
+; SIMD128-VM-FAST-NEXT:    i32.const $push57=, 65535
+; SIMD128-VM-FAST-NEXT:    i32.and $push36=, $pop35, $pop57
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push37=, $2, 5
+; SIMD128-VM-FAST-NEXT:    i32.const $push56=, 15
+; SIMD128-VM-FAST-NEXT:    i32.and $push38=, $pop37, $pop56
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push39=, $pop36, $pop38
+; SIMD128-VM-FAST-NEXT:    i16x8.replace_lane $push40=, $pop34, 5, $pop39
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push41=, $0, 6
+; SIMD128-VM-FAST-NEXT:    i32.const $push55=, 65535
+; SIMD128-VM-FAST-NEXT:    i32.and $push42=, $pop41, $pop55
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push43=, $2, 6
+; SIMD128-VM-FAST-NEXT:    i32.const $push54=, 15
+; SIMD128-VM-FAST-NEXT:    i32.and $push44=, $pop43, $pop54
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push45=, $pop42, $pop44
+; SIMD128-VM-FAST-NEXT:    i16x8.replace_lane $push46=, $pop40, 6, $pop45
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push47=, $0, 7
+; SIMD128-VM-FAST-NEXT:    i32.const $push53=, 65535
+; SIMD128-VM-FAST-NEXT:    i32.and $push48=, $pop47, $pop53
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push49=, $2, 7
+; SIMD128-VM-FAST-NEXT:    i32.const $push52=, 15
+; SIMD128-VM-FAST-NEXT:    i32.and $push50=, $pop49, $pop52
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push51=, $pop48, $pop50
+; SIMD128-VM-FAST-NEXT:    i16x8.replace_lane $push0=, $pop46, 7, $pop51
+; SIMD128-VM-FAST-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: shr_u_v8i16:
+; NO-SIMD128-SLOW:         .functype shr_u_v8i16 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push3=, 14
+; NO-SIMD128-SLOW-NEXT:    i32.add $push4=, $0, $pop3
+; NO-SIMD128-SLOW-NEXT:    i32.const $push0=, 65535
+; NO-SIMD128-SLOW-NEXT:    i32.and $push1=, $8, $pop0
+; NO-SIMD128-SLOW-NEXT:    i32.const $push34=, 65535
+; NO-SIMD128-SLOW-NEXT:    i32.and $push33=, $9, $pop34
+; NO-SIMD128-SLOW-NEXT:    local.tee $push32=, $8=, $pop33
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push2=, $pop1, $pop32
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop4), $pop2
+; NO-SIMD128-SLOW-NEXT:    i32.const $push7=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push8=, $0, $pop7
+; NO-SIMD128-SLOW-NEXT:    i32.const $push31=, 65535
+; NO-SIMD128-SLOW-NEXT:    i32.and $push5=, $7, $pop31
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push6=, $pop5, $8
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop8), $pop6
+; NO-SIMD128-SLOW-NEXT:    i32.const $push11=, 10
+; NO-SIMD128-SLOW-NEXT:    i32.add $push12=, $0, $pop11
+; NO-SIMD128-SLOW-NEXT:    i32.const $push30=, 65535
+; NO-SIMD128-SLOW-NEXT:    i32.and $push9=, $6, $pop30
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push10=, $pop9, $8
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop12), $pop10
+; NO-SIMD128-SLOW-NEXT:    i32.const $push29=, 65535
+; NO-SIMD128-SLOW-NEXT:    i32.and $push13=, $5, $pop29
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push14=, $pop13, $8
+; NO-SIMD128-SLOW-NEXT:    i32.store16 8($0), $pop14
+; NO-SIMD128-SLOW-NEXT:    i32.const $push17=, 6
+; NO-SIMD128-SLOW-NEXT:    i32.add $push18=, $0, $pop17
+; NO-SIMD128-SLOW-NEXT:    i32.const $push28=, 65535
+; NO-SIMD128-SLOW-NEXT:    i32.and $push15=, $4, $pop28
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push16=, $pop15, $8
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop18), $pop16
+; NO-SIMD128-SLOW-NEXT:    i32.const $push27=, 65535
+; NO-SIMD128-SLOW-NEXT:    i32.and $push19=, $3, $pop27
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push20=, $pop19, $8
+; NO-SIMD128-SLOW-NEXT:    i32.store16 4($0), $pop20
+; NO-SIMD128-SLOW-NEXT:    i32.const $push26=, 65535
+; NO-SIMD128-SLOW-NEXT:    i32.and $push21=, $2, $pop26
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push22=, $pop21, $8
+; NO-SIMD128-SLOW-NEXT:    i32.store16 2($0), $pop22
+; NO-SIMD128-SLOW-NEXT:    i32.const $push25=, 65535
+; NO-SIMD128-SLOW-NEXT:    i32.and $push23=, $1, $pop25
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push24=, $pop23, $8
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($0), $pop24
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: shr_u_v8i16:
+; NO-SIMD128-FAST:         .functype shr_u_v8i16 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-FAST-NEXT:    i32.const $push0=, 65535
+; NO-SIMD128-FAST-NEXT:    i32.and $push1=, $1, $pop0
+; NO-SIMD128-FAST-NEXT:    i32.const $push34=, 65535
+; NO-SIMD128-FAST-NEXT:    i32.and $push33=, $9, $pop34
+; NO-SIMD128-FAST-NEXT:    local.tee $push32=, $9=, $pop33
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push2=, $pop1, $pop32
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($0), $pop2
+; NO-SIMD128-FAST-NEXT:    i32.const $push31=, 65535
+; NO-SIMD128-FAST-NEXT:    i32.and $push3=, $2, $pop31
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push4=, $pop3, $9
+; NO-SIMD128-FAST-NEXT:    i32.store16 2($0), $pop4
+; NO-SIMD128-FAST-NEXT:    i32.const $push30=, 65535
+; NO-SIMD128-FAST-NEXT:    i32.and $push5=, $3, $pop30
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push6=, $pop5, $9
+; NO-SIMD128-FAST-NEXT:    i32.store16 4($0), $pop6
+; NO-SIMD128-FAST-NEXT:    i32.const $push9=, 6
+; NO-SIMD128-FAST-NEXT:    i32.add $push10=, $0, $pop9
+; NO-SIMD128-FAST-NEXT:    i32.const $push29=, 65535
+; NO-SIMD128-FAST-NEXT:    i32.and $push7=, $4, $pop29
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push8=, $pop7, $9
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop10), $pop8
+; NO-SIMD128-FAST-NEXT:    i32.const $push28=, 65535
+; NO-SIMD128-FAST-NEXT:    i32.and $push11=, $5, $pop28
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push12=, $pop11, $9
+; NO-SIMD128-FAST-NEXT:    i32.store16 8($0), $pop12
+; NO-SIMD128-FAST-NEXT:    i32.const $push15=, 10
+; NO-SIMD128-FAST-NEXT:    i32.add $push16=, $0, $pop15
+; NO-SIMD128-FAST-NEXT:    i32.const $push27=, 65535
+; NO-SIMD128-FAST-NEXT:    i32.and $push13=, $6, $pop27
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push14=, $pop13, $9
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop16), $pop14
+; NO-SIMD128-FAST-NEXT:    i32.const $push19=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push20=, $0, $pop19
+; NO-SIMD128-FAST-NEXT:    i32.const $push26=, 65535
+; NO-SIMD128-FAST-NEXT:    i32.and $push17=, $7, $pop26
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push18=, $pop17, $9
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop20), $pop18
+; NO-SIMD128-FAST-NEXT:    i32.const $push23=, 14
+; NO-SIMD128-FAST-NEXT:    i32.add $push24=, $0, $pop23
+; NO-SIMD128-FAST-NEXT:    i32.const $push25=, 65535
+; NO-SIMD128-FAST-NEXT:    i32.and $push21=, $8, $pop25
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push22=, $pop21, $9
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop24), $pop22
+; NO-SIMD128-FAST-NEXT:    return
   %t = insertelement <8 x i16> undef, i16 %x, i32 0
   %s = shufflevector <8 x i16> %t, <8 x i16> undef,
     <8 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
   %a = lshr <8 x i16> %v, %s
   ret <8 x i16> %a
 }
 
-; CHECK-LABEL: shr_u_vec_v8i16:
-; NO-SIMD128-NOT: i16x8
-; SIMD128-NEXT: .functype shr_u_vec_v8i16 (v128, v128) -> (v128){{$}}
-; SIMD128-NEXT: i16x8.extract_lane_u $push[[L0:[0-9]+]]=, $0, 0{{$}}
-; SIMD128-NEXT: i32.const $push[[M0:[0-9]+]]=, 15{{$}}
-; SIMD128-NEXT: i16x8.splat $push[[M1:[0-9]+]]=, $pop[[M0]]{{$}}
-; SIMD128-NEXT: v128.and $push[[M2:[0-9]+]]=, $1, $pop[[M1]]{{$}}
-; SIMD128-NEXT: local.tee $push[[M:[0-9]+]]=, $1=, $pop[[M2]]{{$}}
-; SIMD128-NEXT: i16x8.extract_lane_u $push[[L1:[0-9]+]]=, $pop[[M]], 0{{$}}
-; SIMD128-NEXT: i32.shr_u $push[[L2:[0-9]+]]=, $pop[[L0]], $pop[[L1]]{{$}}
-; SIMD128-NEXT: i16x8.splat $push[[L3:[0-9]+]]=, $pop[[L2]]{{$}}
-; Skip 6 lanes
-; SIMD128:      i16x8.extract_lane_u $push[[L4:[0-9]+]]=, $0, 7{{$}}
-; SIMD128-NEXT: i16x8.extract_lane_u $push[[L5:[0-9]+]]=, $1, 7{{$}}
-; SIMD128-NEXT: i32.shr_u $push[[L6:[0-9]+]]=, $pop[[L4]], $pop[[L5]]{{$}}
-; SIMD128-NEXT: i16x8.replace_lane $push[[R:[0-9]+]]=, $pop[[L7:[0-9]+]], 7, $pop[[L6]]{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <8 x i16> @shr_u_vec_v8i16(<8 x i16> %v, <8 x i16> %x) {
+; SIMD128-UNIMPL-SLOW-LABEL: shr_u_vec_v8i16:
+; SIMD128-UNIMPL-SLOW:         .functype shr_u_vec_v8i16 (v128, v128) -> (v128)
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.extract_lane_u $push6=, $0, 0
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.const $push0=, 15
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.splat $push1=, $pop0
+; SIMD128-UNIMPL-SLOW-NEXT:    v128.and $push35=, $1, $pop1
+; SIMD128-UNIMPL-SLOW-NEXT:    local.tee $push34=, $1=, $pop35
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.extract_lane_u $push5=, $pop34, 0
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shr_u $push7=, $pop6, $pop5
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.splat $push8=, $pop7
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.extract_lane_u $push3=, $0, 1
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.extract_lane_u $push2=, $1, 1
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shr_u $push4=, $pop3, $pop2
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.replace_lane $push9=, $pop8, 1, $pop4
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.extract_lane_u $push11=, $0, 2
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.extract_lane_u $push10=, $1, 2
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shr_u $push12=, $pop11, $pop10
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.replace_lane $push13=, $pop9, 2, $pop12
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.extract_lane_u $push15=, $0, 3
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.extract_lane_u $push14=, $1, 3
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shr_u $push16=, $pop15, $pop14
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.replace_lane $push17=, $pop13, 3, $pop16
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.extract_lane_u $push19=, $0, 4
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.extract_lane_u $push18=, $1, 4
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shr_u $push20=, $pop19, $pop18
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.replace_lane $push21=, $pop17, 4, $pop20
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.extract_lane_u $push23=, $0, 5
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.extract_lane_u $push22=, $1, 5
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shr_u $push24=, $pop23, $pop22
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.replace_lane $push25=, $pop21, 5, $pop24
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.extract_lane_u $push27=, $0, 6
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.extract_lane_u $push26=, $1, 6
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shr_u $push28=, $pop27, $pop26
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.replace_lane $push29=, $pop25, 6, $pop28
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.extract_lane_u $push31=, $0, 7
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.extract_lane_u $push30=, $1, 7
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.shr_u $push32=, $pop31, $pop30
+; SIMD128-UNIMPL-SLOW-NEXT:    i16x8.replace_lane $push33=, $pop29, 7, $pop32
+; SIMD128-UNIMPL-SLOW-NEXT:    return $pop33
+;
+; SIMD128-UNIMPL-FAST-LABEL: shr_u_vec_v8i16:
+; SIMD128-UNIMPL-FAST:         .functype shr_u_vec_v8i16 (v128, v128) -> (v128)
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.extract_lane_u $push7=, $0, 0
+; SIMD128-UNIMPL-FAST-NEXT:    i32.const $push1=, 15
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.splat $push2=, $pop1
+; SIMD128-UNIMPL-FAST-NEXT:    v128.and $push35=, $1, $pop2
+; SIMD128-UNIMPL-FAST-NEXT:    local.tee $push34=, $1=, $pop35
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.extract_lane_u $push6=, $pop34, 0
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shr_u $push8=, $pop7, $pop6
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.splat $push9=, $pop8
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.extract_lane_u $push4=, $0, 1
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.extract_lane_u $push3=, $1, 1
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shr_u $push5=, $pop4, $pop3
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.replace_lane $push10=, $pop9, 1, $pop5
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.extract_lane_u $push12=, $0, 2
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.extract_lane_u $push11=, $1, 2
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shr_u $push13=, $pop12, $pop11
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.replace_lane $push14=, $pop10, 2, $pop13
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.extract_lane_u $push16=, $0, 3
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.extract_lane_u $push15=, $1, 3
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shr_u $push17=, $pop16, $pop15
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.replace_lane $push18=, $pop14, 3, $pop17
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.extract_lane_u $push20=, $0, 4
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.extract_lane_u $push19=, $1, 4
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shr_u $push21=, $pop20, $pop19
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.replace_lane $push22=, $pop18, 4, $pop21
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.extract_lane_u $push24=, $0, 5
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.extract_lane_u $push23=, $1, 5
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shr_u $push25=, $pop24, $pop23
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.replace_lane $push26=, $pop22, 5, $pop25
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.extract_lane_u $push28=, $0, 6
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.extract_lane_u $push27=, $1, 6
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shr_u $push29=, $pop28, $pop27
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.replace_lane $push30=, $pop26, 6, $pop29
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.extract_lane_u $push32=, $0, 7
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.extract_lane_u $push31=, $1, 7
+; SIMD128-UNIMPL-FAST-NEXT:    i32.shr_u $push33=, $pop32, $pop31
+; SIMD128-UNIMPL-FAST-NEXT:    i16x8.replace_lane $push0=, $pop30, 7, $pop33
+; SIMD128-UNIMPL-FAST-NEXT:    return $pop0
+;
+; SIMD128-VM-SLOW-LABEL: shr_u_vec_v8i16:
+; SIMD128-VM-SLOW:         .functype shr_u_vec_v8i16 (v128, v128) -> (v128)
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push8=, $0, 0
+; SIMD128-VM-SLOW-NEXT:    i32.const $push1=, 65535
+; SIMD128-VM-SLOW-NEXT:    i32.and $push9=, $pop8, $pop1
+; SIMD128-VM-SLOW-NEXT:    i32.const $push3=, 15
+; SIMD128-VM-SLOW-NEXT:    i16x8.splat $push4=, $pop3
+; SIMD128-VM-SLOW-NEXT:    v128.and $push67=, $1, $pop4
+; SIMD128-VM-SLOW-NEXT:    local.tee $push66=, $1=, $pop67
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push10=, $pop66, 0
+; SIMD128-VM-SLOW-NEXT:    i32.const $push65=, 15
+; SIMD128-VM-SLOW-NEXT:    i32.and $push11=, $pop10, $pop65
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push12=, $pop9, $pop11
+; SIMD128-VM-SLOW-NEXT:    i16x8.splat $push13=, $pop12
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push0=, $0, 1
+; SIMD128-VM-SLOW-NEXT:    i32.const $push64=, 65535
+; SIMD128-VM-SLOW-NEXT:    i32.and $push2=, $pop0, $pop64
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push5=, $1, 1
+; SIMD128-VM-SLOW-NEXT:    i32.const $push63=, 15
+; SIMD128-VM-SLOW-NEXT:    i32.and $push6=, $pop5, $pop63
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push7=, $pop2, $pop6
+; SIMD128-VM-SLOW-NEXT:    i16x8.replace_lane $push14=, $pop13, 1, $pop7
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push15=, $0, 2
+; SIMD128-VM-SLOW-NEXT:    i32.const $push62=, 65535
+; SIMD128-VM-SLOW-NEXT:    i32.and $push16=, $pop15, $pop62
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push17=, $1, 2
+; SIMD128-VM-SLOW-NEXT:    i32.const $push61=, 15
+; SIMD128-VM-SLOW-NEXT:    i32.and $push18=, $pop17, $pop61
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push19=, $pop16, $pop18
+; SIMD128-VM-SLOW-NEXT:    i16x8.replace_lane $push20=, $pop14, 2, $pop19
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push21=, $0, 3
+; SIMD128-VM-SLOW-NEXT:    i32.const $push60=, 65535
+; SIMD128-VM-SLOW-NEXT:    i32.and $push22=, $pop21, $pop60
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push23=, $1, 3
+; SIMD128-VM-SLOW-NEXT:    i32.const $push59=, 15
+; SIMD128-VM-SLOW-NEXT:    i32.and $push24=, $pop23, $pop59
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push25=, $pop22, $pop24
+; SIMD128-VM-SLOW-NEXT:    i16x8.replace_lane $push26=, $pop20, 3, $pop25
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push27=, $0, 4
+; SIMD128-VM-SLOW-NEXT:    i32.const $push58=, 65535
+; SIMD128-VM-SLOW-NEXT:    i32.and $push28=, $pop27, $pop58
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push29=, $1, 4
+; SIMD128-VM-SLOW-NEXT:    i32.const $push57=, 15
+; SIMD128-VM-SLOW-NEXT:    i32.and $push30=, $pop29, $pop57
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push31=, $pop28, $pop30
+; SIMD128-VM-SLOW-NEXT:    i16x8.replace_lane $push32=, $pop26, 4, $pop31
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push33=, $0, 5
+; SIMD128-VM-SLOW-NEXT:    i32.const $push56=, 65535
+; SIMD128-VM-SLOW-NEXT:    i32.and $push34=, $pop33, $pop56
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push35=, $1, 5
+; SIMD128-VM-SLOW-NEXT:    i32.const $push55=, 15
+; SIMD128-VM-SLOW-NEXT:    i32.and $push36=, $pop35, $pop55
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push37=, $pop34, $pop36
+; SIMD128-VM-SLOW-NEXT:    i16x8.replace_lane $push38=, $pop32, 5, $pop37
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push39=, $0, 6
+; SIMD128-VM-SLOW-NEXT:    i32.const $push54=, 65535
+; SIMD128-VM-SLOW-NEXT:    i32.and $push40=, $pop39, $pop54
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push41=, $1, 6
+; SIMD128-VM-SLOW-NEXT:    i32.const $push53=, 15
+; SIMD128-VM-SLOW-NEXT:    i32.and $push42=, $pop41, $pop53
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push43=, $pop40, $pop42
+; SIMD128-VM-SLOW-NEXT:    i16x8.replace_lane $push44=, $pop38, 6, $pop43
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push45=, $0, 7
+; SIMD128-VM-SLOW-NEXT:    i32.const $push52=, 65535
+; SIMD128-VM-SLOW-NEXT:    i32.and $push46=, $pop45, $pop52
+; SIMD128-VM-SLOW-NEXT:    i16x8.extract_lane_s $push47=, $1, 7
+; SIMD128-VM-SLOW-NEXT:    i32.const $push51=, 15
+; SIMD128-VM-SLOW-NEXT:    i32.and $push48=, $pop47, $pop51
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push49=, $pop46, $pop48
+; SIMD128-VM-SLOW-NEXT:    i16x8.replace_lane $push50=, $pop44, 7, $pop49
+; SIMD128-VM-SLOW-NEXT:    return $pop50
+;
+; SIMD128-VM-FAST-LABEL: shr_u_vec_v8i16:
+; SIMD128-VM-FAST:         .functype shr_u_vec_v8i16 (v128, v128) -> (v128)
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push9=, $0, 0
+; SIMD128-VM-FAST-NEXT:    i32.const $push2=, 65535
+; SIMD128-VM-FAST-NEXT:    i32.and $push10=, $pop9, $pop2
+; SIMD128-VM-FAST-NEXT:    i32.const $push4=, 15
+; SIMD128-VM-FAST-NEXT:    i16x8.splat $push5=, $pop4
+; SIMD128-VM-FAST-NEXT:    v128.and $push67=, $1, $pop5
+; SIMD128-VM-FAST-NEXT:    local.tee $push66=, $1=, $pop67
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push11=, $pop66, 0
+; SIMD128-VM-FAST-NEXT:    i32.const $push65=, 15
+; SIMD128-VM-FAST-NEXT:    i32.and $push12=, $pop11, $pop65
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push13=, $pop10, $pop12
+; SIMD128-VM-FAST-NEXT:    i16x8.splat $push14=, $pop13
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push1=, $0, 1
+; SIMD128-VM-FAST-NEXT:    i32.const $push64=, 65535
+; SIMD128-VM-FAST-NEXT:    i32.and $push3=, $pop1, $pop64
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push6=, $1, 1
+; SIMD128-VM-FAST-NEXT:    i32.const $push63=, 15
+; SIMD128-VM-FAST-NEXT:    i32.and $push7=, $pop6, $pop63
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push8=, $pop3, $pop7
+; SIMD128-VM-FAST-NEXT:    i16x8.replace_lane $push15=, $pop14, 1, $pop8
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push16=, $0, 2
+; SIMD128-VM-FAST-NEXT:    i32.const $push62=, 65535
+; SIMD128-VM-FAST-NEXT:    i32.and $push17=, $pop16, $pop62
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push18=, $1, 2
+; SIMD128-VM-FAST-NEXT:    i32.const $push61=, 15
+; SIMD128-VM-FAST-NEXT:    i32.and $push19=, $pop18, $pop61
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push20=, $pop17, $pop19
+; SIMD128-VM-FAST-NEXT:    i16x8.replace_lane $push21=, $pop15, 2, $pop20
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push22=, $0, 3
+; SIMD128-VM-FAST-NEXT:    i32.const $push60=, 65535
+; SIMD128-VM-FAST-NEXT:    i32.and $push23=, $pop22, $pop60
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push24=, $1, 3
+; SIMD128-VM-FAST-NEXT:    i32.const $push59=, 15
+; SIMD128-VM-FAST-NEXT:    i32.and $push25=, $pop24, $pop59
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push26=, $pop23, $pop25
+; SIMD128-VM-FAST-NEXT:    i16x8.replace_lane $push27=, $pop21, 3, $pop26
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push28=, $0, 4
+; SIMD128-VM-FAST-NEXT:    i32.const $push58=, 65535
+; SIMD128-VM-FAST-NEXT:    i32.and $push29=, $pop28, $pop58
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push30=, $1, 4
+; SIMD128-VM-FAST-NEXT:    i32.const $push57=, 15
+; SIMD128-VM-FAST-NEXT:    i32.and $push31=, $pop30, $pop57
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push32=, $pop29, $pop31
+; SIMD128-VM-FAST-NEXT:    i16x8.replace_lane $push33=, $pop27, 4, $pop32
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push34=, $0, 5
+; SIMD128-VM-FAST-NEXT:    i32.const $push56=, 65535
+; SIMD128-VM-FAST-NEXT:    i32.and $push35=, $pop34, $pop56
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push36=, $1, 5
+; SIMD128-VM-FAST-NEXT:    i32.const $push55=, 15
+; SIMD128-VM-FAST-NEXT:    i32.and $push37=, $pop36, $pop55
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push38=, $pop35, $pop37
+; SIMD128-VM-FAST-NEXT:    i16x8.replace_lane $push39=, $pop33, 5, $pop38
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push40=, $0, 6
+; SIMD128-VM-FAST-NEXT:    i32.const $push54=, 65535
+; SIMD128-VM-FAST-NEXT:    i32.and $push41=, $pop40, $pop54
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push42=, $1, 6
+; SIMD128-VM-FAST-NEXT:    i32.const $push53=, 15
+; SIMD128-VM-FAST-NEXT:    i32.and $push43=, $pop42, $pop53
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push44=, $pop41, $pop43
+; SIMD128-VM-FAST-NEXT:    i16x8.replace_lane $push45=, $pop39, 6, $pop44
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push46=, $0, 7
+; SIMD128-VM-FAST-NEXT:    i32.const $push52=, 65535
+; SIMD128-VM-FAST-NEXT:    i32.and $push47=, $pop46, $pop52
+; SIMD128-VM-FAST-NEXT:    i16x8.extract_lane_s $push48=, $1, 7
+; SIMD128-VM-FAST-NEXT:    i32.const $push51=, 15
+; SIMD128-VM-FAST-NEXT:    i32.and $push49=, $pop48, $pop51
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push50=, $pop47, $pop49
+; SIMD128-VM-FAST-NEXT:    i16x8.replace_lane $push0=, $pop45, 7, $pop50
+; SIMD128-VM-FAST-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: shr_u_vec_v8i16:
+; NO-SIMD128-SLOW:         .functype shr_u_vec_v8i16 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push4=, 14
+; NO-SIMD128-SLOW-NEXT:    i32.add $push5=, $0, $pop4
+; NO-SIMD128-SLOW-NEXT:    i32.const $push0=, 65535
+; NO-SIMD128-SLOW-NEXT:    i32.and $push2=, $8, $pop0
+; NO-SIMD128-SLOW-NEXT:    i32.const $push47=, 65535
+; NO-SIMD128-SLOW-NEXT:    i32.and $push1=, $16, $pop47
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push3=, $pop2, $pop1
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop5), $pop3
+; NO-SIMD128-SLOW-NEXT:    i32.const $push9=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push10=, $0, $pop9
+; NO-SIMD128-SLOW-NEXT:    i32.const $push46=, 65535
+; NO-SIMD128-SLOW-NEXT:    i32.and $push7=, $7, $pop46
+; NO-SIMD128-SLOW-NEXT:    i32.const $push45=, 65535
+; NO-SIMD128-SLOW-NEXT:    i32.and $push6=, $15, $pop45
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push8=, $pop7, $pop6
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop10), $pop8
+; NO-SIMD128-SLOW-NEXT:    i32.const $push14=, 10
+; NO-SIMD128-SLOW-NEXT:    i32.add $push15=, $0, $pop14
+; NO-SIMD128-SLOW-NEXT:    i32.const $push44=, 65535
+; NO-SIMD128-SLOW-NEXT:    i32.and $push12=, $6, $pop44
+; NO-SIMD128-SLOW-NEXT:    i32.const $push43=, 65535
+; NO-SIMD128-SLOW-NEXT:    i32.and $push11=, $14, $pop43
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push13=, $pop12, $pop11
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop15), $pop13
+; NO-SIMD128-SLOW-NEXT:    i32.const $push42=, 65535
+; NO-SIMD128-SLOW-NEXT:    i32.and $push17=, $5, $pop42
+; NO-SIMD128-SLOW-NEXT:    i32.const $push41=, 65535
+; NO-SIMD128-SLOW-NEXT:    i32.and $push16=, $13, $pop41
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push18=, $pop17, $pop16
+; NO-SIMD128-SLOW-NEXT:    i32.store16 8($0), $pop18
+; NO-SIMD128-SLOW-NEXT:    i32.const $push22=, 6
+; NO-SIMD128-SLOW-NEXT:    i32.add $push23=, $0, $pop22
+; NO-SIMD128-SLOW-NEXT:    i32.const $push40=, 65535
+; NO-SIMD128-SLOW-NEXT:    i32.and $push20=, $4, $pop40
+; NO-SIMD128-SLOW-NEXT:    i32.const $push39=, 65535
+; NO-SIMD128-SLOW-NEXT:    i32.and $push19=, $12, $pop39
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push21=, $pop20, $pop19
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop23), $pop21
+; NO-SIMD128-SLOW-NEXT:    i32.const $push38=, 65535
+; NO-SIMD128-SLOW-NEXT:    i32.and $push25=, $3, $pop38
+; NO-SIMD128-SLOW-NEXT:    i32.const $push37=, 65535
+; NO-SIMD128-SLOW-NEXT:    i32.and $push24=, $11, $pop37
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push26=, $pop25, $pop24
+; NO-SIMD128-SLOW-NEXT:    i32.store16 4($0), $pop26
+; NO-SIMD128-SLOW-NEXT:    i32.const $push36=, 65535
+; NO-SIMD128-SLOW-NEXT:    i32.and $push28=, $2, $pop36
+; NO-SIMD128-SLOW-NEXT:    i32.const $push35=, 65535
+; NO-SIMD128-SLOW-NEXT:    i32.and $push27=, $10, $pop35
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push29=, $pop28, $pop27
+; NO-SIMD128-SLOW-NEXT:    i32.store16 2($0), $pop29
+; NO-SIMD128-SLOW-NEXT:    i32.const $push34=, 65535
+; NO-SIMD128-SLOW-NEXT:    i32.and $push31=, $1, $pop34
+; NO-SIMD128-SLOW-NEXT:    i32.const $push33=, 65535
+; NO-SIMD128-SLOW-NEXT:    i32.and $push30=, $9, $pop33
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push32=, $pop31, $pop30
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($0), $pop32
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: shr_u_vec_v8i16:
+; NO-SIMD128-FAST:         .functype shr_u_vec_v8i16 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-FAST-NEXT:    i32.const $push0=, 65535
+; NO-SIMD128-FAST-NEXT:    i32.and $push2=, $1, $pop0
+; NO-SIMD128-FAST-NEXT:    i32.const $push47=, 65535
+; NO-SIMD128-FAST-NEXT:    i32.and $push1=, $9, $pop47
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push3=, $pop2, $pop1
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($0), $pop3
+; NO-SIMD128-FAST-NEXT:    i32.const $push46=, 65535
+; NO-SIMD128-FAST-NEXT:    i32.and $push5=, $2, $pop46
+; NO-SIMD128-FAST-NEXT:    i32.const $push45=, 65535
+; NO-SIMD128-FAST-NEXT:    i32.and $push4=, $10, $pop45
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push6=, $pop5, $pop4
+; NO-SIMD128-FAST-NEXT:    i32.store16 2($0), $pop6
+; NO-SIMD128-FAST-NEXT:    i32.const $push44=, 65535
+; NO-SIMD128-FAST-NEXT:    i32.and $push8=, $3, $pop44
+; NO-SIMD128-FAST-NEXT:    i32.const $push43=, 65535
+; NO-SIMD128-FAST-NEXT:    i32.and $push7=, $11, $pop43
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push9=, $pop8, $pop7
+; NO-SIMD128-FAST-NEXT:    i32.store16 4($0), $pop9
+; NO-SIMD128-FAST-NEXT:    i32.const $push13=, 6
+; NO-SIMD128-FAST-NEXT:    i32.add $push14=, $0, $pop13
+; NO-SIMD128-FAST-NEXT:    i32.const $push42=, 65535
+; NO-SIMD128-FAST-NEXT:    i32.and $push11=, $4, $pop42
+; NO-SIMD128-FAST-NEXT:    i32.const $push41=, 65535
+; NO-SIMD128-FAST-NEXT:    i32.and $push10=, $12, $pop41
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push12=, $pop11, $pop10
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop14), $pop12
+; NO-SIMD128-FAST-NEXT:    i32.const $push40=, 65535
+; NO-SIMD128-FAST-NEXT:    i32.and $push16=, $5, $pop40
+; NO-SIMD128-FAST-NEXT:    i32.const $push39=, 65535
+; NO-SIMD128-FAST-NEXT:    i32.and $push15=, $13, $pop39
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push17=, $pop16, $pop15
+; NO-SIMD128-FAST-NEXT:    i32.store16 8($0), $pop17
+; NO-SIMD128-FAST-NEXT:    i32.const $push21=, 10
+; NO-SIMD128-FAST-NEXT:    i32.add $push22=, $0, $pop21
+; NO-SIMD128-FAST-NEXT:    i32.const $push38=, 65535
+; NO-SIMD128-FAST-NEXT:    i32.and $push19=, $6, $pop38
+; NO-SIMD128-FAST-NEXT:    i32.const $push37=, 65535
+; NO-SIMD128-FAST-NEXT:    i32.and $push18=, $14, $pop37
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push20=, $pop19, $pop18
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop22), $pop20
+; NO-SIMD128-FAST-NEXT:    i32.const $push26=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push27=, $0, $pop26
+; NO-SIMD128-FAST-NEXT:    i32.const $push36=, 65535
+; NO-SIMD128-FAST-NEXT:    i32.and $push24=, $7, $pop36
+; NO-SIMD128-FAST-NEXT:    i32.const $push35=, 65535
+; NO-SIMD128-FAST-NEXT:    i32.and $push23=, $15, $pop35
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push25=, $pop24, $pop23
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop27), $pop25
+; NO-SIMD128-FAST-NEXT:    i32.const $push31=, 14
+; NO-SIMD128-FAST-NEXT:    i32.add $push32=, $0, $pop31
+; NO-SIMD128-FAST-NEXT:    i32.const $push34=, 65535
+; NO-SIMD128-FAST-NEXT:    i32.and $push29=, $8, $pop34
+; NO-SIMD128-FAST-NEXT:    i32.const $push33=, 65535
+; NO-SIMD128-FAST-NEXT:    i32.and $push28=, $16, $pop33
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push30=, $pop29, $pop28
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop32), $pop30
+; NO-SIMD128-FAST-NEXT:    return
   %a = lshr <8 x i16> %v, %x
   ret <8 x i16> %a
 }
 
-; CHECK-LABEL: and_v8i16:
-; NO-SIMD128-NOT: v128
-; SIMD128-NEXT: .functype and_v8i16 (v128, v128) -> (v128){{$}}
-; SIMD128-NEXT: v128.and $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <8 x i16> @and_v8i16(<8 x i16> %x, <8 x i16> %y) {
+; SIMD128-LABEL: and_v8i16:
+; SIMD128:         .functype and_v8i16 (v128, v128) -> (v128)
+; SIMD128-NEXT:    v128.and $push0=, $0, $1
+; SIMD128-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: and_v8i16:
+; NO-SIMD128-SLOW:         .functype and_v8i16 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push1=, 14
+; NO-SIMD128-SLOW-NEXT:    i32.add $push2=, $0, $pop1
+; NO-SIMD128-SLOW-NEXT:    i32.and $push0=, $8, $16
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop2), $pop0
+; NO-SIMD128-SLOW-NEXT:    i32.const $push4=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push5=, $0, $pop4
+; NO-SIMD128-SLOW-NEXT:    i32.and $push3=, $7, $15
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop5), $pop3
+; NO-SIMD128-SLOW-NEXT:    i32.const $push7=, 10
+; NO-SIMD128-SLOW-NEXT:    i32.add $push8=, $0, $pop7
+; NO-SIMD128-SLOW-NEXT:    i32.and $push6=, $6, $14
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop8), $pop6
+; NO-SIMD128-SLOW-NEXT:    i32.and $push9=, $5, $13
+; NO-SIMD128-SLOW-NEXT:    i32.store16 8($0), $pop9
+; NO-SIMD128-SLOW-NEXT:    i32.const $push11=, 6
+; NO-SIMD128-SLOW-NEXT:    i32.add $push12=, $0, $pop11
+; NO-SIMD128-SLOW-NEXT:    i32.and $push10=, $4, $12
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop12), $pop10
+; NO-SIMD128-SLOW-NEXT:    i32.and $push13=, $3, $11
+; NO-SIMD128-SLOW-NEXT:    i32.store16 4($0), $pop13
+; NO-SIMD128-SLOW-NEXT:    i32.and $push14=, $2, $10
+; NO-SIMD128-SLOW-NEXT:    i32.store16 2($0), $pop14
+; NO-SIMD128-SLOW-NEXT:    i32.and $push15=, $1, $9
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($0), $pop15
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: and_v8i16:
+; NO-SIMD128-FAST:         .functype and_v8i16 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-FAST-NEXT:    i32.and $push0=, $1, $9
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($0), $pop0
+; NO-SIMD128-FAST-NEXT:    i32.and $push1=, $2, $10
+; NO-SIMD128-FAST-NEXT:    i32.store16 2($0), $pop1
+; NO-SIMD128-FAST-NEXT:    i32.and $push2=, $3, $11
+; NO-SIMD128-FAST-NEXT:    i32.store16 4($0), $pop2
+; NO-SIMD128-FAST-NEXT:    i32.const $push3=, 6
+; NO-SIMD128-FAST-NEXT:    i32.add $push4=, $0, $pop3
+; NO-SIMD128-FAST-NEXT:    i32.and $push5=, $4, $12
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop4), $pop5
+; NO-SIMD128-FAST-NEXT:    i32.and $push6=, $5, $13
+; NO-SIMD128-FAST-NEXT:    i32.store16 8($0), $pop6
+; NO-SIMD128-FAST-NEXT:    i32.const $push7=, 10
+; NO-SIMD128-FAST-NEXT:    i32.add $push8=, $0, $pop7
+; NO-SIMD128-FAST-NEXT:    i32.and $push9=, $6, $14
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop8), $pop9
+; NO-SIMD128-FAST-NEXT:    i32.const $push10=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push11=, $0, $pop10
+; NO-SIMD128-FAST-NEXT:    i32.and $push12=, $7, $15
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop11), $pop12
+; NO-SIMD128-FAST-NEXT:    i32.const $push13=, 14
+; NO-SIMD128-FAST-NEXT:    i32.add $push14=, $0, $pop13
+; NO-SIMD128-FAST-NEXT:    i32.and $push15=, $8, $16
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop14), $pop15
+; NO-SIMD128-FAST-NEXT:    return
   %a = and <8 x i16> %x, %y
   ret <8 x i16> %a
 }
 
