[RISCV] Add VP patterns for vwsll.[vv,vx,vi]

[luke]

Summary

This patch adds patterns for the existing riscv_shl_vl VL node.

[craig.topper]

LGTM

llvm/lib/Target/RISCV/RISCVInstrInfoZvk.td

 }
 
 defm : VPatUnaryVL_V<riscv_bitreverse_vl, "PseudoVBREV">;
 defm : VPatUnaryVL_V<riscv_bswap_vl, "PseudoVREV8">;
 defm : VPatUnaryVL_V<riscv_ctlz_vl, "PseudoVCLZ">;
 defm : VPatUnaryVL_V<riscv_cttz_vl, "PseudoVCTZ">;
 defm : VPatUnaryVL_V<riscv_ctpop_vl, "PseudoVCPOP">;
 
+foreach vtiToWti = AllWidenableIntVectors in {
+  defvar vti = vtiToWti.Vti;
+  defvar wti = vtiToWti.Wti;
+  let Predicates = !listconcat([HasStdExtZvbb],
+                               GetVTypePredicates<vti>.Predicates,
+                               GetVTypePredicates<wti>.Predicates) in {
+    def : Pat<(riscv_shl_vl
+                 (wti.Vector (zext_oneuse (vti.Vector vti.RegClass:$rs2))),
+                 (wti.Vector (ext_oneuse (vti.Vector vti.RegClass:$rs1))),
+                 (wti.Vector wti.RegClass:$merge),
+                 (vti.Mask V0), VLOpFrag),
+              (!cast<Instruction>("PseudoVWSLL_VV_"#vti.LMul.MX#"_MASK")
+                 wti.RegClass:$merge, vti.RegClass:$rs2, vti.RegClass:$rs1,
+                 (vti.Mask V0), GPR:$vl, vti.Log2SEW, TAIL_AGNOSTIC)>;
+
+    def : Pat<(riscv_shl_vl
+                 (wti.Vector (zext_oneuse (vti.Vector vti.RegClass:$rs2))),
+                 (wti.Vector (Low8BitsSplatPat (XLenVT GPR:$rs1))),
+                 (wti.Vector wti.RegClass:$merge),
+                 (vti.Mask V0), VLOpFrag),
+              (!cast<Instruction>("PseudoVWSLL_VX_"#vti.LMul.MX#"_MASK")
+                 wti.RegClass:$merge, vti.RegClass:$rs2, GPR:$rs1,
+                 (vti.Mask V0), GPR:$vl, vti.Log2SEW, TAIL_AGNOSTIC)>;
+
+    def : Pat<(riscv_shl_vl
+                 (wti.Vector (zext_oneuse (vti.Vector vti.RegClass:$rs2))),
+                 (wti.Vector (SplatPat_uimm5 uimm5:$rs1)),
+                 (wti.Vector wti.RegClass:$merge),
+                 (vti.Mask V0), VLOpFrag),
+              (!cast<Instruction>("PseudoVWSLL_VI_"#vti.LMul.MX#"_MASK")
+                 wti.RegClass:$merge, vti.RegClass:$rs2, uimm5:$rs1,
+                 (vti.Mask V0), GPR:$vl, vti.Log2SEW, TAIL_AGNOSTIC)>;
+  }
+}
+
 //===----------------------------------------------------------------------===//
 // Codegen patterns
 //===----------------------------------------------------------------------===//
 
 class VPatUnaryNoMask_Zvk<string intrinsic_name,
                       string inst,
                       string kind,
                       ValueType result_type,

