[clang][NewPM] Add RUNS for tests that produce slightly different IR under new PM

[leonardchan]

Summary

With the new pass manager enabled by default, some tests produce slightly different IR from the legacy PM. This patch just adds separate new PM runs and checks for the new PM IR.

This fixes:

Clang :: CodeGen/avx512-reduceMinMaxIntrin.c
Clang :: CodeGen/avx512f-builtins.c
Clang :: CodeGen/avx512vl-builtins.c
Clang :: CodeGen/avx512vlbw-builtins.c
Clang :: CodeGenOpenCL/convergent.cl

For CodeGenOpenCL/convergent.cl, the new PM produced a slightly different for loop, but this still checks for no loop unrolling as intended.

For the avx-builtins tests, the new PM would produce extra bitcasts that did not affect the overall logic of the function, but I do not know a simple way to get rid of these bitcasts.

For CodeGen/avx512-reduceMinMaxIntrin.c, temporary local variables are produced in the codegen, but the allocas are out of order and the tests were pretty strict on order even though the logic/use of these variables was the same. Nearly all checks in this large file were broken, so I was hoping we could run it with legacy only for now and come back to it later to make this easier.

[chandlerc]

This really confused me. We shouldn't be seeing this kind of difference in the new PM.

But I think I figured it out.

Both PMs have to run *some* inliner at -O0. This is because we need to inline always_inline functions. The new PM has a *super* simple (and fast compile-time wise) inliner designed for this. It's really good at its job. The old PM runs the normal inliner. This thing isn't simple at all. It does complex stuff to incrementally simplify code while inlining because that can have a huge effect on the overall efficiency of the compiler *when doing full optimization passes*. But we're not doing that.

The result is that because all of these tests check code that is emitted by always_inline functions in our builtin headers, and then *inlined* into the test functions, the inliner in the old pass manager did a bunch of "cleanup" of the code that the new PM doesn't do.

Honestly, I think the new PM's behavior is correct for -O0, but I think it makes these tests less easy to understand. I think we should instead change the command running here, and when we do I think we will find a set of checks that work for both PMs.

When we are willing to do something like run opt in addition to clang (the convergent.cl case), instead of just doing -instnamer we should also do -instcombine. I suspect this will remove the differences.

When we're just directly using %clang_cc1 I suspect that rather than checking the completely unoptimized code we should check with -O1 to make the IR much more stable and easy to inspect. It should also result in identical results from the two PMs.

clang/test/CodeGenOpenCL/convergent.cl

2

Probably want to use opt -passes=instnamer or some such to use the new PM in the opt invocation as well...

[leonardchan]

This comment was removed by leonardchan.

[leonardchan]

For the avx tests, I don't suppose you know a simple way to generate these tests? They're about 10k lines long and it's taking a while to go through it by hand to replace the current IR checks with codegen for -O1. (Perhaps something along the lines of utils/update_llc_test_checks.py in llvm?)

[leonardchan]

OK so I instead wrote a script that generated the tests for me since changing each individual case by hand was painful. All the -cc1 tests now match, although I don't know any ways of simplifying the IR further to help reduce test size. I would definitely love to know if there are any ways to simplify the instructions. Unfortunately for convergent.cl, the codegen still differs slightly around the end of the loop when using -instcombine in addition to instnamer:

< for.body:                                         ; preds = %for.body, %entry
<   %i.03 = phi i32 [ 0, %entry ], [ %inc, %for.body ]
---
> for.body:                                         ; preds = %for.body.for.body_crit_edge, %entry
>   %inc.phi = phi i32 [ 1, %entry ], [ %inc.1, %for.body.for.body_crit_edge ]
92,94c92,97
<   %inc = add nuw nsw i32 %i.03, 1
<   %cmp = icmp ult i32 %inc, 10
<   br i1 %cmp, label %for.body, label %for.cond.cleanup
---
>   %cmp = icmp ult i32 %inc.phi, 10
>   br i1 %cmp, label %for.body.for.body_crit_edge, label %for.cond.cleanup
> 
> for.body.for.body_crit_edge:                      ; preds = %for.body
>   %inc.1 = add nuw nsw i32 %inc.phi, 1
>   br label %for.body

Also adding @craig.topper in case he as any insight on the updated tests.

clang/test/CodeGenOpenCL/convergent.cl

2

Unfortunately there doesn't seem to be a new PM equivalent for instnamer (but one does exist for instcombine). For simplicity, should we stick with using the legacy driver for now?

[chandlerc]

OMG, I'm so sorry, I had no idea that the tests would explode like that... Yeah, I don't think that's useful....

Maybe a better approach is to just explicitly run the code through opt -passes=instsimplify before handing it to FileCheck? That should produce almost identical results to the old PM's inliner?

Feel free to just try this out and report back -- if it isn't looking good, don't invest a bunch of time in it, we should talk to Craig and figure out what the right long-term strategy here is.

[leonardchan]

In D63174#1549601, @chandlerc wrote:

OMG, I'm so sorry, I had no idea that the tests would explode like that... Yeah, I don't think that's useful....

Maybe a better approach is to just explicitly run the code through opt -passes=instsimplify before handing it to FileCheck? That should produce almost identical results to the old PM's inliner?

Feel free to just try this out and report back -- if it isn't looking good, don't invest a bunch of time in it, we should talk to Craig and figure out what the right long-term strategy here is.

Hmm so -instsimplify and combinations of other passes like intsnamer and instcombine don't seem to clean this up any further. The IR still differs between PMs (at -O0), and I can provide the diffs if necessary.

In the short term, I imagine an "ok" solution for now would be to just run the original tests with -fno-experimental-new-pas-manager, and either have a separate file with these new PM -O1 checks at or separate runs in the same file to prevent regressions. I emailed Craig for his insight and info on these tests.

For now I reverted back to an early diff for this with the -O0 tests only.

[chandlerc]

Eh, this seems close enough now. I'd like a better approach for the x86 builtins, but no idea what it will end up being.

[craig.topper]

I'm going to try to work on the X86 tests. Can we hold off on committing those?

[leonardchan]

In D63174#1553096, @craig.topper wrote:

I'm going to try to work on the X86 tests. Can we hold off on committing those?

Ok. For now I'll just commit the convergent.cl test separately.

[leonardchan]

Replaced with D65110

clang/test/CodeGen/avx512-reduceMinMaxIntrin.c

-// RUN: %clang_cc1 -ffreestanding %s -O0 -triple=x86_64-apple-darwin -target-cpu skylake-avx512 -emit-llvm -o - -Wall -Werror | FileCheck %s
+// RUN: %clang_cc1 -fno-experimental-new-pass-manager -ffreestanding %s -O0 -triple=x86_64-apple-darwin -target-cpu skylake-avx512 -emit-llvm -o - -Wall -Werror | FileCheck %s
 
 #include <immintrin.h>
 
 // CHECK-LABEL: define i64 @test_mm512_reduce_max_epi64(<8 x i64> %__W) #0 {
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[__A_ADDR_I7_I:%.*]] = alloca <8 x i64>, align 64
 // CHECK-NEXT:    [[__B_ADDR_I8_I:%.*]] = alloca <8 x i64>, align 64
 // CHECK-NEXT:    [[__A_ADDR_I5_I:%.*]] = alloca <8 x i64>, align 64

clang/test/CodeGen/avx512f-builtins.c

-// RUN: %clang_cc1 -ffreestanding %s -triple=x86_64-apple-darwin -target-feature +avx512f -emit-llvm -o - -Wall -Werror | FileCheck %s
-// RUN: %clang_cc1 -fms-extensions -fms-compatibility -ffreestanding %s -triple=x86_64-windows-msvc -target-feature +avx512f -emit-llvm -o - -Wall -Werror | FileCheck %s
+// RUN: %clang_cc1 -fno-experimental-new-pass-manager -ffreestanding %s -triple=x86_64-apple-darwin -target-feature +avx512f -emit-llvm -o - -Wall -Werror | FileCheck %s
+// RUN: %clang_cc1 -fno-experimental-new-pass-manager -fms-extensions -fms-compatibility -ffreestanding %s -triple=x86_64-windows-msvc -target-feature +avx512f -emit-llvm -o - -Wall -Werror | FileCheck %s
+
+// There are a few cases where instead accpeting the result of an instruction
+// directly as an argument to a select, it instead goes through some bitcasts.
+// RUN: %clang_cc1 -fexperimental-new-pass-manager -ffreestanding %s -triple=x86_64-apple-darwin -target-feature +avx512f -emit-llvm -o - -Wall -Werror | FileCheck %s --check-prefixes=CHECK,CHECK-NEWPM
+// RUN: %clang_cc1 -fexperimental-new-pass-manager -fms-extensions -fms-compatibility -ffreestanding %s -triple=x86_64-windows-msvc -target-feature +avx512f -emit-llvm -o - -Wall -Werror | FileCheck %s --check-prefixes=CHECK,CHECK-NEWPM
 