-; CHECK-LABEL: or_v8i16:
-; NO-SIMD128-NOT: v128
-; SIMD128-NEXT: .functype or_v8i16 (v128, v128) -> (v128){{$}}
-; SIMD128-NEXT: v128.or $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <8 x i16> @or_v8i16(<8 x i16> %x, <8 x i16> %y) {
+; SIMD128-LABEL: or_v8i16:
+; SIMD128:         .functype or_v8i16 (v128, v128) -> (v128)
+; SIMD128-NEXT:    v128.or $push0=, $0, $1
+; SIMD128-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: or_v8i16:
+; NO-SIMD128-SLOW:         .functype or_v8i16 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push1=, 14
+; NO-SIMD128-SLOW-NEXT:    i32.add $push2=, $0, $pop1
+; NO-SIMD128-SLOW-NEXT:    i32.or $push0=, $8, $16
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop2), $pop0
+; NO-SIMD128-SLOW-NEXT:    i32.const $push4=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push5=, $0, $pop4
+; NO-SIMD128-SLOW-NEXT:    i32.or $push3=, $7, $15
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop5), $pop3
+; NO-SIMD128-SLOW-NEXT:    i32.const $push7=, 10
+; NO-SIMD128-SLOW-NEXT:    i32.add $push8=, $0, $pop7
+; NO-SIMD128-SLOW-NEXT:    i32.or $push6=, $6, $14
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop8), $pop6
+; NO-SIMD128-SLOW-NEXT:    i32.or $push9=, $5, $13
+; NO-SIMD128-SLOW-NEXT:    i32.store16 8($0), $pop9
+; NO-SIMD128-SLOW-NEXT:    i32.const $push11=, 6
+; NO-SIMD128-SLOW-NEXT:    i32.add $push12=, $0, $pop11
+; NO-SIMD128-SLOW-NEXT:    i32.or $push10=, $4, $12
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop12), $pop10
+; NO-SIMD128-SLOW-NEXT:    i32.or $push13=, $3, $11
+; NO-SIMD128-SLOW-NEXT:    i32.store16 4($0), $pop13
+; NO-SIMD128-SLOW-NEXT:    i32.or $push14=, $2, $10
+; NO-SIMD128-SLOW-NEXT:    i32.store16 2($0), $pop14
+; NO-SIMD128-SLOW-NEXT:    i32.or $push15=, $1, $9
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($0), $pop15
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: or_v8i16:
+; NO-SIMD128-FAST:         .functype or_v8i16 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-FAST-NEXT:    i32.or $push0=, $1, $9
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($0), $pop0
+; NO-SIMD128-FAST-NEXT:    i32.or $push1=, $2, $10
+; NO-SIMD128-FAST-NEXT:    i32.store16 2($0), $pop1
+; NO-SIMD128-FAST-NEXT:    i32.or $push2=, $3, $11
+; NO-SIMD128-FAST-NEXT:    i32.store16 4($0), $pop2
+; NO-SIMD128-FAST-NEXT:    i32.const $push3=, 6
+; NO-SIMD128-FAST-NEXT:    i32.add $push4=, $0, $pop3
+; NO-SIMD128-FAST-NEXT:    i32.or $push5=, $4, $12
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop4), $pop5
+; NO-SIMD128-FAST-NEXT:    i32.or $push6=, $5, $13
+; NO-SIMD128-FAST-NEXT:    i32.store16 8($0), $pop6
+; NO-SIMD128-FAST-NEXT:    i32.const $push7=, 10
+; NO-SIMD128-FAST-NEXT:    i32.add $push8=, $0, $pop7
+; NO-SIMD128-FAST-NEXT:    i32.or $push9=, $6, $14
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop8), $pop9
+; NO-SIMD128-FAST-NEXT:    i32.const $push10=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push11=, $0, $pop10
+; NO-SIMD128-FAST-NEXT:    i32.or $push12=, $7, $15
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop11), $pop12
+; NO-SIMD128-FAST-NEXT:    i32.const $push13=, 14
+; NO-SIMD128-FAST-NEXT:    i32.add $push14=, $0, $pop13
+; NO-SIMD128-FAST-NEXT:    i32.or $push15=, $8, $16
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop14), $pop15
+; NO-SIMD128-FAST-NEXT:    return
   %a = or <8 x i16> %x, %y
   ret <8 x i16> %a
 }
 
-; CHECK-LABEL: xor_v8i16:
-; NO-SIMD128-NOT: v128
-; SIMD128-NEXT: .functype xor_v8i16 (v128, v128) -> (v128){{$}}
-; SIMD128-NEXT: v128.xor $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <8 x i16> @xor_v8i16(<8 x i16> %x, <8 x i16> %y) {
+; SIMD128-LABEL: xor_v8i16:
+; SIMD128:         .functype xor_v8i16 (v128, v128) -> (v128)
+; SIMD128-NEXT:    v128.xor $push0=, $0, $1
+; SIMD128-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: xor_v8i16:
+; NO-SIMD128-SLOW:         .functype xor_v8i16 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push1=, 14
+; NO-SIMD128-SLOW-NEXT:    i32.add $push2=, $0, $pop1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push0=, $8, $16
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop2), $pop0
+; NO-SIMD128-SLOW-NEXT:    i32.const $push4=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push5=, $0, $pop4
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push3=, $7, $15
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop5), $pop3
+; NO-SIMD128-SLOW-NEXT:    i32.const $push7=, 10
+; NO-SIMD128-SLOW-NEXT:    i32.add $push8=, $0, $pop7
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push6=, $6, $14
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop8), $pop6
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push9=, $5, $13
+; NO-SIMD128-SLOW-NEXT:    i32.store16 8($0), $pop9
+; NO-SIMD128-SLOW-NEXT:    i32.const $push11=, 6
+; NO-SIMD128-SLOW-NEXT:    i32.add $push12=, $0, $pop11
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push10=, $4, $12
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop12), $pop10
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push13=, $3, $11
+; NO-SIMD128-SLOW-NEXT:    i32.store16 4($0), $pop13
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push14=, $2, $10
+; NO-SIMD128-SLOW-NEXT:    i32.store16 2($0), $pop14
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push15=, $1, $9
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($0), $pop15
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: xor_v8i16:
+; NO-SIMD128-FAST:         .functype xor_v8i16 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-FAST-NEXT:    i32.xor $push0=, $1, $9
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($0), $pop0
+; NO-SIMD128-FAST-NEXT:    i32.xor $push1=, $2, $10
+; NO-SIMD128-FAST-NEXT:    i32.store16 2($0), $pop1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push2=, $3, $11
+; NO-SIMD128-FAST-NEXT:    i32.store16 4($0), $pop2
+; NO-SIMD128-FAST-NEXT:    i32.const $push3=, 6
+; NO-SIMD128-FAST-NEXT:    i32.add $push4=, $0, $pop3
+; NO-SIMD128-FAST-NEXT:    i32.xor $push5=, $4, $12
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop4), $pop5
+; NO-SIMD128-FAST-NEXT:    i32.xor $push6=, $5, $13
+; NO-SIMD128-FAST-NEXT:    i32.store16 8($0), $pop6
+; NO-SIMD128-FAST-NEXT:    i32.const $push7=, 10
+; NO-SIMD128-FAST-NEXT:    i32.add $push8=, $0, $pop7
+; NO-SIMD128-FAST-NEXT:    i32.xor $push9=, $6, $14
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop8), $pop9
+; NO-SIMD128-FAST-NEXT:    i32.const $push10=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push11=, $0, $pop10
+; NO-SIMD128-FAST-NEXT:    i32.xor $push12=, $7, $15
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop11), $pop12
+; NO-SIMD128-FAST-NEXT:    i32.const $push13=, 14
+; NO-SIMD128-FAST-NEXT:    i32.add $push14=, $0, $pop13
+; NO-SIMD128-FAST-NEXT:    i32.xor $push15=, $8, $16
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop14), $pop15
+; NO-SIMD128-FAST-NEXT:    return
   %a = xor <8 x i16> %x, %y
   ret <8 x i16> %a
 }
 
-; CHECK-LABEL: not_v8i16:
-; NO-SIMD128-NOT: v128
-; SIMD128-NEXT: .functype not_v8i16 (v128) -> (v128){{$}}
-; SIMD128-NEXT: v128.not $push[[R:[0-9]+]]=, $0{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <8 x i16> @not_v8i16(<8 x i16> %x) {
+; SIMD128-LABEL: not_v8i16:
+; SIMD128:         .functype not_v8i16 (v128) -> (v128)
+; SIMD128-NEXT:    v128.not $push0=, $0
+; SIMD128-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: not_v8i16:
+; NO-SIMD128-SLOW:         .functype not_v8i16 (i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push2=, 14
+; NO-SIMD128-SLOW-NEXT:    i32.add $push3=, $0, $pop2
+; NO-SIMD128-SLOW-NEXT:    i32.const $push0=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push1=, $8, $pop0
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop3), $pop1
+; NO-SIMD128-SLOW-NEXT:    i32.const $push5=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push6=, $0, $pop5
+; NO-SIMD128-SLOW-NEXT:    i32.const $push23=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push4=, $7, $pop23
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop6), $pop4
+; NO-SIMD128-SLOW-NEXT:    i32.const $push8=, 10
+; NO-SIMD128-SLOW-NEXT:    i32.add $push9=, $0, $pop8
+; NO-SIMD128-SLOW-NEXT:    i32.const $push22=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push7=, $6, $pop22
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop9), $pop7
+; NO-SIMD128-SLOW-NEXT:    i32.const $push21=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push10=, $5, $pop21
+; NO-SIMD128-SLOW-NEXT:    i32.store16 8($0), $pop10
+; NO-SIMD128-SLOW-NEXT:    i32.const $push12=, 6
+; NO-SIMD128-SLOW-NEXT:    i32.add $push13=, $0, $pop12
+; NO-SIMD128-SLOW-NEXT:    i32.const $push20=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push11=, $4, $pop20
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop13), $pop11
+; NO-SIMD128-SLOW-NEXT:    i32.const $push19=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push14=, $3, $pop19
+; NO-SIMD128-SLOW-NEXT:    i32.store16 4($0), $pop14
+; NO-SIMD128-SLOW-NEXT:    i32.const $push18=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push15=, $2, $pop18
+; NO-SIMD128-SLOW-NEXT:    i32.store16 2($0), $pop15
+; NO-SIMD128-SLOW-NEXT:    i32.const $push17=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push16=, $1, $pop17
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($0), $pop16
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: not_v8i16:
+; NO-SIMD128-FAST:         .functype not_v8i16 (i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-FAST-NEXT:    i32.const $push0=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push1=, $1, $pop0
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($0), $pop1
+; NO-SIMD128-FAST-NEXT:    i32.const $push23=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push2=, $2, $pop23
+; NO-SIMD128-FAST-NEXT:    i32.store16 2($0), $pop2
+; NO-SIMD128-FAST-NEXT:    i32.const $push22=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push3=, $3, $pop22
+; NO-SIMD128-FAST-NEXT:    i32.store16 4($0), $pop3
+; NO-SIMD128-FAST-NEXT:    i32.const $push4=, 6
+; NO-SIMD128-FAST-NEXT:    i32.add $push5=, $0, $pop4
+; NO-SIMD128-FAST-NEXT:    i32.const $push21=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push6=, $4, $pop21
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop5), $pop6
+; NO-SIMD128-FAST-NEXT:    i32.const $push20=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push7=, $5, $pop20
+; NO-SIMD128-FAST-NEXT:    i32.store16 8($0), $pop7
+; NO-SIMD128-FAST-NEXT:    i32.const $push8=, 10
+; NO-SIMD128-FAST-NEXT:    i32.add $push9=, $0, $pop8
+; NO-SIMD128-FAST-NEXT:    i32.const $push19=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push10=, $6, $pop19
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop9), $pop10
+; NO-SIMD128-FAST-NEXT:    i32.const $push11=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push12=, $0, $pop11
+; NO-SIMD128-FAST-NEXT:    i32.const $push18=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push13=, $7, $pop18
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop12), $pop13
+; NO-SIMD128-FAST-NEXT:    i32.const $push14=, 14
+; NO-SIMD128-FAST-NEXT:    i32.add $push15=, $0, $pop14
+; NO-SIMD128-FAST-NEXT:    i32.const $push17=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push16=, $8, $pop17
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop15), $pop16
+; NO-SIMD128-FAST-NEXT:    return
   %a = xor <8 x i16> %x, <i16 -1, i16 -1, i16 -1, i16 -1,
                           i16 -1, i16 -1, i16 -1, i16 -1>
   ret <8 x i16> %a
 }
 
-; CHECK-LABEL: bitselect_v8i16:
-; NO-SIMD128-NOT: v128
-; SIMD128-NEXT: .functype bitselect_v8i16 (v128, v128, v128) -> (v128){{$}}
-; SIMD128-SLOW-NEXT: v128.bitselect $push[[R:[0-9]+]]=, $1, $2, $0{{$}}
-; SIMD128-SLOW-NEXT: return $pop[[R]]{{$}}
-; SIMD128-FAST-NEXT: v128.and
-; SIMD128-FAST-NEXT: v128.not
-; SIMD128-FAST-NEXT: v128.and
-; SIMD128-FAST-NEXT: v128.or
-; SIMD128-FAST-NEXT: return
 define <8 x i16> @bitselect_v8i16(<8 x i16> %c, <8 x i16> %v1, <8 x i16> %v2) {
+; SIMD128-SLOW-LABEL: bitselect_v8i16:
+; SIMD128-SLOW:         .functype bitselect_v8i16 (v128, v128, v128) -> (v128)
+; SIMD128-SLOW-NEXT:    v128.bitselect $push0=, $1, $2, $0
+; SIMD128-SLOW-NEXT:    return $pop0
+;
+; SIMD128-FAST-LABEL: bitselect_v8i16:
+; SIMD128-FAST:         .functype bitselect_v8i16 (v128, v128, v128) -> (v128)
+; SIMD128-FAST-NEXT:    v128.and $push0=, $1, $0
+; SIMD128-FAST-NEXT:    v128.not $push2=, $0
+; SIMD128-FAST-NEXT:    v128.and $push3=, $2, $pop2
+; SIMD128-FAST-NEXT:    v128.or $push1=, $pop0, $pop3
+; SIMD128-FAST-NEXT:    return $pop1
+;
+; NO-SIMD128-SLOW-LABEL: bitselect_v8i16:
+; NO-SIMD128-SLOW:         .functype bitselect_v8i16 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push5=, 14
+; NO-SIMD128-SLOW-NEXT:    i32.add $push6=, $0, $pop5
+; NO-SIMD128-SLOW-NEXT:    i32.and $push0=, $16, $8
+; NO-SIMD128-SLOW-NEXT:    i32.const $push1=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push2=, $8, $pop1
+; NO-SIMD128-SLOW-NEXT:    i32.and $push3=, $24, $pop2
+; NO-SIMD128-SLOW-NEXT:    i32.or $push4=, $pop0, $pop3
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop6), $pop4
+; NO-SIMD128-SLOW-NEXT:    i32.const $push11=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push12=, $0, $pop11
+; NO-SIMD128-SLOW-NEXT:    i32.and $push7=, $15, $7
+; NO-SIMD128-SLOW-NEXT:    i32.const $push47=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push8=, $7, $pop47
+; NO-SIMD128-SLOW-NEXT:    i32.and $push9=, $23, $pop8
+; NO-SIMD128-SLOW-NEXT:    i32.or $push10=, $pop7, $pop9
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop12), $pop10
+; NO-SIMD128-SLOW-NEXT:    i32.const $push17=, 10
+; NO-SIMD128-SLOW-NEXT:    i32.add $push18=, $0, $pop17
+; NO-SIMD128-SLOW-NEXT:    i32.and $push13=, $14, $6
+; NO-SIMD128-SLOW-NEXT:    i32.const $push46=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push14=, $6, $pop46
+; NO-SIMD128-SLOW-NEXT:    i32.and $push15=, $22, $pop14
+; NO-SIMD128-SLOW-NEXT:    i32.or $push16=, $pop13, $pop15
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop18), $pop16
+; NO-SIMD128-SLOW-NEXT:    i32.and $push19=, $13, $5
+; NO-SIMD128-SLOW-NEXT:    i32.const $push45=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push20=, $5, $pop45
+; NO-SIMD128-SLOW-NEXT:    i32.and $push21=, $21, $pop20
+; NO-SIMD128-SLOW-NEXT:    i32.or $push22=, $pop19, $pop21
+; NO-SIMD128-SLOW-NEXT:    i32.store16 8($0), $pop22
+; NO-SIMD128-SLOW-NEXT:    i32.const $push27=, 6
+; NO-SIMD128-SLOW-NEXT:    i32.add $push28=, $0, $pop27
+; NO-SIMD128-SLOW-NEXT:    i32.and $push23=, $12, $4
+; NO-SIMD128-SLOW-NEXT:    i32.const $push44=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push24=, $4, $pop44
+; NO-SIMD128-SLOW-NEXT:    i32.and $push25=, $20, $pop24
+; NO-SIMD128-SLOW-NEXT:    i32.or $push26=, $pop23, $pop25
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($pop28), $pop26
+; NO-SIMD128-SLOW-NEXT:    i32.and $push29=, $11, $3
+; NO-SIMD128-SLOW-NEXT:    i32.const $push43=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push30=, $3, $pop43
+; NO-SIMD128-SLOW-NEXT:    i32.and $push31=, $19, $pop30
+; NO-SIMD128-SLOW-NEXT:    i32.or $push32=, $pop29, $pop31
+; NO-SIMD128-SLOW-NEXT:    i32.store16 4($0), $pop32
+; NO-SIMD128-SLOW-NEXT:    i32.and $push33=, $10, $2
+; NO-SIMD128-SLOW-NEXT:    i32.const $push42=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push34=, $2, $pop42
+; NO-SIMD128-SLOW-NEXT:    i32.and $push35=, $18, $pop34
+; NO-SIMD128-SLOW-NEXT:    i32.or $push36=, $pop33, $pop35
+; NO-SIMD128-SLOW-NEXT:    i32.store16 2($0), $pop36
+; NO-SIMD128-SLOW-NEXT:    i32.and $push37=, $9, $1
+; NO-SIMD128-SLOW-NEXT:    i32.const $push41=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push38=, $1, $pop41
+; NO-SIMD128-SLOW-NEXT:    i32.and $push39=, $17, $pop38
+; NO-SIMD128-SLOW-NEXT:    i32.or $push40=, $pop37, $pop39
+; NO-SIMD128-SLOW-NEXT:    i32.store16 0($0), $pop40
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: bitselect_v8i16:
+; NO-SIMD128-FAST:         .functype bitselect_v8i16 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-FAST-NEXT:    i32.and $push0=, $9, $1
+; NO-SIMD128-FAST-NEXT:    i32.const $push1=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push2=, $1, $pop1
+; NO-SIMD128-FAST-NEXT:    i32.and $push3=, $17, $pop2
+; NO-SIMD128-FAST-NEXT:    i32.or $push4=, $pop0, $pop3
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($0), $pop4
+; NO-SIMD128-FAST-NEXT:    i32.and $push5=, $10, $2
+; NO-SIMD128-FAST-NEXT:    i32.const $push47=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push6=, $2, $pop47
+; NO-SIMD128-FAST-NEXT:    i32.and $push7=, $18, $pop6
+; NO-SIMD128-FAST-NEXT:    i32.or $push8=, $pop5, $pop7
+; NO-SIMD128-FAST-NEXT:    i32.store16 2($0), $pop8
+; NO-SIMD128-FAST-NEXT:    i32.and $push9=, $11, $3
+; NO-SIMD128-FAST-NEXT:    i32.const $push46=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push10=, $3, $pop46
+; NO-SIMD128-FAST-NEXT:    i32.and $push11=, $19, $pop10
+; NO-SIMD128-FAST-NEXT:    i32.or $push12=, $pop9, $pop11
+; NO-SIMD128-FAST-NEXT:    i32.store16 4($0), $pop12
+; NO-SIMD128-FAST-NEXT:    i32.const $push17=, 6
+; NO-SIMD128-FAST-NEXT:    i32.add $push18=, $0, $pop17
+; NO-SIMD128-FAST-NEXT:    i32.and $push13=, $12, $4
+; NO-SIMD128-FAST-NEXT:    i32.const $push45=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push14=, $4, $pop45
+; NO-SIMD128-FAST-NEXT:    i32.and $push15=, $20, $pop14
+; NO-SIMD128-FAST-NEXT:    i32.or $push16=, $pop13, $pop15
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop18), $pop16
+; NO-SIMD128-FAST-NEXT:    i32.and $push19=, $13, $5
+; NO-SIMD128-FAST-NEXT:    i32.const $push44=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push20=, $5, $pop44
+; NO-SIMD128-FAST-NEXT:    i32.and $push21=, $21, $pop20
+; NO-SIMD128-FAST-NEXT:    i32.or $push22=, $pop19, $pop21
+; NO-SIMD128-FAST-NEXT:    i32.store16 8($0), $pop22
+; NO-SIMD128-FAST-NEXT:    i32.const $push27=, 10
+; NO-SIMD128-FAST-NEXT:    i32.add $push28=, $0, $pop27
+; NO-SIMD128-FAST-NEXT:    i32.and $push23=, $14, $6
+; NO-SIMD128-FAST-NEXT:    i32.const $push43=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push24=, $6, $pop43
+; NO-SIMD128-FAST-NEXT:    i32.and $push25=, $22, $pop24
+; NO-SIMD128-FAST-NEXT:    i32.or $push26=, $pop23, $pop25
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop28), $pop26
+; NO-SIMD128-FAST-NEXT:    i32.const $push33=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push34=, $0, $pop33
+; NO-SIMD128-FAST-NEXT:    i32.and $push29=, $15, $7
+; NO-SIMD128-FAST-NEXT:    i32.const $push42=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push30=, $7, $pop42
+; NO-SIMD128-FAST-NEXT:    i32.and $push31=, $23, $pop30
+; NO-SIMD128-FAST-NEXT:    i32.or $push32=, $pop29, $pop31
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop34), $pop32
+; NO-SIMD128-FAST-NEXT:    i32.const $push39=, 14
+; NO-SIMD128-FAST-NEXT:    i32.add $push40=, $0, $pop39
+; NO-SIMD128-FAST-NEXT:    i32.and $push35=, $16, $8
+; NO-SIMD128-FAST-NEXT:    i32.const $push41=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push36=, $8, $pop41
+; NO-SIMD128-FAST-NEXT:    i32.and $push37=, $24, $pop36
+; NO-SIMD128-FAST-NEXT:    i32.or $push38=, $pop35, $pop37
+; NO-SIMD128-FAST-NEXT:    i32.store16 0($pop40), $pop38
+; NO-SIMD128-FAST-NEXT:    return
   %masked_v1 = and <8 x i16> %v1, %c
   %inv_mask = xor <8 x i16>
     <i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1>,
     %c
   %masked_v2 = and <8 x i16> %v2, %inv_mask
   %a = or <8 x i16> %masked_v1, %masked_v2
   ret <8 x i16> %a
 }
 
 ; ==============================================================================
 ; 4 x i32
 ; ==============================================================================
-; CHECK-LABEL: add_v4i32:
-; NO-SIMD128-NOT: i32x4
-; SIMD128-NEXT: .functype add_v4i32 (v128, v128) -> (v128){{$}}
-; SIMD128-NEXT: i32x4.add $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
+
 define <4 x i32> @add_v4i32(<4 x i32> %x, <4 x i32> %y) {
+; SIMD128-LABEL: add_v4i32:
+; SIMD128:         .functype add_v4i32 (v128, v128) -> (v128)
+; SIMD128-NEXT:    i32x4.add $push0=, $0, $1
+; SIMD128-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: add_v4i32:
+; NO-SIMD128-SLOW:         .functype add_v4i32 (i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push1=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push2=, $0, $pop1
+; NO-SIMD128-SLOW-NEXT:    i32.add $push0=, $4, $8
+; NO-SIMD128-SLOW-NEXT:    i32.store 0($pop2), $pop0
+; NO-SIMD128-SLOW-NEXT:    i32.add $push3=, $3, $7
+; NO-SIMD128-SLOW-NEXT:    i32.store 8($0), $pop3
+; NO-SIMD128-SLOW-NEXT:    i32.add $push4=, $2, $6
+; NO-SIMD128-SLOW-NEXT:    i32.store 4($0), $pop4
+; NO-SIMD128-SLOW-NEXT:    i32.add $push5=, $1, $5
+; NO-SIMD128-SLOW-NEXT:    i32.store 0($0), $pop5
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: add_v4i32:
+; NO-SIMD128-FAST:         .functype add_v4i32 (i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-FAST-NEXT:    i32.add $push0=, $1, $5
+; NO-SIMD128-FAST-NEXT:    i32.store 0($0), $pop0
+; NO-SIMD128-FAST-NEXT:    i32.add $push1=, $2, $6
+; NO-SIMD128-FAST-NEXT:    i32.store 4($0), $pop1
+; NO-SIMD128-FAST-NEXT:    i32.add $push2=, $3, $7
+; NO-SIMD128-FAST-NEXT:    i32.store 8($0), $pop2
+; NO-SIMD128-FAST-NEXT:    i32.const $push3=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push4=, $0, $pop3
+; NO-SIMD128-FAST-NEXT:    i32.add $push5=, $4, $8
+; NO-SIMD128-FAST-NEXT:    i32.store 0($pop4), $pop5
+; NO-SIMD128-FAST-NEXT:    return
   %a = add <4 x i32> %x, %y
   ret <4 x i32> %a
 }
 
-; CHECK-LABEL: sub_v4i32:
-; NO-SIMD128-NOT: i32x4
-; SIMD128-NEXT: .functype sub_v4i32 (v128, v128) -> (v128){{$}}
-; SIMD128-NEXT: i32x4.sub $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <4 x i32> @sub_v4i32(<4 x i32> %x, <4 x i32> %y) {
+; SIMD128-LABEL: sub_v4i32:
+; SIMD128:         .functype sub_v4i32 (v128, v128) -> (v128)
+; SIMD128-NEXT:    i32x4.sub $push0=, $0, $1
+; SIMD128-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: sub_v4i32:
+; NO-SIMD128-SLOW:         .functype sub_v4i32 (i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push1=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push2=, $0, $pop1
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push0=, $4, $8
+; NO-SIMD128-SLOW-NEXT:    i32.store 0($pop2), $pop0
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push3=, $3, $7
+; NO-SIMD128-SLOW-NEXT:    i32.store 8($0), $pop3
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push4=, $2, $6
+; NO-SIMD128-SLOW-NEXT:    i32.store 4($0), $pop4
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push5=, $1, $5
+; NO-SIMD128-SLOW-NEXT:    i32.store 0($0), $pop5
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: sub_v4i32:
+; NO-SIMD128-FAST:         .functype sub_v4i32 (i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-FAST-NEXT:    i32.sub $push0=, $1, $5
+; NO-SIMD128-FAST-NEXT:    i32.store 0($0), $pop0
+; NO-SIMD128-FAST-NEXT:    i32.sub $push1=, $2, $6
+; NO-SIMD128-FAST-NEXT:    i32.store 4($0), $pop1
+; NO-SIMD128-FAST-NEXT:    i32.sub $push2=, $3, $7
+; NO-SIMD128-FAST-NEXT:    i32.store 8($0), $pop2
+; NO-SIMD128-FAST-NEXT:    i32.const $push3=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push4=, $0, $pop3
+; NO-SIMD128-FAST-NEXT:    i32.sub $push5=, $4, $8
+; NO-SIMD128-FAST-NEXT:    i32.store 0($pop4), $pop5
+; NO-SIMD128-FAST-NEXT:    return
   %a = sub <4 x i32> %x, %y
   ret <4 x i32> %a
 }
 
-; CHECK-LABEL: mul_v4i32:
-; NO-SIMD128-NOT: i32x4
-; SIMD128-NEXT: .functype mul_v4i32 (v128, v128) -> (v128){{$}}
-; SIMD128-NEXT: i32x4.mul $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <4 x i32> @mul_v4i32(<4 x i32> %x, <4 x i32> %y) {
+; SIMD128-LABEL: mul_v4i32:
+; SIMD128:         .functype mul_v4i32 (v128, v128) -> (v128)
+; SIMD128-NEXT:    i32x4.mul $push0=, $0, $1
+; SIMD128-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: mul_v4i32:
+; NO-SIMD128-SLOW:         .functype mul_v4i32 (i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push1=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push2=, $0, $pop1
+; NO-SIMD128-SLOW-NEXT:    i32.mul $push0=, $4, $8
+; NO-SIMD128-SLOW-NEXT:    i32.store 0($pop2), $pop0
+; NO-SIMD128-SLOW-NEXT:    i32.mul $push3=, $3, $7
+; NO-SIMD128-SLOW-NEXT:    i32.store 8($0), $pop3
+; NO-SIMD128-SLOW-NEXT:    i32.mul $push4=, $2, $6
+; NO-SIMD128-SLOW-NEXT:    i32.store 4($0), $pop4
+; NO-SIMD128-SLOW-NEXT:    i32.mul $push5=, $1, $5
+; NO-SIMD128-SLOW-NEXT:    i32.store 0($0), $pop5
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: mul_v4i32:
+; NO-SIMD128-FAST:         .functype mul_v4i32 (i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-FAST-NEXT:    i32.mul $push0=, $1, $5
+; NO-SIMD128-FAST-NEXT:    i32.store 0($0), $pop0
+; NO-SIMD128-FAST-NEXT:    i32.mul $push1=, $2, $6
+; NO-SIMD128-FAST-NEXT:    i32.store 4($0), $pop1
+; NO-SIMD128-FAST-NEXT:    i32.mul $push2=, $3, $7
+; NO-SIMD128-FAST-NEXT:    i32.store 8($0), $pop2
+; NO-SIMD128-FAST-NEXT:    i32.const $push3=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push4=, $0, $pop3
+; NO-SIMD128-FAST-NEXT:    i32.mul $push5=, $4, $8
+; NO-SIMD128-FAST-NEXT:    i32.store 0($pop4), $pop5
+; NO-SIMD128-FAST-NEXT:    return
   %a = mul <4 x i32> %x, %y
   ret <4 x i32> %a
 }
 
-; CHECK-LABEL: neg_v4i32:
-; NO-SIMD128-NOT: i32x4
-; SIMD128-NEXT: .functype neg_v4i32 (v128) -> (v128){{$}}
-; SIMD128-NEXT: i32x4.neg $push[[R:[0-9]+]]=, $0{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <4 x i32> @neg_v4i32(<4 x i32> %x) {
+; SIMD128-LABEL: neg_v4i32:
+; SIMD128:         .functype neg_v4i32 (v128) -> (v128)
+; SIMD128-NEXT:    i32x4.neg $push0=, $0
+; SIMD128-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: neg_v4i32:
+; NO-SIMD128-SLOW:         .functype neg_v4i32 (i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push2=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push3=, $0, $pop2
+; NO-SIMD128-SLOW-NEXT:    i32.const $push0=, 0
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push1=, $pop0, $4
+; NO-SIMD128-SLOW-NEXT:    i32.store 0($pop3), $pop1
+; NO-SIMD128-SLOW-NEXT:    i32.const $push9=, 0
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push4=, $pop9, $3
+; NO-SIMD128-SLOW-NEXT:    i32.store 8($0), $pop4
+; NO-SIMD128-SLOW-NEXT:    i32.const $push8=, 0
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push5=, $pop8, $2
+; NO-SIMD128-SLOW-NEXT:    i32.store 4($0), $pop5
+; NO-SIMD128-SLOW-NEXT:    i32.const $push7=, 0
+; NO-SIMD128-SLOW-NEXT:    i32.sub $push6=, $pop7, $1
+; NO-SIMD128-SLOW-NEXT:    i32.store 0($0), $pop6
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: neg_v4i32:
+; NO-SIMD128-FAST:         .functype neg_v4i32 (i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-FAST-NEXT:    i32.const $push0=, 0
+; NO-SIMD128-FAST-NEXT:    i32.sub $push1=, $pop0, $1
+; NO-SIMD128-FAST-NEXT:    i32.store 0($0), $pop1
+; NO-SIMD128-FAST-NEXT:    i32.const $push9=, 0
+; NO-SIMD128-FAST-NEXT:    i32.sub $push2=, $pop9, $2
+; NO-SIMD128-FAST-NEXT:    i32.store 4($0), $pop2
+; NO-SIMD128-FAST-NEXT:    i32.const $push8=, 0
+; NO-SIMD128-FAST-NEXT:    i32.sub $push3=, $pop8, $3
+; NO-SIMD128-FAST-NEXT:    i32.store 8($0), $pop3
+; NO-SIMD128-FAST-NEXT:    i32.const $push4=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push5=, $0, $pop4
+; NO-SIMD128-FAST-NEXT:    i32.const $push7=, 0
+; NO-SIMD128-FAST-NEXT:    i32.sub $push6=, $pop7, $4
+; NO-SIMD128-FAST-NEXT:    i32.store 0($pop5), $pop6
+; NO-SIMD128-FAST-NEXT:    return
   %a = sub <4 x i32> <i32 0, i32 0, i32 0, i32 0>, %x
   ret <4 x i32> %a
 }
 
-; CHECK-LABEL: shl_v4i32:
-; NO-SIMD128-NOT: i32x4
-; SIMD128-NEXT: .functype shl_v4i32 (v128, i32) -> (v128){{$}}
-; SIMD128-NEXT: i32x4.shl $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <4 x i32> @shl_v4i32(<4 x i32> %v, i32 %x) {
+; SIMD128-UNIMPL-LABEL: shl_v4i32:
+; SIMD128-UNIMPL:         .functype shl_v4i32 (v128, i32) -> (v128)
+; SIMD128-UNIMPL-NEXT:    i32x4.shl $push0=, $0, $1
+; SIMD128-UNIMPL-NEXT:    return $pop0
+;
+; SIMD128-VM-SLOW-LABEL: shl_v4i32:
+; SIMD128-VM-SLOW:         .functype shl_v4i32 (v128, i32) -> (v128)
+; SIMD128-VM-SLOW-NEXT:    i32x4.extract_lane $push2=, $0, 0
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push3=, $pop2, $1
+; SIMD128-VM-SLOW-NEXT:    i32x4.splat $push4=, $pop3
+; SIMD128-VM-SLOW-NEXT:    i32x4.extract_lane $push0=, $0, 1
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push1=, $pop0, $1
+; SIMD128-VM-SLOW-NEXT:    i32x4.replace_lane $push5=, $pop4, 1, $pop1
+; SIMD128-VM-SLOW-NEXT:    i32x4.extract_lane $push6=, $0, 2
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push7=, $pop6, $1
+; SIMD128-VM-SLOW-NEXT:    i32x4.replace_lane $push8=, $pop5, 2, $pop7
+; SIMD128-VM-SLOW-NEXT:    i32x4.extract_lane $push9=, $0, 3
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push10=, $pop9, $1
+; SIMD128-VM-SLOW-NEXT:    i32x4.replace_lane $push11=, $pop8, 3, $pop10
+; SIMD128-VM-SLOW-NEXT:    return $pop11
+;
+; SIMD128-VM-FAST-LABEL: shl_v4i32:
+; SIMD128-VM-FAST:         .functype shl_v4i32 (v128, i32) -> (v128)
+; SIMD128-VM-FAST-NEXT:    i32x4.extract_lane $push3=, $0, 0
+; SIMD128-VM-FAST-NEXT:    i32.shl $push4=, $pop3, $1
+; SIMD128-VM-FAST-NEXT:    i32x4.splat $push5=, $pop4
+; SIMD128-VM-FAST-NEXT:    i32x4.extract_lane $push1=, $0, 1
+; SIMD128-VM-FAST-NEXT:    i32.shl $push2=, $pop1, $1
+; SIMD128-VM-FAST-NEXT:    i32x4.replace_lane $push6=, $pop5, 1, $pop2
+; SIMD128-VM-FAST-NEXT:    i32x4.extract_lane $push7=, $0, 2
+; SIMD128-VM-FAST-NEXT:    i32.shl $push8=, $pop7, $1
+; SIMD128-VM-FAST-NEXT:    i32x4.replace_lane $push9=, $pop6, 2, $pop8
+; SIMD128-VM-FAST-NEXT:    i32x4.extract_lane $push10=, $0, 3
+; SIMD128-VM-FAST-NEXT:    i32.shl $push11=, $pop10, $1
+; SIMD128-VM-FAST-NEXT:    i32x4.replace_lane $push0=, $pop9, 3, $pop11
+; SIMD128-VM-FAST-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: shl_v4i32:
+; NO-SIMD128-SLOW:         .functype shl_v4i32 (i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push1=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push2=, $0, $pop1
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push0=, $4, $5
+; NO-SIMD128-SLOW-NEXT:    i32.store 0($pop2), $pop0
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push3=, $3, $5
+; NO-SIMD128-SLOW-NEXT:    i32.store 8($0), $pop3
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push4=, $2, $5
+; NO-SIMD128-SLOW-NEXT:    i32.store 4($0), $pop4
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push5=, $1, $5
+; NO-SIMD128-SLOW-NEXT:    i32.store 0($0), $pop5
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: shl_v4i32:
+; NO-SIMD128-FAST:         .functype shl_v4i32 (i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-FAST-NEXT:    i32.shl $push0=, $1, $5
+; NO-SIMD128-FAST-NEXT:    i32.store 0($0), $pop0
+; NO-SIMD128-FAST-NEXT:    i32.shl $push1=, $2, $5
+; NO-SIMD128-FAST-NEXT:    i32.store 4($0), $pop1
+; NO-SIMD128-FAST-NEXT:    i32.shl $push2=, $3, $5
+; NO-SIMD128-FAST-NEXT:    i32.store 8($0), $pop2
+; NO-SIMD128-FAST-NEXT:    i32.const $push3=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push4=, $0, $pop3
+; NO-SIMD128-FAST-NEXT:    i32.shl $push5=, $4, $5
+; NO-SIMD128-FAST-NEXT:    i32.store 0($pop4), $pop5
+; NO-SIMD128-FAST-NEXT:    return
   %t = insertelement <4 x i32> undef, i32 %x, i32 0
   %s = shufflevector <4 x i32> %t, <4 x i32> undef,
     <4 x i32> <i32 0, i32 0, i32 0, i32 0>
   %a = shl <4 x i32> %v, %s
   ret <4 x i32> %a
 }
 