llvm/test/CodeGen/RISCV/rvv/vwsll-vp.ll

+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 2
+; RUN: llc -mtriple=riscv32 -mattr=+v -verify-machineinstrs < %s | FileCheck %s --check-prefixes=CHECK,CHECK-RV32
+; RUN: llc -mtriple=riscv64 -mattr=+v -verify-machineinstrs < %s | FileCheck %s --check-prefixes=CHECK,CHECK-RV64
+; RUN: llc -mtriple=riscv32 -mattr=+v,+experimental-zvbb -verify-machineinstrs < %s | FileCheck %s --check-prefixes=CHECK-ZVBB,CHECK-ZVBB32
+; RUN: llc -mtriple=riscv64 -mattr=+v,+experimental-zvbb -verify-machineinstrs < %s | FileCheck %s --check-prefixes=CHECK-ZVBB,CHECK-ZVBB64
+
+; ==============================================================================
+; i32 -> i64
+; ==============================================================================
+
+declare <vscale x 2 x i64> @llvm.vp.shl.nxv2i64(<vscale x 2 x i64>, <vscale x 2 x i64>, <vscale x 2 x i1>, i32)
+
+define <vscale x 2 x i64> @vwsll_vv_nxv2i64_sext(<vscale x 2 x i32> %a, <vscale x 2 x i32> %b, <vscale x 2 x i1> %m, i32 zeroext %vl) {
+; CHECK-LABEL: vwsll_vv_nxv2i64_sext:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a1, zero, e64, m2, ta, ma
+; CHECK-NEXT:    vzext.vf2 v10, v8
+; CHECK-NEXT:    vsext.vf2 v12, v9
+; CHECK-NEXT:    vsetvli zero, a0, e64, m2, ta, ma
+; CHECK-NEXT:    vsll.vv v8, v10, v12, v0.t
+; CHECK-NEXT:    ret
+;
+; CHECK-ZVBB-LABEL: vwsll_vv_nxv2i64_sext:
+; CHECK-ZVBB:       # %bb.0:
+; CHECK-ZVBB-NEXT:    vsetvli zero, a0, e32, m1, ta, ma
+; CHECK-ZVBB-NEXT:    vwsll.vv v10, v8, v9, v0.t
+; CHECK-ZVBB-NEXT:    vmv2r.v v8, v10
+; CHECK-ZVBB-NEXT:    ret
+  %x = zext <vscale x 2 x i32> %a to <vscale x 2 x i64>
+  %y = sext <vscale x 2 x i32> %b to <vscale x 2 x i64>
+  %z = call <vscale x 2 x i64> @llvm.vp.shl.nxv2i64(<vscale x 2 x i64> %x, <vscale x 2 x i64> %y, <vscale x 2 x i1> %m, i32 %vl)
+  ret <vscale x 2 x i64> %z
+}
+
+define <vscale x 2 x i64> @vwsll_vv_nxv2i64_zext(<vscale x 2 x i32> %a, <vscale x 2 x i32> %b, <vscale x 2 x i1> %m, i32 zeroext %vl) {
+; CHECK-LABEL: vwsll_vv_nxv2i64_zext:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a1, zero, e64, m2, ta, ma
+; CHECK-NEXT:    vzext.vf2 v10, v8
+; CHECK-NEXT:    vzext.vf2 v12, v9
+; CHECK-NEXT:    vsetvli zero, a0, e64, m2, ta, ma
+; CHECK-NEXT:    vsll.vv v8, v10, v12, v0.t
+; CHECK-NEXT:    ret
+;
+; CHECK-ZVBB-LABEL: vwsll_vv_nxv2i64_zext:
+; CHECK-ZVBB:       # %bb.0:
+; CHECK-ZVBB-NEXT:    vsetvli zero, a0, e32, m1, ta, ma