 #include <immintrin.h>
 
 __m512d test_mm512_sqrt_pd(__m512d a)
 {
   // CHECK-LABEL: @test_mm512_sqrt_pd
   // CHECK: call <8 x double> @llvm.sqrt.v8f64(<8 x double> %{{.*}})
   return _mm512_sqrt_pd(a);
__m512i test_mm512_maskz_abs_epi64 (__mmask8 __U, __m512i __A)
   // CHECK: [[CMP:%.*]] = icmp sgt <8 x i64> [[A]], zeroinitializer
   // CHECK: [[SEL:%.*]] = select <8 x i1> [[CMP]], <8 x i64> [[A]], <8 x i64> [[SUB]]
   // CHECK: select <8 x i1> %{{.*}}, <8 x i64> [[SEL]], <8 x i64> %{{.*}}
   return _mm512_maskz_abs_epi64 (__U,__A);
 }
 
 __m512i test_mm512_mask_abs_epi32 (__m512i __W, __mmask16 __U, __m512i __A)
 {
-  // CHECK-LABEL: @test_mm512_mask_abs_epi32 
+  // CHECK-LABEL: @test_mm512_mask_abs_epi32
   // CHECK: [[SUB:%.*]] = sub <16 x i32> zeroinitializer, [[A:%.*]]
   // CHECK: [[CMP:%.*]] = icmp sgt <16 x i32> [[A]], zeroinitializer
   // CHECK: [[SEL:%.*]] = select <16 x i1> [[CMP]], <16 x i32> [[A]], <16 x i32> [[SUB]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast <16 x i32> [[SEL]] to <8 x i64>
+  // CHECK-NEWPM: [[SEL:%.*]] = bitcast <8 x i64> [[TMP]] to <16 x i32>
   // CHECK: select <16 x i1> %{{.*}}, <16 x i32> [[SEL]], <16 x i32> %{{.*}}
   return _mm512_mask_abs_epi32 (__W,__U,__A);
 }
 
 __m512i test_mm512_maskz_abs_epi32 (__mmask16 __U, __m512i __A)
 {
-  // CHECK-LABEL: @test_mm512_maskz_abs_epi32 
+  // CHECK-LABEL: @test_mm512_maskz_abs_epi32
   // CHECK: [[SUB:%.*]] = sub <16 x i32> zeroinitializer, [[A:%.*]]
   // CHECK: [[CMP:%.*]] = icmp sgt <16 x i32> [[A]], zeroinitializer
   // CHECK: [[SEL:%.*]] = select <16 x i1> [[CMP]], <16 x i32> [[A]], <16 x i32> [[SUB]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast <16 x i32> [[SEL]] to <8 x i64>
+  // CHECK-NEWPM: [[SEL:%.*]] = bitcast <8 x i64> [[TMP]] to <16 x i32>
   // CHECK: select <16 x i1> %{{.*}}, <16 x i32> [[SEL]], <16 x i32> %{{.*}}
   return _mm512_maskz_abs_epi32 (__U,__A);
 }
 
 __m512 test_mm512_setr_ps (float __A, float __B, float __C, float __D,
                           float __E, float __F, float __G, float __H,
                           float __I, float __J, float __K, float __L,
                           float __M, float __N, float __O, float __P)

clang/test/CodeGen/avx512vl-builtins.c

-// RUN: %clang_cc1 -ffreestanding %s -triple=x86_64-apple-darwin -target-feature +avx512f -target-feature +avx512vl -emit-llvm -o - -Wall -Werror | FileCheck %s
+// RUN: %clang_cc1 -ffreestanding %s -fno-experimental-new-pass-manager -triple=x86_64-apple-darwin -target-feature +avx512f -target-feature +avx512vl -emit-llvm -o - -Wall -Werror | FileCheck %s
 
+// There are a few cases where instead accpeting the result of an instruction
+// directly as an argument to a select, it instead goes through some bitcasts.
+// RUN: %clang_cc1 -ffreestanding %s -fexperimental-new-pass-manager -triple=x86_64-apple-darwin -target-feature +avx512f -target-feature +avx512vl -emit-llvm -o - -Wall -Werror | FileCheck %s --check-prefixes=CHECK,CHECK-NEWPM
 