-; CHECK-LABEL: shl_const_v4i32:
-; NO-SIMD128-NOT: i32x4
-; SIMD128-NEXT: .functype shl_const_v4i32 (v128) -> (v128){{$}}
-; SIMD128-NEXT: i32.const $push[[L0:[0-9]+]]=, 5
-; SIMD128-NEXT: i32x4.shl $push[[R:[0-9]+]]=, $0, $pop[[L0]]{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <4 x i32> @shl_const_v4i32(<4 x i32> %v) {
+; SIMD128-UNIMPL-SLOW-LABEL: shl_const_v4i32:
+; SIMD128-UNIMPL-SLOW:         .functype shl_const_v4i32 (v128) -> (v128)
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.const $push0=, 5
+; SIMD128-UNIMPL-SLOW-NEXT:    i32x4.shl $push1=, $0, $pop0
+; SIMD128-UNIMPL-SLOW-NEXT:    return $pop1
+;
+; SIMD128-UNIMPL-FAST-LABEL: shl_const_v4i32:
+; SIMD128-UNIMPL-FAST:         .functype shl_const_v4i32 (v128) -> (v128)
+; SIMD128-UNIMPL-FAST-NEXT:    i32.const $push1=, 5
+; SIMD128-UNIMPL-FAST-NEXT:    i32x4.shl $push0=, $0, $pop1
+; SIMD128-UNIMPL-FAST-NEXT:    return $pop0
+;
+; SIMD128-VM-SLOW-LABEL: shl_const_v4i32:
+; SIMD128-VM-SLOW:         .functype shl_const_v4i32 (v128) -> (v128)
+; SIMD128-VM-SLOW-NEXT:    i32x4.extract_lane $push3=, $0, 0
+; SIMD128-VM-SLOW-NEXT:    i32.const $push1=, 5
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push4=, $pop3, $pop1
+; SIMD128-VM-SLOW-NEXT:    i32x4.splat $push5=, $pop4
+; SIMD128-VM-SLOW-NEXT:    i32x4.extract_lane $push0=, $0, 1
+; SIMD128-VM-SLOW-NEXT:    i32.const $push15=, 5
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push2=, $pop0, $pop15
+; SIMD128-VM-SLOW-NEXT:    i32x4.replace_lane $push6=, $pop5, 1, $pop2
+; SIMD128-VM-SLOW-NEXT:    i32x4.extract_lane $push7=, $0, 2
+; SIMD128-VM-SLOW-NEXT:    i32.const $push14=, 5
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push8=, $pop7, $pop14
+; SIMD128-VM-SLOW-NEXT:    i32x4.replace_lane $push9=, $pop6, 2, $pop8
+; SIMD128-VM-SLOW-NEXT:    i32x4.extract_lane $push10=, $0, 3
+; SIMD128-VM-SLOW-NEXT:    i32.const $push13=, 5
+; SIMD128-VM-SLOW-NEXT:    i32.shl $push11=, $pop10, $pop13
+; SIMD128-VM-SLOW-NEXT:    i32x4.replace_lane $push12=, $pop9, 3, $pop11
+; SIMD128-VM-SLOW-NEXT:    return $pop12
+;
+; SIMD128-VM-FAST-LABEL: shl_const_v4i32:
+; SIMD128-VM-FAST:         .functype shl_const_v4i32 (v128) -> (v128)
+; SIMD128-VM-FAST-NEXT:    i32x4.extract_lane $push4=, $0, 0
+; SIMD128-VM-FAST-NEXT:    i32.const $push2=, 5
+; SIMD128-VM-FAST-NEXT:    i32.shl $push5=, $pop4, $pop2
+; SIMD128-VM-FAST-NEXT:    i32x4.splat $push6=, $pop5
+; SIMD128-VM-FAST-NEXT:    i32x4.extract_lane $push1=, $0, 1
+; SIMD128-VM-FAST-NEXT:    i32.const $push15=, 5
+; SIMD128-VM-FAST-NEXT:    i32.shl $push3=, $pop1, $pop15
+; SIMD128-VM-FAST-NEXT:    i32x4.replace_lane $push7=, $pop6, 1, $pop3
+; SIMD128-VM-FAST-NEXT:    i32x4.extract_lane $push8=, $0, 2
+; SIMD128-VM-FAST-NEXT:    i32.const $push14=, 5
+; SIMD128-VM-FAST-NEXT:    i32.shl $push9=, $pop8, $pop14
+; SIMD128-VM-FAST-NEXT:    i32x4.replace_lane $push10=, $pop7, 2, $pop9
+; SIMD128-VM-FAST-NEXT:    i32x4.extract_lane $push11=, $0, 3
+; SIMD128-VM-FAST-NEXT:    i32.const $push13=, 5
+; SIMD128-VM-FAST-NEXT:    i32.shl $push12=, $pop11, $pop13
+; SIMD128-VM-FAST-NEXT:    i32x4.replace_lane $push0=, $pop10, 3, $pop12
+; SIMD128-VM-FAST-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: shl_const_v4i32:
+; NO-SIMD128-SLOW:         .functype shl_const_v4i32 (i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push2=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push3=, $0, $pop2
+; NO-SIMD128-SLOW-NEXT:    i32.const $push0=, 5
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push1=, $4, $pop0
+; NO-SIMD128-SLOW-NEXT:    i32.store 0($pop3), $pop1
+; NO-SIMD128-SLOW-NEXT:    i32.const $push9=, 5
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push4=, $3, $pop9
+; NO-SIMD128-SLOW-NEXT:    i32.store 8($0), $pop4
+; NO-SIMD128-SLOW-NEXT:    i32.const $push8=, 5
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push5=, $2, $pop8
+; NO-SIMD128-SLOW-NEXT:    i32.store 4($0), $pop5
+; NO-SIMD128-SLOW-NEXT:    i32.const $push7=, 5
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push6=, $1, $pop7
+; NO-SIMD128-SLOW-NEXT:    i32.store 0($0), $pop6
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: shl_const_v4i32:
+; NO-SIMD128-FAST:         .functype shl_const_v4i32 (i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-FAST-NEXT:    i32.const $push0=, 5
+; NO-SIMD128-FAST-NEXT:    i32.shl $push1=, $1, $pop0
+; NO-SIMD128-FAST-NEXT:    i32.store 0($0), $pop1
+; NO-SIMD128-FAST-NEXT:    i32.const $push9=, 5
+; NO-SIMD128-FAST-NEXT:    i32.shl $push2=, $2, $pop9
+; NO-SIMD128-FAST-NEXT:    i32.store 4($0), $pop2
+; NO-SIMD128-FAST-NEXT:    i32.const $push8=, 5
+; NO-SIMD128-FAST-NEXT:    i32.shl $push3=, $3, $pop8
+; NO-SIMD128-FAST-NEXT:    i32.store 8($0), $pop3
+; NO-SIMD128-FAST-NEXT:    i32.const $push4=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push5=, $0, $pop4
+; NO-SIMD128-FAST-NEXT:    i32.const $push7=, 5
+; NO-SIMD128-FAST-NEXT:    i32.shl $push6=, $4, $pop7
+; NO-SIMD128-FAST-NEXT:    i32.store 0($pop5), $pop6
+; NO-SIMD128-FAST-NEXT:    return
   %a = shl <4 x i32> %v, <i32 5, i32 5, i32 5, i32 5>
   ret <4 x i32> %a
 }
 
-; CHECK-LABEL: shl_vec_v4i32:
-; NO-SIMD128-NOT: i32x4
-; SIMD128-NEXT: .functype shl_vec_v4i32 (v128, v128) -> (v128){{$}}
-; SIMD128-NEXT: i32x4.extract_lane $push[[L0:[0-9]+]]=, $0, 0{{$}}
-; SIMD128-NEXT: i32x4.extract_lane $push[[L1:[0-9]+]]=, $1, 0{{$}}
-; SIMD128-NEXT: i32.shl $push[[L2:[0-9]+]]=, $pop[[L0]], $pop[[L1]]{{$}}
-; SIMD128-NEXT: i32x4.splat $push[[L3:[0-9]+]]=, $pop[[L2]]{{$}}
-; Skip 2 lanes
-; SIMD128:      i32x4.extract_lane $push[[L4:[0-9]+]]=, $0, 3{{$}}
-; SIMD128-NEXT: i32x4.extract_lane $push[[L5:[0-9]+]]=, $1, 3{{$}}
-; SIMD128-NEXT: i32.shl $push[[L6:[0-9]+]]=, $pop[[L4]], $pop[[L5]]{{$}}
-; SIMD128-NEXT: i32x4.replace_lane $push[[R:[0-9]+]]=, $pop[[L7:[0-9]+]], 3, $pop[[L6]]{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <4 x i32> @shl_vec_v4i32(<4 x i32> %v, <4 x i32> %x) {
+; SIMD128-SLOW-LABEL: shl_vec_v4i32:
+; SIMD128-SLOW:         .functype shl_vec_v4i32 (v128, v128) -> (v128)
+; SIMD128-SLOW-NEXT:    i32x4.extract_lane $push4=, $0, 0
+; SIMD128-SLOW-NEXT:    i32x4.extract_lane $push3=, $1, 0
+; SIMD128-SLOW-NEXT:    i32.shl $push5=, $pop4, $pop3
+; SIMD128-SLOW-NEXT:    i32x4.splat $push6=, $pop5
+; SIMD128-SLOW-NEXT:    i32x4.extract_lane $push1=, $0, 1
+; SIMD128-SLOW-NEXT:    i32x4.extract_lane $push0=, $1, 1
+; SIMD128-SLOW-NEXT:    i32.shl $push2=, $pop1, $pop0
+; SIMD128-SLOW-NEXT:    i32x4.replace_lane $push7=, $pop6, 1, $pop2
+; SIMD128-SLOW-NEXT:    i32x4.extract_lane $push9=, $0, 2
+; SIMD128-SLOW-NEXT:    i32x4.extract_lane $push8=, $1, 2
+; SIMD128-SLOW-NEXT:    i32.shl $push10=, $pop9, $pop8
+; SIMD128-SLOW-NEXT:    i32x4.replace_lane $push11=, $pop7, 2, $pop10
+; SIMD128-SLOW-NEXT:    i32x4.extract_lane $push13=, $0, 3
+; SIMD128-SLOW-NEXT:    i32x4.extract_lane $push12=, $1, 3
+; SIMD128-SLOW-NEXT:    i32.shl $push14=, $pop13, $pop12
+; SIMD128-SLOW-NEXT:    i32x4.replace_lane $push15=, $pop11, 3, $pop14
+; SIMD128-SLOW-NEXT:    return $pop15
+;
+; SIMD128-FAST-LABEL: shl_vec_v4i32:
+; SIMD128-FAST:         .functype shl_vec_v4i32 (v128, v128) -> (v128)
+; SIMD128-FAST-NEXT:    i32x4.extract_lane $push5=, $0, 0
+; SIMD128-FAST-NEXT:    i32x4.extract_lane $push4=, $1, 0
+; SIMD128-FAST-NEXT:    i32.shl $push6=, $pop5, $pop4
+; SIMD128-FAST-NEXT:    i32x4.splat $push7=, $pop6
+; SIMD128-FAST-NEXT:    i32x4.extract_lane $push2=, $0, 1
+; SIMD128-FAST-NEXT:    i32x4.extract_lane $push1=, $1, 1
+; SIMD128-FAST-NEXT:    i32.shl $push3=, $pop2, $pop1
+; SIMD128-FAST-NEXT:    i32x4.replace_lane $push8=, $pop7, 1, $pop3
+; SIMD128-FAST-NEXT:    i32x4.extract_lane $push10=, $0, 2
+; SIMD128-FAST-NEXT:    i32x4.extract_lane $push9=, $1, 2
+; SIMD128-FAST-NEXT:    i32.shl $push11=, $pop10, $pop9
+; SIMD128-FAST-NEXT:    i32x4.replace_lane $push12=, $pop8, 2, $pop11
+; SIMD128-FAST-NEXT:    i32x4.extract_lane $push14=, $0, 3
+; SIMD128-FAST-NEXT:    i32x4.extract_lane $push13=, $1, 3
+; SIMD128-FAST-NEXT:    i32.shl $push15=, $pop14, $pop13
+; SIMD128-FAST-NEXT:    i32x4.replace_lane $push0=, $pop12, 3, $pop15
+; SIMD128-FAST-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: shl_vec_v4i32:
+; NO-SIMD128-SLOW:         .functype shl_vec_v4i32 (i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push1=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push2=, $0, $pop1
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push0=, $4, $8
+; NO-SIMD128-SLOW-NEXT:    i32.store 0($pop2), $pop0
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push3=, $3, $7
+; NO-SIMD128-SLOW-NEXT:    i32.store 8($0), $pop3
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push4=, $2, $6
+; NO-SIMD128-SLOW-NEXT:    i32.store 4($0), $pop4
+; NO-SIMD128-SLOW-NEXT:    i32.shl $push5=, $1, $5
+; NO-SIMD128-SLOW-NEXT:    i32.store 0($0), $pop5
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: shl_vec_v4i32:
+; NO-SIMD128-FAST:         .functype shl_vec_v4i32 (i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-FAST-NEXT:    i32.shl $push0=, $1, $5
+; NO-SIMD128-FAST-NEXT:    i32.store 0($0), $pop0
+; NO-SIMD128-FAST-NEXT:    i32.shl $push1=, $2, $6
+; NO-SIMD128-FAST-NEXT:    i32.store 4($0), $pop1
+; NO-SIMD128-FAST-NEXT:    i32.shl $push2=, $3, $7
+; NO-SIMD128-FAST-NEXT:    i32.store 8($0), $pop2
+; NO-SIMD128-FAST-NEXT:    i32.const $push3=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push4=, $0, $pop3
+; NO-SIMD128-FAST-NEXT:    i32.shl $push5=, $4, $8
+; NO-SIMD128-FAST-NEXT:    i32.store 0($pop4), $pop5
+; NO-SIMD128-FAST-NEXT:    return
   %a = shl <4 x i32> %v, %x
   ret <4 x i32> %a
 }
 
-; CHECK-LABEL: shr_s_v4i32:
-; NO-SIMD128-NOT: i32x4
-; SIMD128-NEXT: .functype shr_s_v4i32 (v128, i32) -> (v128){{$}}
-; SIMD128-NEXT: i32x4.shr_s $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <4 x i32> @shr_s_v4i32(<4 x i32> %v, i32 %x) {
+; SIMD128-UNIMPL-LABEL: shr_s_v4i32:
+; SIMD128-UNIMPL:         .functype shr_s_v4i32 (v128, i32) -> (v128)
+; SIMD128-UNIMPL-NEXT:    i32x4.shr_s $push0=, $0, $1
+; SIMD128-UNIMPL-NEXT:    return $pop0
+;
+; SIMD128-VM-SLOW-LABEL: shr_s_v4i32:
+; SIMD128-VM-SLOW:         .functype shr_s_v4i32 (v128, i32) -> (v128)
+; SIMD128-VM-SLOW-NEXT:    i32x4.extract_lane $push2=, $0, 0
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push3=, $pop2, $1
+; SIMD128-VM-SLOW-NEXT:    i32x4.splat $push4=, $pop3
+; SIMD128-VM-SLOW-NEXT:    i32x4.extract_lane $push0=, $0, 1
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push1=, $pop0, $1
+; SIMD128-VM-SLOW-NEXT:    i32x4.replace_lane $push5=, $pop4, 1, $pop1
+; SIMD128-VM-SLOW-NEXT:    i32x4.extract_lane $push6=, $0, 2
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push7=, $pop6, $1
+; SIMD128-VM-SLOW-NEXT:    i32x4.replace_lane $push8=, $pop5, 2, $pop7
+; SIMD128-VM-SLOW-NEXT:    i32x4.extract_lane $push9=, $0, 3
+; SIMD128-VM-SLOW-NEXT:    i32.shr_s $push10=, $pop9, $1
+; SIMD128-VM-SLOW-NEXT:    i32x4.replace_lane $push11=, $pop8, 3, $pop10
+; SIMD128-VM-SLOW-NEXT:    return $pop11
+;
+; SIMD128-VM-FAST-LABEL: shr_s_v4i32:
+; SIMD128-VM-FAST:         .functype shr_s_v4i32 (v128, i32) -> (v128)
+; SIMD128-VM-FAST-NEXT:    i32x4.extract_lane $push3=, $0, 0
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push4=, $pop3, $1
+; SIMD128-VM-FAST-NEXT:    i32x4.splat $push5=, $pop4
+; SIMD128-VM-FAST-NEXT:    i32x4.extract_lane $push1=, $0, 1
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push2=, $pop1, $1
+; SIMD128-VM-FAST-NEXT:    i32x4.replace_lane $push6=, $pop5, 1, $pop2
+; SIMD128-VM-FAST-NEXT:    i32x4.extract_lane $push7=, $0, 2
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push8=, $pop7, $1
+; SIMD128-VM-FAST-NEXT:    i32x4.replace_lane $push9=, $pop6, 2, $pop8
+; SIMD128-VM-FAST-NEXT:    i32x4.extract_lane $push10=, $0, 3
+; SIMD128-VM-FAST-NEXT:    i32.shr_s $push11=, $pop10, $1
+; SIMD128-VM-FAST-NEXT:    i32x4.replace_lane $push0=, $pop9, 3, $pop11
+; SIMD128-VM-FAST-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: shr_s_v4i32:
+; NO-SIMD128-SLOW:         .functype shr_s_v4i32 (i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push1=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push2=, $0, $pop1
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push0=, $4, $5
+; NO-SIMD128-SLOW-NEXT:    i32.store 0($pop2), $pop0
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push3=, $3, $5
+; NO-SIMD128-SLOW-NEXT:    i32.store 8($0), $pop3
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push4=, $2, $5
+; NO-SIMD128-SLOW-NEXT:    i32.store 4($0), $pop4
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push5=, $1, $5
+; NO-SIMD128-SLOW-NEXT:    i32.store 0($0), $pop5
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: shr_s_v4i32:
+; NO-SIMD128-FAST:         .functype shr_s_v4i32 (i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push0=, $1, $5
+; NO-SIMD128-FAST-NEXT:    i32.store 0($0), $pop0
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push1=, $2, $5
+; NO-SIMD128-FAST-NEXT:    i32.store 4($0), $pop1
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push2=, $3, $5
+; NO-SIMD128-FAST-NEXT:    i32.store 8($0), $pop2
+; NO-SIMD128-FAST-NEXT:    i32.const $push3=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push4=, $0, $pop3
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push5=, $4, $5
+; NO-SIMD128-FAST-NEXT:    i32.store 0($pop4), $pop5
+; NO-SIMD128-FAST-NEXT:    return
   %t = insertelement <4 x i32> undef, i32 %x, i32 0
   %s = shufflevector <4 x i32> %t, <4 x i32> undef,
     <4 x i32> <i32 0, i32 0, i32 0, i32 0>
   %a = ashr <4 x i32> %v, %s
   ret <4 x i32> %a
 }
 
-; CHECK-LABEL: shr_s_vec_v4i32:
-; NO-SIMD128-NOT: i32x4
-; SIMD128-NEXT: .functype shr_s_vec_v4i32 (v128, v128) -> (v128){{$}}
-; SIMD128-NEXT: i32x4.extract_lane $push[[L0:[0-9]+]]=, $0, 0{{$}}
-; SIMD128-NEXT: i32x4.extract_lane $push[[L1:[0-9]+]]=, $1, 0{{$}}
-; SIMD128-NEXT: i32.shr_s $push[[L2:[0-9]+]]=, $pop[[L0]], $pop[[L1]]{{$}}
-; SIMD128-NEXT: i32x4.splat $push[[L3:[0-9]+]]=, $pop[[L2]]{{$}}
-; Skip 2 lanes
-; SIMD128:      i32x4.extract_lane $push[[L4:[0-9]+]]=, $0, 3{{$}}
-; SIMD128-NEXT: i32x4.extract_lane $push[[L5:[0-9]+]]=, $1, 3{{$}}
-; SIMD128-NEXT: i32.shr_s $push[[L6:[0-9]+]]=, $pop[[L4]], $pop[[L5]]{{$}}
-; SIMD128-NEXT: i32x4.replace_lane $push[[R:[0-9]+]]=, $pop[[L7:[0-9]+]], 3, $pop[[L6]]{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <4 x i32> @shr_s_vec_v4i32(<4 x i32> %v, <4 x i32> %x) {
+; SIMD128-SLOW-LABEL: shr_s_vec_v4i32:
+; SIMD128-SLOW:         .functype shr_s_vec_v4i32 (v128, v128) -> (v128)
+; SIMD128-SLOW-NEXT:    i32x4.extract_lane $push4=, $0, 0
+; SIMD128-SLOW-NEXT:    i32x4.extract_lane $push3=, $1, 0
+; SIMD128-SLOW-NEXT:    i32.shr_s $push5=, $pop4, $pop3
+; SIMD128-SLOW-NEXT:    i32x4.splat $push6=, $pop5
+; SIMD128-SLOW-NEXT:    i32x4.extract_lane $push1=, $0, 1
+; SIMD128-SLOW-NEXT:    i32x4.extract_lane $push0=, $1, 1
+; SIMD128-SLOW-NEXT:    i32.shr_s $push2=, $pop1, $pop0
+; SIMD128-SLOW-NEXT:    i32x4.replace_lane $push7=, $pop6, 1, $pop2
+; SIMD128-SLOW-NEXT:    i32x4.extract_lane $push9=, $0, 2
+; SIMD128-SLOW-NEXT:    i32x4.extract_lane $push8=, $1, 2
+; SIMD128-SLOW-NEXT:    i32.shr_s $push10=, $pop9, $pop8
+; SIMD128-SLOW-NEXT:    i32x4.replace_lane $push11=, $pop7, 2, $pop10
+; SIMD128-SLOW-NEXT:    i32x4.extract_lane $push13=, $0, 3
+; SIMD128-SLOW-NEXT:    i32x4.extract_lane $push12=, $1, 3
+; SIMD128-SLOW-NEXT:    i32.shr_s $push14=, $pop13, $pop12
+; SIMD128-SLOW-NEXT:    i32x4.replace_lane $push15=, $pop11, 3, $pop14
+; SIMD128-SLOW-NEXT:    return $pop15
+;
+; SIMD128-FAST-LABEL: shr_s_vec_v4i32:
+; SIMD128-FAST:         .functype shr_s_vec_v4i32 (v128, v128) -> (v128)
+; SIMD128-FAST-NEXT:    i32x4.extract_lane $push5=, $0, 0
+; SIMD128-FAST-NEXT:    i32x4.extract_lane $push4=, $1, 0
+; SIMD128-FAST-NEXT:    i32.shr_s $push6=, $pop5, $pop4
+; SIMD128-FAST-NEXT:    i32x4.splat $push7=, $pop6
+; SIMD128-FAST-NEXT:    i32x4.extract_lane $push2=, $0, 1
+; SIMD128-FAST-NEXT:    i32x4.extract_lane $push1=, $1, 1
+; SIMD128-FAST-NEXT:    i32.shr_s $push3=, $pop2, $pop1
+; SIMD128-FAST-NEXT:    i32x4.replace_lane $push8=, $pop7, 1, $pop3
+; SIMD128-FAST-NEXT:    i32x4.extract_lane $push10=, $0, 2
+; SIMD128-FAST-NEXT:    i32x4.extract_lane $push9=, $1, 2
+; SIMD128-FAST-NEXT:    i32.shr_s $push11=, $pop10, $pop9
+; SIMD128-FAST-NEXT:    i32x4.replace_lane $push12=, $pop8, 2, $pop11
+; SIMD128-FAST-NEXT:    i32x4.extract_lane $push14=, $0, 3
+; SIMD128-FAST-NEXT:    i32x4.extract_lane $push13=, $1, 3
+; SIMD128-FAST-NEXT:    i32.shr_s $push15=, $pop14, $pop13
+; SIMD128-FAST-NEXT:    i32x4.replace_lane $push0=, $pop12, 3, $pop15
+; SIMD128-FAST-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: shr_s_vec_v4i32:
+; NO-SIMD128-SLOW:         .functype shr_s_vec_v4i32 (i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push1=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push2=, $0, $pop1
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push0=, $4, $8
+; NO-SIMD128-SLOW-NEXT:    i32.store 0($pop2), $pop0
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push3=, $3, $7
+; NO-SIMD128-SLOW-NEXT:    i32.store 8($0), $pop3
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push4=, $2, $6
+; NO-SIMD128-SLOW-NEXT:    i32.store 4($0), $pop4
+; NO-SIMD128-SLOW-NEXT:    i32.shr_s $push5=, $1, $5
+; NO-SIMD128-SLOW-NEXT:    i32.store 0($0), $pop5
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: shr_s_vec_v4i32:
+; NO-SIMD128-FAST:         .functype shr_s_vec_v4i32 (i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push0=, $1, $5
+; NO-SIMD128-FAST-NEXT:    i32.store 0($0), $pop0
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push1=, $2, $6
+; NO-SIMD128-FAST-NEXT:    i32.store 4($0), $pop1
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push2=, $3, $7
+; NO-SIMD128-FAST-NEXT:    i32.store 8($0), $pop2
+; NO-SIMD128-FAST-NEXT:    i32.const $push3=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push4=, $0, $pop3
+; NO-SIMD128-FAST-NEXT:    i32.shr_s $push5=, $4, $8
+; NO-SIMD128-FAST-NEXT:    i32.store 0($pop4), $pop5
+; NO-SIMD128-FAST-NEXT:    return
   %a = ashr <4 x i32> %v, %x
   ret <4 x i32> %a
 }
 
-; CHECK-LABEL: shr_u_v4i32:
-; NO-SIMD128-NOT: i32x4
-; SIMD128-NEXT: .functype shr_u_v4i32 (v128, i32) -> (v128){{$}}
-; SIMD128-NEXT: i32x4.shr_u $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <4 x i32> @shr_u_v4i32(<4 x i32> %v, i32 %x) {
+; SIMD128-UNIMPL-LABEL: shr_u_v4i32:
+; SIMD128-UNIMPL:         .functype shr_u_v4i32 (v128, i32) -> (v128)
+; SIMD128-UNIMPL-NEXT:    i32x4.shr_u $push0=, $0, $1
+; SIMD128-UNIMPL-NEXT:    return $pop0
+;
+; SIMD128-VM-SLOW-LABEL: shr_u_v4i32:
+; SIMD128-VM-SLOW:         .functype shr_u_v4i32 (v128, i32) -> (v128)
+; SIMD128-VM-SLOW-NEXT:    i32x4.extract_lane $push2=, $0, 0
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push3=, $pop2, $1
+; SIMD128-VM-SLOW-NEXT:    i32x4.splat $push4=, $pop3
+; SIMD128-VM-SLOW-NEXT:    i32x4.extract_lane $push0=, $0, 1
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push1=, $pop0, $1
+; SIMD128-VM-SLOW-NEXT:    i32x4.replace_lane $push5=, $pop4, 1, $pop1
+; SIMD128-VM-SLOW-NEXT:    i32x4.extract_lane $push6=, $0, 2
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push7=, $pop6, $1
+; SIMD128-VM-SLOW-NEXT:    i32x4.replace_lane $push8=, $pop5, 2, $pop7
+; SIMD128-VM-SLOW-NEXT:    i32x4.extract_lane $push9=, $0, 3
+; SIMD128-VM-SLOW-NEXT:    i32.shr_u $push10=, $pop9, $1
+; SIMD128-VM-SLOW-NEXT:    i32x4.replace_lane $push11=, $pop8, 3, $pop10
+; SIMD128-VM-SLOW-NEXT:    return $pop11
+;
+; SIMD128-VM-FAST-LABEL: shr_u_v4i32:
+; SIMD128-VM-FAST:         .functype shr_u_v4i32 (v128, i32) -> (v128)
+; SIMD128-VM-FAST-NEXT:    i32x4.extract_lane $push3=, $0, 0
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push4=, $pop3, $1
+; SIMD128-VM-FAST-NEXT:    i32x4.splat $push5=, $pop4
+; SIMD128-VM-FAST-NEXT:    i32x4.extract_lane $push1=, $0, 1
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push2=, $pop1, $1
+; SIMD128-VM-FAST-NEXT:    i32x4.replace_lane $push6=, $pop5, 1, $pop2
+; SIMD128-VM-FAST-NEXT:    i32x4.extract_lane $push7=, $0, 2
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push8=, $pop7, $1
+; SIMD128-VM-FAST-NEXT:    i32x4.replace_lane $push9=, $pop6, 2, $pop8
+; SIMD128-VM-FAST-NEXT:    i32x4.extract_lane $push10=, $0, 3
+; SIMD128-VM-FAST-NEXT:    i32.shr_u $push11=, $pop10, $1
+; SIMD128-VM-FAST-NEXT:    i32x4.replace_lane $push0=, $pop9, 3, $pop11
+; SIMD128-VM-FAST-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: shr_u_v4i32:
+; NO-SIMD128-SLOW:         .functype shr_u_v4i32 (i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push1=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push2=, $0, $pop1
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push0=, $4, $5
+; NO-SIMD128-SLOW-NEXT:    i32.store 0($pop2), $pop0
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push3=, $3, $5
+; NO-SIMD128-SLOW-NEXT:    i32.store 8($0), $pop3
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push4=, $2, $5
+; NO-SIMD128-SLOW-NEXT:    i32.store 4($0), $pop4
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push5=, $1, $5
+; NO-SIMD128-SLOW-NEXT:    i32.store 0($0), $pop5
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: shr_u_v4i32:
+; NO-SIMD128-FAST:         .functype shr_u_v4i32 (i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push0=, $1, $5
+; NO-SIMD128-FAST-NEXT:    i32.store 0($0), $pop0
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push1=, $2, $5
+; NO-SIMD128-FAST-NEXT:    i32.store 4($0), $pop1
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push2=, $3, $5
+; NO-SIMD128-FAST-NEXT:    i32.store 8($0), $pop2
+; NO-SIMD128-FAST-NEXT:    i32.const $push3=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push4=, $0, $pop3
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push5=, $4, $5
+; NO-SIMD128-FAST-NEXT:    i32.store 0($pop4), $pop5
+; NO-SIMD128-FAST-NEXT:    return
   %t = insertelement <4 x i32> undef, i32 %x, i32 0
   %s = shufflevector <4 x i32> %t, <4 x i32> undef,
     <4 x i32> <i32 0, i32 0, i32 0, i32 0>
   %a = lshr <4 x i32> %v, %s
   ret <4 x i32> %a
 }
 
-; CHECK-LABEL: shr_u_vec_v4i32:
-; NO-SIMD128-NOT: i32x4
-; SIMD128-NEXT: .functype shr_u_vec_v4i32 (v128, v128) -> (v128){{$}}
-; SIMD128-NEXT: i32x4.extract_lane $push[[L0:[0-9]+]]=, $0, 0{{$}}
-; SIMD128-NEXT: i32x4.extract_lane $push[[L1:[0-9]+]]=, $1, 0{{$}}
-; SIMD128-NEXT: i32.shr_u $push[[L2:[0-9]+]]=, $pop[[L0]], $pop[[L1]]{{$}}
-; SIMD128-NEXT: i32x4.splat $push[[L3:[0-9]+]]=, $pop[[L2]]{{$}}
-; Skip 2 lanes
-; SIMD128:      i32x4.extract_lane $push[[L4:[0-9]+]]=, $0, 3{{$}}
-; SIMD128-NEXT: i32x4.extract_lane $push[[L5:[0-9]+]]=, $1, 3{{$}}
-; SIMD128-NEXT: i32.shr_u $push[[L6:[0-9]+]]=, $pop[[L4]], $pop[[L5]]{{$}}
-; SIMD128-NEXT: i32x4.replace_lane $push[[R:[0-9]+]]=, $pop[[L7:[0-9]+]], 3, $pop[[L6]]{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <4 x i32> @shr_u_vec_v4i32(<4 x i32> %v, <4 x i32> %x) {
+; SIMD128-SLOW-LABEL: shr_u_vec_v4i32:
+; SIMD128-SLOW:         .functype shr_u_vec_v4i32 (v128, v128) -> (v128)
+; SIMD128-SLOW-NEXT:    i32x4.extract_lane $push4=, $0, 0
+; SIMD128-SLOW-NEXT:    i32x4.extract_lane $push3=, $1, 0
+; SIMD128-SLOW-NEXT:    i32.shr_u $push5=, $pop4, $pop3
+; SIMD128-SLOW-NEXT:    i32x4.splat $push6=, $pop5
+; SIMD128-SLOW-NEXT:    i32x4.extract_lane $push1=, $0, 1
+; SIMD128-SLOW-NEXT:    i32x4.extract_lane $push0=, $1, 1
+; SIMD128-SLOW-NEXT:    i32.shr_u $push2=, $pop1, $pop0
+; SIMD128-SLOW-NEXT:    i32x4.replace_lane $push7=, $pop6, 1, $pop2
+; SIMD128-SLOW-NEXT:    i32x4.extract_lane $push9=, $0, 2
+; SIMD128-SLOW-NEXT:    i32x4.extract_lane $push8=, $1, 2
+; SIMD128-SLOW-NEXT:    i32.shr_u $push10=, $pop9, $pop8
+; SIMD128-SLOW-NEXT:    i32x4.replace_lane $push11=, $pop7, 2, $pop10
+; SIMD128-SLOW-NEXT:    i32x4.extract_lane $push13=, $0, 3
+; SIMD128-SLOW-NEXT:    i32x4.extract_lane $push12=, $1, 3
+; SIMD128-SLOW-NEXT:    i32.shr_u $push14=, $pop13, $pop12
+; SIMD128-SLOW-NEXT:    i32x4.replace_lane $push15=, $pop11, 3, $pop14
+; SIMD128-SLOW-NEXT:    return $pop15
+;
+; SIMD128-FAST-LABEL: shr_u_vec_v4i32:
+; SIMD128-FAST:         .functype shr_u_vec_v4i32 (v128, v128) -> (v128)
+; SIMD128-FAST-NEXT:    i32x4.extract_lane $push5=, $0, 0
+; SIMD128-FAST-NEXT:    i32x4.extract_lane $push4=, $1, 0
+; SIMD128-FAST-NEXT:    i32.shr_u $push6=, $pop5, $pop4
+; SIMD128-FAST-NEXT:    i32x4.splat $push7=, $pop6
+; SIMD128-FAST-NEXT:    i32x4.extract_lane $push2=, $0, 1
+; SIMD128-FAST-NEXT:    i32x4.extract_lane $push1=, $1, 1
+; SIMD128-FAST-NEXT:    i32.shr_u $push3=, $pop2, $pop1
+; SIMD128-FAST-NEXT:    i32x4.replace_lane $push8=, $pop7, 1, $pop3
+; SIMD128-FAST-NEXT:    i32x4.extract_lane $push10=, $0, 2
+; SIMD128-FAST-NEXT:    i32x4.extract_lane $push9=, $1, 2
+; SIMD128-FAST-NEXT:    i32.shr_u $push11=, $pop10, $pop9
+; SIMD128-FAST-NEXT:    i32x4.replace_lane $push12=, $pop8, 2, $pop11
+; SIMD128-FAST-NEXT:    i32x4.extract_lane $push14=, $0, 3
+; SIMD128-FAST-NEXT:    i32x4.extract_lane $push13=, $1, 3
+; SIMD128-FAST-NEXT:    i32.shr_u $push15=, $pop14, $pop13
+; SIMD128-FAST-NEXT:    i32x4.replace_lane $push0=, $pop12, 3, $pop15
+; SIMD128-FAST-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: shr_u_vec_v4i32:
+; NO-SIMD128-SLOW:         .functype shr_u_vec_v4i32 (i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push1=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push2=, $0, $pop1
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push0=, $4, $8
+; NO-SIMD128-SLOW-NEXT:    i32.store 0($pop2), $pop0
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push3=, $3, $7
+; NO-SIMD128-SLOW-NEXT:    i32.store 8($0), $pop3
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push4=, $2, $6
+; NO-SIMD128-SLOW-NEXT:    i32.store 4($0), $pop4
+; NO-SIMD128-SLOW-NEXT:    i32.shr_u $push5=, $1, $5
+; NO-SIMD128-SLOW-NEXT:    i32.store 0($0), $pop5
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: shr_u_vec_v4i32:
+; NO-SIMD128-FAST:         .functype shr_u_vec_v4i32 (i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push0=, $1, $5
+; NO-SIMD128-FAST-NEXT:    i32.store 0($0), $pop0
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push1=, $2, $6
+; NO-SIMD128-FAST-NEXT:    i32.store 4($0), $pop1
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push2=, $3, $7
+; NO-SIMD128-FAST-NEXT:    i32.store 8($0), $pop2
+; NO-SIMD128-FAST-NEXT:    i32.const $push3=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push4=, $0, $pop3
+; NO-SIMD128-FAST-NEXT:    i32.shr_u $push5=, $4, $8
+; NO-SIMD128-FAST-NEXT:    i32.store 0($pop4), $pop5
+; NO-SIMD128-FAST-NEXT:    return
   %a = lshr <4 x i32> %v, %x
   ret <4 x i32> %a
 }
 