+; CHECK-ZVBB-NEXT:    vwsll.vv v10, v8, v9, v0.t
+; CHECK-ZVBB-NEXT:    vmv2r.v v8, v10
+; CHECK-ZVBB-NEXT:    ret
+  %x = zext <vscale x 2 x i32> %a to <vscale x 2 x i64>
+  %y = zext <vscale x 2 x i32> %b to <vscale x 2 x i64>
+  %z = call <vscale x 2 x i64> @llvm.vp.shl.nxv2i64(<vscale x 2 x i64> %x, <vscale x 2 x i64> %y, <vscale x 2 x i1> %m, i32 %vl)
+  ret <vscale x 2 x i64> %z
+}
+
+define <vscale x 2 x i64> @vwsll_vx_i64_nxv2i64(<vscale x 2 x i32> %a, i64 %b, <vscale x 2 x i1> %m, i32 zeroext %vl) {
+; CHECK-RV32-LABEL: vwsll_vx_i64_nxv2i64:
+; CHECK-RV32:       # %bb.0:
+; CHECK-RV32-NEXT:    vsetvli a1, zero, e64, m2, ta, ma
+; CHECK-RV32-NEXT:    vzext.vf2 v10, v8
+; CHECK-RV32-NEXT:    vsetvli zero, a2, e64, m2, ta, ma
+; CHECK-RV32-NEXT:    vsll.vx v8, v10, a0, v0.t
+; CHECK-RV32-NEXT:    ret
+;
+; CHECK-RV64-LABEL: vwsll_vx_i64_nxv2i64:
+; CHECK-RV64:       # %bb.0:
+; CHECK-RV64-NEXT:    vsetvli a2, zero, e64, m2, ta, ma
+; CHECK-RV64-NEXT:    vzext.vf2 v10, v8
+; CHECK-RV64-NEXT:    vsetvli zero, a1, e64, m2, ta, ma
+; CHECK-RV64-NEXT:    vsll.vx v8, v10, a0, v0.t
+; CHECK-RV64-NEXT:    ret
+;
+; CHECK-ZVBB32-LABEL: vwsll_vx_i64_nxv2i64:
+; CHECK-ZVBB32:       # %bb.0:
+; CHECK-ZVBB32-NEXT:    vsetvli zero, a2, e32, m1, ta, ma
+; CHECK-ZVBB32-NEXT:    vwsll.vx v10, v8, a0, v0.t
+; CHECK-ZVBB32-NEXT:    vmv2r.v v8, v10
+; CHECK-ZVBB32-NEXT:    ret
+;
+; CHECK-ZVBB64-LABEL: vwsll_vx_i64_nxv2i64:
+; CHECK-ZVBB64:       # %bb.0:
+; CHECK-ZVBB64-NEXT:    vsetvli zero, a1, e32, m1, ta, ma
+; CHECK-ZVBB64-NEXT:    vwsll.vx v10, v8, a0, v0.t
+; CHECK-ZVBB64-NEXT:    vmv2r.v v8, v10
+; CHECK-ZVBB64-NEXT:    ret
+  %head = insertelement <vscale x 2 x i64> poison, i64 %b, i32 0
+  %splat = shufflevector <vscale x 2 x i64> %head, <vscale x 2 x i64> poison, <vscale x 2 x i32> zeroinitializer
+  %x = zext <vscale x 2 x i32> %a to <vscale x 2 x i64>
+  %z = call <vscale x 2 x i64> @llvm.vp.shl.nxv2i64(<vscale x 2 x i64> %x, <vscale x 2 x i64> %splat, <vscale x 2 x i1> %m, i32 %vl)
+  ret <vscale x 2 x i64> %z
+}
+
+define <vscale x 2 x i64> @vwsll_vx_i32_nxv2i64_sext(<vscale x 2 x i32> %a, i32 %b, <vscale x 2 x i1> %m, i32 zeroext %vl) {
+; CHECK-LABEL: vwsll_vx_i32_nxv2i64_sext:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a2, zero, e32, m1, ta, ma
+; CHECK-NEXT:    vmv.v.x v9, a0
+; CHECK-NEXT:    vsetvli zero, zero, e64, m2, ta, ma
+; CHECK-NEXT:    vzext.vf2 v10, v8
+; CHECK-NEXT:    vsext.vf2 v12, v9