 #include <immintrin.h>
 
 __mmask8 test_mm_cmpeq_epu32_mask(__m128i __a, __m128i __b) {
   // CHECK-LABEL: @test_mm_cmpeq_epu32_mask
   // CHECK: icmp eq <4 x i32> %{{.*}}, %{{.*}}
   // CHECK: shufflevector <4 x i1> %{{.*}}, <4 x i1> zeroinitializer, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
   return (__mmask8)_mm_cmpeq_epu32_mask(__a, __b);
__m256 test_mm256_maskz_mul_ps(__mmask8 __U, __m256 __A, __m256 __B) {
   // CHECK: select <8 x i1> %{{.*}}, <8 x float> %{{.*}}, <8 x float> %{{.*}}
   return _mm256_maskz_mul_ps(__U,__A,__B); 
 }
 __m128i test_mm_mask_abs_epi32(__m128i __W, __mmask8 __U, __m128i __A) {
   // CHECK-LABEL: @test_mm_mask_abs_epi32
   // CHECK: [[SUB:%.*]] = sub <4 x i32> zeroinitializer, [[A:%.*]]
   // CHECK: [[CMP:%.*]] = icmp sgt <4 x i32> [[A]], zeroinitializer
   // CHECK: [[SEL:%.*]] = select <4 x i1> [[CMP]], <4 x i32> [[A]], <4 x i32> [[SUB]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast <4 x i32> [[SEL]] to <2 x i64>
+  // CHECK-NEWPM: [[SEL:%.*]] = bitcast <2 x i64> [[TMP]] to <4 x i32>
   // CHECK: select <4 x i1> %{{.*}}, <4 x i32> [[SEL]], <4 x i32> %{{.*}}
   return _mm_mask_abs_epi32(__W,__U,__A); 
 }
 __m128i test_mm_maskz_abs_epi32(__mmask8 __U, __m128i __A) {
   // CHECK-LABEL: @test_mm_maskz_abs_epi32
   // CHECK: [[SUB:%.*]] = sub <4 x i32> zeroinitializer, [[A:%.*]]
   // CHECK: [[CMP:%.*]] = icmp sgt <4 x i32> [[A]], zeroinitializer
   // CHECK: [[SEL:%.*]] = select <4 x i1> [[CMP]], <4 x i32> [[A]], <4 x i32> [[SUB]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast <4 x i32> [[SEL]] to <2 x i64>
+  // CHECK-NEWPM: [[SEL:%.*]] = bitcast <2 x i64> [[TMP]] to <4 x i32>
   // CHECK: select <4 x i1> %{{.*}}, <4 x i32> [[SEL]], <4 x i32> %{{.*}}
   return _mm_maskz_abs_epi32(__U,__A); 
 }
 __m256i test_mm256_mask_abs_epi32(__m256i __W, __mmask8 __U, __m256i __A) {
   // CHECK-LABEL: @test_mm256_mask_abs_epi32
   // CHECK: [[SUB:%.*]] = sub <8 x i32> zeroinitializer, [[A:%.*]]
   // CHECK: [[CMP:%.*]] = icmp sgt <8 x i32> [[A]], zeroinitializer
   // CHECK: [[SEL:%.*]] = select <8 x i1> [[CMP]], <8 x i32> [[A]], <8 x i32> [[SUB]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast <8 x i32> [[SEL]] to <4 x i64>
+  // CHECK-NEWPM: [[SEL:%.*]] = bitcast <4 x i64> [[TMP]] to <8 x i32>
   // CHECK: select <8 x i1> %{{.*}}, <8 x i32> [[SEL]], <8 x i32> %{{.*}}
   return _mm256_mask_abs_epi32(__W,__U,__A); 
 }
 __m256i test_mm256_maskz_abs_epi32(__mmask8 __U, __m256i __A) {
   // CHECK-LABEL: @test_mm256_maskz_abs_epi32
   // CHECK: [[SUB:%.*]] = sub <8 x i32> zeroinitializer, [[A:%.*]]
   // CHECK: [[CMP:%.*]] = icmp sgt <8 x i32> [[A]], zeroinitializer
   // CHECK: [[SEL:%.*]] = select <8 x i1> [[CMP]], <8 x i32> [[A]], <8 x i32> [[SUB]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast <8 x i32> [[SEL]] to <4 x i64>
+  // CHECK-NEWPM: [[SEL:%.*]] = bitcast <4 x i64> [[TMP]] to <8 x i32>
   // CHECK: select <8 x i1> %{{.*}}, <8 x i32> [[SEL]], <8 x i32> %{{.*}}
   return _mm256_maskz_abs_epi32(__U,__A); 
 }
 __m128i test_mm_abs_epi64(__m128i __A) {
   // CHECK-LABEL: @test_mm_abs_epi64
   // CHECK: [[SUB:%.*]] = sub <2 x i64> zeroinitializer, [[A:%.*]]
   // CHECK: [[CMP:%.*]] = icmp sgt <2 x i64> [[A]], zeroinitializer
   // CHECK: select <2 x i1> [[CMP]], <2 x i64> [[A]], <2 x i64> [[SUB]]
__m256i test_mm256_maskz_abs_epi64(__mmask8 __U, __m256i __A) {
   // CHECK: [[SEL:%.*]] = select <4 x i1> [[CMP]], <4 x i64> [[A]], <4 x i64> [[SUB]]
   // CHECK: select <4 x i1> %{{.*}}, <4 x i64> [[SEL]], <4 x i64> %{{.*}}
   return _mm256_maskz_abs_epi64(__U,__A); 
 }
 __m128i test_mm_maskz_max_epi32(__mmask8 __M, __m128i __A, __m128i __B) {
   // CHECK-LABEL: @test_mm_maskz_max_epi32
   // CHECK:       [[CMP:%.*]] = icmp sgt <4 x i32> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <4 x i1> [[CMP]], <4 x i32> [[X]], <4 x i32> [[Y]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast <4 x i32> [[RES]] to <2 x i64>
+  // CHECK-NEWPM: [[RES:%.*]] = bitcast <2 x i64> [[TMP]] to <4 x i32>
   // CHECK:       select <4 x i1> {{.*}}, <4 x i32> [[RES]], <4 x i32> {{.*}}
   return _mm_maskz_max_epi32(__M,__A,__B); 
 }
 __m128i test_mm_mask_max_epi32(__m128i __W, __mmask8 __M, __m128i __A, __m128i __B) {
   // CHECK-LABEL: @test_mm_mask_max_epi32
   // CHECK:       [[CMP:%.*]] = icmp sgt <4 x i32> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <4 x i1> [[CMP]], <4 x i32> [[X]], <4 x i32> [[Y]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast <4 x i32> [[RES]] to <2 x i64>
+  // CHECK-NEWPM: [[RES:%.*]] = bitcast <2 x i64> [[TMP]] to <4 x i32>
   // CHECK:       select <4 x i1> {{.*}}, <4 x i32> [[RES]], <4 x i32> {{.*}}
   return _mm_mask_max_epi32(__W,__M,__A,__B); 
 }
 __m256i test_mm256_maskz_max_epi32(__mmask8 __M, __m256i __A, __m256i __B) {
   // CHECK-LABEL: @test_mm256_maskz_max_epi32
   // CHECK:       [[CMP:%.*]] = icmp sgt <8 x i32> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <8 x i1> [[CMP]], <8 x i32> [[X]], <8 x i32> [[Y]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast <8 x i32> [[RES]] to <4 x i64>
+  // CHECK-NEWPM: [[RES:%.*]] = bitcast <4 x i64> [[TMP]] to <8 x i32>
   // CHECK:       select <8 x i1> {{.*}}, <8 x i32> [[RES]], <8 x i32> {{.*}}
   return _mm256_maskz_max_epi32(__M,__A,__B); 
 }
 __m256i test_mm256_mask_max_epi32(__m256i __W, __mmask8 __M, __m256i __A, __m256i __B) {
   // CHECK-LABEL: @test_mm256_mask_max_epi32
   // CHECK:       [[CMP:%.*]] = icmp sgt <8 x i32> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <8 x i1> [[CMP]], <8 x i32> [[X]], <8 x i32> [[Y]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast <8 x i32> [[RES]] to <4 x i64>
+  // CHECK-NEWPM: [[RES:%.*]] = bitcast <4 x i64> [[TMP]] to <8 x i32>
   // CHECK:       select <8 x i1> {{.*}}, <8 x i32> [[RES]], <8 x i32> {{.*}}
   return _mm256_mask_max_epi32(__W,__M,__A,__B); 
 }
 __m128i test_mm_maskz_max_epi64(__mmask8 __M, __m128i __A, __m128i __B) {
   // CHECK-LABEL: @test_mm_maskz_max_epi64
   // CHECK:       [[CMP:%.*]] = icmp sgt <2 x i64> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <2 x i1> [[CMP]], <2 x i64> [[X]], <2 x i64> [[Y]]
   // CHECK:       select <2 x i1> {{.*}}, <2 x i64> [[RES]], <2 x i64> {{.*}}
__m256i test_mm256_max_epi64(__m256i __A, __m256i __B) {
   // CHECK:       [[CMP:%.*]] = icmp sgt <4 x i64> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <4 x i1> [[CMP]], <4 x i64> [[X]], <4 x i64> [[Y]]
   return _mm256_max_epi64(__A,__B); 
 }
 __m128i test_mm_maskz_max_epu32(__mmask8 __M, __m128i __A, __m128i __B) {
   // CHECK-LABEL: @test_mm_maskz_max_epu32
   // CHECK:       [[CMP:%.*]] = icmp ugt <4 x i32> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <4 x i1> [[CMP]], <4 x i32> [[X]], <4 x i32> [[Y]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast <4 x i32> [[RES]] to <2 x i64>
+  // CHECK-NEWPM: [[RES:%.*]] = bitcast <2 x i64> [[TMP]] to <4 x i32>
   // CHECK:       select <4 x i1> {{.*}}, <4 x i32> [[RES]], <4 x i32> {{.*}}
   return _mm_maskz_max_epu32(__M,__A,__B); 
 }
 __m128i test_mm_mask_max_epu32(__m128i __W, __mmask8 __M, __m128i __A, __m128i __B) {
   // CHECK-LABEL: @test_mm_mask_max_epu32
   // CHECK:       [[CMP:%.*]] = icmp ugt <4 x i32> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <4 x i1> [[CMP]], <4 x i32> [[X]], <4 x i32> [[Y]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast <4 x i32> [[RES]] to <2 x i64>
+  // CHECK-NEWPM: [[RES:%.*]] = bitcast <2 x i64> [[TMP]] to <4 x i32>
   // CHECK:       select <4 x i1> {{.*}}, <4 x i32> [[RES]], <4 x i32> {{.*}}
   return _mm_mask_max_epu32(__W,__M,__A,__B); 
 }
 __m256i test_mm256_maskz_max_epu32(__mmask8 __M, __m256i __A, __m256i __B) {
   // CHECK-LABEL: @test_mm256_maskz_max_epu32
   // CHECK:       [[CMP:%.*]] = icmp ugt <8 x i32> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <8 x i1> [[CMP]], <8 x i32> [[X]], <8 x i32> [[Y]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast <8 x i32> [[RES]] to <4 x i64>
+  // CHECK-NEWPM: [[RES:%.*]] = bitcast <4 x i64> [[TMP]] to <8 x i32>
   // CHECK:       select <8 x i1> {{.*}}, <8 x i32> [[RES]], <8 x i32> {{.*}}
   return _mm256_maskz_max_epu32(__M,__A,__B); 
 }
 __m256i test_mm256_mask_max_epu32(__m256i __W, __mmask8 __M, __m256i __A, __m256i __B) {
   // CHECK-LABEL: @test_mm256_mask_max_epu32
   // CHECK:       [[CMP:%.*]] = icmp ugt <8 x i32> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <8 x i1> [[CMP]], <8 x i32> [[X]], <8 x i32> [[Y]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast <8 x i32> [[RES]] to <4 x i64>
+  // CHECK-NEWPM: [[RES:%.*]] = bitcast <4 x i64> [[TMP]] to <8 x i32>
   // CHECK:       select <8 x i1> {{.*}}, <8 x i32> [[RES]], <8 x i32> {{.*}}
   return _mm256_mask_max_epu32(__W,__M,__A,__B); 
 }
 __m128i test_mm_maskz_max_epu64(__mmask8 __M, __m128i __A, __m128i __B) {
   // CHECK-LABEL: @test_mm_maskz_max_epu64
   // CHECK:       [[CMP:%.*]] = icmp ugt <2 x i64> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <2 x i1> [[CMP]], <2 x i64> [[X]], <2 x i64> [[Y]]
   // CHECK:       select <2 x i1> {{.*}}, <2 x i64> [[RES]], <2 x i64> {{.*}}
__m256i test_mm256_mask_max_epu64(__m256i __W, __mmask8 __M, __m256i __A, __m256i __B) {
   // CHECK-NEXT:  [[RES:%.*]] = select <4 x i1> [[CMP]], <4 x i64> [[X]], <4 x i64> [[Y]]
   // CHECK:       select <4 x i1> {{.*}}, <4 x i64> [[RES]], <4 x i64> {{.*}}
   return _mm256_mask_max_epu64(__W,__M,__A,__B); 
 }
 __m128i test_mm_maskz_min_epi32(__mmask8 __M, __m128i __A, __m128i __B) {
   // CHECK-LABEL: @test_mm_maskz_min_epi32
   // CHECK:       [[CMP:%.*]] = icmp slt <4 x i32> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <4 x i1> [[CMP]], <4 x i32> [[X]], <4 x i32> [[Y]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast <4 x i32> [[RES]] to <2 x i64>
+  // CHECK-NEWPM: [[RES:%.*]] = bitcast <2 x i64> [[TMP]] to <4 x i32>
   // CHECK:       select <4 x i1> {{.*}}, <4 x i32> [[RES]], <4 x i32> {{.*}}
   return _mm_maskz_min_epi32(__M,__A,__B); 
 }
 __m128i test_mm_mask_min_epi32(__m128i __W, __mmask8 __M, __m128i __A, __m128i __B) {
   // CHECK-LABEL: @test_mm_mask_min_epi32
   // CHECK:       [[CMP:%.*]] = icmp slt <4 x i32> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <4 x i1> [[CMP]], <4 x i32> [[X]], <4 x i32> [[Y]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast <4 x i32> [[RES]] to <2 x i64>
+  // CHECK-NEWPM: [[RES:%.*]] = bitcast <2 x i64> [[TMP]] to <4 x i32>
   // CHECK:       select <4 x i1> {{.*}}, <4 x i32> [[RES]], <4 x i32> {{.*}}
   return _mm_mask_min_epi32(__W,__M,__A,__B); 
 }
 __m256i test_mm256_maskz_min_epi32(__mmask8 __M, __m256i __A, __m256i __B) {
   // CHECK-LABEL: @test_mm256_maskz_min_epi32
   // CHECK:       [[CMP:%.*]] = icmp slt <8 x i32> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <8 x i1> [[CMP]], <8 x i32> [[X]], <8 x i32> [[Y]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast <8 x i32> [[RES]] to <4 x i64>
+  // CHECK-NEWPM: [[RES:%.*]] = bitcast <4 x i64> [[TMP]] to <8 x i32>
   // CHECK:       select <8 x i1> {{.*}}, <8 x i32> [[RES]], <8 x i32> {{.*}}
   return _mm256_maskz_min_epi32(__M,__A,__B); 
 }
 __m256i test_mm256_mask_min_epi32(__m256i __W, __mmask8 __M, __m256i __A, __m256i __B) {
   // CHECK-LABEL: @test_mm256_mask_min_epi32
   // CHECK:       [[CMP:%.*]] = icmp slt <8 x i32> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <8 x i1> [[CMP]], <8 x i32> [[X]], <8 x i32> [[Y]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast <8 x i32> [[RES]] to <4 x i64>
+  // CHECK-NEWPM: [[RES:%.*]] = bitcast <4 x i64> [[TMP]] to <8 x i32>
   // CHECK:       select <8 x i1> {{.*}}, <8 x i32> [[RES]], <8 x i32> {{.*}}
   return _mm256_mask_min_epi32(__W,__M,__A,__B); 
 }
 __m128i test_mm_min_epi64(__m128i __A, __m128i __B) {
   // CHECK-LABEL: @test_mm_min_epi64
   // CHECK:       [[CMP:%.*]] = icmp slt <2 x i64> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <2 x i1> [[CMP]], <2 x i64> [[X]], <2 x i64> [[Y]]
   return _mm_min_epi64(__A,__B); 
__m256i test_mm256_maskz_min_epi64(__mmask8 __M, __m256i __A, __m256i __B) {
   // CHECK-NEXT:  [[RES:%.*]] = select <4 x i1> [[CMP]], <4 x i64> [[X]], <4 x i64> [[Y]]
   // CHECK:       select <4 x i1> {{.*}}, <4 x i64> [[RES]], <4 x i64> {{.*}}
   return _mm256_maskz_min_epi64(__M,__A,__B); 
 }
 __m128i test_mm_maskz_min_epu32(__mmask8 __M, __m128i __A, __m128i __B) {
   // CHECK-LABEL: @test_mm_maskz_min_epu32
   // CHECK:       [[CMP:%.*]] = icmp ult <4 x i32> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <4 x i1> [[CMP]], <4 x i32> [[X]], <4 x i32> [[Y]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast <4 x i32> [[RES]] to <2 x i64>
+  // CHECK-NEWPM: [[RES:%.*]] = bitcast <2 x i64> [[TMP]] to <4 x i32>
   // CHECK:       select <4 x i1> {{.*}}, <4 x i32> [[RES]], <4 x i32> {{.*}}
   return _mm_maskz_min_epu32(__M,__A,__B); 
 }
 __m128i test_mm_mask_min_epu32(__m128i __W, __mmask8 __M, __m128i __A, __m128i __B) {
   // CHECK-LABEL: @test_mm_mask_min_epu32
   // CHECK:       [[CMP:%.*]] = icmp ult <4 x i32> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <4 x i1> [[CMP]], <4 x i32> [[X]], <4 x i32> [[Y]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast <4 x i32> [[RES]] to <2 x i64>
+  // CHECK-NEWPM: [[RES:%.*]] = bitcast <2 x i64> [[TMP]] to <4 x i32>
   // CHECK:       select <4 x i1> {{.*}}, <4 x i32> [[RES]], <4 x i32> {{.*}}
   return _mm_mask_min_epu32(__W,__M,__A,__B); 
 }
 __m256i test_mm256_maskz_min_epu32(__mmask8 __M, __m256i __A, __m256i __B) {
   // CHECK-LABEL: @test_mm256_maskz_min_epu32
   // CHECK:       [[CMP:%.*]] = icmp ult <8 x i32> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <8 x i1> [[CMP]], <8 x i32> [[X]], <8 x i32> [[Y]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast <8 x i32> [[RES]] to <4 x i64>
+  // CHECK-NEWPM: [[RES:%.*]] = bitcast <4 x i64> [[TMP]] to <8 x i32>
   // CHECK:       select <8 x i1> {{.*}}, <8 x i32> [[RES]], <8 x i32> {{.*}}
   return _mm256_maskz_min_epu32(__M,__A,__B); 
 }
 __m256i test_mm256_mask_min_epu32(__m256i __W, __mmask8 __M, __m256i __A, __m256i __B) {
   // CHECK-LABEL: @test_mm256_mask_min_epu32
   // CHECK:       [[CMP:%.*]] = icmp ult <8 x i32> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <8 x i1> [[CMP]], <8 x i32> [[X]], <8 x i32> [[Y]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast <8 x i32> [[RES]] to <4 x i64>
+  // CHECK-NEWPM: [[RES:%.*]] = bitcast <4 x i64> [[TMP]] to <8 x i32>
   // CHECK:       select <8 x i1> {{.*}}, <8 x i32> [[RES]], <8 x i32> {{.*}}
   return _mm256_mask_min_epu32(__W,__M,__A,__B); 
 }
 __m128i test_mm_min_epu64(__m128i __A, __m128i __B) {
   // CHECK-LABEL: @test_mm_min_epu64
   // CHECK:       [[CMP:%.*]] = icmp ult <2 x i64> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <2 x i1> [[CMP]], <2 x i64> [[X]], <2 x i64> [[Y]]
   return _mm_min_epu64(__A,__B); 