-; CHECK-LABEL: and_v4i32:
-; NO-SIMD128-NOT: v128
-; SIMD128-NEXT: .functype and_v4i32 (v128, v128) -> (v128){{$}}
-; SIMD128-NEXT: v128.and $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <4 x i32> @and_v4i32(<4 x i32> %x, <4 x i32> %y) {
+; SIMD128-LABEL: and_v4i32:
+; SIMD128:         .functype and_v4i32 (v128, v128) -> (v128)
+; SIMD128-NEXT:    v128.and $push0=, $0, $1
+; SIMD128-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: and_v4i32:
+; NO-SIMD128-SLOW:         .functype and_v4i32 (i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push1=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push2=, $0, $pop1
+; NO-SIMD128-SLOW-NEXT:    i32.and $push0=, $4, $8
+; NO-SIMD128-SLOW-NEXT:    i32.store 0($pop2), $pop0
+; NO-SIMD128-SLOW-NEXT:    i32.and $push3=, $3, $7
+; NO-SIMD128-SLOW-NEXT:    i32.store 8($0), $pop3
+; NO-SIMD128-SLOW-NEXT:    i32.and $push4=, $2, $6
+; NO-SIMD128-SLOW-NEXT:    i32.store 4($0), $pop4
+; NO-SIMD128-SLOW-NEXT:    i32.and $push5=, $1, $5
+; NO-SIMD128-SLOW-NEXT:    i32.store 0($0), $pop5
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: and_v4i32:
+; NO-SIMD128-FAST:         .functype and_v4i32 (i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-FAST-NEXT:    i32.and $push0=, $1, $5
+; NO-SIMD128-FAST-NEXT:    i32.store 0($0), $pop0
+; NO-SIMD128-FAST-NEXT:    i32.and $push1=, $2, $6
+; NO-SIMD128-FAST-NEXT:    i32.store 4($0), $pop1
+; NO-SIMD128-FAST-NEXT:    i32.and $push2=, $3, $7
+; NO-SIMD128-FAST-NEXT:    i32.store 8($0), $pop2
+; NO-SIMD128-FAST-NEXT:    i32.const $push3=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push4=, $0, $pop3
+; NO-SIMD128-FAST-NEXT:    i32.and $push5=, $4, $8
+; NO-SIMD128-FAST-NEXT:    i32.store 0($pop4), $pop5
+; NO-SIMD128-FAST-NEXT:    return
   %a = and <4 x i32> %x, %y
   ret <4 x i32> %a
 }
 
-; CHECK-LABEL: or_v4i32:
-; NO-SIMD128-NOT: v128
-; SIMD128-NEXT: .functype or_v4i32 (v128, v128) -> (v128){{$}}
-; SIMD128-NEXT: v128.or $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <4 x i32> @or_v4i32(<4 x i32> %x, <4 x i32> %y) {
+; SIMD128-LABEL: or_v4i32:
+; SIMD128:         .functype or_v4i32 (v128, v128) -> (v128)
+; SIMD128-NEXT:    v128.or $push0=, $0, $1
+; SIMD128-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: or_v4i32:
+; NO-SIMD128-SLOW:         .functype or_v4i32 (i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push1=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push2=, $0, $pop1
+; NO-SIMD128-SLOW-NEXT:    i32.or $push0=, $4, $8
+; NO-SIMD128-SLOW-NEXT:    i32.store 0($pop2), $pop0
+; NO-SIMD128-SLOW-NEXT:    i32.or $push3=, $3, $7
+; NO-SIMD128-SLOW-NEXT:    i32.store 8($0), $pop3
+; NO-SIMD128-SLOW-NEXT:    i32.or $push4=, $2, $6
+; NO-SIMD128-SLOW-NEXT:    i32.store 4($0), $pop4
+; NO-SIMD128-SLOW-NEXT:    i32.or $push5=, $1, $5
+; NO-SIMD128-SLOW-NEXT:    i32.store 0($0), $pop5
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: or_v4i32:
+; NO-SIMD128-FAST:         .functype or_v4i32 (i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-FAST-NEXT:    i32.or $push0=, $1, $5
+; NO-SIMD128-FAST-NEXT:    i32.store 0($0), $pop0
+; NO-SIMD128-FAST-NEXT:    i32.or $push1=, $2, $6
+; NO-SIMD128-FAST-NEXT:    i32.store 4($0), $pop1
+; NO-SIMD128-FAST-NEXT:    i32.or $push2=, $3, $7
+; NO-SIMD128-FAST-NEXT:    i32.store 8($0), $pop2
+; NO-SIMD128-FAST-NEXT:    i32.const $push3=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push4=, $0, $pop3
+; NO-SIMD128-FAST-NEXT:    i32.or $push5=, $4, $8
+; NO-SIMD128-FAST-NEXT:    i32.store 0($pop4), $pop5
+; NO-SIMD128-FAST-NEXT:    return
   %a = or <4 x i32> %x, %y
   ret <4 x i32> %a
 }
 
-; CHECK-LABEL: xor_v4i32:
-; NO-SIMD128-NOT: v128
-; SIMD128-NEXT: .functype xor_v4i32 (v128, v128) -> (v128){{$}}
-; SIMD128-NEXT: v128.xor $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <4 x i32> @xor_v4i32(<4 x i32> %x, <4 x i32> %y) {
+; SIMD128-LABEL: xor_v4i32:
+; SIMD128:         .functype xor_v4i32 (v128, v128) -> (v128)
+; SIMD128-NEXT:    v128.xor $push0=, $0, $1
+; SIMD128-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: xor_v4i32:
+; NO-SIMD128-SLOW:         .functype xor_v4i32 (i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push1=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push2=, $0, $pop1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push0=, $4, $8
+; NO-SIMD128-SLOW-NEXT:    i32.store 0($pop2), $pop0
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push3=, $3, $7
+; NO-SIMD128-SLOW-NEXT:    i32.store 8($0), $pop3
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push4=, $2, $6
+; NO-SIMD128-SLOW-NEXT:    i32.store 4($0), $pop4
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push5=, $1, $5
+; NO-SIMD128-SLOW-NEXT:    i32.store 0($0), $pop5
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: xor_v4i32:
+; NO-SIMD128-FAST:         .functype xor_v4i32 (i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-FAST-NEXT:    i32.xor $push0=, $1, $5
+; NO-SIMD128-FAST-NEXT:    i32.store 0($0), $pop0
+; NO-SIMD128-FAST-NEXT:    i32.xor $push1=, $2, $6
+; NO-SIMD128-FAST-NEXT:    i32.store 4($0), $pop1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push2=, $3, $7
+; NO-SIMD128-FAST-NEXT:    i32.store 8($0), $pop2
+; NO-SIMD128-FAST-NEXT:    i32.const $push3=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push4=, $0, $pop3
+; NO-SIMD128-FAST-NEXT:    i32.xor $push5=, $4, $8
+; NO-SIMD128-FAST-NEXT:    i32.store 0($pop4), $pop5
+; NO-SIMD128-FAST-NEXT:    return
   %a = xor <4 x i32> %x, %y
   ret <4 x i32> %a
 }
 
-; CHECK-LABEL: not_v4i32:
-; NO-SIMD128-NOT: v128
-; SIMD128-NEXT: .functype not_v4i32 (v128) -> (v128){{$}}
-; SIMD128-NEXT: v128.not $push[[R:[0-9]+]]=, $0{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <4 x i32> @not_v4i32(<4 x i32> %x) {
+; SIMD128-LABEL: not_v4i32:
+; SIMD128:         .functype not_v4i32 (v128) -> (v128)
+; SIMD128-NEXT:    v128.not $push0=, $0
+; SIMD128-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: not_v4i32:
+; NO-SIMD128-SLOW:         .functype not_v4i32 (i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push2=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push3=, $0, $pop2
+; NO-SIMD128-SLOW-NEXT:    i32.const $push0=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push1=, $4, $pop0
+; NO-SIMD128-SLOW-NEXT:    i32.store 0($pop3), $pop1
+; NO-SIMD128-SLOW-NEXT:    i32.const $push9=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push4=, $3, $pop9
+; NO-SIMD128-SLOW-NEXT:    i32.store 8($0), $pop4
+; NO-SIMD128-SLOW-NEXT:    i32.const $push8=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push5=, $2, $pop8
+; NO-SIMD128-SLOW-NEXT:    i32.store 4($0), $pop5
+; NO-SIMD128-SLOW-NEXT:    i32.const $push7=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push6=, $1, $pop7
+; NO-SIMD128-SLOW-NEXT:    i32.store 0($0), $pop6
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: not_v4i32:
+; NO-SIMD128-FAST:         .functype not_v4i32 (i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-FAST-NEXT:    i32.const $push0=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push1=, $1, $pop0
+; NO-SIMD128-FAST-NEXT:    i32.store 0($0), $pop1
+; NO-SIMD128-FAST-NEXT:    i32.const $push9=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push2=, $2, $pop9
+; NO-SIMD128-FAST-NEXT:    i32.store 4($0), $pop2
+; NO-SIMD128-FAST-NEXT:    i32.const $push8=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push3=, $3, $pop8
+; NO-SIMD128-FAST-NEXT:    i32.store 8($0), $pop3
+; NO-SIMD128-FAST-NEXT:    i32.const $push4=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push5=, $0, $pop4
+; NO-SIMD128-FAST-NEXT:    i32.const $push7=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push6=, $4, $pop7
+; NO-SIMD128-FAST-NEXT:    i32.store 0($pop5), $pop6
+; NO-SIMD128-FAST-NEXT:    return
   %a = xor <4 x i32> %x, <i32 -1, i32 -1, i32 -1, i32 -1>
   ret <4 x i32> %a
 }
 
-; CHECK-LABEL: bitselect_v4i32:
-; NO-SIMD128-NOT: v128
-; SIMD128-NEXT: .functype bitselect_v4i32 (v128, v128, v128) -> (v128){{$}}
-; SIMD128-SLOW-NEXT: v128.bitselect $push[[R:[0-9]+]]=, $1, $2, $0{{$}}
-; SIMD128-SLOW-NEXT: return $pop[[R]]{{$}}
-; SIMD128-FAST-NEXT: v128.not
-; SIMD128-FAST-NEXT: v128.and
-; SIMD128-FAST-NEXT: v128.and
-; SIMD128-FAST-NEXT: v128.or
-; SIMD128-FAST-NEXT: return
 define <4 x i32> @bitselect_v4i32(<4 x i32> %c, <4 x i32> %v1, <4 x i32> %v2) {
+; SIMD128-SLOW-LABEL: bitselect_v4i32:
+; SIMD128-SLOW:         .functype bitselect_v4i32 (v128, v128, v128) -> (v128)
+; SIMD128-SLOW-NEXT:    v128.bitselect $push0=, $1, $2, $0
+; SIMD128-SLOW-NEXT:    return $pop0
+;
+; SIMD128-FAST-LABEL: bitselect_v4i32:
+; SIMD128-FAST:         .functype bitselect_v4i32 (v128, v128, v128) -> (v128)
+; SIMD128-FAST-NEXT:    v128.not $push2=, $0
+; SIMD128-FAST-NEXT:    v128.and $push3=, $pop2, $2
+; SIMD128-FAST-NEXT:    v128.and $push0=, $0, $1
+; SIMD128-FAST-NEXT:    v128.or $push1=, $pop3, $pop0
+; SIMD128-FAST-NEXT:    return $pop1
+;
+; NO-SIMD128-SLOW-LABEL: bitselect_v4i32:
+; NO-SIMD128-SLOW:         .functype bitselect_v4i32 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push5=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push6=, $0, $pop5
+; NO-SIMD128-SLOW-NEXT:    i32.const $push1=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push2=, $4, $pop1
+; NO-SIMD128-SLOW-NEXT:    i32.and $push3=, $pop2, $12
+; NO-SIMD128-SLOW-NEXT:    i32.and $push0=, $4, $8
+; NO-SIMD128-SLOW-NEXT:    i32.or $push4=, $pop3, $pop0
+; NO-SIMD128-SLOW-NEXT:    i32.store 0($pop6), $pop4
+; NO-SIMD128-SLOW-NEXT:    i32.const $push21=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push8=, $3, $pop21
+; NO-SIMD128-SLOW-NEXT:    i32.and $push9=, $pop8, $11
+; NO-SIMD128-SLOW-NEXT:    i32.and $push7=, $3, $7
+; NO-SIMD128-SLOW-NEXT:    i32.or $push10=, $pop9, $pop7
+; NO-SIMD128-SLOW-NEXT:    i32.store 8($0), $pop10
+; NO-SIMD128-SLOW-NEXT:    i32.const $push20=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push12=, $2, $pop20
+; NO-SIMD128-SLOW-NEXT:    i32.and $push13=, $pop12, $10
+; NO-SIMD128-SLOW-NEXT:    i32.and $push11=, $2, $6
+; NO-SIMD128-SLOW-NEXT:    i32.or $push14=, $pop13, $pop11
+; NO-SIMD128-SLOW-NEXT:    i32.store 4($0), $pop14
+; NO-SIMD128-SLOW-NEXT:    i32.const $push19=, -1
+; NO-SIMD128-SLOW-NEXT:    i32.xor $push16=, $1, $pop19
+; NO-SIMD128-SLOW-NEXT:    i32.and $push17=, $pop16, $9
+; NO-SIMD128-SLOW-NEXT:    i32.and $push15=, $1, $5
+; NO-SIMD128-SLOW-NEXT:    i32.or $push18=, $pop17, $pop15
+; NO-SIMD128-SLOW-NEXT:    i32.store 0($0), $pop18
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: bitselect_v4i32:
+; NO-SIMD128-FAST:         .functype bitselect_v4i32 (i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32) -> ()
+; NO-SIMD128-FAST-NEXT:    i32.const $push1=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push2=, $1, $pop1
+; NO-SIMD128-FAST-NEXT:    i32.and $push3=, $pop2, $9
+; NO-SIMD128-FAST-NEXT:    i32.and $push0=, $1, $5
+; NO-SIMD128-FAST-NEXT:    i32.or $push4=, $pop3, $pop0
+; NO-SIMD128-FAST-NEXT:    i32.store 0($0), $pop4
+; NO-SIMD128-FAST-NEXT:    i32.const $push21=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push6=, $2, $pop21
+; NO-SIMD128-FAST-NEXT:    i32.and $push7=, $pop6, $10
+; NO-SIMD128-FAST-NEXT:    i32.and $push5=, $2, $6
+; NO-SIMD128-FAST-NEXT:    i32.or $push8=, $pop7, $pop5
+; NO-SIMD128-FAST-NEXT:    i32.store 4($0), $pop8
+; NO-SIMD128-FAST-NEXT:    i32.const $push20=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push10=, $3, $pop20
+; NO-SIMD128-FAST-NEXT:    i32.and $push11=, $pop10, $11
+; NO-SIMD128-FAST-NEXT:    i32.and $push9=, $3, $7
+; NO-SIMD128-FAST-NEXT:    i32.or $push12=, $pop11, $pop9
+; NO-SIMD128-FAST-NEXT:    i32.store 8($0), $pop12
+; NO-SIMD128-FAST-NEXT:    i32.const $push17=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push18=, $0, $pop17
+; NO-SIMD128-FAST-NEXT:    i32.const $push19=, -1
+; NO-SIMD128-FAST-NEXT:    i32.xor $push14=, $4, $pop19
+; NO-SIMD128-FAST-NEXT:    i32.and $push15=, $pop14, $12
+; NO-SIMD128-FAST-NEXT:    i32.and $push13=, $4, $8
+; NO-SIMD128-FAST-NEXT:    i32.or $push16=, $pop15, $pop13
+; NO-SIMD128-FAST-NEXT:    i32.store 0($pop18), $pop16
+; NO-SIMD128-FAST-NEXT:    return
   %masked_v1 = and <4 x i32> %c, %v1
   %inv_mask = xor <4 x i32> <i32 -1, i32 -1, i32 -1, i32 -1>, %c
   %masked_v2 = and <4 x i32> %inv_mask, %v2
   %a = or <4 x i32> %masked_v2, %masked_v1
   ret <4 x i32> %a
 }
 
 ; ==============================================================================
 ; 2 x i64
 ; ==============================================================================
-; CHECK-LABEL: add_v2i64:
-; NO-SIMD128-NOT: i64x2
-; SIMD128-VM-NOT: i64x2
-; SIMD128-NEXT: .functype add_v2i64 (v128, v128) -> (v128){{$}}
-; SIMD128-NEXT: i64x2.add $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
+
 define <2 x i64> @add_v2i64(<2 x i64> %x, <2 x i64> %y) {
+; SIMD128-UNIMPL-LABEL: add_v2i64:
+; SIMD128-UNIMPL:         .functype add_v2i64 (v128, v128) -> (v128)
+; SIMD128-UNIMPL-NEXT:    i64x2.add $push0=, $0, $1
+; SIMD128-UNIMPL-NEXT:    return $pop0
+;
+; NO-SIMD64-SLOW-LABEL: add_v2i64:
+; NO-SIMD64-SLOW:         .functype add_v2i64 (i32, i64, i64, i64, i64) -> ()
+; NO-SIMD64-SLOW-NEXT:    i64.add $push0=, $2, $4
+; NO-SIMD64-SLOW-NEXT:    i64.store 8($0), $pop0
+; NO-SIMD64-SLOW-NEXT:    i64.add $push1=, $1, $3
+; NO-SIMD64-SLOW-NEXT:    i64.store 0($0), $pop1
+; NO-SIMD64-SLOW-NEXT:    return
+;
+; NO-SIMD64-FAST-LABEL: add_v2i64:
+; NO-SIMD64-FAST:         .functype add_v2i64 (i32, i64, i64, i64, i64) -> ()
+; NO-SIMD64-FAST-NEXT:    i64.add $push0=, $1, $3
+; NO-SIMD64-FAST-NEXT:    i64.store 0($0), $pop0
+; NO-SIMD64-FAST-NEXT:    i64.add $push1=, $2, $4
+; NO-SIMD64-FAST-NEXT:    i64.store 8($0), $pop1
+; NO-SIMD64-FAST-NEXT:    return
   %a = add <2 x i64> %x, %y
   ret <2 x i64> %a
 }
 
-; CHECK-LABEL: sub_v2i64:
-; NO-SIMD128-NOT: i64x2
-; SIMD128-VM-NOT: i64x2
-; SIMD128-NEXT: .functype sub_v2i64 (v128, v128) -> (v128){{$}}
-; SIMD128-NEXT: i64x2.sub $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <2 x i64> @sub_v2i64(<2 x i64> %x, <2 x i64> %y) {
+; SIMD128-UNIMPL-LABEL: sub_v2i64:
+; SIMD128-UNIMPL:         .functype sub_v2i64 (v128, v128) -> (v128)
+; SIMD128-UNIMPL-NEXT:    i64x2.sub $push0=, $0, $1
+; SIMD128-UNIMPL-NEXT:    return $pop0
+;
+; NO-SIMD64-SLOW-LABEL: sub_v2i64:
+; NO-SIMD64-SLOW:         .functype sub_v2i64 (i32, i64, i64, i64, i64) -> ()
+; NO-SIMD64-SLOW-NEXT:    i64.sub $push0=, $2, $4
+; NO-SIMD64-SLOW-NEXT:    i64.store 8($0), $pop0
+; NO-SIMD64-SLOW-NEXT:    i64.sub $push1=, $1, $3
+; NO-SIMD64-SLOW-NEXT:    i64.store 0($0), $pop1
+; NO-SIMD64-SLOW-NEXT:    return
+;
+; NO-SIMD64-FAST-LABEL: sub_v2i64:
+; NO-SIMD64-FAST:         .functype sub_v2i64 (i32, i64, i64, i64, i64) -> ()
+; NO-SIMD64-FAST-NEXT:    i64.sub $push0=, $1, $3
+; NO-SIMD64-FAST-NEXT:    i64.store 0($0), $pop0
+; NO-SIMD64-FAST-NEXT:    i64.sub $push1=, $2, $4
+; NO-SIMD64-FAST-NEXT:    i64.store 8($0), $pop1
+; NO-SIMD64-FAST-NEXT:    return
   %a = sub <2 x i64> %x, %y
   ret <2 x i64> %a
 }
 
-; v2i64.mul is not in spec
-; CHECK-LABEL: mul_v2i64:
-; NO-SIMD128-NOT: i64x2
-; SIMD128-VM-NOT: i64x2
-; SIMD128-NOT: i64x2.mul
-; SIMD128: i64x2.extract_lane
-; SIMD128: i64.mul
 define <2 x i64> @mul_v2i64(<2 x i64> %x, <2 x i64> %y) {
+; SIMD128-UNIMPL-SLOW-LABEL: mul_v2i64:
+; SIMD128-UNIMPL-SLOW:         .functype mul_v2i64 (v128, v128) -> (v128)
+; SIMD128-UNIMPL-SLOW-NEXT:    i64x2.extract_lane $push4=, $0, 0
+; SIMD128-UNIMPL-SLOW-NEXT:    i64x2.extract_lane $push3=, $1, 0
+; SIMD128-UNIMPL-SLOW-NEXT:    i64.mul $push5=, $pop4, $pop3
+; SIMD128-UNIMPL-SLOW-NEXT:    i64x2.splat $push6=, $pop5
+; SIMD128-UNIMPL-SLOW-NEXT:    i64x2.extract_lane $push1=, $0, 1
+; SIMD128-UNIMPL-SLOW-NEXT:    i64x2.extract_lane $push0=, $1, 1
+; SIMD128-UNIMPL-SLOW-NEXT:    i64.mul $push2=, $pop1, $pop0
+; SIMD128-UNIMPL-SLOW-NEXT:    i64x2.replace_lane $push7=, $pop6, 1, $pop2
+; SIMD128-UNIMPL-SLOW-NEXT:    return $pop7
+;
+; SIMD128-UNIMPL-FAST-LABEL: mul_v2i64:
+; SIMD128-UNIMPL-FAST:         .functype mul_v2i64 (v128, v128) -> (v128)
+; SIMD128-UNIMPL-FAST-NEXT:    i64x2.extract_lane $push5=, $0, 0
+; SIMD128-UNIMPL-FAST-NEXT:    i64x2.extract_lane $push4=, $1, 0
+; SIMD128-UNIMPL-FAST-NEXT:    i64.mul $push6=, $pop5, $pop4
+; SIMD128-UNIMPL-FAST-NEXT:    i64x2.splat $push7=, $pop6
+; SIMD128-UNIMPL-FAST-NEXT:    i64x2.extract_lane $push2=, $0, 1
+; SIMD128-UNIMPL-FAST-NEXT:    i64x2.extract_lane $push1=, $1, 1
+; SIMD128-UNIMPL-FAST-NEXT:    i64.mul $push3=, $pop2, $pop1
+; SIMD128-UNIMPL-FAST-NEXT:    i64x2.replace_lane $push0=, $pop7, 1, $pop3
+; SIMD128-UNIMPL-FAST-NEXT:    return $pop0
+;
+; NO-SIMD64-SLOW-LABEL: mul_v2i64:
+; NO-SIMD64-SLOW:         .functype mul_v2i64 (i32, i64, i64, i64, i64) -> ()
+; NO-SIMD64-SLOW-NEXT:    i64.mul $push0=, $2, $4
+; NO-SIMD64-SLOW-NEXT:    i64.store 8($0), $pop0
+; NO-SIMD64-SLOW-NEXT:    i64.mul $push1=, $1, $3
+; NO-SIMD64-SLOW-NEXT:    i64.store 0($0), $pop1
+; NO-SIMD64-SLOW-NEXT:    return
+;
+; NO-SIMD64-FAST-LABEL: mul_v2i64:
+; NO-SIMD64-FAST:         .functype mul_v2i64 (i32, i64, i64, i64, i64) -> ()
+; NO-SIMD64-FAST-NEXT:    i64.mul $push0=, $1, $3
+; NO-SIMD64-FAST-NEXT:    i64.store 0($0), $pop0
+; NO-SIMD64-FAST-NEXT:    i64.mul $push1=, $2, $4
+; NO-SIMD64-FAST-NEXT:    i64.store 8($0), $pop1
+; NO-SIMD64-FAST-NEXT:    return
   %a = mul <2 x i64> %x, %y
   ret <2 x i64> %a
 }
 
-; CHECK-LABEL: neg_v2i64:
-; NO-SIMD128-NOT: i64x2
-; SIMD128-NEXT: .functype neg_v2i64 (v128) -> (v128){{$}}
-; SIMD128-NEXT: i64x2.neg $push[[R:[0-9]+]]=, $0{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <2 x i64> @neg_v2i64(<2 x i64> %x) {
+; SIMD128-UNIMPL-LABEL: neg_v2i64:
+; SIMD128-UNIMPL:         .functype neg_v2i64 (v128) -> (v128)
+; SIMD128-UNIMPL-NEXT:    i64x2.neg $push0=, $0
+; SIMD128-UNIMPL-NEXT:    return $pop0
+;
+; NO-SIMD64-SLOW-LABEL: neg_v2i64:
+; NO-SIMD64-SLOW:         .functype neg_v2i64 (i32, i64, i64) -> ()
+; NO-SIMD64-SLOW-NEXT:    i64.const $push0=, 0
+; NO-SIMD64-SLOW-NEXT:    i64.sub $push1=, $pop0, $2
+; NO-SIMD64-SLOW-NEXT:    i64.store 8($0), $pop1
+; NO-SIMD64-SLOW-NEXT:    i64.const $push3=, 0
+; NO-SIMD64-SLOW-NEXT:    i64.sub $push2=, $pop3, $1
+; NO-SIMD64-SLOW-NEXT:    i64.store 0($0), $pop2
+; NO-SIMD64-SLOW-NEXT:    return
+;
+; NO-SIMD64-FAST-LABEL: neg_v2i64:
+; NO-SIMD64-FAST:         .functype neg_v2i64 (i32, i64, i64) -> ()
+; NO-SIMD64-FAST-NEXT:    i64.const $push0=, 0
+; NO-SIMD64-FAST-NEXT:    i64.sub $push1=, $pop0, $1
+; NO-SIMD64-FAST-NEXT:    i64.store 0($0), $pop1
+; NO-SIMD64-FAST-NEXT:    i64.const $push3=, 0
+; NO-SIMD64-FAST-NEXT:    i64.sub $push2=, $pop3, $2
+; NO-SIMD64-FAST-NEXT:    i64.store 8($0), $pop2
+; NO-SIMD64-FAST-NEXT:    return
   %a = sub <2 x i64> <i64 0, i64 0>, %x
   ret <2 x i64> %a
 }
 
-; CHECK-LABEL: shl_v2i64:
-; NO-SIMD128-NOT: i64x2
-; SIMD128-NEXT: .functype shl_v2i64 (v128, i32) -> (v128){{$}}
-; SIMD128-NEXT: i64x2.shl $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <2 x i64> @shl_v2i64(<2 x i64> %v, i32 %x) {
+; SIMD128-UNIMPL-LABEL: shl_v2i64:
+; SIMD128-UNIMPL:         .functype shl_v2i64 (v128, i32) -> (v128)
+; SIMD128-UNIMPL-NEXT:    i64x2.shl $push0=, $0, $1
+; SIMD128-UNIMPL-NEXT:    return $pop0
+;
+; NO-SIMD64-LABEL: shl_v2i64:
+; NO-SIMD64:         .functype shl_v2i64 (i32, i64, i64, i32) -> ()
+; NO-SIMD64-NEXT:    i64.extend_i32_u $push3=, $3
+; NO-SIMD64-NEXT:    local.tee $push2=, $4=, $pop3
+; NO-SIMD64-NEXT:    i64.shl $push0=, $2, $pop2
+; NO-SIMD64-NEXT:    i64.store 8($0), $pop0
+; NO-SIMD64-NEXT:    i64.shl $push1=, $1, $4
+; NO-SIMD64-NEXT:    i64.store 0($0), $pop1
+; NO-SIMD64-NEXT:    return
   %x2 = zext i32 %x to i64
   %t = insertelement <2 x i64> undef, i64 %x2, i32 0
   %s = shufflevector <2 x i64> %t, <2 x i64> undef, <2 x i32> <i32 0, i32 0>
   %a = shl <2 x i64> %v, %s
   ret <2 x i64> %a
 }
 
-; CHECK-LABEL: shl_sext_v2i64:
-; NO-SIMD128-NOT: i64x2
-; SIMD128-NEXT: .functype shl_sext_v2i64 (v128, i32) -> (v128){{$}}
-; SIMD128-NEXT: i64x2.shl $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <2 x i64> @shl_sext_v2i64(<2 x i64> %v, i32 %x) {
+; SIMD128-UNIMPL-LABEL: shl_sext_v2i64:
+; SIMD128-UNIMPL:         .functype shl_sext_v2i64 (v128, i32) -> (v128)
+; SIMD128-UNIMPL-NEXT:    i64x2.shl $push0=, $0, $1
+; SIMD128-UNIMPL-NEXT:    return $pop0
+;
+; NO-SIMD64-LABEL: shl_sext_v2i64:
+; NO-SIMD64:         .functype shl_sext_v2i64 (i32, i64, i64, i32) -> ()
+; NO-SIMD64-NEXT:    i64.extend_i32_s $push3=, $3
+; NO-SIMD64-NEXT:    local.tee $push2=, $4=, $pop3
+; NO-SIMD64-NEXT:    i64.shl $push0=, $2, $pop2
+; NO-SIMD64-NEXT:    i64.store 8($0), $pop0
+; NO-SIMD64-NEXT:    i64.shl $push1=, $1, $4
+; NO-SIMD64-NEXT:    i64.store 0($0), $pop1
+; NO-SIMD64-NEXT:    return
   %x2 = sext i32 %x to i64
   %t = insertelement <2 x i64> undef, i64 %x2, i32 0
   %s = shufflevector <2 x i64> %t, <2 x i64> undef, <2 x i32> <i32 0, i32 0>
   %a = shl <2 x i64> %v, %s
   ret <2 x i64> %a
 }
 
-; CHECK-LABEL: shl_noext_v2i64:
-; NO-SIMD128-NOT: i64x2
-; SIMD128-NEXT: .functype shl_noext_v2i64 (v128, i64) -> (v128){{$}}
-; SIMD128-NEXT: i32.wrap_i64 $push[[L0:[0-9]+]]=, $1{{$}}
-; SIMD128-NEXT: i64x2.shl $push[[R:[0-9]+]]=, $0, $pop[[L0]]{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <2 x i64> @shl_noext_v2i64(<2 x i64> %v, i64 %x) {
+; SIMD128-UNIMPL-SLOW-LABEL: shl_noext_v2i64:
+; SIMD128-UNIMPL-SLOW:         .functype shl_noext_v2i64 (v128, i64) -> (v128)
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.wrap_i64 $push0=, $1
+; SIMD128-UNIMPL-SLOW-NEXT:    i64x2.shl $push1=, $0, $pop0
+; SIMD128-UNIMPL-SLOW-NEXT:    return $pop1
+;
+; SIMD128-UNIMPL-FAST-LABEL: shl_noext_v2i64:
+; SIMD128-UNIMPL-FAST:         .functype shl_noext_v2i64 (v128, i64) -> (v128)
+; SIMD128-UNIMPL-FAST-NEXT:    i32.wrap_i64 $push1=, $1
+; SIMD128-UNIMPL-FAST-NEXT:    i64x2.shl $push0=, $0, $pop1
+; SIMD128-UNIMPL-FAST-NEXT:    return $pop0
+;
+; NO-SIMD64-SLOW-LABEL: shl_noext_v2i64:
+; NO-SIMD64-SLOW:         .functype shl_noext_v2i64 (i32, i64, i64, i64) -> ()
+; NO-SIMD64-SLOW-NEXT:    i64.shl $push0=, $2, $3
+; NO-SIMD64-SLOW-NEXT:    i64.store 8($0), $pop0
+; NO-SIMD64-SLOW-NEXT:    i64.shl $push1=, $1, $3
+; NO-SIMD64-SLOW-NEXT:    i64.store 0($0), $pop1
+; NO-SIMD64-SLOW-NEXT:    return
+;
+; NO-SIMD64-FAST-LABEL: shl_noext_v2i64:
+; NO-SIMD64-FAST:         .functype shl_noext_v2i64 (i32, i64, i64, i64) -> ()
+; NO-SIMD64-FAST-NEXT:    i64.shl $push0=, $1, $3
+; NO-SIMD64-FAST-NEXT:    i64.store 0($0), $pop0
+; NO-SIMD64-FAST-NEXT:    i64.shl $push1=, $2, $3
+; NO-SIMD64-FAST-NEXT:    i64.store 8($0), $pop1
+; NO-SIMD64-FAST-NEXT:    return
   %t = insertelement <2 x i64> undef, i64 %x, i32 0
   %s = shufflevector <2 x i64> %t, <2 x i64> undef, <2 x i32> <i32 0, i32 0>
   %a = shl <2 x i64> %v, %s
   ret <2 x i64> %a
 }
 
-; CHECK-LABEL: shl_const_v2i64:
-; NO-SIMD128-NOT: i64x2
-; SIMD128-NEXT: .functype shl_const_v2i64 (v128) -> (v128){{$}}
-; SIMD128-NEXT: i32.const $push[[L0:[0-9]+]]=, 5{{$}}
-; SIMD128-NEXT: i64x2.shl $push[[R:[0-9]+]]=, $0, $pop[[L0]]{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <2 x i64> @shl_const_v2i64(<2 x i64> %v) {
+; SIMD128-UNIMPL-SLOW-LABEL: shl_const_v2i64:
+; SIMD128-UNIMPL-SLOW:         .functype shl_const_v2i64 (v128) -> (v128)
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.const $push0=, 5
+; SIMD128-UNIMPL-SLOW-NEXT:    i64x2.shl $push1=, $0, $pop0
+; SIMD128-UNIMPL-SLOW-NEXT:    return $pop1
+;
+; SIMD128-UNIMPL-FAST-LABEL: shl_const_v2i64:
+; SIMD128-UNIMPL-FAST:         .functype shl_const_v2i64 (v128) -> (v128)
+; SIMD128-UNIMPL-FAST-NEXT:    i32.const $push1=, 5
+; SIMD128-UNIMPL-FAST-NEXT:    i64x2.shl $push0=, $0, $pop1
+; SIMD128-UNIMPL-FAST-NEXT:    return $pop0
+;
+; NO-SIMD64-SLOW-LABEL: shl_const_v2i64:
+; NO-SIMD64-SLOW:         .functype shl_const_v2i64 (i32, i64, i64) -> ()
+; NO-SIMD64-SLOW-NEXT:    i64.const $push0=, 5
+; NO-SIMD64-SLOW-NEXT:    i64.shl $push1=, $2, $pop0
+; NO-SIMD64-SLOW-NEXT:    i64.store 8($0), $pop1
+; NO-SIMD64-SLOW-NEXT:    i64.const $push3=, 5
+; NO-SIMD64-SLOW-NEXT:    i64.shl $push2=, $1, $pop3
+; NO-SIMD64-SLOW-NEXT:    i64.store 0($0), $pop2
+; NO-SIMD64-SLOW-NEXT:    return
+;
+; NO-SIMD64-FAST-LABEL: shl_const_v2i64:
+; NO-SIMD64-FAST:         .functype shl_const_v2i64 (i32, i64, i64) -> ()
+; NO-SIMD64-FAST-NEXT:    i64.const $push0=, 5
+; NO-SIMD64-FAST-NEXT:    i64.shl $push1=, $1, $pop0
+; NO-SIMD64-FAST-NEXT:    i64.store 0($0), $pop1
+; NO-SIMD64-FAST-NEXT:    i64.const $push3=, 5
+; NO-SIMD64-FAST-NEXT:    i64.shl $push2=, $2, $pop3
+; NO-SIMD64-FAST-NEXT:    i64.store 8($0), $pop2
+; NO-SIMD64-FAST-NEXT:    return
   %a = shl <2 x i64> %v, <i64 5, i64 5>
   ret <2 x i64> %a
 }
 
-; CHECK-LABEL: shl_vec_v2i64:
-; NO-SIMD128-NOT: i64x2
-; SIMD128-NEXT: .functype shl_vec_v2i64 (v128, v128) -> (v128){{$}}
-; SIMD128-NEXT: i64x2.extract_lane $push[[L0:[0-9]+]]=, $0, 0{{$}}
-; SIMD128-NEXT: i64x2.extract_lane $push[[L1:[0-9]+]]=, $1, 0{{$}}
-; SIMD128-NEXT: i64.shl $push[[L2:[0-9]+]]=, $pop[[L0]], $pop[[L1]]{{$}}
-; SIMD128-NEXT: i64x2.splat $push[[L3:[0-9]+]]=, $pop[[L2]]{{$}}
-; SIMD128-NEXT: i64x2.extract_lane $push[[L4:[0-9]+]]=, $0, 1{{$}}
-; SIMD128-NEXT: i64x2.extract_lane $push[[L5:[0-9]+]]=, $1, 1{{$}}
-; SIMD128-NEXT: i64.shl $push[[L6:[0-9]+]]=, $pop[[L4]], $pop[[L5]]{{$}}
-; SIMD128-NEXT: i64x2.replace_lane $push[[R:[0-9]+]]=, $pop[[L3]], 1, $pop[[L6]]{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <2 x i64> @shl_vec_v2i64(<2 x i64> %v, <2 x i64> %x) {
+; SIMD128-UNIMPL-SLOW-LABEL: shl_vec_v2i64:
+; SIMD128-UNIMPL-SLOW:         .functype shl_vec_v2i64 (v128, v128) -> (v128)
+; SIMD128-UNIMPL-SLOW-NEXT:    i64x2.extract_lane $push4=, $0, 0
+; SIMD128-UNIMPL-SLOW-NEXT:    i64x2.extract_lane $push3=, $1, 0
+; SIMD128-UNIMPL-SLOW-NEXT:    i64.shl $push5=, $pop4, $pop3
+; SIMD128-UNIMPL-SLOW-NEXT:    i64x2.splat $push6=, $pop5
+; SIMD128-UNIMPL-SLOW-NEXT:    i64x2.extract_lane $push1=, $0, 1
+; SIMD128-UNIMPL-SLOW-NEXT:    i64x2.extract_lane $push0=, $1, 1
+; SIMD128-UNIMPL-SLOW-NEXT:    i64.shl $push2=, $pop1, $pop0
+; SIMD128-UNIMPL-SLOW-NEXT:    i64x2.replace_lane $push7=, $pop6, 1, $pop2
+; SIMD128-UNIMPL-SLOW-NEXT:    return $pop7
+;
+; SIMD128-UNIMPL-FAST-LABEL: shl_vec_v2i64:
+; SIMD128-UNIMPL-FAST:         .functype shl_vec_v2i64 (v128, v128) -> (v128)
+; SIMD128-UNIMPL-FAST-NEXT:    i64x2.extract_lane $push5=, $0, 0
+; SIMD128-UNIMPL-FAST-NEXT:    i64x2.extract_lane $push4=, $1, 0
+; SIMD128-UNIMPL-FAST-NEXT:    i64.shl $push6=, $pop5, $pop4
+; SIMD128-UNIMPL-FAST-NEXT:    i64x2.splat $push7=, $pop6
+; SIMD128-UNIMPL-FAST-NEXT:    i64x2.extract_lane $push2=, $0, 1
+; SIMD128-UNIMPL-FAST-NEXT:    i64x2.extract_lane $push1=, $1, 1
+; SIMD128-UNIMPL-FAST-NEXT:    i64.shl $push3=, $pop2, $pop1
+; SIMD128-UNIMPL-FAST-NEXT:    i64x2.replace_lane $push0=, $pop7, 1, $pop3
+; SIMD128-UNIMPL-FAST-NEXT:    return $pop0
+;
+; NO-SIMD64-SLOW-LABEL: shl_vec_v2i64:
+; NO-SIMD64-SLOW:         .functype shl_vec_v2i64 (i32, i64, i64, i64, i64) -> ()
+; NO-SIMD64-SLOW-NEXT:    i64.shl $push0=, $2, $4
+; NO-SIMD64-SLOW-NEXT:    i64.store 8($0), $pop0
+; NO-SIMD64-SLOW-NEXT:    i64.shl $push1=, $1, $3
+; NO-SIMD64-SLOW-NEXT:    i64.store 0($0), $pop1
+; NO-SIMD64-SLOW-NEXT:    return
+;
+; NO-SIMD64-FAST-LABEL: shl_vec_v2i64:
+; NO-SIMD64-FAST:         .functype shl_vec_v2i64 (i32, i64, i64, i64, i64) -> ()
+; NO-SIMD64-FAST-NEXT:    i64.shl $push0=, $1, $3
+; NO-SIMD64-FAST-NEXT:    i64.store 0($0), $pop0
+; NO-SIMD64-FAST-NEXT:    i64.shl $push1=, $2, $4
+; NO-SIMD64-FAST-NEXT:    i64.store 8($0), $pop1
+; NO-SIMD64-FAST-NEXT:    return
   %a = shl <2 x i64> %v, %x
   ret <2 x i64> %a
 }
 
-; CHECK-LABEL: shr_s_v2i64:
-; NO-SIMD128-NOT: i64x2
-; SIMD128-NEXT: .functype shr_s_v2i64 (v128, i32) -> (v128){{$}}
-; SIMD128-NEXT: i64x2.shr_s $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <2 x i64> @shr_s_v2i64(<2 x i64> %v, i32 %x) {
+; SIMD128-UNIMPL-LABEL: shr_s_v2i64:
+; SIMD128-UNIMPL:         .functype shr_s_v2i64 (v128, i32) -> (v128)
+; SIMD128-UNIMPL-NEXT:    i64x2.shr_s $push0=, $0, $1
+; SIMD128-UNIMPL-NEXT:    return $pop0
+;
+; NO-SIMD64-LABEL: shr_s_v2i64:
+; NO-SIMD64:         .functype shr_s_v2i64 (i32, i64, i64, i32) -> ()
+; NO-SIMD64-NEXT:    i64.extend_i32_u $push3=, $3
+; NO-SIMD64-NEXT:    local.tee $push2=, $4=, $pop3
+; NO-SIMD64-NEXT:    i64.shr_s $push0=, $2, $pop2
+; NO-SIMD64-NEXT:    i64.store 8($0), $pop0
+; NO-SIMD64-NEXT:    i64.shr_s $push1=, $1, $4
+; NO-SIMD64-NEXT:    i64.store 0($0), $pop1
+; NO-SIMD64-NEXT:    return
   %x2 = zext i32 %x to i64
   %t = insertelement <2 x i64> undef, i64 %x2, i32 0
   %s = shufflevector <2 x i64> %t, <2 x i64> undef, <2 x i32> <i32 0, i32 0>
   %a = ashr <2 x i64> %v, %s
   ret <2 x i64> %a
 }
 
-; CHECK-LABEL: shr_s_sext_v2i64:
-; NO-SIMD128-NOT: i64x2
-; SIMD128-NEXT: .functype shr_s_sext_v2i64 (v128, i32) -> (v128){{$}}
-; SIMD128-NEXT: i64x2.shr_s $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <2 x i64> @shr_s_sext_v2i64(<2 x i64> %v, i32 %x) {
+; SIMD128-UNIMPL-LABEL: shr_s_sext_v2i64:
+; SIMD128-UNIMPL:         .functype shr_s_sext_v2i64 (v128, i32) -> (v128)
+; SIMD128-UNIMPL-NEXT:    i64x2.shr_s $push0=, $0, $1
+; SIMD128-UNIMPL-NEXT:    return $pop0
+;
+; NO-SIMD64-LABEL: shr_s_sext_v2i64:
+; NO-SIMD64:         .functype shr_s_sext_v2i64 (i32, i64, i64, i32) -> ()
+; NO-SIMD64-NEXT:    i64.extend_i32_s $push3=, $3
+; NO-SIMD64-NEXT:    local.tee $push2=, $4=, $pop3
+; NO-SIMD64-NEXT:    i64.shr_s $push0=, $2, $pop2
+; NO-SIMD64-NEXT:    i64.store 8($0), $pop0
+; NO-SIMD64-NEXT:    i64.shr_s $push1=, $1, $4
+; NO-SIMD64-NEXT:    i64.store 0($0), $pop1
+; NO-SIMD64-NEXT:    return
   %x2 = sext i32 %x to i64
   %t = insertelement <2 x i64> undef, i64 %x2, i32 0
   %s = shufflevector <2 x i64> %t, <2 x i64> undef, <2 x i32> <i32 0, i32 0>
   %a = ashr <2 x i64> %v, %s
   ret <2 x i64> %a
 }
 
-; CHECK-LABEL: shr_s_noext_v2i64:
-; NO-SIMD128-NOT: i64x2
-; SIMD128-NEXT: .functype shr_s_noext_v2i64 (v128, i64) -> (v128){{$}}
-; SIMD128-NEXT: i32.wrap_i64 $push[[L0:[0-9]+]]=, $1{{$}}
-; SIMD128-NEXT: i64x2.shr_s $push[[R:[0-9]+]]=, $0, $pop[[L0]]{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <2 x i64> @shr_s_noext_v2i64(<2 x i64> %v, i64 %x) {
+; SIMD128-UNIMPL-SLOW-LABEL: shr_s_noext_v2i64:
+; SIMD128-UNIMPL-SLOW:         .functype shr_s_noext_v2i64 (v128, i64) -> (v128)
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.wrap_i64 $push0=, $1
+; SIMD128-UNIMPL-SLOW-NEXT:    i64x2.shr_s $push1=, $0, $pop0
+; SIMD128-UNIMPL-SLOW-NEXT:    return $pop1
+;
+; SIMD128-UNIMPL-FAST-LABEL: shr_s_noext_v2i64:
+; SIMD128-UNIMPL-FAST:         .functype shr_s_noext_v2i64 (v128, i64) -> (v128)
+; SIMD128-UNIMPL-FAST-NEXT:    i32.wrap_i64 $push1=, $1
+; SIMD128-UNIMPL-FAST-NEXT:    i64x2.shr_s $push0=, $0, $pop1
+; SIMD128-UNIMPL-FAST-NEXT:    return $pop0
+;
+; NO-SIMD64-SLOW-LABEL: shr_s_noext_v2i64:
+; NO-SIMD64-SLOW:         .functype shr_s_noext_v2i64 (i32, i64, i64, i64) -> ()
+; NO-SIMD64-SLOW-NEXT:    i64.shr_s $push0=, $2, $3
+; NO-SIMD64-SLOW-NEXT:    i64.store 8($0), $pop0
+; NO-SIMD64-SLOW-NEXT:    i64.shr_s $push1=, $1, $3
+; NO-SIMD64-SLOW-NEXT:    i64.store 0($0), $pop1
+; NO-SIMD64-SLOW-NEXT:    return
+;
+; NO-SIMD64-FAST-LABEL: shr_s_noext_v2i64:
+; NO-SIMD64-FAST:         .functype shr_s_noext_v2i64 (i32, i64, i64, i64) -> ()
+; NO-SIMD64-FAST-NEXT:    i64.shr_s $push0=, $1, $3
+; NO-SIMD64-FAST-NEXT:    i64.store 0($0), $pop0
+; NO-SIMD64-FAST-NEXT:    i64.shr_s $push1=, $2, $3
+; NO-SIMD64-FAST-NEXT:    i64.store 8($0), $pop1
+; NO-SIMD64-FAST-NEXT:    return
   %t = insertelement <2 x i64> undef, i64 %x, i32 0
   %s = shufflevector <2 x i64> %t, <2 x i64> undef, <2 x i32> <i32 0, i32 0>
   %a = ashr <2 x i64> %v, %s
   ret <2 x i64> %a
 }
 
-; CHECK-LABEL: shr_s_const_v2i64:
-; NO-SIMD128-NOT: i64x2
-; SIMD128-NEXT: .functype shr_s_const_v2i64 (v128) -> (v128){{$}}
-; SIMD128-NEXT: i32.const $push[[L0:[0-9]+]]=, 5{{$}}
-; SIMD128-NEXT: i64x2.shr_s $push[[R:[0-9]+]]=, $0, $pop[[L0]]{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <2 x i64> @shr_s_const_v2i64(<2 x i64> %v) {
+; SIMD128-UNIMPL-SLOW-LABEL: shr_s_const_v2i64:
+; SIMD128-UNIMPL-SLOW:         .functype shr_s_const_v2i64 (v128) -> (v128)
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.const $push0=, 5
+; SIMD128-UNIMPL-SLOW-NEXT:    i64x2.shr_s $push1=, $0, $pop0
+; SIMD128-UNIMPL-SLOW-NEXT:    return $pop1
+;
+; SIMD128-UNIMPL-FAST-LABEL: shr_s_const_v2i64:
+; SIMD128-UNIMPL-FAST:         .functype shr_s_const_v2i64 (v128) -> (v128)
+; SIMD128-UNIMPL-FAST-NEXT:    i32.const $push1=, 5
+; SIMD128-UNIMPL-FAST-NEXT:    i64x2.shr_s $push0=, $0, $pop1
+; SIMD128-UNIMPL-FAST-NEXT:    return $pop0
+;
+; NO-SIMD64-SLOW-LABEL: shr_s_const_v2i64:
+; NO-SIMD64-SLOW:         .functype shr_s_const_v2i64 (i32, i64, i64) -> ()
+; NO-SIMD64-SLOW-NEXT:    i64.const $push0=, 5
+; NO-SIMD64-SLOW-NEXT:    i64.shr_s $push1=, $2, $pop0
+; NO-SIMD64-SLOW-NEXT:    i64.store 8($0), $pop1
+; NO-SIMD64-SLOW-NEXT:    i64.const $push3=, 5
+; NO-SIMD64-SLOW-NEXT:    i64.shr_s $push2=, $1, $pop3
+; NO-SIMD64-SLOW-NEXT:    i64.store 0($0), $pop2
+; NO-SIMD64-SLOW-NEXT:    return
+;
+; NO-SIMD64-FAST-LABEL: shr_s_const_v2i64:
+; NO-SIMD64-FAST:         .functype shr_s_const_v2i64 (i32, i64, i64) -> ()
+; NO-SIMD64-FAST-NEXT:    i64.const $push0=, 5
+; NO-SIMD64-FAST-NEXT:    i64.shr_s $push1=, $1, $pop0
+; NO-SIMD64-FAST-NEXT:    i64.store 0($0), $pop1
+; NO-SIMD64-FAST-NEXT:    i64.const $push3=, 5
+; NO-SIMD64-FAST-NEXT:    i64.shr_s $push2=, $2, $pop3
+; NO-SIMD64-FAST-NEXT:    i64.store 8($0), $pop2
+; NO-SIMD64-FAST-NEXT:    return
   %a = ashr <2 x i64> %v, <i64 5, i64 5>
   ret <2 x i64> %a
 }
 
-; CHECK-LABEL: shr_s_vec_v2i64:
-; NO-SIMD128-NOT: i64x2
-; SIMD128-NEXT: .functype shr_s_vec_v2i64 (v128, v128) -> (v128){{$}}
-; SIMD128-NEXT: i64x2.extract_lane $push[[L0:[0-9]+]]=, $0, 0{{$}}
-; SIMD128-NEXT: i64x2.extract_lane $push[[L1:[0-9]+]]=, $1, 0{{$}}
-; SIMD128-NEXT: i64.shr_s $push[[L2:[0-9]+]]=, $pop[[L0]], $pop[[L1]]{{$}}
-; SIMD128-NEXT: i64x2.splat $push[[L3:[0-9]+]]=, $pop[[L2]]{{$}}
-; SIMD128-NEXT: i64x2.extract_lane $push[[L4:[0-9]+]]=, $0, 1{{$}}
-; SIMD128-NEXT: i64x2.extract_lane $push[[L5:[0-9]+]]=, $1, 1{{$}}
-; SIMD128-NEXT: i64.shr_s $push[[L6:[0-9]+]]=, $pop[[L4]], $pop[[L5]]{{$}}
-; SIMD128-NEXT: i64x2.replace_lane $push[[R:[0-9]+]]=, $pop[[L3]], 1, $pop[[L6]]{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <2 x i64> @shr_s_vec_v2i64(<2 x i64> %v, <2 x i64> %x) {
+; SIMD128-UNIMPL-SLOW-LABEL: shr_s_vec_v2i64:
+; SIMD128-UNIMPL-SLOW:         .functype shr_s_vec_v2i64 (v128, v128) -> (v128)
+; SIMD128-UNIMPL-SLOW-NEXT:    i64x2.extract_lane $push4=, $0, 0
+; SIMD128-UNIMPL-SLOW-NEXT:    i64x2.extract_lane $push3=, $1, 0
+; SIMD128-UNIMPL-SLOW-NEXT:    i64.shr_s $push5=, $pop4, $pop3
+; SIMD128-UNIMPL-SLOW-NEXT:    i64x2.splat $push6=, $pop5
+; SIMD128-UNIMPL-SLOW-NEXT:    i64x2.extract_lane $push1=, $0, 1
+; SIMD128-UNIMPL-SLOW-NEXT:    i64x2.extract_lane $push0=, $1, 1
+; SIMD128-UNIMPL-SLOW-NEXT:    i64.shr_s $push2=, $pop1, $pop0
+; SIMD128-UNIMPL-SLOW-NEXT:    i64x2.replace_lane $push7=, $pop6, 1, $pop2
+; SIMD128-UNIMPL-SLOW-NEXT:    return $pop7
+;
+; SIMD128-UNIMPL-FAST-LABEL: shr_s_vec_v2i64:
+; SIMD128-UNIMPL-FAST:         .functype shr_s_vec_v2i64 (v128, v128) -> (v128)
+; SIMD128-UNIMPL-FAST-NEXT:    i64x2.extract_lane $push5=, $0, 0
+; SIMD128-UNIMPL-FAST-NEXT:    i64x2.extract_lane $push4=, $1, 0
+; SIMD128-UNIMPL-FAST-NEXT:    i64.shr_s $push6=, $pop5, $pop4
+; SIMD128-UNIMPL-FAST-NEXT:    i64x2.splat $push7=, $pop6
+; SIMD128-UNIMPL-FAST-NEXT:    i64x2.extract_lane $push2=, $0, 1
+; SIMD128-UNIMPL-FAST-NEXT:    i64x2.extract_lane $push1=, $1, 1
+; SIMD128-UNIMPL-FAST-NEXT:    i64.shr_s $push3=, $pop2, $pop1
+; SIMD128-UNIMPL-FAST-NEXT:    i64x2.replace_lane $push0=, $pop7, 1, $pop3
+; SIMD128-UNIMPL-FAST-NEXT:    return $pop0
+;
+; NO-SIMD64-SLOW-LABEL: shr_s_vec_v2i64:
+; NO-SIMD64-SLOW:         .functype shr_s_vec_v2i64 (i32, i64, i64, i64, i64) -> ()
+; NO-SIMD64-SLOW-NEXT:    i64.shr_s $push0=, $2, $4
+; NO-SIMD64-SLOW-NEXT:    i64.store 8($0), $pop0
+; NO-SIMD64-SLOW-NEXT:    i64.shr_s $push1=, $1, $3
+; NO-SIMD64-SLOW-NEXT:    i64.store 0($0), $pop1
+; NO-SIMD64-SLOW-NEXT:    return
+;
+; NO-SIMD64-FAST-LABEL: shr_s_vec_v2i64:
+; NO-SIMD64-FAST:         .functype shr_s_vec_v2i64 (i32, i64, i64, i64, i64) -> ()
+; NO-SIMD64-FAST-NEXT:    i64.shr_s $push0=, $1, $3
+; NO-SIMD64-FAST-NEXT:    i64.store 0($0), $pop0
+; NO-SIMD64-FAST-NEXT:    i64.shr_s $push1=, $2, $4
+; NO-SIMD64-FAST-NEXT:    i64.store 8($0), $pop1
+; NO-SIMD64-FAST-NEXT:    return
   %a = ashr <2 x i64> %v, %x
   ret <2 x i64> %a
 }
 
-; CHECK-LABEL: shr_u_v2i64:
-; NO-SIMD128-NOT: i64x2
-; SIMD128-NEXT: .functype shr_u_v2i64 (v128, i32) -> (v128){{$}}
-; SIMD128-NEXT: i64x2.shr_u $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <2 x i64> @shr_u_v2i64(<2 x i64> %v, i32 %x) {
+; SIMD128-UNIMPL-LABEL: shr_u_v2i64:
+; SIMD128-UNIMPL:         .functype shr_u_v2i64 (v128, i32) -> (v128)
+; SIMD128-UNIMPL-NEXT:    i64x2.shr_u $push0=, $0, $1
+; SIMD128-UNIMPL-NEXT:    return $pop0
+;
+; NO-SIMD64-LABEL: shr_u_v2i64:
+; NO-SIMD64:         .functype shr_u_v2i64 (i32, i64, i64, i32) -> ()
+; NO-SIMD64-NEXT:    i64.extend_i32_u $push3=, $3
+; NO-SIMD64-NEXT:    local.tee $push2=, $4=, $pop3
+; NO-SIMD64-NEXT:    i64.shr_u $push0=, $2, $pop2
+; NO-SIMD64-NEXT:    i64.store 8($0), $pop0
+; NO-SIMD64-NEXT:    i64.shr_u $push1=, $1, $4
+; NO-SIMD64-NEXT:    i64.store 0($0), $pop1
+; NO-SIMD64-NEXT:    return
   %x2 = zext i32 %x to i64
   %t = insertelement <2 x i64> undef, i64 %x2, i32 0
   %s = shufflevector <2 x i64> %t, <2 x i64> undef, <2 x i32> <i32 0, i32 0>
   %a = lshr <2 x i64> %v, %s
   ret <2 x i64> %a
 }
 
-; CHECK-LABEL: shr_u_sext_v2i64:
-; NO-SIMD128-NOT: i64x2
-; SIMD128-NEXT: .functype shr_u_sext_v2i64 (v128, i32) -> (v128){{$}}
-; SIMD128-NEXT: i64x2.shr_u $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <2 x i64> @shr_u_sext_v2i64(<2 x i64> %v, i32 %x) {
+; SIMD128-UNIMPL-LABEL: shr_u_sext_v2i64:
+; SIMD128-UNIMPL:         .functype shr_u_sext_v2i64 (v128, i32) -> (v128)
+; SIMD128-UNIMPL-NEXT:    i64x2.shr_u $push0=, $0, $1
+; SIMD128-UNIMPL-NEXT:    return $pop0
+;
+; NO-SIMD64-LABEL: shr_u_sext_v2i64:
+; NO-SIMD64:         .functype shr_u_sext_v2i64 (i32, i64, i64, i32) -> ()
+; NO-SIMD64-NEXT:    i64.extend_i32_s $push3=, $3
+; NO-SIMD64-NEXT:    local.tee $push2=, $4=, $pop3
+; NO-SIMD64-NEXT:    i64.shr_u $push0=, $2, $pop2
+; NO-SIMD64-NEXT:    i64.store 8($0), $pop0
+; NO-SIMD64-NEXT:    i64.shr_u $push1=, $1, $4
+; NO-SIMD64-NEXT:    i64.store 0($0), $pop1
+; NO-SIMD64-NEXT:    return
   %x2 = sext i32 %x to i64
   %t = insertelement <2 x i64> undef, i64 %x2, i32 0
   %s = shufflevector <2 x i64> %t, <2 x i64> undef, <2 x i32> <i32 0, i32 0>
   %a = lshr <2 x i64> %v, %s
   ret <2 x i64> %a
 }
 
-; CHECK-LABEL: shr_u_noext_v2i64:
-; NO-SIMD128-NOT: i64x2
-; SIMD128-NEXT: .functype shr_u_noext_v2i64 (v128, i64) -> (v128){{$}}
-; SIMD128-NEXT: i32.wrap_i64 $push[[L0:[0-9]+]]=, $1{{$}}
-; SIMD128-NEXT: i64x2.shr_u $push[[R:[0-9]+]]=, $0, $pop[[L0]]{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <2 x i64> @shr_u_noext_v2i64(<2 x i64> %v, i64 %x) {
+; SIMD128-UNIMPL-SLOW-LABEL: shr_u_noext_v2i64:
+; SIMD128-UNIMPL-SLOW:         .functype shr_u_noext_v2i64 (v128, i64) -> (v128)
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.wrap_i64 $push0=, $1
+; SIMD128-UNIMPL-SLOW-NEXT:    i64x2.shr_u $push1=, $0, $pop0
+; SIMD128-UNIMPL-SLOW-NEXT:    return $pop1
+;
+; SIMD128-UNIMPL-FAST-LABEL: shr_u_noext_v2i64:
+; SIMD128-UNIMPL-FAST:         .functype shr_u_noext_v2i64 (v128, i64) -> (v128)
+; SIMD128-UNIMPL-FAST-NEXT:    i32.wrap_i64 $push1=, $1
+; SIMD128-UNIMPL-FAST-NEXT:    i64x2.shr_u $push0=, $0, $pop1
+; SIMD128-UNIMPL-FAST-NEXT:    return $pop0
+;
+; NO-SIMD64-SLOW-LABEL: shr_u_noext_v2i64:
+; NO-SIMD64-SLOW:         .functype shr_u_noext_v2i64 (i32, i64, i64, i64) -> ()
+; NO-SIMD64-SLOW-NEXT:    i64.shr_u $push0=, $2, $3
+; NO-SIMD64-SLOW-NEXT:    i64.store 8($0), $pop0
+; NO-SIMD64-SLOW-NEXT:    i64.shr_u $push1=, $1, $3
+; NO-SIMD64-SLOW-NEXT:    i64.store 0($0), $pop1
+; NO-SIMD64-SLOW-NEXT:    return
+;
+; NO-SIMD64-FAST-LABEL: shr_u_noext_v2i64:
+; NO-SIMD64-FAST:         .functype shr_u_noext_v2i64 (i32, i64, i64, i64) -> ()
+; NO-SIMD64-FAST-NEXT:    i64.shr_u $push0=, $1, $3
+; NO-SIMD64-FAST-NEXT:    i64.store 0($0), $pop0
+; NO-SIMD64-FAST-NEXT:    i64.shr_u $push1=, $2, $3
+; NO-SIMD64-FAST-NEXT:    i64.store 8($0), $pop1
+; NO-SIMD64-FAST-NEXT:    return
   %t = insertelement <2 x i64> undef, i64 %x, i32 0
   %s = shufflevector <2 x i64> %t, <2 x i64> undef, <2 x i32> <i32 0, i32 0>
   %a = lshr <2 x i64> %v, %s
   ret <2 x i64> %a
 }
 
-; CHECK-LABEL: shr_u_const_v2i64:
-; NO-SIMD128-NOT: i64x2
-; SIMD128-NEXT: .functype shr_u_const_v2i64 (v128) -> (v128){{$}}
-; SIMD128-NEXT: i32.const $push[[L0:[0-9]+]]=, 5{{$}}
-; SIMD128-NEXT: i64x2.shr_u $push[[R:[0-9]+]]=, $0, $pop[[L0]]{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <2 x i64> @shr_u_const_v2i64(<2 x i64> %v) {
+; SIMD128-UNIMPL-SLOW-LABEL: shr_u_const_v2i64:
+; SIMD128-UNIMPL-SLOW:         .functype shr_u_const_v2i64 (v128) -> (v128)
+; SIMD128-UNIMPL-SLOW-NEXT:    i32.const $push0=, 5
+; SIMD128-UNIMPL-SLOW-NEXT:    i64x2.shr_u $push1=, $0, $pop0
+; SIMD128-UNIMPL-SLOW-NEXT:    return $pop1
+;
+; SIMD128-UNIMPL-FAST-LABEL: shr_u_const_v2i64:
+; SIMD128-UNIMPL-FAST:         .functype shr_u_const_v2i64 (v128) -> (v128)
+; SIMD128-UNIMPL-FAST-NEXT:    i32.const $push1=, 5
+; SIMD128-UNIMPL-FAST-NEXT:    i64x2.shr_u $push0=, $0, $pop1
+; SIMD128-UNIMPL-FAST-NEXT:    return $pop0
+;
+; NO-SIMD64-SLOW-LABEL: shr_u_const_v2i64:
+; NO-SIMD64-SLOW:         .functype shr_u_const_v2i64 (i32, i64, i64) -> ()
+; NO-SIMD64-SLOW-NEXT:    i64.const $push0=, 5
+; NO-SIMD64-SLOW-NEXT:    i64.shr_u $push1=, $2, $pop0
+; NO-SIMD64-SLOW-NEXT:    i64.store 8($0), $pop1
+; NO-SIMD64-SLOW-NEXT:    i64.const $push3=, 5
+; NO-SIMD64-SLOW-NEXT:    i64.shr_u $push2=, $1, $pop3
+; NO-SIMD64-SLOW-NEXT:    i64.store 0($0), $pop2
+; NO-SIMD64-SLOW-NEXT:    return
+;
+; NO-SIMD64-FAST-LABEL: shr_u_const_v2i64:
+; NO-SIMD64-FAST:         .functype shr_u_const_v2i64 (i32, i64, i64) -> ()
+; NO-SIMD64-FAST-NEXT:    i64.const $push0=, 5
+; NO-SIMD64-FAST-NEXT:    i64.shr_u $push1=, $1, $pop0
+; NO-SIMD64-FAST-NEXT:    i64.store 0($0), $pop1
+; NO-SIMD64-FAST-NEXT:    i64.const $push3=, 5
+; NO-SIMD64-FAST-NEXT:    i64.shr_u $push2=, $2, $pop3
+; NO-SIMD64-FAST-NEXT:    i64.store 8($0), $pop2
+; NO-SIMD64-FAST-NEXT:    return
   %a = lshr <2 x i64> %v, <i64 5, i64 5>
   ret <2 x i64> %a
 }
 
-; CHECK-LABEL: shr_u_vec_v2i64:
-; NO-SIMD128-NOT: i64x2
-; SIMD128-NEXT: .functype shr_u_vec_v2i64 (v128, v128) -> (v128){{$}}
-; SIMD128-NEXT: i64x2.extract_lane $push[[L0:[0-9]+]]=, $0, 0{{$}}
-; SIMD128-NEXT: i64x2.extract_lane $push[[L1:[0-9]+]]=, $1, 0{{$}}
-; SIMD128-NEXT: i64.shr_u $push[[L2:[0-9]+]]=, $pop[[L0]], $pop[[L1]]{{$}}
-; SIMD128-NEXT: i64x2.splat $push[[L3:[0-9]+]]=, $pop[[L2]]{{$}}
-; SIMD128-NEXT: i64x2.extract_lane $push[[L4:[0-9]+]]=, $0, 1{{$}}
-; SIMD128-NEXT: i64x2.extract_lane $push[[L5:[0-9]+]]=, $1, 1{{$}}
-; SIMD128-NEXT: i64.shr_u $push[[L6:[0-9]+]]=, $pop[[L4]], $pop[[L5]]{{$}}
-; SIMD128-NEXT: i64x2.replace_lane $push[[R:[0-9]+]]=, $pop[[L3]], 1, $pop[[L6]]{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <2 x i64> @shr_u_vec_v2i64(<2 x i64> %v, <2 x i64> %x) {
+; SIMD128-UNIMPL-SLOW-LABEL: shr_u_vec_v2i64:
+; SIMD128-UNIMPL-SLOW:         .functype shr_u_vec_v2i64 (v128, v128) -> (v128)
+; SIMD128-UNIMPL-SLOW-NEXT:    i64x2.extract_lane $push4=, $0, 0
+; SIMD128-UNIMPL-SLOW-NEXT:    i64x2.extract_lane $push3=, $1, 0
+; SIMD128-UNIMPL-SLOW-NEXT:    i64.shr_u $push5=, $pop4, $pop3
+; SIMD128-UNIMPL-SLOW-NEXT:    i64x2.splat $push6=, $pop5
+; SIMD128-UNIMPL-SLOW-NEXT:    i64x2.extract_lane $push1=, $0, 1
+; SIMD128-UNIMPL-SLOW-NEXT:    i64x2.extract_lane $push0=, $1, 1
+; SIMD128-UNIMPL-SLOW-NEXT:    i64.shr_u $push2=, $pop1, $pop0
+; SIMD128-UNIMPL-SLOW-NEXT:    i64x2.replace_lane $push7=, $pop6, 1, $pop2
+; SIMD128-UNIMPL-SLOW-NEXT:    return $pop7
+;
+; SIMD128-UNIMPL-FAST-LABEL: shr_u_vec_v2i64:
+; SIMD128-UNIMPL-FAST:         .functype shr_u_vec_v2i64 (v128, v128) -> (v128)
+; SIMD128-UNIMPL-FAST-NEXT:    i64x2.extract_lane $push5=, $0, 0
+; SIMD128-UNIMPL-FAST-NEXT:    i64x2.extract_lane $push4=, $1, 0
+; SIMD128-UNIMPL-FAST-NEXT:    i64.shr_u $push6=, $pop5, $pop4
+; SIMD128-UNIMPL-FAST-NEXT:    i64x2.splat $push7=, $pop6
+; SIMD128-UNIMPL-FAST-NEXT:    i64x2.extract_lane $push2=, $0, 1
+; SIMD128-UNIMPL-FAST-NEXT:    i64x2.extract_lane $push1=, $1, 1
+; SIMD128-UNIMPL-FAST-NEXT:    i64.shr_u $push3=, $pop2, $pop1
+; SIMD128-UNIMPL-FAST-NEXT:    i64x2.replace_lane $push0=, $pop7, 1, $pop3
+; SIMD128-UNIMPL-FAST-NEXT:    return $pop0
+;
+; NO-SIMD64-SLOW-LABEL: shr_u_vec_v2i64:
+; NO-SIMD64-SLOW:         .functype shr_u_vec_v2i64 (i32, i64, i64, i64, i64) -> ()
+; NO-SIMD64-SLOW-NEXT:    i64.shr_u $push0=, $2, $4
+; NO-SIMD64-SLOW-NEXT:    i64.store 8($0), $pop0
+; NO-SIMD64-SLOW-NEXT:    i64.shr_u $push1=, $1, $3
+; NO-SIMD64-SLOW-NEXT:    i64.store 0($0), $pop1
+; NO-SIMD64-SLOW-NEXT:    return
+;
+; NO-SIMD64-FAST-LABEL: shr_u_vec_v2i64:
+; NO-SIMD64-FAST:         .functype shr_u_vec_v2i64 (i32, i64, i64, i64, i64) -> ()
+; NO-SIMD64-FAST-NEXT:    i64.shr_u $push0=, $1, $3
+; NO-SIMD64-FAST-NEXT:    i64.store 0($0), $pop0
+; NO-SIMD64-FAST-NEXT:    i64.shr_u $push1=, $2, $4
+; NO-SIMD64-FAST-NEXT:    i64.store 8($0), $pop1
+; NO-SIMD64-FAST-NEXT:    return
   %a = lshr <2 x i64> %v, %x
   ret <2 x i64> %a
 }
 