clang/test/CodeGen/avx512vlbw-builtins.c

-// RUN: %clang_cc1 -ffreestanding %s -triple=x86_64-apple-darwin -target-feature +avx512bw -target-feature +avx512vl -emit-llvm -o - -Wall -Werror | FileCheck %s
-// RUN: %clang_cc1 -ffreestanding %s -triple=x86_64-apple-darwin -target-feature +avx512bw -target-feature +avx512vl -fno-signed-char -emit-llvm -o - -Wall -Werror | FileCheck %s
+// RUN: %clang_cc1 -ffreestanding %s -fno-experimental-new-pass-manager -triple=x86_64-apple-darwin -target-feature +avx512bw -target-feature +avx512vl -emit-llvm -o - -Wall -Werror | FileCheck %s
+// RUN: %clang_cc1 -ffreestanding %s -fno-experimental-new-pass-manager -triple=x86_64-apple-darwin -target-feature +avx512bw -target-feature +avx512vl -fno-signed-char -emit-llvm -o - -Wall -Werror | FileCheck %s
 
+// There are a few cases where instead accpeting the result of an instruction
+// directly as an argument to a select, it instead goes through some bitcasts.
+// RUN: %clang_cc1 -ffreestanding %s -fexperimental-new-pass-manager -triple=x86_64-apple-darwin -target-feature +avx512bw -target-feature +avx512vl -emit-llvm -o - -Wall -Werror | FileCheck %s --check-prefixes=CHECK,CHECK-NEWPM
+// RUN: %clang_cc1 -ffreestanding %s -fexperimental-new-pass-manager -triple=x86_64-apple-darwin -target-feature +avx512bw -target-feature +avx512vl -fno-signed-char -emit-llvm -o - -Wall -Werror | FileCheck %s --check-prefixes=CHECK,CHECK-NEWPM
 
 #include <immintrin.h>
 
 __mmask32 test_mm256_cmpeq_epi8_mask(__m256i __a, __m256i __b) {
   // CHECK-LABEL: @test_mm256_cmpeq_epi8_mask
   // CHECK: icmp eq <32 x i8> %{{.*}}, %{{.*}}
   return (__mmask32)_mm256_cmpeq_epi8_mask(__a, __b);
 }
__m256i test_mm256_mask_blend_epi16(__mmask16 __U, __m256i __A, __m256i __W) {
   return _mm256_mask_blend_epi16(__U,__A,__W); 
 }
 
 __m128i test_mm_mask_abs_epi8(__m128i __W, __mmask16 __U, __m128i __A) {
   // CHECK-LABEL: @test_mm_mask_abs_epi8
   // CHECK: [[SUB:%.*]] = sub <16 x i8> zeroinitializer, [[A:%.*]]
   // CHECK: [[CMP:%.*]] = icmp sgt <16 x i8> [[A]], zeroinitializer
   // CHECK: [[SEL:%.*]] = select <16 x i1> [[CMP]], <16 x i8> [[A]], <16 x i8> [[SUB]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast [[SRCTY:<.*>]] [[SEL]] to [[DSTTY:<.*>]]
+  // CHECK-NEWPM: [[SEL:%.*]] = bitcast [[DSTTY]] [[TMP]] to [[SRCTY]]
   // CHECK: select <16 x i1> %{{.*}}, <16 x i8> [[SEL]], <16 x i8> %{{.*}}
   return _mm_mask_abs_epi8(__W,__U,__A); 
 }
 