-; CHECK-LABEL: and_v2i64:
-; NO-SIMD128-NOT: v128
-; SIMD128-VM-NOT: v128
-; SIMD128-NEXT: .functype and_v2i64 (v128, v128) -> (v128){{$}}
-; SIMD128-NEXT: v128.and $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <2 x i64> @and_v2i64(<2 x i64> %x, <2 x i64> %y) {
+; SIMD128-UNIMPL-LABEL: and_v2i64:
+; SIMD128-UNIMPL:         .functype and_v2i64 (v128, v128) -> (v128)
+; SIMD128-UNIMPL-NEXT:    v128.and $push0=, $0, $1
+; SIMD128-UNIMPL-NEXT:    return $pop0
+;
+; NO-SIMD64-SLOW-LABEL: and_v2i64:
+; NO-SIMD64-SLOW:         .functype and_v2i64 (i32, i64, i64, i64, i64) -> ()
+; NO-SIMD64-SLOW-NEXT:    i64.and $push0=, $2, $4
+; NO-SIMD64-SLOW-NEXT:    i64.store 8($0), $pop0
+; NO-SIMD64-SLOW-NEXT:    i64.and $push1=, $1, $3
+; NO-SIMD64-SLOW-NEXT:    i64.store 0($0), $pop1
+; NO-SIMD64-SLOW-NEXT:    return
+;
+; NO-SIMD64-FAST-LABEL: and_v2i64:
+; NO-SIMD64-FAST:         .functype and_v2i64 (i32, i64, i64, i64, i64) -> ()
+; NO-SIMD64-FAST-NEXT:    i64.and $push0=, $1, $3
+; NO-SIMD64-FAST-NEXT:    i64.store 0($0), $pop0
+; NO-SIMD64-FAST-NEXT:    i64.and $push1=, $2, $4
+; NO-SIMD64-FAST-NEXT:    i64.store 8($0), $pop1
+; NO-SIMD64-FAST-NEXT:    return
   %a = and <2 x i64> %x, %y
   ret <2 x i64> %a
 }
 
-; CHECK-LABEL: or_v2i64:
-; NO-SIMD128-NOT: v128
-; SIMD128-VM-NOT: v128
-; SIMD128-NEXT: .functype or_v2i64 (v128, v128) -> (v128){{$}}
-; SIMD128-NEXT: v128.or $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <2 x i64> @or_v2i64(<2 x i64> %x, <2 x i64> %y) {
+; SIMD128-UNIMPL-LABEL: or_v2i64:
+; SIMD128-UNIMPL:         .functype or_v2i64 (v128, v128) -> (v128)
+; SIMD128-UNIMPL-NEXT:    v128.or $push0=, $0, $1
+; SIMD128-UNIMPL-NEXT:    return $pop0
+;
+; NO-SIMD64-SLOW-LABEL: or_v2i64:
+; NO-SIMD64-SLOW:         .functype or_v2i64 (i32, i64, i64, i64, i64) -> ()
+; NO-SIMD64-SLOW-NEXT:    i64.or $push0=, $2, $4
+; NO-SIMD64-SLOW-NEXT:    i64.store 8($0), $pop0
+; NO-SIMD64-SLOW-NEXT:    i64.or $push1=, $1, $3
+; NO-SIMD64-SLOW-NEXT:    i64.store 0($0), $pop1
+; NO-SIMD64-SLOW-NEXT:    return
+;
+; NO-SIMD64-FAST-LABEL: or_v2i64:
+; NO-SIMD64-FAST:         .functype or_v2i64 (i32, i64, i64, i64, i64) -> ()
+; NO-SIMD64-FAST-NEXT:    i64.or $push0=, $1, $3
+; NO-SIMD64-FAST-NEXT:    i64.store 0($0), $pop0
+; NO-SIMD64-FAST-NEXT:    i64.or $push1=, $2, $4
+; NO-SIMD64-FAST-NEXT:    i64.store 8($0), $pop1
+; NO-SIMD64-FAST-NEXT:    return
   %a = or <2 x i64> %x, %y
   ret <2 x i64> %a
 }
 
-; CHECK-LABEL: xor_v2i64:
-; NO-SIMD128-NOT: v128
-; SIMD128-VM-NOT: v128
-; SIMD128-NEXT: .functype xor_v2i64 (v128, v128) -> (v128){{$}}
-; SIMD128-NEXT: v128.xor $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <2 x i64> @xor_v2i64(<2 x i64> %x, <2 x i64> %y) {
+; SIMD128-UNIMPL-LABEL: xor_v2i64:
+; SIMD128-UNIMPL:         .functype xor_v2i64 (v128, v128) -> (v128)
+; SIMD128-UNIMPL-NEXT:    v128.xor $push0=, $0, $1
+; SIMD128-UNIMPL-NEXT:    return $pop0
+;
+; NO-SIMD64-SLOW-LABEL: xor_v2i64:
+; NO-SIMD64-SLOW:         .functype xor_v2i64 (i32, i64, i64, i64, i64) -> ()
+; NO-SIMD64-SLOW-NEXT:    i64.xor $push0=, $2, $4
+; NO-SIMD64-SLOW-NEXT:    i64.store 8($0), $pop0
+; NO-SIMD64-SLOW-NEXT:    i64.xor $push1=, $1, $3
+; NO-SIMD64-SLOW-NEXT:    i64.store 0($0), $pop1
+; NO-SIMD64-SLOW-NEXT:    return
+;
+; NO-SIMD64-FAST-LABEL: xor_v2i64:
+; NO-SIMD64-FAST:         .functype xor_v2i64 (i32, i64, i64, i64, i64) -> ()
+; NO-SIMD64-FAST-NEXT:    i64.xor $push0=, $1, $3
+; NO-SIMD64-FAST-NEXT:    i64.store 0($0), $pop0
+; NO-SIMD64-FAST-NEXT:    i64.xor $push1=, $2, $4
+; NO-SIMD64-FAST-NEXT:    i64.store 8($0), $pop1
+; NO-SIMD64-FAST-NEXT:    return
   %a = xor <2 x i64> %x, %y
   ret <2 x i64> %a
 }
 
-; CHECK-LABEL: not_v2i64:
-; NO-SIMD128-NOT: v128
-; SIMD128-VM-NOT: v128
-; SIMD128-NEXT: .functype not_v2i64 (v128) -> (v128){{$}}
-; SIMD128-NEXT: v128.not $push[[R:[0-9]+]]=, $0{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <2 x i64> @not_v2i64(<2 x i64> %x) {
+; SIMD128-UNIMPL-LABEL: not_v2i64:
+; SIMD128-UNIMPL:         .functype not_v2i64 (v128) -> (v128)
+; SIMD128-UNIMPL-NEXT:    v128.not $push0=, $0
+; SIMD128-UNIMPL-NEXT:    return $pop0
+;
+; NO-SIMD64-SLOW-LABEL: not_v2i64:
+; NO-SIMD64-SLOW:         .functype not_v2i64 (i32, i64, i64) -> ()
+; NO-SIMD64-SLOW-NEXT:    i64.const $push0=, -1
+; NO-SIMD64-SLOW-NEXT:    i64.xor $push1=, $2, $pop0
+; NO-SIMD64-SLOW-NEXT:    i64.store 8($0), $pop1
+; NO-SIMD64-SLOW-NEXT:    i64.const $push3=, -1
+; NO-SIMD64-SLOW-NEXT:    i64.xor $push2=, $1, $pop3
+; NO-SIMD64-SLOW-NEXT:    i64.store 0($0), $pop2
+; NO-SIMD64-SLOW-NEXT:    return
+;
+; NO-SIMD64-FAST-LABEL: not_v2i64:
+; NO-SIMD64-FAST:         .functype not_v2i64 (i32, i64, i64) -> ()
+; NO-SIMD64-FAST-NEXT:    i64.const $push0=, -1
+; NO-SIMD64-FAST-NEXT:    i64.xor $push1=, $1, $pop0
+; NO-SIMD64-FAST-NEXT:    i64.store 0($0), $pop1
+; NO-SIMD64-FAST-NEXT:    i64.const $push3=, -1
+; NO-SIMD64-FAST-NEXT:    i64.xor $push2=, $2, $pop3
+; NO-SIMD64-FAST-NEXT:    i64.store 8($0), $pop2
+; NO-SIMD64-FAST-NEXT:    return
   %a = xor <2 x i64> %x, <i64 -1, i64 -1>
   ret <2 x i64> %a
 }
 
-; CHECK-LABEL: bitselect_v2i64:
-; NO-SIMD128-NOT: v128
-; SIMD128-VM-NOT: v128
-; SIMD128-NEXT: .functype bitselect_v2i64 (v128, v128, v128) -> (v128){{$}}
-; SIMD128-SLOW-NEXT: v128.bitselect $push[[R:[0-9]+]]=, $1, $2, $0{{$}}
-; SIMD128-SLOW-NEXT: return $pop[[R]]{{$}}
-; SIMD128-FAST-NEXT: v128.not
-; SIMD128-FAST-NEXT: v128.and
-; SIMD128-FAST-NEXT: v128.and
-; SIMD128-FAST-NEXT: v128.or
-; SIMD128-FAST-NEXT: return
 define <2 x i64> @bitselect_v2i64(<2 x i64> %c, <2 x i64> %v1, <2 x i64> %v2) {
+; SIMD128-UNIMPL-SLOW-LABEL: bitselect_v2i64:
+; SIMD128-UNIMPL-SLOW:         .functype bitselect_v2i64 (v128, v128, v128) -> (v128)
+; SIMD128-UNIMPL-SLOW-NEXT:    v128.bitselect $push0=, $1, $2, $0
+; SIMD128-UNIMPL-SLOW-NEXT:    return $pop0
+;
+; SIMD128-UNIMPL-FAST-LABEL: bitselect_v2i64:
+; SIMD128-UNIMPL-FAST:         .functype bitselect_v2i64 (v128, v128, v128) -> (v128)
+; SIMD128-UNIMPL-FAST-NEXT:    v128.not $push2=, $0
+; SIMD128-UNIMPL-FAST-NEXT:    v128.and $push3=, $2, $pop2
+; SIMD128-UNIMPL-FAST-NEXT:    v128.and $push0=, $1, $0
+; SIMD128-UNIMPL-FAST-NEXT:    v128.or $push1=, $pop3, $pop0
+; SIMD128-UNIMPL-FAST-NEXT:    return $pop1
+;
+; NO-SIMD64-SLOW-LABEL: bitselect_v2i64:
+; NO-SIMD64-SLOW:         .functype bitselect_v2i64 (i32, i64, i64, i64, i64, i64, i64) -> ()
+; NO-SIMD64-SLOW-NEXT:    i64.const $push1=, -1
+; NO-SIMD64-SLOW-NEXT:    i64.xor $push2=, $2, $pop1
+; NO-SIMD64-SLOW-NEXT:    i64.and $push3=, $6, $pop2
+; NO-SIMD64-SLOW-NEXT:    i64.and $push0=, $4, $2
+; NO-SIMD64-SLOW-NEXT:    i64.or $push4=, $pop3, $pop0
+; NO-SIMD64-SLOW-NEXT:    i64.store 8($0), $pop4
+; NO-SIMD64-SLOW-NEXT:    i64.const $push9=, -1
+; NO-SIMD64-SLOW-NEXT:    i64.xor $push6=, $1, $pop9
+; NO-SIMD64-SLOW-NEXT:    i64.and $push7=, $5, $pop6
+; NO-SIMD64-SLOW-NEXT:    i64.and $push5=, $3, $1
+; NO-SIMD64-SLOW-NEXT:    i64.or $push8=, $pop7, $pop5
+; NO-SIMD64-SLOW-NEXT:    i64.store 0($0), $pop8
+; NO-SIMD64-SLOW-NEXT:    return
+;
+; NO-SIMD64-FAST-LABEL: bitselect_v2i64:
+; NO-SIMD64-FAST:         .functype bitselect_v2i64 (i32, i64, i64, i64, i64, i64, i64) -> ()
+; NO-SIMD64-FAST-NEXT:    i64.const $push1=, -1
+; NO-SIMD64-FAST-NEXT:    i64.xor $push2=, $1, $pop1
+; NO-SIMD64-FAST-NEXT:    i64.and $push3=, $5, $pop2
+; NO-SIMD64-FAST-NEXT:    i64.and $push0=, $3, $1
+; NO-SIMD64-FAST-NEXT:    i64.or $push4=, $pop3, $pop0
+; NO-SIMD64-FAST-NEXT:    i64.store 0($0), $pop4
+; NO-SIMD64-FAST-NEXT:    i64.const $push9=, -1
+; NO-SIMD64-FAST-NEXT:    i64.xor $push6=, $2, $pop9
+; NO-SIMD64-FAST-NEXT:    i64.and $push7=, $6, $pop6
+; NO-SIMD64-FAST-NEXT:    i64.and $push5=, $4, $2
+; NO-SIMD64-FAST-NEXT:    i64.or $push8=, $pop7, $pop5
+; NO-SIMD64-FAST-NEXT:    i64.store 8($0), $pop8
+; NO-SIMD64-FAST-NEXT:    return
   %masked_v1 = and <2 x i64> %v1, %c
   %inv_mask = xor <2 x i64> <i64 -1, i64 -1>, %c
   %masked_v2 = and <2 x i64> %v2, %inv_mask
   %a = or <2 x i64> %masked_v2, %masked_v1
   ret <2 x i64> %a
 }
 
 ; ==============================================================================
 ; 4 x float
 ; ==============================================================================
-; CHECK-LABEL: neg_v4f32:
-; NO-SIMD128-NOT: f32x4
-; SIMD128-NEXT: .functype neg_v4f32 (v128) -> (v128){{$}}
-; SIMD128-NEXT: f32x4.neg $push[[R:[0-9]+]]=, $0{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
-define <4 x float> @neg_v4f32(<4 x float> %x) {
-  ; nsz makes this semantically equivalent to flipping sign bit
+
+define <4 x float> @neg_v4f32(<4 x float> %x) {
+; SIMD128-LABEL: neg_v4f32:
+; SIMD128:         .functype neg_v4f32 (v128) -> (v128)
+; SIMD128-NEXT:    f32x4.neg $push0=, $0
+; SIMD128-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: neg_v4f32:
+; NO-SIMD128-SLOW:         .functype neg_v4f32 (i32, f32, f32, f32, f32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push0=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push1=, $0, $pop0
+; NO-SIMD128-SLOW-NEXT:    f32.neg $push2=, $4
+; NO-SIMD128-SLOW-NEXT:    f32.store 0($pop1), $pop2
+; NO-SIMD128-SLOW-NEXT:    f32.neg $push3=, $3
+; NO-SIMD128-SLOW-NEXT:    f32.store 8($0), $pop3
+; NO-SIMD128-SLOW-NEXT:    f32.neg $push4=, $2
+; NO-SIMD128-SLOW-NEXT:    f32.store 4($0), $pop4
+; NO-SIMD128-SLOW-NEXT:    f32.neg $push5=, $1
+; NO-SIMD128-SLOW-NEXT:    f32.store 0($0), $pop5
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: neg_v4f32:
+; NO-SIMD128-FAST:         .functype neg_v4f32 (i32, f32, f32, f32, f32) -> ()
+; NO-SIMD128-FAST-NEXT:    f32.neg $push0=, $1
+; NO-SIMD128-FAST-NEXT:    f32.store 0($0), $pop0
+; NO-SIMD128-FAST-NEXT:    f32.neg $push1=, $2
+; NO-SIMD128-FAST-NEXT:    f32.store 4($0), $pop1
+; NO-SIMD128-FAST-NEXT:    f32.neg $push2=, $3
+; NO-SIMD128-FAST-NEXT:    f32.store 8($0), $pop2
+; NO-SIMD128-FAST-NEXT:    i32.const $push3=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push4=, $0, $pop3
+; NO-SIMD128-FAST-NEXT:    f32.neg $push5=, $4
+; NO-SIMD128-FAST-NEXT:    f32.store 0($pop4), $pop5
+; NO-SIMD128-FAST-NEXT:    return
   %a = fsub nsz <4 x float> <float 0.0, float 0.0, float 0.0, float 0.0>, %x
   ret <4 x float> %a
 }
 
-; CHECK-LABEL: abs_v4f32:
-; NO-SIMD128-NOT: f32x4
-; SIMD128-NEXT: .functype abs_v4f32 (v128) -> (v128){{$}}
-; SIMD128-NEXT: f32x4.abs $push[[R:[0-9]+]]=, $0{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 declare <4 x float> @llvm.fabs.v4f32(<4 x float>) nounwind readnone
 define <4 x float> @abs_v4f32(<4 x float> %x) {
+; SIMD128-LABEL: abs_v4f32:
+; SIMD128:         .functype abs_v4f32 (v128) -> (v128)
+; SIMD128-NEXT:    f32x4.abs $push0=, $0
+; SIMD128-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: abs_v4f32:
+; NO-SIMD128-SLOW:         .functype abs_v4f32 (i32, f32, f32, f32, f32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push0=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push1=, $0, $pop0
+; NO-SIMD128-SLOW-NEXT:    f32.abs $push2=, $4
+; NO-SIMD128-SLOW-NEXT:    f32.store 0($pop1), $pop2
+; NO-SIMD128-SLOW-NEXT:    f32.abs $push3=, $3
+; NO-SIMD128-SLOW-NEXT:    f32.store 8($0), $pop3
+; NO-SIMD128-SLOW-NEXT:    f32.abs $push4=, $2
+; NO-SIMD128-SLOW-NEXT:    f32.store 4($0), $pop4
+; NO-SIMD128-SLOW-NEXT:    f32.abs $push5=, $1
+; NO-SIMD128-SLOW-NEXT:    f32.store 0($0), $pop5
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: abs_v4f32:
+; NO-SIMD128-FAST:         .functype abs_v4f32 (i32, f32, f32, f32, f32) -> ()
+; NO-SIMD128-FAST-NEXT:    f32.abs $push0=, $1
+; NO-SIMD128-FAST-NEXT:    f32.store 0($0), $pop0
+; NO-SIMD128-FAST-NEXT:    f32.abs $push1=, $2
+; NO-SIMD128-FAST-NEXT:    f32.store 4($0), $pop1
+; NO-SIMD128-FAST-NEXT:    f32.abs $push2=, $3
+; NO-SIMD128-FAST-NEXT:    f32.store 8($0), $pop2
+; NO-SIMD128-FAST-NEXT:    i32.const $push3=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push4=, $0, $pop3
+; NO-SIMD128-FAST-NEXT:    f32.abs $push5=, $4
+; NO-SIMD128-FAST-NEXT:    f32.store 0($pop4), $pop5
+; NO-SIMD128-FAST-NEXT:    return
   %a = call <4 x float> @llvm.fabs.v4f32(<4 x float> %x)
   ret <4 x float> %a
 }
 
-; CHECK-LABEL: min_unordered_v4f32:
-; NO-SIMD128-NOT: f32x4
-; SIMD128-NEXT: .functype min_unordered_v4f32 (v128) -> (v128){{$}}
-; SIMD128-NEXT: f32.const $push[[L0:[0-9]+]]=, 0x1.4p2
-; SIMD128-NEXT: f32x4.splat $push[[L1:[0-9]+]]=, $pop[[L0]]
-; SIMD128-NEXT: f32x4.min $push[[R:[0-9]+]]=, $0, $pop[[L1]]{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <4 x float> @min_unordered_v4f32(<4 x float> %x) {
+; SIMD128-SLOW-LABEL: min_unordered_v4f32:
+; SIMD128-SLOW:         .functype min_unordered_v4f32 (v128) -> (v128)
+; SIMD128-SLOW-NEXT:    f32.const $push0=, 0x1.4p2
+; SIMD128-SLOW-NEXT:    f32x4.splat $push1=, $pop0
+; SIMD128-SLOW-NEXT:    f32x4.min $push2=, $0, $pop1
+; SIMD128-SLOW-NEXT:    return $pop2
+;
+; SIMD128-FAST-LABEL: min_unordered_v4f32:
+; SIMD128-FAST:         .functype min_unordered_v4f32 (v128) -> (v128)
+; SIMD128-FAST-NEXT:    f32.const $push1=, 0x1.4p2
+; SIMD128-FAST-NEXT:    f32x4.splat $push2=, $pop1
+; SIMD128-FAST-NEXT:    f32x4.min $push0=, $0, $pop2
+; SIMD128-FAST-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: min_unordered_v4f32:
+; NO-SIMD128-SLOW:         .functype min_unordered_v4f32 (i32, f32, f32, f32, f32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push2=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push3=, $0, $pop2
+; NO-SIMD128-SLOW-NEXT:    f32.const $push0=, 0x1.4p2
+; NO-SIMD128-SLOW-NEXT:    f32.min $push1=, $4, $pop0
+; NO-SIMD128-SLOW-NEXT:    f32.store 0($pop3), $pop1
+; NO-SIMD128-SLOW-NEXT:    f32.const $push9=, 0x1.4p2
+; NO-SIMD128-SLOW-NEXT:    f32.min $push4=, $3, $pop9
+; NO-SIMD128-SLOW-NEXT:    f32.store 8($0), $pop4
+; NO-SIMD128-SLOW-NEXT:    f32.const $push8=, 0x1.4p2
+; NO-SIMD128-SLOW-NEXT:    f32.min $push5=, $2, $pop8
+; NO-SIMD128-SLOW-NEXT:    f32.store 4($0), $pop5
+; NO-SIMD128-SLOW-NEXT:    f32.const $push7=, 0x1.4p2
+; NO-SIMD128-SLOW-NEXT:    f32.min $push6=, $1, $pop7
+; NO-SIMD128-SLOW-NEXT:    f32.store 0($0), $pop6
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: min_unordered_v4f32:
+; NO-SIMD128-FAST:         .functype min_unordered_v4f32 (i32, f32, f32, f32, f32) -> ()
+; NO-SIMD128-FAST-NEXT:    f32.const $push0=, 0x1.4p2
+; NO-SIMD128-FAST-NEXT:    f32.min $push1=, $1, $pop0
+; NO-SIMD128-FAST-NEXT:    f32.store 0($0), $pop1
+; NO-SIMD128-FAST-NEXT:    f32.const $push9=, 0x1.4p2
+; NO-SIMD128-FAST-NEXT:    f32.min $push2=, $2, $pop9
+; NO-SIMD128-FAST-NEXT:    f32.store 4($0), $pop2
+; NO-SIMD128-FAST-NEXT:    f32.const $push8=, 0x1.4p2
+; NO-SIMD128-FAST-NEXT:    f32.min $push3=, $3, $pop8
+; NO-SIMD128-FAST-NEXT:    f32.store 8($0), $pop3
+; NO-SIMD128-FAST-NEXT:    i32.const $push4=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push5=, $0, $pop4
+; NO-SIMD128-FAST-NEXT:    f32.const $push7=, 0x1.4p2
+; NO-SIMD128-FAST-NEXT:    f32.min $push6=, $4, $pop7
+; NO-SIMD128-FAST-NEXT:    f32.store 0($pop5), $pop6
+; NO-SIMD128-FAST-NEXT:    return
   %cmps = fcmp ule <4 x float> %x, <float 5., float 5., float 5., float 5.>
   %a = select <4 x i1> %cmps, <4 x float> %x,
     <4 x float> <float 5., float 5., float 5., float 5.>
   ret <4 x float> %a
 }
 
-; CHECK-LABEL: max_unordered_v4f32:
-; NO-SIMD128-NOT: f32x4
-; SIMD128-NEXT: .functype max_unordered_v4f32 (v128) -> (v128){{$}}
-; SIMD128-NEXT: f32.const $push[[L0:[0-9]+]]=, 0x1.4p2
-; SIMD128-NEXT: f32x4.splat $push[[L1:[0-9]+]]=, $pop[[L0]]
-; SIMD128-NEXT: f32x4.max $push[[R:[0-9]+]]=, $0, $pop[[L1]]{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <4 x float> @max_unordered_v4f32(<4 x float> %x) {
+; SIMD128-SLOW-LABEL: max_unordered_v4f32:
+; SIMD128-SLOW:         .functype max_unordered_v4f32 (v128) -> (v128)
+; SIMD128-SLOW-NEXT:    f32.const $push0=, 0x1.4p2
+; SIMD128-SLOW-NEXT:    f32x4.splat $push1=, $pop0
+; SIMD128-SLOW-NEXT:    f32x4.max $push2=, $0, $pop1
+; SIMD128-SLOW-NEXT:    return $pop2
+;
+; SIMD128-FAST-LABEL: max_unordered_v4f32:
+; SIMD128-FAST:         .functype max_unordered_v4f32 (v128) -> (v128)
+; SIMD128-FAST-NEXT:    f32.const $push1=, 0x1.4p2
+; SIMD128-FAST-NEXT:    f32x4.splat $push2=, $pop1
+; SIMD128-FAST-NEXT:    f32x4.max $push0=, $0, $pop2
+; SIMD128-FAST-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: max_unordered_v4f32:
+; NO-SIMD128-SLOW:         .functype max_unordered_v4f32 (i32, f32, f32, f32, f32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push2=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push3=, $0, $pop2
+; NO-SIMD128-SLOW-NEXT:    f32.const $push0=, 0x1.4p2
+; NO-SIMD128-SLOW-NEXT:    f32.max $push1=, $4, $pop0
+; NO-SIMD128-SLOW-NEXT:    f32.store 0($pop3), $pop1
+; NO-SIMD128-SLOW-NEXT:    f32.const $push9=, 0x1.4p2
+; NO-SIMD128-SLOW-NEXT:    f32.max $push4=, $3, $pop9
+; NO-SIMD128-SLOW-NEXT:    f32.store 8($0), $pop4
+; NO-SIMD128-SLOW-NEXT:    f32.const $push8=, 0x1.4p2
+; NO-SIMD128-SLOW-NEXT:    f32.max $push5=, $2, $pop8
+; NO-SIMD128-SLOW-NEXT:    f32.store 4($0), $pop5
+; NO-SIMD128-SLOW-NEXT:    f32.const $push7=, 0x1.4p2
+; NO-SIMD128-SLOW-NEXT:    f32.max $push6=, $1, $pop7
+; NO-SIMD128-SLOW-NEXT:    f32.store 0($0), $pop6
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: max_unordered_v4f32:
+; NO-SIMD128-FAST:         .functype max_unordered_v4f32 (i32, f32, f32, f32, f32) -> ()
+; NO-SIMD128-FAST-NEXT:    f32.const $push0=, 0x1.4p2
+; NO-SIMD128-FAST-NEXT:    f32.max $push1=, $1, $pop0
+; NO-SIMD128-FAST-NEXT:    f32.store 0($0), $pop1
+; NO-SIMD128-FAST-NEXT:    f32.const $push9=, 0x1.4p2
+; NO-SIMD128-FAST-NEXT:    f32.max $push2=, $2, $pop9
+; NO-SIMD128-FAST-NEXT:    f32.store 4($0), $pop2
+; NO-SIMD128-FAST-NEXT:    f32.const $push8=, 0x1.4p2
+; NO-SIMD128-FAST-NEXT:    f32.max $push3=, $3, $pop8
+; NO-SIMD128-FAST-NEXT:    f32.store 8($0), $pop3
+; NO-SIMD128-FAST-NEXT:    i32.const $push4=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push5=, $0, $pop4
+; NO-SIMD128-FAST-NEXT:    f32.const $push7=, 0x1.4p2
+; NO-SIMD128-FAST-NEXT:    f32.max $push6=, $4, $pop7
+; NO-SIMD128-FAST-NEXT:    f32.store 0($pop5), $pop6
+; NO-SIMD128-FAST-NEXT:    return
   %cmps = fcmp uge <4 x float> %x, <float 5., float 5., float 5., float 5.>
   %a = select <4 x i1> %cmps, <4 x float> %x,
     <4 x float> <float 5., float 5., float 5., float 5.>
   ret <4 x float> %a
 }
 
-; CHECK-LABEL: min_ordered_v4f32:
-; NO-SIMD128-NOT: f32x4
-; SIMD128-NEXT: .functype min_ordered_v4f32 (v128) -> (v128){{$}}
-; SIMD128-NEXT: f32.const $push[[L0:[0-9]+]]=, 0x1.4p2
-; SIMD128-NEXT: f32x4.splat $push[[L1:[0-9]+]]=, $pop[[L0]]
-; SIMD128-NEXT: f32x4.min $push[[R:[0-9]+]]=, $0, $pop[[L1]]{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <4 x float> @min_ordered_v4f32(<4 x float> %x) {
+; SIMD128-SLOW-LABEL: min_ordered_v4f32:
+; SIMD128-SLOW:         .functype min_ordered_v4f32 (v128) -> (v128)
+; SIMD128-SLOW-NEXT:    f32.const $push0=, 0x1.4p2
+; SIMD128-SLOW-NEXT:    f32x4.splat $push1=, $pop0
+; SIMD128-SLOW-NEXT:    f32x4.min $push2=, $0, $pop1
+; SIMD128-SLOW-NEXT:    return $pop2
+;
+; SIMD128-FAST-LABEL: min_ordered_v4f32:
+; SIMD128-FAST:         .functype min_ordered_v4f32 (v128) -> (v128)
+; SIMD128-FAST-NEXT:    f32.const $push1=, 0x1.4p2
+; SIMD128-FAST-NEXT:    f32x4.splat $push2=, $pop1
+; SIMD128-FAST-NEXT:    f32x4.min $push0=, $0, $pop2
+; SIMD128-FAST-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: min_ordered_v4f32:
+; NO-SIMD128-SLOW:         .functype min_ordered_v4f32 (i32, f32, f32, f32, f32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push2=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push3=, $0, $pop2
+; NO-SIMD128-SLOW-NEXT:    f32.const $push0=, 0x1.4p2
+; NO-SIMD128-SLOW-NEXT:    f32.min $push1=, $4, $pop0
+; NO-SIMD128-SLOW-NEXT:    f32.store 0($pop3), $pop1
+; NO-SIMD128-SLOW-NEXT:    f32.const $push9=, 0x1.4p2
+; NO-SIMD128-SLOW-NEXT:    f32.min $push4=, $3, $pop9
+; NO-SIMD128-SLOW-NEXT:    f32.store 8($0), $pop4
+; NO-SIMD128-SLOW-NEXT:    f32.const $push8=, 0x1.4p2
+; NO-SIMD128-SLOW-NEXT:    f32.min $push5=, $2, $pop8
+; NO-SIMD128-SLOW-NEXT:    f32.store 4($0), $pop5
+; NO-SIMD128-SLOW-NEXT:    f32.const $push7=, 0x1.4p2
+; NO-SIMD128-SLOW-NEXT:    f32.min $push6=, $1, $pop7
+; NO-SIMD128-SLOW-NEXT:    f32.store 0($0), $pop6
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: min_ordered_v4f32:
+; NO-SIMD128-FAST:         .functype min_ordered_v4f32 (i32, f32, f32, f32, f32) -> ()
+; NO-SIMD128-FAST-NEXT:    f32.const $push0=, 0x1.4p2
+; NO-SIMD128-FAST-NEXT:    f32.min $push1=, $1, $pop0
+; NO-SIMD128-FAST-NEXT:    f32.store 0($0), $pop1
+; NO-SIMD128-FAST-NEXT:    f32.const $push9=, 0x1.4p2
+; NO-SIMD128-FAST-NEXT:    f32.min $push2=, $2, $pop9
+; NO-SIMD128-FAST-NEXT:    f32.store 4($0), $pop2
+; NO-SIMD128-FAST-NEXT:    f32.const $push8=, 0x1.4p2
+; NO-SIMD128-FAST-NEXT:    f32.min $push3=, $3, $pop8
+; NO-SIMD128-FAST-NEXT:    f32.store 8($0), $pop3
+; NO-SIMD128-FAST-NEXT:    i32.const $push4=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push5=, $0, $pop4
+; NO-SIMD128-FAST-NEXT:    f32.const $push7=, 0x1.4p2
+; NO-SIMD128-FAST-NEXT:    f32.min $push6=, $4, $pop7
+; NO-SIMD128-FAST-NEXT:    f32.store 0($pop5), $pop6
+; NO-SIMD128-FAST-NEXT:    return
   %cmps = fcmp ole <4 x float> <float 5., float 5., float 5., float 5.>, %x
   %a = select <4 x i1> %cmps,
     <4 x float> <float 5., float 5., float 5., float 5.>, <4 x float> %x
   ret <4 x float> %a
 }
 
-; CHECK-LABEL: max_ordered_v4f32:
-; NO-SIMD128-NOT: f32x4
-; SIMD128-NEXT: .functype max_ordered_v4f32 (v128) -> (v128){{$}}
-; SIMD128-NEXT: f32.const $push[[L0:[0-9]+]]=, 0x1.4p2
-; SIMD128-NEXT: f32x4.splat $push[[L1:[0-9]+]]=, $pop[[L0]]
-; SIMD128-NEXT: f32x4.max $push[[R:[0-9]+]]=, $0, $pop[[L1]]{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <4 x float> @max_ordered_v4f32(<4 x float> %x) {
+; SIMD128-SLOW-LABEL: max_ordered_v4f32:
+; SIMD128-SLOW:         .functype max_ordered_v4f32 (v128) -> (v128)
+; SIMD128-SLOW-NEXT:    f32.const $push0=, 0x1.4p2
+; SIMD128-SLOW-NEXT:    f32x4.splat $push1=, $pop0
+; SIMD128-SLOW-NEXT:    f32x4.max $push2=, $0, $pop1
+; SIMD128-SLOW-NEXT:    return $pop2
+;
+; SIMD128-FAST-LABEL: max_ordered_v4f32:
+; SIMD128-FAST:         .functype max_ordered_v4f32 (v128) -> (v128)
+; SIMD128-FAST-NEXT:    f32.const $push1=, 0x1.4p2
+; SIMD128-FAST-NEXT:    f32x4.splat $push2=, $pop1
+; SIMD128-FAST-NEXT:    f32x4.max $push0=, $0, $pop2
+; SIMD128-FAST-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: max_ordered_v4f32:
+; NO-SIMD128-SLOW:         .functype max_ordered_v4f32 (i32, f32, f32, f32, f32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push2=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push3=, $0, $pop2
+; NO-SIMD128-SLOW-NEXT:    f32.const $push0=, 0x1.4p2
+; NO-SIMD128-SLOW-NEXT:    f32.max $push1=, $4, $pop0
+; NO-SIMD128-SLOW-NEXT:    f32.store 0($pop3), $pop1
+; NO-SIMD128-SLOW-NEXT:    f32.const $push9=, 0x1.4p2
+; NO-SIMD128-SLOW-NEXT:    f32.max $push4=, $3, $pop9
+; NO-SIMD128-SLOW-NEXT:    f32.store 8($0), $pop4
+; NO-SIMD128-SLOW-NEXT:    f32.const $push8=, 0x1.4p2
+; NO-SIMD128-SLOW-NEXT:    f32.max $push5=, $2, $pop8
+; NO-SIMD128-SLOW-NEXT:    f32.store 4($0), $pop5
+; NO-SIMD128-SLOW-NEXT:    f32.const $push7=, 0x1.4p2
+; NO-SIMD128-SLOW-NEXT:    f32.max $push6=, $1, $pop7
+; NO-SIMD128-SLOW-NEXT:    f32.store 0($0), $pop6
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: max_ordered_v4f32:
+; NO-SIMD128-FAST:         .functype max_ordered_v4f32 (i32, f32, f32, f32, f32) -> ()
+; NO-SIMD128-FAST-NEXT:    f32.const $push0=, 0x1.4p2
+; NO-SIMD128-FAST-NEXT:    f32.max $push1=, $1, $pop0
+; NO-SIMD128-FAST-NEXT:    f32.store 0($0), $pop1
+; NO-SIMD128-FAST-NEXT:    f32.const $push9=, 0x1.4p2
+; NO-SIMD128-FAST-NEXT:    f32.max $push2=, $2, $pop9
+; NO-SIMD128-FAST-NEXT:    f32.store 4($0), $pop2
+; NO-SIMD128-FAST-NEXT:    f32.const $push8=, 0x1.4p2
+; NO-SIMD128-FAST-NEXT:    f32.max $push3=, $3, $pop8
+; NO-SIMD128-FAST-NEXT:    f32.store 8($0), $pop3
+; NO-SIMD128-FAST-NEXT:    i32.const $push4=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push5=, $0, $pop4
+; NO-SIMD128-FAST-NEXT:    f32.const $push7=, 0x1.4p2
+; NO-SIMD128-FAST-NEXT:    f32.max $push6=, $4, $pop7
+; NO-SIMD128-FAST-NEXT:    f32.store 0($pop5), $pop6
+; NO-SIMD128-FAST-NEXT:    return
   %cmps = fcmp oge <4 x float> <float 5., float 5., float 5., float 5.>, %x
   %a = select <4 x i1> %cmps,
     <4 x float> <float 5., float 5., float 5., float 5.>, <4 x float> %x
   ret <4 x float> %a
 }
 
-; CHECK-LABEL: min_intrinsic_v4f32:
-; NO-SIMD128-NOT: f32x4
-; SIMD128-NEXT: .functype min_intrinsic_v4f32 (v128, v128) -> (v128){{$}}
-; SIMD128-NEXT: f32x4.min $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 declare <4 x float> @llvm.minimum.v4f32(<4 x float>, <4 x float>)
 define <4 x float> @min_intrinsic_v4f32(<4 x float> %x, <4 x float> %y) {
+; SIMD128-LABEL: min_intrinsic_v4f32:
+; SIMD128:         .functype min_intrinsic_v4f32 (v128, v128) -> (v128)
+; SIMD128-NEXT:    f32x4.min $push0=, $0, $1
+; SIMD128-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: min_intrinsic_v4f32:
+; NO-SIMD128-SLOW:         .functype min_intrinsic_v4f32 (i32, f32, f32, f32, f32, f32, f32, f32, f32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push1=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push2=, $0, $pop1
+; NO-SIMD128-SLOW-NEXT:    f32.min $push0=, $4, $8
+; NO-SIMD128-SLOW-NEXT:    f32.store 0($pop2), $pop0
+; NO-SIMD128-SLOW-NEXT:    f32.min $push3=, $3, $7
+; NO-SIMD128-SLOW-NEXT:    f32.store 8($0), $pop3
+; NO-SIMD128-SLOW-NEXT:    f32.min $push4=, $2, $6
+; NO-SIMD128-SLOW-NEXT:    f32.store 4($0), $pop4
+; NO-SIMD128-SLOW-NEXT:    f32.min $push5=, $1, $5
+; NO-SIMD128-SLOW-NEXT:    f32.store 0($0), $pop5
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: min_intrinsic_v4f32:
+; NO-SIMD128-FAST:         .functype min_intrinsic_v4f32 (i32, f32, f32, f32, f32, f32, f32, f32, f32) -> ()
+; NO-SIMD128-FAST-NEXT:    f32.min $push0=, $1, $5
+; NO-SIMD128-FAST-NEXT:    f32.store 0($0), $pop0
+; NO-SIMD128-FAST-NEXT:    f32.min $push1=, $2, $6
+; NO-SIMD128-FAST-NEXT:    f32.store 4($0), $pop1
+; NO-SIMD128-FAST-NEXT:    f32.min $push2=, $3, $7
+; NO-SIMD128-FAST-NEXT:    f32.store 8($0), $pop2
+; NO-SIMD128-FAST-NEXT:    i32.const $push3=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push4=, $0, $pop3
+; NO-SIMD128-FAST-NEXT:    f32.min $push5=, $4, $8
+; NO-SIMD128-FAST-NEXT:    f32.store 0($pop4), $pop5
+; NO-SIMD128-FAST-NEXT:    return
   %a = call <4 x float> @llvm.minimum.v4f32(<4 x float> %x, <4 x float> %y)
   ret <4 x float> %a
 }
 
-; CHECK-LABEL: minnum_intrinsic_v4f32:
-; NO-SIMD128-NOT: f32x4
-; SIMD128-NEXT: .functype minnum_intrinsic_v4f32 (v128, v128) -> (v128){{$}}
-; SIMD128-NEXT: f32x4.min $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 declare <4 x float> @llvm.minnum.v4f32(<4 x float>, <4 x float>)
 define <4 x float> @minnum_intrinsic_v4f32(<4 x float> %x, <4 x float> %y) {
+; SIMD128-LABEL: minnum_intrinsic_v4f32:
+; SIMD128:         .functype minnum_intrinsic_v4f32 (v128, v128) -> (v128)
+; SIMD128-NEXT:    f32x4.min $push0=, $0, $1
+; SIMD128-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: minnum_intrinsic_v4f32:
+; NO-SIMD128-SLOW:         .functype minnum_intrinsic_v4f32 (i32, f32, f32, f32, f32, f32, f32, f32, f32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push1=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push2=, $0, $pop1
+; NO-SIMD128-SLOW-NEXT:    f32.call $push0=, fminf, $4, $8
+; NO-SIMD128-SLOW-NEXT:    f32.store 0($pop2), $pop0
+; NO-SIMD128-SLOW-NEXT:    f32.call $push3=, fminf, $3, $7
+; NO-SIMD128-SLOW-NEXT:    f32.store 8($0), $pop3
+; NO-SIMD128-SLOW-NEXT:    f32.call $push4=, fminf, $2, $6
+; NO-SIMD128-SLOW-NEXT:    f32.store 4($0), $pop4
+; NO-SIMD128-SLOW-NEXT:    f32.call $push5=, fminf, $1, $5
+; NO-SIMD128-SLOW-NEXT:    f32.store 0($0), $pop5
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: minnum_intrinsic_v4f32:
+; NO-SIMD128-FAST:         .functype minnum_intrinsic_v4f32 (i32, f32, f32, f32, f32, f32, f32, f32, f32) -> ()
+; NO-SIMD128-FAST-NEXT:    f32.call $push0=, fminf, $1, $5
+; NO-SIMD128-FAST-NEXT:    f32.store 0($0), $pop0
+; NO-SIMD128-FAST-NEXT:    f32.call $push1=, fminf, $2, $6
+; NO-SIMD128-FAST-NEXT:    f32.store 4($0), $pop1
+; NO-SIMD128-FAST-NEXT:    f32.call $push2=, fminf, $3, $7
+; NO-SIMD128-FAST-NEXT:    f32.store 8($0), $pop2
+; NO-SIMD128-FAST-NEXT:    i32.const $push3=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push4=, $0, $pop3
+; NO-SIMD128-FAST-NEXT:    f32.call $push5=, fminf, $4, $8
+; NO-SIMD128-FAST-NEXT:    f32.store 0($pop4), $pop5
+; NO-SIMD128-FAST-NEXT:    return
   %a = call nnan <4 x float> @llvm.minnum.v4f32(<4 x float> %x, <4 x float> %y)
   ret <4 x float> %a
 }
 
-; CHECK-LABEL: max_intrinsic_v4f32:
-; NO-SIMD128-NOT: f32x4
-; SIMD128-NEXT: .functype max_intrinsic_v4f32 (v128, v128) -> (v128){{$}}
-; SIMD128-NEXT: f32x4.max $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 declare <4 x float> @llvm.maximum.v4f32(<4 x float>, <4 x float>)
 define <4 x float> @max_intrinsic_v4f32(<4 x float> %x, <4 x float> %y) {
+; SIMD128-LABEL: max_intrinsic_v4f32:
+; SIMD128:         .functype max_intrinsic_v4f32 (v128, v128) -> (v128)
+; SIMD128-NEXT:    f32x4.max $push0=, $0, $1
+; SIMD128-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: max_intrinsic_v4f32:
+; NO-SIMD128-SLOW:         .functype max_intrinsic_v4f32 (i32, f32, f32, f32, f32, f32, f32, f32, f32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push1=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push2=, $0, $pop1
+; NO-SIMD128-SLOW-NEXT:    f32.max $push0=, $4, $8
+; NO-SIMD128-SLOW-NEXT:    f32.store 0($pop2), $pop0
+; NO-SIMD128-SLOW-NEXT:    f32.max $push3=, $3, $7
+; NO-SIMD128-SLOW-NEXT:    f32.store 8($0), $pop3
+; NO-SIMD128-SLOW-NEXT:    f32.max $push4=, $2, $6
+; NO-SIMD128-SLOW-NEXT:    f32.store 4($0), $pop4
+; NO-SIMD128-SLOW-NEXT:    f32.max $push5=, $1, $5
+; NO-SIMD128-SLOW-NEXT:    f32.store 0($0), $pop5
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: max_intrinsic_v4f32:
+; NO-SIMD128-FAST:         .functype max_intrinsic_v4f32 (i32, f32, f32, f32, f32, f32, f32, f32, f32) -> ()
+; NO-SIMD128-FAST-NEXT:    f32.max $push0=, $1, $5
+; NO-SIMD128-FAST-NEXT:    f32.store 0($0), $pop0
+; NO-SIMD128-FAST-NEXT:    f32.max $push1=, $2, $6
+; NO-SIMD128-FAST-NEXT:    f32.store 4($0), $pop1
+; NO-SIMD128-FAST-NEXT:    f32.max $push2=, $3, $7
+; NO-SIMD128-FAST-NEXT:    f32.store 8($0), $pop2
+; NO-SIMD128-FAST-NEXT:    i32.const $push3=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push4=, $0, $pop3
+; NO-SIMD128-FAST-NEXT:    f32.max $push5=, $4, $8
+; NO-SIMD128-FAST-NEXT:    f32.store 0($pop4), $pop5
+; NO-SIMD128-FAST-NEXT:    return
   %a = call <4 x float> @llvm.maximum.v4f32(<4 x float> %x, <4 x float> %y)
   ret <4 x float> %a
 }
 
-; CHECK-LABEL: maxnum_intrinsic_v4f32:
-; NO-SIMD128-NOT: f32x4
-; SIMD128-NEXT: .functype maxnum_intrinsic_v4f32 (v128, v128) -> (v128){{$}}
-; SIMD128-NEXT: f32x4.max $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 declare <4 x float> @llvm.maxnum.v4f32(<4 x float>, <4 x float>)
 define <4 x float> @maxnum_intrinsic_v4f32(<4 x float> %x, <4 x float> %y) {
+; SIMD128-LABEL: maxnum_intrinsic_v4f32:
+; SIMD128:         .functype maxnum_intrinsic_v4f32 (v128, v128) -> (v128)
+; SIMD128-NEXT:    f32x4.max $push0=, $0, $1
+; SIMD128-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: maxnum_intrinsic_v4f32:
+; NO-SIMD128-SLOW:         .functype maxnum_intrinsic_v4f32 (i32, f32, f32, f32, f32, f32, f32, f32, f32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push1=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push2=, $0, $pop1
+; NO-SIMD128-SLOW-NEXT:    f32.call $push0=, fmaxf, $4, $8
+; NO-SIMD128-SLOW-NEXT:    f32.store 0($pop2), $pop0
+; NO-SIMD128-SLOW-NEXT:    f32.call $push3=, fmaxf, $3, $7
+; NO-SIMD128-SLOW-NEXT:    f32.store 8($0), $pop3
+; NO-SIMD128-SLOW-NEXT:    f32.call $push4=, fmaxf, $2, $6
+; NO-SIMD128-SLOW-NEXT:    f32.store 4($0), $pop4
+; NO-SIMD128-SLOW-NEXT:    f32.call $push5=, fmaxf, $1, $5
+; NO-SIMD128-SLOW-NEXT:    f32.store 0($0), $pop5
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: maxnum_intrinsic_v4f32:
+; NO-SIMD128-FAST:         .functype maxnum_intrinsic_v4f32 (i32, f32, f32, f32, f32, f32, f32, f32, f32) -> ()
+; NO-SIMD128-FAST-NEXT:    f32.call $push0=, fmaxf, $1, $5
+; NO-SIMD128-FAST-NEXT:    f32.store 0($0), $pop0
+; NO-SIMD128-FAST-NEXT:    f32.call $push1=, fmaxf, $2, $6
+; NO-SIMD128-FAST-NEXT:    f32.store 4($0), $pop1
+; NO-SIMD128-FAST-NEXT:    f32.call $push2=, fmaxf, $3, $7
+; NO-SIMD128-FAST-NEXT:    f32.store 8($0), $pop2
+; NO-SIMD128-FAST-NEXT:    i32.const $push3=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push4=, $0, $pop3
+; NO-SIMD128-FAST-NEXT:    f32.call $push5=, fmaxf, $4, $8
+; NO-SIMD128-FAST-NEXT:    f32.store 0($pop4), $pop5
+; NO-SIMD128-FAST-NEXT:    return
   %a = call nnan <4 x float> @llvm.maxnum.v4f32(<4 x float> %x, <4 x float> %y)
   ret <4 x float> %a
 }
 
-; CHECK-LABEL: min_const_intrinsic_v4f32:
-; NO-SIMD128-NOT: f32x4
-; SIMD128-NEXT: .functype min_const_intrinsic_v4f32 () -> (v128){{$}}
-; SIMD128-NEXT: f32.const $push[[L:[0-9]+]]=, 0x1.4p2{{$}}
-; SIMD128-NEXT: f32x4.splat $push[[R:[0-9]+]]=, $pop[[L]]{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <4 x float> @min_const_intrinsic_v4f32() {
+; SIMD128-SLOW-LABEL: min_const_intrinsic_v4f32:
+; SIMD128-SLOW:         .functype min_const_intrinsic_v4f32 () -> (v128)
+; SIMD128-SLOW-NEXT:    f32.const $push0=, 0x1.4p2
+; SIMD128-SLOW-NEXT:    f32x4.splat $push1=, $pop0
+; SIMD128-SLOW-NEXT:    return $pop1
+;
+; SIMD128-FAST-LABEL: min_const_intrinsic_v4f32:
+; SIMD128-FAST:         .functype min_const_intrinsic_v4f32 () -> (v128)
+; SIMD128-FAST-NEXT:    f32.const $push1=, 0x1.4p2
+; SIMD128-FAST-NEXT:    f32x4.splat $push0=, $pop1
+; SIMD128-FAST-NEXT:    return $pop0
+;
+; NO-SIMD128-LABEL: min_const_intrinsic_v4f32:
+; NO-SIMD128:         .functype min_const_intrinsic_v4f32 (i32) -> ()
+; NO-SIMD128-NEXT:    i64.const $push0=, 4656722015785320448
+; NO-SIMD128-NEXT:    i64.store 8($0), $pop0
+; NO-SIMD128-NEXT:    i64.const $push1=, 4656722015785320448
+; NO-SIMD128-NEXT:    i64.store 0($0), $pop1
+; NO-SIMD128-NEXT:    return
   %a = call <4 x float> @llvm.minimum.v4f32(
     <4 x float> <float 42., float 42., float 42., float 42.>,
     <4 x float> <float 5., float 5., float 5., float 5.>
   )
   ret <4 x float> %a
 }
 
-; CHECK-LABEL: max_const_intrinsic_v4f32:
-; NO-SIMD128-NOT: f32x4
-; SIMD128-NEXT: .functype max_const_intrinsic_v4f32 () -> (v128){{$}}
-; SIMD128-NEXT: f32.const $push[[L:[0-9]+]]=, 0x1.5p5{{$}}
-; SIMD128-NEXT: f32x4.splat $push[[R:[0-9]+]]=, $pop[[L]]{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <4 x float> @max_const_intrinsic_v4f32() {
+; SIMD128-SLOW-LABEL: max_const_intrinsic_v4f32:
+; SIMD128-SLOW:         .functype max_const_intrinsic_v4f32 () -> (v128)
+; SIMD128-SLOW-NEXT:    f32.const $push0=, 0x1.5p5
+; SIMD128-SLOW-NEXT:    f32x4.splat $push1=, $pop0
+; SIMD128-SLOW-NEXT:    return $pop1
+;
+; SIMD128-FAST-LABEL: max_const_intrinsic_v4f32:
+; SIMD128-FAST:         .functype max_const_intrinsic_v4f32 () -> (v128)
+; SIMD128-FAST-NEXT:    f32.const $push1=, 0x1.5p5
+; SIMD128-FAST-NEXT:    f32x4.splat $push0=, $pop1
+; SIMD128-FAST-NEXT:    return $pop0
+;
+; NO-SIMD128-LABEL: max_const_intrinsic_v4f32:
+; NO-SIMD128:         .functype max_const_intrinsic_v4f32 (i32) -> ()
+; NO-SIMD128-NEXT:    i64.const $push0=, 4767060206681587712
+; NO-SIMD128-NEXT:    i64.store 8($0), $pop0
+; NO-SIMD128-NEXT:    i64.const $push1=, 4767060206681587712
+; NO-SIMD128-NEXT:    i64.store 0($0), $pop1
+; NO-SIMD128-NEXT:    return
   %a = call <4 x float> @llvm.maximum.v4f32(
     <4 x float> <float 42., float 42., float 42., float 42.>,
     <4 x float> <float 5., float 5., float 5., float 5.>
   )
   ret <4 x float> %a
 }
 
-; CHECK-LABEL: add_v4f32:
-; NO-SIMD128-NOT: f32x4
-; SIMD128-NEXT: .functype add_v4f32 (v128, v128) -> (v128){{$}}
-; SIMD128-NEXT: f32x4.add $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <4 x float> @add_v4f32(<4 x float> %x, <4 x float> %y) {
+; SIMD128-LABEL: add_v4f32:
+; SIMD128:         .functype add_v4f32 (v128, v128) -> (v128)
+; SIMD128-NEXT:    f32x4.add $push0=, $0, $1
+; SIMD128-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: add_v4f32:
+; NO-SIMD128-SLOW:         .functype add_v4f32 (i32, f32, f32, f32, f32, f32, f32, f32, f32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push1=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push2=, $0, $pop1
+; NO-SIMD128-SLOW-NEXT:    f32.add $push0=, $4, $8
+; NO-SIMD128-SLOW-NEXT:    f32.store 0($pop2), $pop0
+; NO-SIMD128-SLOW-NEXT:    f32.add $push3=, $3, $7
+; NO-SIMD128-SLOW-NEXT:    f32.store 8($0), $pop3
+; NO-SIMD128-SLOW-NEXT:    f32.add $push4=, $2, $6
+; NO-SIMD128-SLOW-NEXT:    f32.store 4($0), $pop4
+; NO-SIMD128-SLOW-NEXT:    f32.add $push5=, $1, $5
+; NO-SIMD128-SLOW-NEXT:    f32.store 0($0), $pop5
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: add_v4f32:
+; NO-SIMD128-FAST:         .functype add_v4f32 (i32, f32, f32, f32, f32, f32, f32, f32, f32) -> ()
+; NO-SIMD128-FAST-NEXT:    f32.add $push0=, $1, $5
+; NO-SIMD128-FAST-NEXT:    f32.store 0($0), $pop0
+; NO-SIMD128-FAST-NEXT:    f32.add $push1=, $2, $6
+; NO-SIMD128-FAST-NEXT:    f32.store 4($0), $pop1
+; NO-SIMD128-FAST-NEXT:    f32.add $push2=, $3, $7
+; NO-SIMD128-FAST-NEXT:    f32.store 8($0), $pop2
+; NO-SIMD128-FAST-NEXT:    i32.const $push3=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push4=, $0, $pop3
+; NO-SIMD128-FAST-NEXT:    f32.add $push5=, $4, $8
+; NO-SIMD128-FAST-NEXT:    f32.store 0($pop4), $pop5
+; NO-SIMD128-FAST-NEXT:    return
   %a = fadd <4 x float> %x, %y
   ret <4 x float> %a
 }
 
-; CHECK-LABEL: sub_v4f32:
-; NO-SIMD128-NOT: f32x4
-; SIMD128-NEXT: .functype sub_v4f32 (v128, v128) -> (v128){{$}}
-; SIMD128-NEXT: f32x4.sub $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <4 x float> @sub_v4f32(<4 x float> %x, <4 x float> %y) {
+; SIMD128-LABEL: sub_v4f32:
+; SIMD128:         .functype sub_v4f32 (v128, v128) -> (v128)
+; SIMD128-NEXT:    f32x4.sub $push0=, $0, $1
+; SIMD128-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: sub_v4f32:
+; NO-SIMD128-SLOW:         .functype sub_v4f32 (i32, f32, f32, f32, f32, f32, f32, f32, f32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push1=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push2=, $0, $pop1
+; NO-SIMD128-SLOW-NEXT:    f32.sub $push0=, $4, $8
+; NO-SIMD128-SLOW-NEXT:    f32.store 0($pop2), $pop0
+; NO-SIMD128-SLOW-NEXT:    f32.sub $push3=, $3, $7
+; NO-SIMD128-SLOW-NEXT:    f32.store 8($0), $pop3
+; NO-SIMD128-SLOW-NEXT:    f32.sub $push4=, $2, $6
+; NO-SIMD128-SLOW-NEXT:    f32.store 4($0), $pop4
+; NO-SIMD128-SLOW-NEXT:    f32.sub $push5=, $1, $5
+; NO-SIMD128-SLOW-NEXT:    f32.store 0($0), $pop5
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: sub_v4f32:
+; NO-SIMD128-FAST:         .functype sub_v4f32 (i32, f32, f32, f32, f32, f32, f32, f32, f32) -> ()
+; NO-SIMD128-FAST-NEXT:    f32.sub $push0=, $1, $5
+; NO-SIMD128-FAST-NEXT:    f32.store 0($0), $pop0
+; NO-SIMD128-FAST-NEXT:    f32.sub $push1=, $2, $6
+; NO-SIMD128-FAST-NEXT:    f32.store 4($0), $pop1
+; NO-SIMD128-FAST-NEXT:    f32.sub $push2=, $3, $7
+; NO-SIMD128-FAST-NEXT:    f32.store 8($0), $pop2
+; NO-SIMD128-FAST-NEXT:    i32.const $push3=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push4=, $0, $pop3
+; NO-SIMD128-FAST-NEXT:    f32.sub $push5=, $4, $8
+; NO-SIMD128-FAST-NEXT:    f32.store 0($pop4), $pop5
+; NO-SIMD128-FAST-NEXT:    return
   %a = fsub <4 x float> %x, %y
   ret <4 x float> %a
 }
 
-; CHECK-LABEL: div_v4f32:
-; NO-SIMD128-NOT: f32x4
-; SIMD128-VM-NOT: f32x4.div
-; SIMD128-NEXT: .functype div_v4f32 (v128, v128) -> (v128){{$}}
-; SIMD128-NEXT: f32x4.div $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <4 x float> @div_v4f32(<4 x float> %x, <4 x float> %y) {
+; SIMD128-UNIMPL-LABEL: div_v4f32:
+; SIMD128-UNIMPL:         .functype div_v4f32 (v128, v128) -> (v128)
+; SIMD128-UNIMPL-NEXT:    f32x4.div $push0=, $0, $1
+; SIMD128-UNIMPL-NEXT:    return $pop0
+;
+; SIMD128-VM-SLOW-LABEL: div_v4f32:
+; SIMD128-VM-SLOW:         .functype div_v4f32 (v128, v128) -> (v128)
+; SIMD128-VM-SLOW-NEXT:    f32x4.extract_lane $push4=, $0, 0
+; SIMD128-VM-SLOW-NEXT:    f32x4.extract_lane $push3=, $1, 0
+; SIMD128-VM-SLOW-NEXT:    f32.div $push5=, $pop4, $pop3
+; SIMD128-VM-SLOW-NEXT:    f32x4.splat $push6=, $pop5
+; SIMD128-VM-SLOW-NEXT:    f32x4.extract_lane $push1=, $0, 1
+; SIMD128-VM-SLOW-NEXT:    f32x4.extract_lane $push0=, $1, 1
+; SIMD128-VM-SLOW-NEXT:    f32.div $push2=, $pop1, $pop0
+; SIMD128-VM-SLOW-NEXT:    f32x4.replace_lane $push7=, $pop6, 1, $pop2
+; SIMD128-VM-SLOW-NEXT:    f32x4.extract_lane $push9=, $0, 2
+; SIMD128-VM-SLOW-NEXT:    f32x4.extract_lane $push8=, $1, 2
+; SIMD128-VM-SLOW-NEXT:    f32.div $push10=, $pop9, $pop8
+; SIMD128-VM-SLOW-NEXT:    f32x4.replace_lane $push11=, $pop7, 2, $pop10
+; SIMD128-VM-SLOW-NEXT:    f32x4.extract_lane $push13=, $0, 3
+; SIMD128-VM-SLOW-NEXT:    f32x4.extract_lane $push12=, $1, 3
+; SIMD128-VM-SLOW-NEXT:    f32.div $push14=, $pop13, $pop12
+; SIMD128-VM-SLOW-NEXT:    f32x4.replace_lane $push15=, $pop11, 3, $pop14
+; SIMD128-VM-SLOW-NEXT:    return $pop15
+;
+; SIMD128-VM-FAST-LABEL: div_v4f32:
+; SIMD128-VM-FAST:         .functype div_v4f32 (v128, v128) -> (v128)
+; SIMD128-VM-FAST-NEXT:    f32x4.extract_lane $push5=, $0, 0
+; SIMD128-VM-FAST-NEXT:    f32x4.extract_lane $push4=, $1, 0
+; SIMD128-VM-FAST-NEXT:    f32.div $push6=, $pop5, $pop4
+; SIMD128-VM-FAST-NEXT:    f32x4.splat $push7=, $pop6
+; SIMD128-VM-FAST-NEXT:    f32x4.extract_lane $push2=, $0, 1
+; SIMD128-VM-FAST-NEXT:    f32x4.extract_lane $push1=, $1, 1
+; SIMD128-VM-FAST-NEXT:    f32.div $push3=, $pop2, $pop1
+; SIMD128-VM-FAST-NEXT:    f32x4.replace_lane $push8=, $pop7, 1, $pop3
+; SIMD128-VM-FAST-NEXT:    f32x4.extract_lane $push10=, $0, 2
+; SIMD128-VM-FAST-NEXT:    f32x4.extract_lane $push9=, $1, 2
+; SIMD128-VM-FAST-NEXT:    f32.div $push11=, $pop10, $pop9
+; SIMD128-VM-FAST-NEXT:    f32x4.replace_lane $push12=, $pop8, 2, $pop11
+; SIMD128-VM-FAST-NEXT:    f32x4.extract_lane $push14=, $0, 3
+; SIMD128-VM-FAST-NEXT:    f32x4.extract_lane $push13=, $1, 3
+; SIMD128-VM-FAST-NEXT:    f32.div $push15=, $pop14, $pop13
+; SIMD128-VM-FAST-NEXT:    f32x4.replace_lane $push0=, $pop12, 3, $pop15
+; SIMD128-VM-FAST-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: div_v4f32:
+; NO-SIMD128-SLOW:         .functype div_v4f32 (i32, f32, f32, f32, f32, f32, f32, f32, f32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push1=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push2=, $0, $pop1
+; NO-SIMD128-SLOW-NEXT:    f32.div $push0=, $4, $8
+; NO-SIMD128-SLOW-NEXT:    f32.store 0($pop2), $pop0
+; NO-SIMD128-SLOW-NEXT:    f32.div $push3=, $3, $7
+; NO-SIMD128-SLOW-NEXT:    f32.store 8($0), $pop3
+; NO-SIMD128-SLOW-NEXT:    f32.div $push4=, $2, $6
+; NO-SIMD128-SLOW-NEXT:    f32.store 4($0), $pop4
+; NO-SIMD128-SLOW-NEXT:    f32.div $push5=, $1, $5
+; NO-SIMD128-SLOW-NEXT:    f32.store 0($0), $pop5
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: div_v4f32:
+; NO-SIMD128-FAST:         .functype div_v4f32 (i32, f32, f32, f32, f32, f32, f32, f32, f32) -> ()
+; NO-SIMD128-FAST-NEXT:    f32.div $push0=, $1, $5
+; NO-SIMD128-FAST-NEXT:    f32.store 0($0), $pop0
+; NO-SIMD128-FAST-NEXT:    f32.div $push1=, $2, $6
+; NO-SIMD128-FAST-NEXT:    f32.store 4($0), $pop1
+; NO-SIMD128-FAST-NEXT:    f32.div $push2=, $3, $7
+; NO-SIMD128-FAST-NEXT:    f32.store 8($0), $pop2
+; NO-SIMD128-FAST-NEXT:    i32.const $push3=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push4=, $0, $pop3
+; NO-SIMD128-FAST-NEXT:    f32.div $push5=, $4, $8
+; NO-SIMD128-FAST-NEXT:    f32.store 0($pop4), $pop5
+; NO-SIMD128-FAST-NEXT:    return
   %a = fdiv <4 x float> %x, %y
   ret <4 x float> %a
 }
 
-; CHECK-LABEL: mul_v4f32:
-; NO-SIMD128-NOT: f32x4
-; SIMD128-NEXT: .functype mul_v4f32 (v128, v128) -> (v128){{$}}
-; SIMD128-NEXT: f32x4.mul $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <4 x float> @mul_v4f32(<4 x float> %x, <4 x float> %y) {
+; SIMD128-LABEL: mul_v4f32:
+; SIMD128:         .functype mul_v4f32 (v128, v128) -> (v128)
+; SIMD128-NEXT:    f32x4.mul $push0=, $0, $1
+; SIMD128-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: mul_v4f32:
+; NO-SIMD128-SLOW:         .functype mul_v4f32 (i32, f32, f32, f32, f32, f32, f32, f32, f32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push1=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push2=, $0, $pop1
+; NO-SIMD128-SLOW-NEXT:    f32.mul $push0=, $4, $8
+; NO-SIMD128-SLOW-NEXT:    f32.store 0($pop2), $pop0
+; NO-SIMD128-SLOW-NEXT:    f32.mul $push3=, $3, $7
+; NO-SIMD128-SLOW-NEXT:    f32.store 8($0), $pop3
+; NO-SIMD128-SLOW-NEXT:    f32.mul $push4=, $2, $6
+; NO-SIMD128-SLOW-NEXT:    f32.store 4($0), $pop4
+; NO-SIMD128-SLOW-NEXT:    f32.mul $push5=, $1, $5
+; NO-SIMD128-SLOW-NEXT:    f32.store 0($0), $pop5
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: mul_v4f32:
+; NO-SIMD128-FAST:         .functype mul_v4f32 (i32, f32, f32, f32, f32, f32, f32, f32, f32) -> ()
+; NO-SIMD128-FAST-NEXT:    f32.mul $push0=, $1, $5
+; NO-SIMD128-FAST-NEXT:    f32.store 0($0), $pop0
+; NO-SIMD128-FAST-NEXT:    f32.mul $push1=, $2, $6
+; NO-SIMD128-FAST-NEXT:    f32.store 4($0), $pop1
+; NO-SIMD128-FAST-NEXT:    f32.mul $push2=, $3, $7
+; NO-SIMD128-FAST-NEXT:    f32.store 8($0), $pop2
+; NO-SIMD128-FAST-NEXT:    i32.const $push3=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push4=, $0, $pop3
+; NO-SIMD128-FAST-NEXT:    f32.mul $push5=, $4, $8
+; NO-SIMD128-FAST-NEXT:    f32.store 0($pop4), $pop5
+; NO-SIMD128-FAST-NEXT:    return
   %a = fmul <4 x float> %x, %y
   ret <4 x float> %a
 }
 
-; CHECK-LABEL: sqrt_v4f32:
-; NO-SIMD128-NOT: f32x4
-; SIMD128-VM-NOT: f32x4.sqrt
-; SIMD128-NEXT: .functype sqrt_v4f32 (v128) -> (v128){{$}}
-; SIMD128-NEXT: f32x4.sqrt $push[[R:[0-9]+]]=, $0{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 declare <4 x float> @llvm.sqrt.v4f32(<4 x float> %x)
 define <4 x float> @sqrt_v4f32(<4 x float> %x) {
+; SIMD128-UNIMPL-LABEL: sqrt_v4f32:
+; SIMD128-UNIMPL:         .functype sqrt_v4f32 (v128) -> (v128)
+; SIMD128-UNIMPL-NEXT:    f32x4.sqrt $push0=, $0
+; SIMD128-UNIMPL-NEXT:    return $pop0
+;
+; SIMD128-VM-SLOW-LABEL: sqrt_v4f32:
+; SIMD128-VM-SLOW:         .functype sqrt_v4f32 (v128) -> (v128)
+; SIMD128-VM-SLOW-NEXT:    f32x4.extract_lane $push2=, $0, 0
+; SIMD128-VM-SLOW-NEXT:    f32.sqrt $push3=, $pop2
+; SIMD128-VM-SLOW-NEXT:    f32x4.splat $push4=, $pop3
+; SIMD128-VM-SLOW-NEXT:    f32x4.extract_lane $push0=, $0, 1
+; SIMD128-VM-SLOW-NEXT:    f32.sqrt $push1=, $pop0
+; SIMD128-VM-SLOW-NEXT:    f32x4.replace_lane $push5=, $pop4, 1, $pop1
+; SIMD128-VM-SLOW-NEXT:    f32x4.extract_lane $push6=, $0, 2
+; SIMD128-VM-SLOW-NEXT:    f32.sqrt $push7=, $pop6
+; SIMD128-VM-SLOW-NEXT:    f32x4.replace_lane $push8=, $pop5, 2, $pop7
+; SIMD128-VM-SLOW-NEXT:    f32x4.extract_lane $push9=, $0, 3
+; SIMD128-VM-SLOW-NEXT:    f32.sqrt $push10=, $pop9
+; SIMD128-VM-SLOW-NEXT:    f32x4.replace_lane $push11=, $pop8, 3, $pop10
+; SIMD128-VM-SLOW-NEXT:    return $pop11
+;
+; SIMD128-VM-FAST-LABEL: sqrt_v4f32:
+; SIMD128-VM-FAST:         .functype sqrt_v4f32 (v128) -> (v128)
+; SIMD128-VM-FAST-NEXT:    f32x4.extract_lane $push3=, $0, 0
+; SIMD128-VM-FAST-NEXT:    f32.sqrt $push4=, $pop3
+; SIMD128-VM-FAST-NEXT:    f32x4.splat $push5=, $pop4
+; SIMD128-VM-FAST-NEXT:    f32x4.extract_lane $push1=, $0, 1
+; SIMD128-VM-FAST-NEXT:    f32.sqrt $push2=, $pop1
+; SIMD128-VM-FAST-NEXT:    f32x4.replace_lane $push6=, $pop5, 1, $pop2
+; SIMD128-VM-FAST-NEXT:    f32x4.extract_lane $push7=, $0, 2
+; SIMD128-VM-FAST-NEXT:    f32.sqrt $push8=, $pop7
+; SIMD128-VM-FAST-NEXT:    f32x4.replace_lane $push9=, $pop6, 2, $pop8
+; SIMD128-VM-FAST-NEXT:    f32x4.extract_lane $push10=, $0, 3
+; SIMD128-VM-FAST-NEXT:    f32.sqrt $push11=, $pop10
+; SIMD128-VM-FAST-NEXT:    f32x4.replace_lane $push0=, $pop9, 3, $pop11
+; SIMD128-VM-FAST-NEXT:    return $pop0
+;
+; NO-SIMD128-SLOW-LABEL: sqrt_v4f32:
+; NO-SIMD128-SLOW:         .functype sqrt_v4f32 (i32, f32, f32, f32, f32) -> ()
+; NO-SIMD128-SLOW-NEXT:    i32.const $push0=, 12
+; NO-SIMD128-SLOW-NEXT:    i32.add $push1=, $0, $pop0
+; NO-SIMD128-SLOW-NEXT:    f32.sqrt $push2=, $4
+; NO-SIMD128-SLOW-NEXT:    f32.store 0($pop1), $pop2
+; NO-SIMD128-SLOW-NEXT:    f32.sqrt $push3=, $3
+; NO-SIMD128-SLOW-NEXT:    f32.store 8($0), $pop3
+; NO-SIMD128-SLOW-NEXT:    f32.sqrt $push4=, $2
+; NO-SIMD128-SLOW-NEXT:    f32.store 4($0), $pop4
+; NO-SIMD128-SLOW-NEXT:    f32.sqrt $push5=, $1
+; NO-SIMD128-SLOW-NEXT:    f32.store 0($0), $pop5
+; NO-SIMD128-SLOW-NEXT:    return
+;
+; NO-SIMD128-FAST-LABEL: sqrt_v4f32:
+; NO-SIMD128-FAST:         .functype sqrt_v4f32 (i32, f32, f32, f32, f32) -> ()
+; NO-SIMD128-FAST-NEXT:    f32.sqrt $push0=, $1
+; NO-SIMD128-FAST-NEXT:    f32.store 0($0), $pop0
+; NO-SIMD128-FAST-NEXT:    f32.sqrt $push1=, $2
+; NO-SIMD128-FAST-NEXT:    f32.store 4($0), $pop1
+; NO-SIMD128-FAST-NEXT:    f32.sqrt $push2=, $3
+; NO-SIMD128-FAST-NEXT:    f32.store 8($0), $pop2
+; NO-SIMD128-FAST-NEXT:    i32.const $push3=, 12
+; NO-SIMD128-FAST-NEXT:    i32.add $push4=, $0, $pop3
+; NO-SIMD128-FAST-NEXT:    f32.sqrt $push5=, $4
+; NO-SIMD128-FAST-NEXT:    f32.store 0($pop4), $pop5
+; NO-SIMD128-FAST-NEXT:    return
   %a = call <4 x float> @llvm.sqrt.v4f32(<4 x float> %x)
   ret <4 x float> %a
 }
 