 __m128i test_mm_maskz_abs_epi8(__mmask16 __U, __m128i __A) {
   // CHECK-LABEL: @test_mm_maskz_abs_epi8
   // CHECK: [[SUB:%.*]] = sub <16 x i8> zeroinitializer, [[A:%.*]]
   // CHECK: [[CMP:%.*]] = icmp sgt <16 x i8> [[A]], zeroinitializer
   // CHECK: [[SEL:%.*]] = select <16 x i1> [[CMP]], <16 x i8> [[A]], <16 x i8> [[SUB]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast [[SRCTY:<.*>]] [[SEL]] to [[DSTTY:<.*>]]
+  // CHECK-NEWPM: [[SEL:%.*]] = bitcast [[DSTTY]] [[TMP]] to [[SRCTY]]
   // CHECK: select <16 x i1> %{{.*}}, <16 x i8> [[SEL]], <16 x i8> %{{.*}}
   return _mm_maskz_abs_epi8(__U,__A); 
 }
 
 __m256i test_mm256_mask_abs_epi8(__m256i __W, __mmask32 __U, __m256i __A) {
   // CHECK-LABEL: @test_mm256_mask_abs_epi8
   // CHECK: [[SUB:%.*]] = sub <32 x i8> zeroinitializer, [[A:%.*]]
   // CHECK: [[CMP:%.*]] = icmp sgt <32 x i8> [[A]], zeroinitializer
   // CHECK: [[SEL:%.*]] = select <32 x i1> [[CMP]], <32 x i8> [[A]], <32 x i8> [[SUB]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast [[SRCTY:<.*>]] [[SEL]] to [[DSTTY:<.*>]]
+  // CHECK-NEWPM: [[SEL:%.*]] = bitcast [[DSTTY]] [[TMP]] to [[SRCTY]]
   // CHECK: select <32 x i1> %{{.*}}, <32 x i8> [[SEL]], <32 x i8> %{{.*}}
   return _mm256_mask_abs_epi8(__W,__U,__A); 
 }
 
 __m256i test_mm256_maskz_abs_epi8(__mmask32 __U, __m256i __A) {
   // CHECK-LABEL: @test_mm256_maskz_abs_epi8
   // CHECK: [[SUB:%.*]] = sub <32 x i8> zeroinitializer, [[A:%.*]]
   // CHECK: [[CMP:%.*]] = icmp sgt <32 x i8> [[A]], zeroinitializer
   // CHECK: [[SEL:%.*]] = select <32 x i1> [[CMP]], <32 x i8> [[A]], <32 x i8> [[SUB]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast [[SRCTY:<.*>]] [[SEL]] to [[DSTTY:<.*>]]
+  // CHECK-NEWPM: [[SEL:%.*]] = bitcast [[DSTTY]] [[TMP]] to [[SRCTY]]
   // CHECK: select <32 x i1> %{{.*}}, <32 x i8> [[SEL]], <32 x i8> %{{.*}}
   return _mm256_maskz_abs_epi8(__U,__A); 
 }
 
 __m128i test_mm_mask_abs_epi16(__m128i __W, __mmask8 __U, __m128i __A) {
   // CHECK-LABEL: @test_mm_mask_abs_epi16
   // CHECK: [[SUB:%.*]] = sub <8 x i16> zeroinitializer, [[A:%.*]]
   // CHECK: [[CMP:%.*]] = icmp sgt <8 x i16> [[A]], zeroinitializer
   // CHECK: [[SEL:%.*]] = select <8 x i1> [[CMP]], <8 x i16> [[A]], <8 x i16> [[SUB]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast [[SRCTY:<.*>]] [[SEL]] to [[DSTTY:<.*>]]
+  // CHECK-NEWPM: [[SEL:%.*]] = bitcast [[DSTTY]] [[TMP]] to [[SRCTY]]
   // CHECK: select <8 x i1> %{{.*}}, <8 x i16> [[SEL]], <8 x i16> %{{.*}}
   return _mm_mask_abs_epi16(__W,__U,__A); 
 }
 
 __m128i test_mm_maskz_abs_epi16(__mmask8 __U, __m128i __A) {
   // CHECK-LABEL: @test_mm_maskz_abs_epi16
   // CHECK: [[SUB:%.*]] = sub <8 x i16> zeroinitializer, [[A:%.*]]
   // CHECK: [[CMP:%.*]] = icmp sgt <8 x i16> [[A]], zeroinitializer
   // CHECK: [[SEL:%.*]] = select <8 x i1> [[CMP]], <8 x i16> [[A]], <8 x i16> [[SUB]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast [[SRCTY:<.*>]] [[SEL]] to [[DSTTY:<.*>]]
+  // CHECK-NEWPM: [[SEL:%.*]] = bitcast [[DSTTY]] [[TMP]] to [[SRCTY]]
   // CHECK: select <8 x i1> %{{.*}}, <8 x i16> [[SEL]], <8 x i16> %{{.*}}
   return _mm_maskz_abs_epi16(__U,__A); 
 }
 
 __m256i test_mm256_mask_abs_epi16(__m256i __W, __mmask16 __U, __m256i __A) {
   // CHECK-LABEL: @test_mm256_mask_abs_epi16
   // CHECK: [[SUB:%.*]] = sub <16 x i16> zeroinitializer, [[A:%.*]]
   // CHECK: [[CMP:%.*]] = icmp sgt <16 x i16> [[A]], zeroinitializer
   // CHECK: [[SEL:%.*]] = select <16 x i1> [[CMP]], <16 x i16> [[A]], <16 x i16> [[SUB]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast [[SRCTY:<.*>]] [[SEL]] to [[DSTTY:<.*>]]
+  // CHECK-NEWPM: [[SEL:%.*]] = bitcast [[DSTTY]] [[TMP]] to [[SRCTY]]
   // CHECK: select <16 x i1> %{{.*}}, <16 x i16> [[SEL]], <16 x i16> %{{.*}}
   return _mm256_mask_abs_epi16(__W,__U,__A); 
 }
 
 __m256i test_mm256_maskz_abs_epi16(__mmask16 __U, __m256i __A) {
   // CHECK-LABEL: @test_mm256_maskz_abs_epi16
   // CHECK: [[SUB:%.*]] = sub <16 x i16> zeroinitializer, [[A:%.*]]
   // CHECK: [[CMP:%.*]] = icmp sgt <16 x i16> [[A]], zeroinitializer
   // CHECK: [[SEL:%.*]] = select <16 x i1> [[CMP]], <16 x i16> [[A]], <16 x i16> [[SUB]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast [[SRCTY:<.*>]] [[SEL]] to [[DSTTY:<.*>]]
+  // CHECK-NEWPM: [[SEL:%.*]] = bitcast [[DSTTY]] [[TMP]] to [[SRCTY]]
   // CHECK: select <16 x i1> %{{.*}}, <16 x i16> [[SEL]], <16 x i16> %{{.*}}
   return _mm256_maskz_abs_epi16(__U,__A); 
 }
 