 ; ==============================================================================
 ; 2 x double
 ; ==============================================================================
-; CHECK-LABEL: neg_v2f64:
-; NO-SIMD128-NOT: f64x2
-; SIMD128-NEXT: .functype neg_v2f64 (v128) -> (v128){{$}}
-; SIMD128-NEXT: f64x2.neg $push[[R:[0-9]+]]=, $0{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
+
 define <2 x double> @neg_v2f64(<2 x double> %x) {
   ; nsz makes this semantically equivalent to flipping sign bit
+; SIMD128-UNIMPL-LABEL: neg_v2f64:
+; SIMD128-UNIMPL:         .functype neg_v2f64 (v128) -> (v128)
+; SIMD128-UNIMPL-NEXT:    f64x2.neg $push0=, $0
+; SIMD128-UNIMPL-NEXT:    return $pop0
+;
+; NO-SIMD64-SLOW-LABEL: neg_v2f64:
+; NO-SIMD64-SLOW:         .functype neg_v2f64 (i32, f64, f64) -> ()
+; NO-SIMD64-SLOW-NEXT:    f64.neg $push0=, $2
+; NO-SIMD64-SLOW-NEXT:    f64.store 8($0), $pop0
+; NO-SIMD64-SLOW-NEXT:    f64.neg $push1=, $1
+; NO-SIMD64-SLOW-NEXT:    f64.store 0($0), $pop1
+; NO-SIMD64-SLOW-NEXT:    return
+;
+; NO-SIMD64-FAST-LABEL: neg_v2f64:
+; NO-SIMD64-FAST:         .functype neg_v2f64 (i32, f64, f64) -> ()
+; NO-SIMD64-FAST-NEXT:    f64.neg $push0=, $1
+; NO-SIMD64-FAST-NEXT:    f64.store 0($0), $pop0
+; NO-SIMD64-FAST-NEXT:    f64.neg $push1=, $2
+; NO-SIMD64-FAST-NEXT:    f64.store 8($0), $pop1
+; NO-SIMD64-FAST-NEXT:    return
   %a = fsub nsz <2 x double> <double 0., double 0.>, %x
   ret <2 x double> %a
 }
 
-; CHECK-LABEL: abs_v2f64:
-; NO-SIMD128-NOT: f64x2
-; SIMD128-NEXT: .functype abs_v2f64 (v128) -> (v128){{$}}
-; SIMD128-NEXT: f64x2.abs $push[[R:[0-9]+]]=, $0{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 declare <2 x double> @llvm.fabs.v2f64(<2 x double>) nounwind readnone
 define <2 x double> @abs_v2f64(<2 x double> %x) {
+; SIMD128-UNIMPL-LABEL: abs_v2f64:
+; SIMD128-UNIMPL:         .functype abs_v2f64 (v128) -> (v128)
+; SIMD128-UNIMPL-NEXT:    f64x2.abs $push0=, $0
+; SIMD128-UNIMPL-NEXT:    return $pop0
+;
+; NO-SIMD64-SLOW-LABEL: abs_v2f64:
+; NO-SIMD64-SLOW:         .functype abs_v2f64 (i32, f64, f64) -> ()
+; NO-SIMD64-SLOW-NEXT:    f64.abs $push0=, $2
+; NO-SIMD64-SLOW-NEXT:    f64.store 8($0), $pop0
+; NO-SIMD64-SLOW-NEXT:    f64.abs $push1=, $1
+; NO-SIMD64-SLOW-NEXT:    f64.store 0($0), $pop1
+; NO-SIMD64-SLOW-NEXT:    return
+;
+; NO-SIMD64-FAST-LABEL: abs_v2f64:
+; NO-SIMD64-FAST:         .functype abs_v2f64 (i32, f64, f64) -> ()
+; NO-SIMD64-FAST-NEXT:    f64.abs $push0=, $1
+; NO-SIMD64-FAST-NEXT:    f64.store 0($0), $pop0
+; NO-SIMD64-FAST-NEXT:    f64.abs $push1=, $2
+; NO-SIMD64-FAST-NEXT:    f64.store 8($0), $pop1
+; NO-SIMD64-FAST-NEXT:    return
   %a = call <2 x double> @llvm.fabs.v2f64(<2 x double> %x)
   ret <2 x double> %a
 }
 
-; CHECK-LABEL: min_unordered_v2f64:
-; NO-SIMD128-NOT: f64x2
-; SIMD128-NEXT: .functype min_unordered_v2f64 (v128) -> (v128){{$}}
-; SIMD128-NEXT: f64.const $push[[L0:[0-9]+]]=, 0x1.4p2
-; SIMD128-NEXT: f64x2.splat $push[[L1:[0-9]+]]=, $pop[[L0]]
-; SIMD128-NEXT: f64x2.min $push[[R:[0-9]+]]=, $0, $pop[[L1]]{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <2 x double> @min_unordered_v2f64(<2 x double> %x) {
+; SIMD128-UNIMPL-SLOW-LABEL: min_unordered_v2f64:
+; SIMD128-UNIMPL-SLOW:         .functype min_unordered_v2f64 (v128) -> (v128)
+; SIMD128-UNIMPL-SLOW-NEXT:    f64.const $push0=, 0x1.4p2
+; SIMD128-UNIMPL-SLOW-NEXT:    f64x2.splat $push1=, $pop0
+; SIMD128-UNIMPL-SLOW-NEXT:    f64x2.min $push2=, $0, $pop1
+; SIMD128-UNIMPL-SLOW-NEXT:    return $pop2
+;
+; SIMD128-UNIMPL-FAST-LABEL: min_unordered_v2f64:
+; SIMD128-UNIMPL-FAST:         .functype min_unordered_v2f64 (v128) -> (v128)
+; SIMD128-UNIMPL-FAST-NEXT:    f64.const $push1=, 0x1.4p2
+; SIMD128-UNIMPL-FAST-NEXT:    f64x2.splat $push2=, $pop1
+; SIMD128-UNIMPL-FAST-NEXT:    f64x2.min $push0=, $0, $pop2
+; SIMD128-UNIMPL-FAST-NEXT:    return $pop0
+;
+; NO-SIMD64-SLOW-LABEL: min_unordered_v2f64:
+; NO-SIMD64-SLOW:         .functype min_unordered_v2f64 (i32, f64, f64) -> ()
+; NO-SIMD64-SLOW-NEXT:    f64.const $push0=, 0x1.4p2
+; NO-SIMD64-SLOW-NEXT:    f64.min $push1=, $2, $pop0
+; NO-SIMD64-SLOW-NEXT:    f64.store 8($0), $pop1
+; NO-SIMD64-SLOW-NEXT:    f64.const $push3=, 0x1.4p2
+; NO-SIMD64-SLOW-NEXT:    f64.min $push2=, $1, $pop3
+; NO-SIMD64-SLOW-NEXT:    f64.store 0($0), $pop2
+; NO-SIMD64-SLOW-NEXT:    return
+;
+; NO-SIMD64-FAST-LABEL: min_unordered_v2f64:
+; NO-SIMD64-FAST:         .functype min_unordered_v2f64 (i32, f64, f64) -> ()
+; NO-SIMD64-FAST-NEXT:    f64.const $push0=, 0x1.4p2
+; NO-SIMD64-FAST-NEXT:    f64.min $push1=, $1, $pop0
+; NO-SIMD64-FAST-NEXT:    f64.store 0($0), $pop1
+; NO-SIMD64-FAST-NEXT:    f64.const $push3=, 0x1.4p2
+; NO-SIMD64-FAST-NEXT:    f64.min $push2=, $2, $pop3
+; NO-SIMD64-FAST-NEXT:    f64.store 8($0), $pop2
+; NO-SIMD64-FAST-NEXT:    return
   %cmps = fcmp ule <2 x double> %x, <double 5., double 5.>
   %a = select <2 x i1> %cmps, <2 x double> %x,
     <2 x double> <double 5., double 5.>
   ret <2 x double> %a
 }
 
-; CHECK-LABEL: max_unordered_v2f64:
-; NO-SIMD128-NOT: f64x2
-; SIMD128-NEXT: .functype max_unordered_v2f64 (v128) -> (v128){{$}}
-; SIMD128-NEXT: f64.const $push[[L0:[0-9]+]]=, 0x1.4p2
-; SIMD128-NEXT: f64x2.splat $push[[L1:[0-9]+]]=, $pop[[L0]]
-; SIMD128-NEXT: f64x2.max $push[[R:[0-9]+]]=, $0, $pop[[L1]]{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <2 x double> @max_unordered_v2f64(<2 x double> %x) {
+; SIMD128-UNIMPL-SLOW-LABEL: max_unordered_v2f64:
+; SIMD128-UNIMPL-SLOW:         .functype max_unordered_v2f64 (v128) -> (v128)
+; SIMD128-UNIMPL-SLOW-NEXT:    f64.const $push0=, 0x1.4p2
+; SIMD128-UNIMPL-SLOW-NEXT:    f64x2.splat $push1=, $pop0
+; SIMD128-UNIMPL-SLOW-NEXT:    f64x2.max $push2=, $0, $pop1
+; SIMD128-UNIMPL-SLOW-NEXT:    return $pop2
+;
+; SIMD128-UNIMPL-FAST-LABEL: max_unordered_v2f64:
+; SIMD128-UNIMPL-FAST:         .functype max_unordered_v2f64 (v128) -> (v128)
+; SIMD128-UNIMPL-FAST-NEXT:    f64.const $push1=, 0x1.4p2
+; SIMD128-UNIMPL-FAST-NEXT:    f64x2.splat $push2=, $pop1
+; SIMD128-UNIMPL-FAST-NEXT:    f64x2.max $push0=, $0, $pop2
+; SIMD128-UNIMPL-FAST-NEXT:    return $pop0
+;
+; NO-SIMD64-SLOW-LABEL: max_unordered_v2f64:
+; NO-SIMD64-SLOW:         .functype max_unordered_v2f64 (i32, f64, f64) -> ()
+; NO-SIMD64-SLOW-NEXT:    f64.const $push0=, 0x1.4p2
+; NO-SIMD64-SLOW-NEXT:    f64.max $push1=, $2, $pop0
+; NO-SIMD64-SLOW-NEXT:    f64.store 8($0), $pop1
+; NO-SIMD64-SLOW-NEXT:    f64.const $push3=, 0x1.4p2
+; NO-SIMD64-SLOW-NEXT:    f64.max $push2=, $1, $pop3
+; NO-SIMD64-SLOW-NEXT:    f64.store 0($0), $pop2
+; NO-SIMD64-SLOW-NEXT:    return
+;
+; NO-SIMD64-FAST-LABEL: max_unordered_v2f64:
+; NO-SIMD64-FAST:         .functype max_unordered_v2f64 (i32, f64, f64) -> ()
+; NO-SIMD64-FAST-NEXT:    f64.const $push0=, 0x1.4p2
+; NO-SIMD64-FAST-NEXT:    f64.max $push1=, $1, $pop0
+; NO-SIMD64-FAST-NEXT:    f64.store 0($0), $pop1
+; NO-SIMD64-FAST-NEXT:    f64.const $push3=, 0x1.4p2
+; NO-SIMD64-FAST-NEXT:    f64.max $push2=, $2, $pop3
+; NO-SIMD64-FAST-NEXT:    f64.store 8($0), $pop2
+; NO-SIMD64-FAST-NEXT:    return
   %cmps = fcmp uge <2 x double> %x, <double 5., double 5.>
   %a = select <2 x i1> %cmps, <2 x double> %x,
     <2 x double> <double 5., double 5.>
   ret <2 x double> %a
 }
 
-; CHECK-LABEL: min_ordered_v2f64:
-; NO-SIMD128-NOT: f64x2
-; SIMD128-NEXT: .functype min_ordered_v2f64 (v128) -> (v128){{$}}
-; SIMD128-NEXT: f64.const $push[[L0:[0-9]+]]=, 0x1.4p2
-; SIMD128-NEXT: f64x2.splat $push[[L1:[0-9]+]]=, $pop[[L0]]
-; SIMD128-NEXT: f64x2.min $push[[R:[0-9]+]]=, $0, $pop[[L1]]{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <2 x double> @min_ordered_v2f64(<2 x double> %x) {
+; SIMD128-UNIMPL-SLOW-LABEL: min_ordered_v2f64:
+; SIMD128-UNIMPL-SLOW:         .functype min_ordered_v2f64 (v128) -> (v128)
+; SIMD128-UNIMPL-SLOW-NEXT:    f64.const $push0=, 0x1.4p2
+; SIMD128-UNIMPL-SLOW-NEXT:    f64x2.splat $push1=, $pop0
+; SIMD128-UNIMPL-SLOW-NEXT:    f64x2.min $push2=, $0, $pop1
+; SIMD128-UNIMPL-SLOW-NEXT:    return $pop2
+;
+; SIMD128-UNIMPL-FAST-LABEL: min_ordered_v2f64:
+; SIMD128-UNIMPL-FAST:         .functype min_ordered_v2f64 (v128) -> (v128)
+; SIMD128-UNIMPL-FAST-NEXT:    f64.const $push1=, 0x1.4p2
+; SIMD128-UNIMPL-FAST-NEXT:    f64x2.splat $push2=, $pop1
+; SIMD128-UNIMPL-FAST-NEXT:    f64x2.min $push0=, $0, $pop2
+; SIMD128-UNIMPL-FAST-NEXT:    return $pop0
+;
+; NO-SIMD64-SLOW-LABEL: min_ordered_v2f64:
+; NO-SIMD64-SLOW:         .functype min_ordered_v2f64 (i32, f64, f64) -> ()
+; NO-SIMD64-SLOW-NEXT:    f64.const $push0=, 0x1.4p2
+; NO-SIMD64-SLOW-NEXT:    f64.min $push1=, $2, $pop0
+; NO-SIMD64-SLOW-NEXT:    f64.store 8($0), $pop1
+; NO-SIMD64-SLOW-NEXT:    f64.const $push3=, 0x1.4p2
+; NO-SIMD64-SLOW-NEXT:    f64.min $push2=, $1, $pop3
+; NO-SIMD64-SLOW-NEXT:    f64.store 0($0), $pop2
+; NO-SIMD64-SLOW-NEXT:    return
+;
+; NO-SIMD64-FAST-LABEL: min_ordered_v2f64:
+; NO-SIMD64-FAST:         .functype min_ordered_v2f64 (i32, f64, f64) -> ()
+; NO-SIMD64-FAST-NEXT:    f64.const $push0=, 0x1.4p2
+; NO-SIMD64-FAST-NEXT:    f64.min $push1=, $1, $pop0
+; NO-SIMD64-FAST-NEXT:    f64.store 0($0), $pop1
+; NO-SIMD64-FAST-NEXT:    f64.const $push3=, 0x1.4p2
+; NO-SIMD64-FAST-NEXT:    f64.min $push2=, $2, $pop3
+; NO-SIMD64-FAST-NEXT:    f64.store 8($0), $pop2
+; NO-SIMD64-FAST-NEXT:    return
   %cmps = fcmp ole <2 x double> <double 5., double 5.>, %x
   %a = select <2 x i1> %cmps, <2 x double> <double 5., double 5.>,
     <2 x double> %x
   ret <2 x double> %a
 }
 
-; CHECK-LABEL: max_ordered_v2f64:
-; NO-SIMD128-NOT: f64x2
-; SIMD128-NEXT: .functype max_ordered_v2f64 (v128) -> (v128){{$}}
-; SIMD128-NEXT: f64.const $push[[L0:[0-9]+]]=, 0x1.4p2
-; SIMD128-NEXT: f64x2.splat $push[[L1:[0-9]+]]=, $pop[[L0]]
-; SIMD128-NEXT: f64x2.max $push[[R:[0-9]+]]=, $0, $pop[[L1]]{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <2 x double> @max_ordered_v2f64(<2 x double> %x) {
+; SIMD128-UNIMPL-SLOW-LABEL: max_ordered_v2f64:
+; SIMD128-UNIMPL-SLOW:         .functype max_ordered_v2f64 (v128) -> (v128)
+; SIMD128-UNIMPL-SLOW-NEXT:    f64.const $push0=, 0x1.4p2
+; SIMD128-UNIMPL-SLOW-NEXT:    f64x2.splat $push1=, $pop0
+; SIMD128-UNIMPL-SLOW-NEXT:    f64x2.max $push2=, $0, $pop1
+; SIMD128-UNIMPL-SLOW-NEXT:    return $pop2
+;
+; SIMD128-UNIMPL-FAST-LABEL: max_ordered_v2f64:
+; SIMD128-UNIMPL-FAST:         .functype max_ordered_v2f64 (v128) -> (v128)
+; SIMD128-UNIMPL-FAST-NEXT:    f64.const $push1=, 0x1.4p2
+; SIMD128-UNIMPL-FAST-NEXT:    f64x2.splat $push2=, $pop1
+; SIMD128-UNIMPL-FAST-NEXT:    f64x2.max $push0=, $0, $pop2
+; SIMD128-UNIMPL-FAST-NEXT:    return $pop0
+;
+; NO-SIMD64-SLOW-LABEL: max_ordered_v2f64:
+; NO-SIMD64-SLOW:         .functype max_ordered_v2f64 (i32, f64, f64) -> ()
+; NO-SIMD64-SLOW-NEXT:    f64.const $push0=, 0x1.4p2
+; NO-SIMD64-SLOW-NEXT:    f64.max $push1=, $2, $pop0
+; NO-SIMD64-SLOW-NEXT:    f64.store 8($0), $pop1
+; NO-SIMD64-SLOW-NEXT:    f64.const $push3=, 0x1.4p2
+; NO-SIMD64-SLOW-NEXT:    f64.max $push2=, $1, $pop3
+; NO-SIMD64-SLOW-NEXT:    f64.store 0($0), $pop2
+; NO-SIMD64-SLOW-NEXT:    return
+;
+; NO-SIMD64-FAST-LABEL: max_ordered_v2f64:
+; NO-SIMD64-FAST:         .functype max_ordered_v2f64 (i32, f64, f64) -> ()
+; NO-SIMD64-FAST-NEXT:    f64.const $push0=, 0x1.4p2
+; NO-SIMD64-FAST-NEXT:    f64.max $push1=, $1, $pop0
+; NO-SIMD64-FAST-NEXT:    f64.store 0($0), $pop1
+; NO-SIMD64-FAST-NEXT:    f64.const $push3=, 0x1.4p2
+; NO-SIMD64-FAST-NEXT:    f64.max $push2=, $2, $pop3
+; NO-SIMD64-FAST-NEXT:    f64.store 8($0), $pop2
+; NO-SIMD64-FAST-NEXT:    return
   %cmps = fcmp oge <2 x double> <double 5., double 5.>, %x
   %a = select <2 x i1> %cmps, <2 x double> <double 5., double 5.>,
     <2 x double> %x
   ret <2 x double> %a
 }
 
-; CHECK-LABEL: min_intrinsic_v2f64:
-; NO-SIMD128-NOT: f64x2
-; SIMD128-NEXT: .functype min_intrinsic_v2f64 (v128, v128) -> (v128){{$}}
-; SIMD128-NEXT: f64x2.min $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 declare <2 x double> @llvm.minimum.v2f64(<2 x double>, <2 x double>)
 define <2 x double> @min_intrinsic_v2f64(<2 x double> %x, <2 x double> %y) {
+; SIMD128-UNIMPL-LABEL: min_intrinsic_v2f64:
+; SIMD128-UNIMPL:         .functype min_intrinsic_v2f64 (v128, v128) -> (v128)
+; SIMD128-UNIMPL-NEXT:    f64x2.min $push0=, $0, $1
+; SIMD128-UNIMPL-NEXT:    return $pop0
+;
+; NO-SIMD64-SLOW-LABEL: min_intrinsic_v2f64:
+; NO-SIMD64-SLOW:         .functype min_intrinsic_v2f64 (i32, f64, f64, f64, f64) -> ()
+; NO-SIMD64-SLOW-NEXT:    f64.min $push0=, $2, $4
+; NO-SIMD64-SLOW-NEXT:    f64.store 8($0), $pop0
+; NO-SIMD64-SLOW-NEXT:    f64.min $push1=, $1, $3
+; NO-SIMD64-SLOW-NEXT:    f64.store 0($0), $pop1
+; NO-SIMD64-SLOW-NEXT:    return
+;
+; NO-SIMD64-FAST-LABEL: min_intrinsic_v2f64:
+; NO-SIMD64-FAST:         .functype min_intrinsic_v2f64 (i32, f64, f64, f64, f64) -> ()
+; NO-SIMD64-FAST-NEXT:    f64.min $push0=, $1, $3
+; NO-SIMD64-FAST-NEXT:    f64.store 0($0), $pop0
+; NO-SIMD64-FAST-NEXT:    f64.min $push1=, $2, $4
+; NO-SIMD64-FAST-NEXT:    f64.store 8($0), $pop1
+; NO-SIMD64-FAST-NEXT:    return
   %a = call <2 x double> @llvm.minimum.v2f64(<2 x double> %x, <2 x double> %y)
   ret <2 x double> %a
 }
 
-; CHECK-LABEL: max_intrinsic_v2f64:
-; NO-SIMD128-NOT: f64x2
-; SIMD128-NEXT: .functype max_intrinsic_v2f64 (v128, v128) -> (v128){{$}}
-; SIMD128-NEXT: f64x2.max $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 declare <2 x double> @llvm.maximum.v2f64(<2 x double>, <2 x double>)
 define <2 x double> @max_intrinsic_v2f64(<2 x double> %x, <2 x double> %y) {
+; SIMD128-UNIMPL-LABEL: max_intrinsic_v2f64:
+; SIMD128-UNIMPL:         .functype max_intrinsic_v2f64 (v128, v128) -> (v128)
+; SIMD128-UNIMPL-NEXT:    f64x2.max $push0=, $0, $1
+; SIMD128-UNIMPL-NEXT:    return $pop0
+;
+; NO-SIMD64-SLOW-LABEL: max_intrinsic_v2f64:
+; NO-SIMD64-SLOW:         .functype max_intrinsic_v2f64 (i32, f64, f64, f64, f64) -> ()
+; NO-SIMD64-SLOW-NEXT:    f64.max $push0=, $2, $4
+; NO-SIMD64-SLOW-NEXT:    f64.store 8($0), $pop0
+; NO-SIMD64-SLOW-NEXT:    f64.max $push1=, $1, $3
+; NO-SIMD64-SLOW-NEXT:    f64.store 0($0), $pop1
+; NO-SIMD64-SLOW-NEXT:    return
+;
+; NO-SIMD64-FAST-LABEL: max_intrinsic_v2f64:
+; NO-SIMD64-FAST:         .functype max_intrinsic_v2f64 (i32, f64, f64, f64, f64) -> ()
+; NO-SIMD64-FAST-NEXT:    f64.max $push0=, $1, $3
+; NO-SIMD64-FAST-NEXT:    f64.store 0($0), $pop0
+; NO-SIMD64-FAST-NEXT:    f64.max $push1=, $2, $4
+; NO-SIMD64-FAST-NEXT:    f64.store 8($0), $pop1
+; NO-SIMD64-FAST-NEXT:    return
   %a = call <2 x double> @llvm.maximum.v2f64(<2 x double> %x, <2 x double> %y)
   ret <2 x double> %a
 }
 
-; CHECK-LABEL: min_const_intrinsic_v2f64:
-; NO-SIMD128-NOT: f64x2
-; SIMD128-NEXT: .functype min_const_intrinsic_v2f64 () -> (v128){{$}}
-; SIMD128-NEXT: f64.const $push[[L:[0-9]+]]=, 0x1.4p2{{$}}
-; SIMD128-NEXT: f64x2.splat $push[[R:[0-9]+]]=, $pop[[L]]{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <2 x double> @min_const_intrinsic_v2f64() {
+; SIMD128-UNIMPL-SLOW-LABEL: min_const_intrinsic_v2f64:
+; SIMD128-UNIMPL-SLOW:         .functype min_const_intrinsic_v2f64 () -> (v128)
+; SIMD128-UNIMPL-SLOW-NEXT:    f64.const $push0=, 0x1.4p2
+; SIMD128-UNIMPL-SLOW-NEXT:    f64x2.splat $push1=, $pop0
+; SIMD128-UNIMPL-SLOW-NEXT:    return $pop1
+;
+; SIMD128-UNIMPL-FAST-LABEL: min_const_intrinsic_v2f64:
+; SIMD128-UNIMPL-FAST:         .functype min_const_intrinsic_v2f64 () -> (v128)
+; SIMD128-UNIMPL-FAST-NEXT:    f64.const $push1=, 0x1.4p2
+; SIMD128-UNIMPL-FAST-NEXT:    f64x2.splat $push0=, $pop1
+; SIMD128-UNIMPL-FAST-NEXT:    return $pop0
+;
+; NO-SIMD64-LABEL: min_const_intrinsic_v2f64:
+; NO-SIMD64:         .functype min_const_intrinsic_v2f64 (i32) -> ()
+; NO-SIMD64-NEXT:    i64.const $push0=, 4617315517961601024
+; NO-SIMD64-NEXT:    i64.store 8($0), $pop0
+; NO-SIMD64-NEXT:    i64.const $push1=, 4617315517961601024
+; NO-SIMD64-NEXT:    i64.store 0($0), $pop1
+; NO-SIMD64-NEXT:    return
   %a = call <2 x double> @llvm.minimum.v2f64(
     <2 x double> <double 42., double 42.>,
     <2 x double> <double 5., double 5.>
   )
   ret <2 x double> %a
 }
 
-; CHECK-LABEL: max_const_intrinsic_v2f64:
-; NO-SIMD128-NOT: f64x2
-; SIMD128-NEXT: .functype max_const_intrinsic_v2f64 () -> (v128){{$}}
-; SIMD128-NEXT: f64.const $push[[L:[0-9]+]]=, 0x1.5p5{{$}}
-; SIMD128-NEXT: f64x2.splat $push[[R:[0-9]+]]=, $pop[[L]]{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <2 x double> @max_const_intrinsic_v2f64() {
+; SIMD128-UNIMPL-SLOW-LABEL: max_const_intrinsic_v2f64:
+; SIMD128-UNIMPL-SLOW:         .functype max_const_intrinsic_v2f64 () -> (v128)
+; SIMD128-UNIMPL-SLOW-NEXT:    f64.const $push0=, 0x1.5p5
+; SIMD128-UNIMPL-SLOW-NEXT:    f64x2.splat $push1=, $pop0
+; SIMD128-UNIMPL-SLOW-NEXT:    return $pop1
+;
+; SIMD128-UNIMPL-FAST-LABEL: max_const_intrinsic_v2f64:
+; SIMD128-UNIMPL-FAST:         .functype max_const_intrinsic_v2f64 () -> (v128)
+; SIMD128-UNIMPL-FAST-NEXT:    f64.const $push1=, 0x1.5p5
+; SIMD128-UNIMPL-FAST-NEXT:    f64x2.splat $push0=, $pop1
+; SIMD128-UNIMPL-FAST-NEXT:    return $pop0
+;
+; NO-SIMD64-LABEL: max_const_intrinsic_v2f64:
+; NO-SIMD64:         .functype max_const_intrinsic_v2f64 (i32) -> ()
+; NO-SIMD64-NEXT:    i64.const $push0=, 4631107791820423168
+; NO-SIMD64-NEXT:    i64.store 8($0), $pop0
+; NO-SIMD64-NEXT:    i64.const $push1=, 4631107791820423168
+; NO-SIMD64-NEXT:    i64.store 0($0), $pop1
+; NO-SIMD64-NEXT:    return
   %a = call <2 x double> @llvm.maximum.v2f64(
     <2 x double> <double 42., double 42.>,
     <2 x double> <double 5., double 5.>
   )
   ret <2 x double> %a
 }
 
-; CHECK-LABEL: add_v2f64:
-; NO-SIMD128-NOT: f64x2
-; SIMD128-VM-NOT: f62x2
-; SIMD128-NEXT: .functype add_v2f64 (v128, v128) -> (v128){{$}}
-; SIMD128-NEXT: f64x2.add $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <2 x double> @add_v2f64(<2 x double> %x, <2 x double> %y) {
+; SIMD128-UNIMPL-LABEL: add_v2f64:
+; SIMD128-UNIMPL:         .functype add_v2f64 (v128, v128) -> (v128)
+; SIMD128-UNIMPL-NEXT:    f64x2.add $push0=, $0, $1
+; SIMD128-UNIMPL-NEXT:    return $pop0
+;
+; NO-SIMD64-SLOW-LABEL: add_v2f64:
+; NO-SIMD64-SLOW:         .functype add_v2f64 (i32, f64, f64, f64, f64) -> ()
+; NO-SIMD64-SLOW-NEXT:    f64.add $push0=, $2, $4
+; NO-SIMD64-SLOW-NEXT:    f64.store 8($0), $pop0
+; NO-SIMD64-SLOW-NEXT:    f64.add $push1=, $1, $3
+; NO-SIMD64-SLOW-NEXT:    f64.store 0($0), $pop1
+; NO-SIMD64-SLOW-NEXT:    return
+;
+; NO-SIMD64-FAST-LABEL: add_v2f64:
+; NO-SIMD64-FAST:         .functype add_v2f64 (i32, f64, f64, f64, f64) -> ()
+; NO-SIMD64-FAST-NEXT:    f64.add $push0=, $1, $3
+; NO-SIMD64-FAST-NEXT:    f64.store 0($0), $pop0
+; NO-SIMD64-FAST-NEXT:    f64.add $push1=, $2, $4
+; NO-SIMD64-FAST-NEXT:    f64.store 8($0), $pop1
+; NO-SIMD64-FAST-NEXT:    return
   %a = fadd <2 x double> %x, %y
   ret <2 x double> %a
 }
 
-; CHECK-LABEL: sub_v2f64:
-; NO-SIMD128-NOT: f64x2
-; SIMD128-VM-NOT: f62x2
-; SIMD128-NEXT: .functype sub_v2f64 (v128, v128) -> (v128){{$}}
-; SIMD128-NEXT: f64x2.sub $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <2 x double> @sub_v2f64(<2 x double> %x, <2 x double> %y) {
+; SIMD128-UNIMPL-LABEL: sub_v2f64:
+; SIMD128-UNIMPL:         .functype sub_v2f64 (v128, v128) -> (v128)
+; SIMD128-UNIMPL-NEXT:    f64x2.sub $push0=, $0, $1
+; SIMD128-UNIMPL-NEXT:    return $pop0
+;
+; NO-SIMD64-SLOW-LABEL: sub_v2f64:
+; NO-SIMD64-SLOW:         .functype sub_v2f64 (i32, f64, f64, f64, f64) -> ()
+; NO-SIMD64-SLOW-NEXT:    f64.sub $push0=, $2, $4
+; NO-SIMD64-SLOW-NEXT:    f64.store 8($0), $pop0
+; NO-SIMD64-SLOW-NEXT:    f64.sub $push1=, $1, $3
+; NO-SIMD64-SLOW-NEXT:    f64.store 0($0), $pop1
+; NO-SIMD64-SLOW-NEXT:    return
+;
+; NO-SIMD64-FAST-LABEL: sub_v2f64:
+; NO-SIMD64-FAST:         .functype sub_v2f64 (i32, f64, f64, f64, f64) -> ()
+; NO-SIMD64-FAST-NEXT:    f64.sub $push0=, $1, $3
+; NO-SIMD64-FAST-NEXT:    f64.store 0($0), $pop0
+; NO-SIMD64-FAST-NEXT:    f64.sub $push1=, $2, $4
+; NO-SIMD64-FAST-NEXT:    f64.store 8($0), $pop1
+; NO-SIMD64-FAST-NEXT:    return
   %a = fsub <2 x double> %x, %y
   ret <2 x double> %a
 }
 
-; CHECK-LABEL: div_v2f64:
-; NO-SIMD128-NOT: f64x2
-; SIMD128-VM-NOT: f62x2
-; SIMD128-NEXT: .functype div_v2f64 (v128, v128) -> (v128){{$}}
-; SIMD128-NEXT: f64x2.div $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <2 x double> @div_v2f64(<2 x double> %x, <2 x double> %y) {
+; SIMD128-UNIMPL-LABEL: div_v2f64:
+; SIMD128-UNIMPL:         .functype div_v2f64 (v128, v128) -> (v128)
+; SIMD128-UNIMPL-NEXT:    f64x2.div $push0=, $0, $1
+; SIMD128-UNIMPL-NEXT:    return $pop0
+;
+; NO-SIMD64-SLOW-LABEL: div_v2f64:
+; NO-SIMD64-SLOW:         .functype div_v2f64 (i32, f64, f64, f64, f64) -> ()
+; NO-SIMD64-SLOW-NEXT:    f64.div $push0=, $2, $4
+; NO-SIMD64-SLOW-NEXT:    f64.store 8($0), $pop0
+; NO-SIMD64-SLOW-NEXT:    f64.div $push1=, $1, $3
+; NO-SIMD64-SLOW-NEXT:    f64.store 0($0), $pop1
+; NO-SIMD64-SLOW-NEXT:    return
+;
+; NO-SIMD64-FAST-LABEL: div_v2f64:
+; NO-SIMD64-FAST:         .functype div_v2f64 (i32, f64, f64, f64, f64) -> ()
+; NO-SIMD64-FAST-NEXT:    f64.div $push0=, $1, $3
+; NO-SIMD64-FAST-NEXT:    f64.store 0($0), $pop0
+; NO-SIMD64-FAST-NEXT:    f64.div $push1=, $2, $4
+; NO-SIMD64-FAST-NEXT:    f64.store 8($0), $pop1
+; NO-SIMD64-FAST-NEXT:    return
   %a = fdiv <2 x double> %x, %y
   ret <2 x double> %a
 }
 
-; CHECK-LABEL: mul_v2f64:
-; NO-SIMD128-NOT: f64x2
-; SIMD128-VM-NOT: f62x2
-; SIMD128-NEXT: .functype mul_v2f64 (v128, v128) -> (v128){{$}}
-; SIMD128-NEXT: f64x2.mul $push[[R:[0-9]+]]=, $0, $1{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 define <2 x double> @mul_v2f64(<2 x double> %x, <2 x double> %y) {
+; SIMD128-UNIMPL-LABEL: mul_v2f64:
+; SIMD128-UNIMPL:         .functype mul_v2f64 (v128, v128) -> (v128)
+; SIMD128-UNIMPL-NEXT:    f64x2.mul $push0=, $0, $1
+; SIMD128-UNIMPL-NEXT:    return $pop0
+;
+; NO-SIMD64-SLOW-LABEL: mul_v2f64:
+; NO-SIMD64-SLOW:         .functype mul_v2f64 (i32, f64, f64, f64, f64) -> ()
+; NO-SIMD64-SLOW-NEXT:    f64.mul $push0=, $2, $4
+; NO-SIMD64-SLOW-NEXT:    f64.store 8($0), $pop0
+; NO-SIMD64-SLOW-NEXT:    f64.mul $push1=, $1, $3
+; NO-SIMD64-SLOW-NEXT:    f64.store 0($0), $pop1
+; NO-SIMD64-SLOW-NEXT:    return
+;
+; NO-SIMD64-FAST-LABEL: mul_v2f64:
+; NO-SIMD64-FAST:         .functype mul_v2f64 (i32, f64, f64, f64, f64) -> ()
+; NO-SIMD64-FAST-NEXT:    f64.mul $push0=, $1, $3
+; NO-SIMD64-FAST-NEXT:    f64.store 0($0), $pop0
+; NO-SIMD64-FAST-NEXT:    f64.mul $push1=, $2, $4
+; NO-SIMD64-FAST-NEXT:    f64.store 8($0), $pop1
+; NO-SIMD64-FAST-NEXT:    return
   %a = fmul <2 x double> %x, %y
   ret <2 x double> %a
 }
 
-; CHECK-LABEL: sqrt_v2f64:
-; NO-SIMD128-NOT: f64x2
-; SIMD128-NEXT: .functype sqrt_v2f64 (v128) -> (v128){{$}}
-; SIMD128-NEXT: f64x2.sqrt $push[[R:[0-9]+]]=, $0{{$}}
-; SIMD128-NEXT: return $pop[[R]]{{$}}
 declare <2 x double> @llvm.sqrt.v2f64(<2 x double> %x)
 define <2 x double> @sqrt_v2f64(<2 x double> %x) {
+; SIMD128-UNIMPL-LABEL: sqrt_v2f64:
+; SIMD128-UNIMPL:         .functype sqrt_v2f64 (v128) -> (v128)
+; SIMD128-UNIMPL-NEXT:    f64x2.sqrt $push0=, $0
+; SIMD128-UNIMPL-NEXT:    return $pop0
+;
+; NO-SIMD64-SLOW-LABEL: sqrt_v2f64:
+; NO-SIMD64-SLOW:         .functype sqrt_v2f64 (i32, f64, f64) -> ()
+; NO-SIMD64-SLOW-NEXT:    f64.sqrt $push0=, $2
+; NO-SIMD64-SLOW-NEXT:    f64.store 8($0), $pop0
+; NO-SIMD64-SLOW-NEXT:    f64.sqrt $push1=, $1
+; NO-SIMD64-SLOW-NEXT:    f64.store 0($0), $pop1
+; NO-SIMD64-SLOW-NEXT:    return
+;
+; NO-SIMD64-FAST-LABEL: sqrt_v2f64:
+; NO-SIMD64-FAST:         .functype sqrt_v2f64 (i32, f64, f64) -> ()
+; NO-SIMD64-FAST-NEXT:    f64.sqrt $push0=, $1
+; NO-SIMD64-FAST-NEXT:    f64.store 0($0), $pop0
+; NO-SIMD64-FAST-NEXT:    f64.sqrt $push1=, $2
+; NO-SIMD64-FAST-NEXT:    f64.store 8($0), $pop1
+; NO-SIMD64-FAST-NEXT:    return
   %a = call <2 x double> @llvm.sqrt.v2f64(<2 x double> %x)
   ret <2 x double> %a
 }