 __m128i test_mm_maskz_packs_epi32(__mmask8 __M, __m128i __A, __m128i __B) {
   // CHECK-LABEL: @test_mm_maskz_packs_epi32
   // CHECK: @llvm.x86.sse2.packssdw
   // CHECK: select <8 x i1> %{{.*}}, <8 x i16> %{{.*}}, <8 x i16> %{{.*}}
__m256i test_mm256_maskz_avg_epu16(__mmask16 __U, __m256i __A, __m256i __B) {
   // CHECK: @llvm.x86.avx2.pavg.w
   // CHECK: select <16 x i1> %{{.*}}, <16 x i16> %{{.*}}, <16 x i16> %{{.*}}
   return _mm256_maskz_avg_epu16(__U,__A,__B); 
 }
 __m128i test_mm_maskz_max_epi8(__mmask16 __M, __m128i __A, __m128i __B) {
   // CHECK-LABEL: @test_mm_maskz_max_epi8
   // CHECK:       [[CMP:%.*]] = icmp sgt <16 x i8> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <16 x i1> [[CMP]], <16 x i8> [[X]], <16 x i8> [[Y]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast [[SRCTY:<.*>]] [[RES]] to [[DSTTY:<.*>]]
+  // CHECK-NEWPM: [[RES:%.*]] = bitcast [[DSTTY]] [[TMP]] to [[SRCTY]]
   // CHECK:       select <16 x i1> {{.*}}, <16 x i8> [[RES]], <16 x i8> {{.*}}
   return _mm_maskz_max_epi8(__M,__A,__B); 
 }
 __m128i test_mm_mask_max_epi8(__m128i __W, __mmask16 __M, __m128i __A, __m128i __B) {
   // CHECK-LABEL: @test_mm_mask_max_epi8
   // CHECK:       [[CMP:%.*]] = icmp sgt <16 x i8> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <16 x i1> [[CMP]], <16 x i8> [[X]], <16 x i8> [[Y]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast [[SRCTY:<.*>]] [[RES]] to [[DSTTY:<.*>]]
+  // CHECK-NEWPM: [[RES:%.*]] = bitcast [[DSTTY]] [[TMP]] to [[SRCTY]]
   // CHECK:       select <16 x i1> {{.*}}, <16 x i8> [[RES]], <16 x i8> {{.*}}
   return _mm_mask_max_epi8(__W,__M,__A,__B); 
 }
 __m256i test_mm256_maskz_max_epi8(__mmask32 __M, __m256i __A, __m256i __B) {
   // CHECK-LABEL: @test_mm256_maskz_max_epi8
   // CHECK:       [[CMP:%.*]] = icmp sgt <32 x i8> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <32 x i1> [[CMP]], <32 x i8> [[X]], <32 x i8> [[Y]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast [[SRCTY:<.*>]] [[RES]] to [[DSTTY:<.*>]]
+  // CHECK-NEWPM: [[RES:%.*]] = bitcast [[DSTTY]] [[TMP]] to [[SRCTY]]
   // CHECK:       select <32 x i1> {{.*}}, <32 x i8> [[RES]], <32 x i8> {{.*}}
   return _mm256_maskz_max_epi8(__M,__A,__B); 
 }
 __m256i test_mm256_mask_max_epi8(__m256i __W, __mmask32 __M, __m256i __A, __m256i __B) {
   // CHECK-LABEL: @test_mm256_mask_max_epi8
   // CHECK:       [[CMP:%.*]] = icmp sgt <32 x i8> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <32 x i1> [[CMP]], <32 x i8> [[X]], <32 x i8> [[Y]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast [[SRCTY:<.*>]] [[RES]] to [[DSTTY:<.*>]]
+  // CHECK-NEWPM: [[RES:%.*]] = bitcast [[DSTTY]] [[TMP]] to [[SRCTY]]
   // CHECK:       select <32 x i1> {{.*}}, <32 x i8> [[RES]], <32 x i8> {{.*}}
   return _mm256_mask_max_epi8(__W,__M,__A,__B); 
 }
 __m128i test_mm_maskz_max_epi16(__mmask8 __M, __m128i __A, __m128i __B) {
   // CHECK-LABEL: @test_mm_maskz_max_epi16
   // CHECK:       [[CMP:%.*]] = icmp sgt <8 x i16> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <8 x i1> [[CMP]], <8 x i16> [[X]], <8 x i16> [[Y]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast [[SRCTY:<.*>]] [[RES]] to [[DSTTY:<.*>]]
+  // CHECK-NEWPM: [[RES:%.*]] = bitcast [[DSTTY]] [[TMP]] to [[SRCTY]]
   // CHECK:       select <8 x i1> {{.*}}, <8 x i16> [[RES]], <8 x i16> {{.*}}
   return _mm_maskz_max_epi16(__M,__A,__B); 
 }
 __m128i test_mm_mask_max_epi16(__m128i __W, __mmask8 __M, __m128i __A, __m128i __B) {
   // CHECK-LABEL: @test_mm_mask_max_epi16
   // CHECK:       [[CMP:%.*]] = icmp sgt <8 x i16> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <8 x i1> [[CMP]], <8 x i16> [[X]], <8 x i16> [[Y]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast [[SRCTY:<.*>]] [[RES]] to [[DSTTY:<.*>]]
+  // CHECK-NEWPM: [[RES:%.*]] = bitcast [[DSTTY]] [[TMP]] to [[SRCTY]]
   // CHECK:       select <8 x i1> {{.*}}, <8 x i16> [[RES]], <8 x i16> {{.*}}
   return _mm_mask_max_epi16(__W,__M,__A,__B); 
 }
 __m256i test_mm256_maskz_max_epi16(__mmask16 __M, __m256i __A, __m256i __B) {
   // CHECK-LABEL: @test_mm256_maskz_max_epi16
   // CHECK:       [[CMP:%.*]] = icmp sgt <16 x i16> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <16 x i1> [[CMP]], <16 x i16> [[X]], <16 x i16> [[Y]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast [[SRCTY:<.*>]] [[RES]] to [[DSTTY:<.*>]]
+  // CHECK-NEWPM: [[RES:%.*]] = bitcast [[DSTTY]] [[TMP]] to [[SRCTY]]
   // CHECK:       select <16 x i1> {{.*}}, <16 x i16> [[RES]], <16 x i16> {{.*}}
   return _mm256_maskz_max_epi16(__M,__A,__B); 
 }
 __m256i test_mm256_mask_max_epi16(__m256i __W, __mmask16 __M, __m256i __A, __m256i __B) {
   // CHECK-LABEL: @test_mm256_mask_max_epi16
   // CHECK:       [[CMP:%.*]] = icmp sgt <16 x i16> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <16 x i1> [[CMP]], <16 x i16> [[X]], <16 x i16> [[Y]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast [[SRCTY:<.*>]] [[RES]] to [[DSTTY:<.*>]]
+  // CHECK-NEWPM: [[RES:%.*]] = bitcast [[DSTTY]] [[TMP]] to [[SRCTY]]
   // CHECK:       select <16 x i1> {{.*}}, <16 x i16> [[RES]], <16 x i16> {{.*}}
   return _mm256_mask_max_epi16(__W,__M,__A,__B); 
 }
 __m128i test_mm_maskz_max_epu8(__mmask16 __M, __m128i __A, __m128i __B) {
   // CHECK-LABEL: @test_mm_maskz_max_epu8
   // CHECK:       [[CMP:%.*]] = icmp ugt <16 x i8> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <16 x i1> [[CMP]], <16 x i8> [[X]], <16 x i8> [[Y]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast [[SRCTY:<.*>]] [[RES]] to [[DSTTY:<.*>]]
+  // CHECK-NEWPM: [[RES:%.*]] = bitcast [[DSTTY]] [[TMP]] to [[SRCTY]]
   // CHECK:       select <16 x i1> {{.*}}, <16 x i8> [[RES]], <16 x i8> {{.*}}
   return _mm_maskz_max_epu8(__M,__A,__B); 
 }
 __m128i test_mm_mask_max_epu8(__m128i __W, __mmask16 __M, __m128i __A, __m128i __B) {
   // CHECK-LABEL: @test_mm_mask_max_epu8
   // CHECK:       [[CMP:%.*]] = icmp ugt <16 x i8> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <16 x i1> [[CMP]], <16 x i8> [[X]], <16 x i8> [[Y]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast [[SRCTY:<.*>]] [[RES]] to [[DSTTY:<.*>]]
+  // CHECK-NEWPM: [[RES:%.*]] = bitcast [[DSTTY]] [[TMP]] to [[SRCTY]]
   // CHECK:       select <16 x i1> {{.*}}, <16 x i8> [[RES]], <16 x i8> {{.*}}
   return _mm_mask_max_epu8(__W,__M,__A,__B); 
 }
 __m256i test_mm256_maskz_max_epu8(__mmask32 __M, __m256i __A, __m256i __B) {
   // CHECK-LABEL: @test_mm256_maskz_max_epu8
   // CHECK:       [[CMP:%.*]] = icmp ugt <32 x i8> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <32 x i1> [[CMP]], <32 x i8> [[X]], <32 x i8> [[Y]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast [[SRCTY:<.*>]] [[RES]] to [[DSTTY:<.*>]]
+  // CHECK-NEWPM: [[RES:%.*]] = bitcast [[DSTTY]] [[TMP]] to [[SRCTY]]
   // CHECK:       select <32 x i1> {{.*}}, <32 x i8> [[RES]], <32 x i8> {{.*}}
   return _mm256_maskz_max_epu8(__M,__A,__B); 
 }
 __m256i test_mm256_mask_max_epu8(__m256i __W, __mmask32 __M, __m256i __A, __m256i __B) {
   // CHECK-LABEL: @test_mm256_mask_max_epu8
   // CHECK:       [[CMP:%.*]] = icmp ugt <32 x i8> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <32 x i1> [[CMP]], <32 x i8> [[X]], <32 x i8> [[Y]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast [[SRCTY:<.*>]] [[RES]] to [[DSTTY:<.*>]]
+  // CHECK-NEWPM: [[RES:%.*]] = bitcast [[DSTTY]] [[TMP]] to [[SRCTY]]
   // CHECK:       select <32 x i1> {{.*}}, <32 x i8> [[RES]], <32 x i8> {{.*}}
   return _mm256_mask_max_epu8(__W,__M,__A,__B); 
 }
 __m128i test_mm_maskz_max_epu16(__mmask8 __M, __m128i __A, __m128i __B) {
   // CHECK-LABEL: @test_mm_maskz_max_epu16
   // CHECK:       [[CMP:%.*]] = icmp ugt <8 x i16> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <8 x i1> [[CMP]], <8 x i16> [[X]], <8 x i16> [[Y]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast [[SRCTY:<.*>]] [[RES]] to [[DSTTY:<.*>]]
+  // CHECK-NEWPM: [[RES:%.*]] = bitcast [[DSTTY]] [[TMP]] to [[SRCTY]]
   // CHECK:       select <8 x i1> {{.*}}, <8 x i16> [[RES]], <8 x i16> {{.*}}
   return _mm_maskz_max_epu16(__M,__A,__B); 
 }
 __m128i test_mm_mask_max_epu16(__m128i __W, __mmask8 __M, __m128i __A, __m128i __B) {
   // CHECK-LABEL: @test_mm_mask_max_epu16
   // CHECK:       [[CMP:%.*]] = icmp ugt <8 x i16> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <8 x i1> [[CMP]], <8 x i16> [[X]], <8 x i16> [[Y]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast [[SRCTY:<.*>]] [[RES]] to [[DSTTY:<.*>]]
+  // CHECK-NEWPM: [[RES:%.*]] = bitcast [[DSTTY]] [[TMP]] to [[SRCTY]]
   // CHECK:       select <8 x i1> {{.*}}, <8 x i16> [[RES]], <8 x i16> {{.*}}
   return _mm_mask_max_epu16(__W,__M,__A,__B); 
 }
 __m256i test_mm256_maskz_max_epu16(__mmask16 __M, __m256i __A, __m256i __B) {
   // CHECK-LABEL: @test_mm256_maskz_max_epu16
   // CHECK:       [[CMP:%.*]] = icmp ugt <16 x i16> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <16 x i1> [[CMP]], <16 x i16> [[X]], <16 x i16> [[Y]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast [[SRCTY:<.*>]] [[RES]] to [[DSTTY:<.*>]]
+  // CHECK-NEWPM: [[RES:%.*]] = bitcast [[DSTTY]] [[TMP]] to [[SRCTY]]
   // CHECK:       select <16 x i1> {{.*}}, <16 x i16> [[RES]], <16 x i16> {{.*}}
   return _mm256_maskz_max_epu16(__M,__A,__B); 
 }
 __m256i test_mm256_mask_max_epu16(__m256i __W, __mmask16 __M, __m256i __A, __m256i __B) {
   // CHECK-LABEL: @test_mm256_mask_max_epu16
   // CHECK:       [[CMP:%.*]] = icmp ugt <16 x i16> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <16 x i1> [[CMP]], <16 x i16> [[X]], <16 x i16> [[Y]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast [[SRCTY:<.*>]] [[RES]] to [[DSTTY:<.*>]]
+  // CHECK-NEWPM: [[RES:%.*]] = bitcast [[DSTTY]] [[TMP]] to [[SRCTY]]
   // CHECK:       select <16 x i1> {{.*}}, <16 x i16> [[RES]], <16 x i16> {{.*}}
   return _mm256_mask_max_epu16(__W,__M,__A,__B); 
 }
 __m128i test_mm_maskz_min_epi8(__mmask16 __M, __m128i __A, __m128i __B) {
   // CHECK-LABEL: @test_mm_maskz_min_epi8
   // CHECK:       [[CMP:%.*]] = icmp slt <16 x i8> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <16 x i1> [[CMP]], <16 x i8> [[X]], <16 x i8> [[Y]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast [[SRCTY:<.*>]] [[RES]] to [[DSTTY:<.*>]]
+  // CHECK-NEWPM: [[RES:%.*]] = bitcast [[DSTTY]] [[TMP]] to [[SRCTY]]
   // CHECK:       select <16 x i1> {{.*}}, <16 x i8> [[RES]], <16 x i8> {{.*}}
   return _mm_maskz_min_epi8(__M,__A,__B); 
 }
 __m128i test_mm_mask_min_epi8(__m128i __W, __mmask16 __M, __m128i __A, __m128i __B) {
   // CHECK-LABEL: @test_mm_mask_min_epi8
   // CHECK:       [[CMP:%.*]] = icmp slt <16 x i8> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <16 x i1> [[CMP]], <16 x i8> [[X]], <16 x i8> [[Y]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast [[SRCTY:<.*>]] [[RES]] to [[DSTTY:<.*>]]
+  // CHECK-NEWPM: [[RES:%.*]] = bitcast [[DSTTY]] [[TMP]] to [[SRCTY]]
   // CHECK:       select <16 x i1> {{.*}}, <16 x i8> [[RES]], <16 x i8> {{.*}}
   return _mm_mask_min_epi8(__W,__M,__A,__B); 
 }
 __m256i test_mm256_maskz_min_epi8(__mmask32 __M, __m256i __A, __m256i __B) {
   // CHECK-LABEL: @test_mm256_maskz_min_epi8
   // CHECK:       [[CMP:%.*]] = icmp slt <32 x i8> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <32 x i1> [[CMP]], <32 x i8> [[X]], <32 x i8> [[Y]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast [[SRCTY:<.*>]] [[RES]] to [[DSTTY:<.*>]]
+  // CHECK-NEWPM: [[RES:%.*]] = bitcast [[DSTTY]] [[TMP]] to [[SRCTY]]
   // CHECK:       select <32 x i1> {{.*}}, <32 x i8> [[RES]], <32 x i8> {{.*}}
   return _mm256_maskz_min_epi8(__M,__A,__B); 
 }
 __m256i test_mm256_mask_min_epi8(__m256i __W, __mmask32 __M, __m256i __A, __m256i __B) {
   // CHECK-LABEL: @test_mm256_mask_min_epi8
   // CHECK:       [[CMP:%.*]] = icmp slt <32 x i8> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <32 x i1> [[CMP]], <32 x i8> [[X]], <32 x i8> [[Y]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast [[SRCTY:<.*>]] [[RES]] to [[DSTTY:<.*>]]
+  // CHECK-NEWPM: [[RES:%.*]] = bitcast [[DSTTY]] [[TMP]] to [[SRCTY]]
   // CHECK:       select <32 x i1> {{.*}}, <32 x i8> [[RES]], <32 x i8> {{.*}}
   return _mm256_mask_min_epi8(__W,__M,__A,__B); 
 }
 __m128i test_mm_maskz_min_epi16(__mmask8 __M, __m128i __A, __m128i __B) {
   // CHECK-LABEL: @test_mm_maskz_min_epi16
   // CHECK:       [[CMP:%.*]] = icmp slt <8 x i16> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <8 x i1> [[CMP]], <8 x i16> [[X]], <8 x i16> [[Y]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast [[SRCTY:<.*>]] [[RES]] to [[DSTTY:<.*>]]
+  // CHECK-NEWPM: [[RES:%.*]] = bitcast [[DSTTY]] [[TMP]] to [[SRCTY]]
   // CHECK:       select <8 x i1> {{.*}}, <8 x i16> [[RES]], <8 x i16> {{.*}}
   return _mm_maskz_min_epi16(__M,__A,__B); 
 }
 __m128i test_mm_mask_min_epi16(__m128i __W, __mmask8 __M, __m128i __A, __m128i __B) {
   // CHECK-LABEL: @test_mm_mask_min_epi16
   // CHECK:       [[CMP:%.*]] = icmp slt <8 x i16> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <8 x i1> [[CMP]], <8 x i16> [[X]], <8 x i16> [[Y]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast [[SRCTY:<.*>]] [[RES]] to [[DSTTY:<.*>]]
+  // CHECK-NEWPM: [[RES:%.*]] = bitcast [[DSTTY]] [[TMP]] to [[SRCTY]]
   // CHECK:       select <8 x i1> {{.*}}, <8 x i16> [[RES]], <8 x i16> {{.*}}
   return _mm_mask_min_epi16(__W,__M,__A,__B); 
 }
 __m256i test_mm256_maskz_min_epi16(__mmask16 __M, __m256i __A, __m256i __B) {
   // CHECK-LABEL: @test_mm256_maskz_min_epi16
   // CHECK:       [[CMP:%.*]] = icmp slt <16 x i16> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <16 x i1> [[CMP]], <16 x i16> [[X]], <16 x i16> [[Y]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast [[SRCTY:<.*>]] [[RES]] to [[DSTTY:<.*>]]
+  // CHECK-NEWPM: [[RES:%.*]] = bitcast [[DSTTY]] [[TMP]] to [[SRCTY]]
   // CHECK:       select <16 x i1> {{.*}}, <16 x i16> [[RES]], <16 x i16> {{.*}}
   return _mm256_maskz_min_epi16(__M,__A,__B); 
 }
 __m256i test_mm256_mask_min_epi16(__m256i __W, __mmask16 __M, __m256i __A, __m256i __B) {
   // CHECK-LABEL: @test_mm256_mask_min_epi16
   // CHECK:       [[CMP:%.*]] = icmp slt <16 x i16> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <16 x i1> [[CMP]], <16 x i16> [[X]], <16 x i16> [[Y]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast [[SRCTY:<.*>]] [[RES]] to [[DSTTY:<.*>]]
+  // CHECK-NEWPM: [[RES:%.*]] = bitcast [[DSTTY]] [[TMP]] to [[SRCTY]]
   // CHECK:       select <16 x i1> {{.*}}, <16 x i16> [[RES]], <16 x i16> {{.*}}
   return _mm256_mask_min_epi16(__W,__M,__A,__B); 
 }
 __m128i test_mm_maskz_min_epu8(__mmask16 __M, __m128i __A, __m128i __B) {
   // CHECK-LABEL: @test_mm_maskz_min_epu8
   // CHECK:       [[CMP:%.*]] = icmp ult <16 x i8> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <16 x i1> [[CMP]], <16 x i8> [[X]], <16 x i8> [[Y]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast [[SRCTY:<.*>]] [[RES]] to [[DSTTY:<.*>]]
+  // CHECK-NEWPM: [[RES:%.*]] = bitcast [[DSTTY]] [[TMP]] to [[SRCTY]]
   // CHECK:       select <16 x i1> {{.*}}, <16 x i8> [[RES]], <16 x i8> {{.*}}
   return _mm_maskz_min_epu8(__M,__A,__B); 
 }
 __m128i test_mm_mask_min_epu8(__m128i __W, __mmask16 __M, __m128i __A, __m128i __B) {
   // CHECK-LABEL: @test_mm_mask_min_epu8
   // CHECK:       [[CMP:%.*]] = icmp ult <16 x i8> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <16 x i1> [[CMP]], <16 x i8> [[X]], <16 x i8> [[Y]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast [[SRCTY:<.*>]] [[RES]] to [[DSTTY:<.*>]]
+  // CHECK-NEWPM: [[RES:%.*]] = bitcast [[DSTTY]] [[TMP]] to [[SRCTY]]
   // CHECK:       select <16 x i1> {{.*}}, <16 x i8> [[RES]], <16 x i8> {{.*}}
   return _mm_mask_min_epu8(__W,__M,__A,__B); 
 }
 __m256i test_mm256_maskz_min_epu8(__mmask32 __M, __m256i __A, __m256i __B) {
   // CHECK-LABEL: @test_mm256_maskz_min_epu8
   // CHECK:       [[CMP:%.*]] = icmp ult <32 x i8> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <32 x i1> [[CMP]], <32 x i8> [[X]], <32 x i8> [[Y]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast [[SRCTY:<.*>]] [[RES]] to [[DSTTY:<.*>]]
+  // CHECK-NEWPM: [[RES:%.*]] = bitcast [[DSTTY]] [[TMP]] to [[SRCTY]]
   // CHECK:       select <32 x i1> {{.*}}, <32 x i8> [[RES]], <32 x i8> {{.*}}
   return _mm256_maskz_min_epu8(__M,__A,__B); 
 }
 __m256i test_mm256_mask_min_epu8(__m256i __W, __mmask32 __M, __m256i __A, __m256i __B) {
   // CHECK-LABEL: @test_mm256_mask_min_epu8
   // CHECK:       [[CMP:%.*]] = icmp ult <32 x i8> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <32 x i1> [[CMP]], <32 x i8> [[X]], <32 x i8> [[Y]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast [[SRCTY:<.*>]] [[RES]] to [[DSTTY:<.*>]]
+  // CHECK-NEWPM: [[RES:%.*]] = bitcast [[DSTTY]] [[TMP]] to [[SRCTY]]
   // CHECK:       select <32 x i1> {{.*}}, <32 x i8> [[RES]], <32 x i8> {{.*}}
   return _mm256_mask_min_epu8(__W,__M,__A,__B); 
 }
 __m128i test_mm_maskz_min_epu16(__mmask8 __M, __m128i __A, __m128i __B) {
   // CHECK-LABEL: @test_mm_maskz_min_epu16
   // CHECK:       [[CMP:%.*]] = icmp ult <8 x i16> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <8 x i1> [[CMP]], <8 x i16> [[X]], <8 x i16> [[Y]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast [[SRCTY:<.*>]] [[RES]] to [[DSTTY:<.*>]]
+  // CHECK-NEWPM: [[RES:%.*]] = bitcast [[DSTTY]] [[TMP]] to [[SRCTY]]
   // CHECK:       select <8 x i1> {{.*}}, <8 x i16> [[RES]], <8 x i16> {{.*}}
   return _mm_maskz_min_epu16(__M,__A,__B); 
 }
 __m128i test_mm_mask_min_epu16(__m128i __W, __mmask8 __M, __m128i __A, __m128i __B) {
   // CHECK-LABEL: @test_mm_mask_min_epu16
   // CHECK:       [[CMP:%.*]] = icmp ult <8 x i16> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <8 x i1> [[CMP]], <8 x i16> [[X]], <8 x i16> [[Y]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast [[SRCTY:<.*>]] [[RES]] to [[DSTTY:<.*>]]
+  // CHECK-NEWPM: [[RES:%.*]] = bitcast [[DSTTY]] [[TMP]] to [[SRCTY]]
   // CHECK:       select <8 x i1> {{.*}}, <8 x i16> [[RES]], <8 x i16> {{.*}}
   return _mm_mask_min_epu16(__W,__M,__A,__B); 
 }
 __m256i test_mm256_maskz_min_epu16(__mmask16 __M, __m256i __A, __m256i __B) {
   // CHECK-LABEL: @test_mm256_maskz_min_epu16
   // CHECK:       [[CMP:%.*]] = icmp ult <16 x i16> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <16 x i1> [[CMP]], <16 x i16> [[X]], <16 x i16> [[Y]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast [[SRCTY:<.*>]] [[RES]] to [[DSTTY:<.*>]]
+  // CHECK-NEWPM: [[RES:%.*]] = bitcast [[DSTTY]] [[TMP]] to [[SRCTY]]
   // CHECK:       select <16 x i1> {{.*}}, <16 x i16> [[RES]], <16 x i16> {{.*}}
   return _mm256_maskz_min_epu16(__M,__A,__B); 
 }
 __m256i test_mm256_mask_min_epu16(__m256i __W, __mmask16 __M, __m256i __A, __m256i __B) {
   // CHECK-LABEL: @test_mm256_mask_min_epu16
   // CHECK:       [[CMP:%.*]] = icmp ult <16 x i16> [[X:%.*]], [[Y:%.*]]
   // CHECK-NEXT:  [[RES:%.*]] = select <16 x i1> [[CMP]], <16 x i16> [[X]], <16 x i16> [[Y]]
+  // CHECK-NEWPM: [[TMP:%.*]] = bitcast [[SRCTY:<.*>]] [[RES]] to [[DSTTY:<.*>]]
+  // CHECK-NEWPM: [[RES:%.*]] = bitcast [[DSTTY]] [[TMP]] to [[SRCTY]]
   // CHECK:       select <16 x i1> {{.*}}, <16 x i16> [[RES]], <16 x i16> {{.*}}
   return _mm256_mask_min_epu16(__W,__M,__A,__B); 
 }
 __m128i test_mm_mask_shuffle_epi8(__m128i __W, __mmask16 __U, __m128i __A, __m128i __B) {
   // CHECK-LABEL: @test_mm_mask_shuffle_epi8
   // CHECK: @llvm.x86.ssse3.pshuf.b
   // CHECK: select <16 x i1> %{{.*}}, <16 x i8> %{{.*}}, <16 x i8> %{{.*}}
   return _mm_mask_shuffle_epi8(__W,__U,__A,__B); 

clang/test/CodeGenOpenCL/convergent.cl

-// RUN: %clang_cc1 -triple spir-unknown-unknown -emit-llvm %s -o - | opt -instnamer -S | FileCheck -enable-var-scope %s
+// RUN: %clang_cc1 -triple spir-unknown-unknown -emit-llvm %s -o - -fno-experimental-new-pass-manager | opt -instnamer -S | FileCheck -enable-var-scope %s --check-prefixes=CHECK,CHECK-LEGACY
+// RUN: %clang_cc1 -triple spir-unknown-unknown -emit-llvm %s -o - -fexperimental-new-pass-manager | opt -instnamer -S | FileCheck -enable-var-scope %s --check-prefixes=CHECK,CHECK-NEWPM
 
 // This is initially assumed convergent, but can be deduced to not require it.
 
 // CHECK-LABEL: define spir_func void @non_convfun() local_unnamed_addr #0
 // CHECK: ret void
 __attribute__((noinline))
 void non_convfun(void) {
   volatile int* p;

 // Test loop is not unrolled for noduplicate function.
 // CHECK-LABEL: define spir_func void @test_not_unroll()
 // CHECK:  br label %[[for_body:.+]]
 // CHECK: [[for_cond_cleanup:.+]]:
 // CHECK:  ret void
 // CHECK: [[for_body]]:
 // CHECK:  tail call spir_func void @nodupfun() #[[attr5:[0-9]+]]
 // CHECK-NOT: call spir_func void @nodupfun()
-// CHECK:  br i1 %{{.+}}, label %[[for_body]], label %[[for_cond_cleanup]]
+
+// The new PM produces a slightly different IR for the loop from the legacy PM,
+// but the test still checks that the loop is not unrolled.
+// CHECK-LEGACY:  br i1 %{{.+}}, label %[[for_body]], label %[[for_cond_cleanup]]
+// CHECK-NEW:     br i1 %{{.+}}, label %[[for_body_crit_edge:.+]], label %[[for_cond_cleanup]]
+// CHECK-NEW:     [[for_body_crit_edge]]:
 
 void test_not_unroll() {
   for (int i = 0; i < 10; i++)
     nodupfun();
 }
 
 // CHECK: declare spir_func void @nodupfun(){{[^#]*}} #[[attr3:[0-9]+]]