+; CHECK-NEXT:    vsetvli zero, a1, e64, m2, ta, ma
+; CHECK-NEXT:    vsll.vv v8, v10, v12, v0.t
+; CHECK-NEXT:    ret
+;
+; CHECK-ZVBB-LABEL: vwsll_vx_i32_nxv2i64_sext:
+; CHECK-ZVBB:       # %bb.0:
+; CHECK-ZVBB-NEXT:    vsetvli zero, a1, e32, m1, ta, ma
+; CHECK-ZVBB-NEXT:    vwsll.vx v10, v8, a0, v0.t
+; CHECK-ZVBB-NEXT:    vmv2r.v v8, v10
+; CHECK-ZVBB-NEXT:    ret
+  %head = insertelement <vscale x 2 x i32> poison, i32 %b, i32 0
+  %splat = shufflevector <vscale x 2 x i32> %head, <vscale x 2 x i32> poison, <vscale x 2 x i32> zeroinitializer
+  %x = zext <vscale x 2 x i32> %a to <vscale x 2 x i64>
+  %y = sext <vscale x 2 x i32> %splat to <vscale x 2 x i64>
+  %z = call <vscale x 2 x i64> @llvm.vp.shl.nxv2i64(<vscale x 2 x i64> %x, <vscale x 2 x i64> %y, <vscale x 2 x i1> %m, i32 %vl)
+  ret <vscale x 2 x i64> %z
+}
+
+define <vscale x 2 x i64> @vwsll_vx_i32_nxv2i64_zext(<vscale x 2 x i32> %a, i32 %b, <vscale x 2 x i1> %m, i32 zeroext %vl) {
+; CHECK-LABEL: vwsll_vx_i32_nxv2i64_zext:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a2, zero, e32, m1, ta, ma
+; CHECK-NEXT:    vmv.v.x v9, a0
+; CHECK-NEXT:    vsetvli zero, zero, e64, m2, ta, ma
+; CHECK-NEXT:    vzext.vf2 v10, v8
+; CHECK-NEXT:    vzext.vf2 v12, v9
+; CHECK-NEXT:    vsetvli zero, a1, e64, m2, ta, ma
+; CHECK-NEXT:    vsll.vv v8, v10, v12, v0.t
+; CHECK-NEXT:    ret
+;
+; CHECK-ZVBB-LABEL: vwsll_vx_i32_nxv2i64_zext:
+; CHECK-ZVBB:       # %bb.0:
+; CHECK-ZVBB-NEXT:    vsetvli zero, a1, e32, m1, ta, ma
+; CHECK-ZVBB-NEXT:    vwsll.vx v10, v8, a0, v0.t
+; CHECK-ZVBB-NEXT:    vmv2r.v v8, v10
+; CHECK-ZVBB-NEXT:    ret
+  %head = insertelement <vscale x 2 x i32> poison, i32 %b, i32 0
+  %splat = shufflevector <vscale x 2 x i32> %head, <vscale x 2 x i32> poison, <vscale x 2 x i32> zeroinitializer
+  %x = zext <vscale x 2 x i32> %a to <vscale x 2 x i64>
+  %y = zext <vscale x 2 x i32> %splat to <vscale x 2 x i64>
+  %z = call <vscale x 2 x i64> @llvm.vp.shl.nxv2i64(<vscale x 2 x i64> %x, <vscale x 2 x i64> %y, <vscale x 2 x i1> %m, i32 %vl)
+  ret <vscale x 2 x i64> %z
+}
+
+define <vscale x 2 x i64> @vwsll_vx_i16_nxv2i64_sext(<vscale x 2 x i32> %a, i16 %b, <vscale x 2 x i1> %m, i32 zeroext %vl) {
+; CHECK-LABEL: vwsll_vx_i16_nxv2i64_sext:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a2, zero, e16, mf2, ta, ma
+; CHECK-NEXT:    vmv.v.x v9, a0
+; CHECK-NEXT:    vsetvli zero, zero, e64, m2, ta, ma
+; CHECK-NEXT:    vzext.vf2 v10, v8
+; CHECK-NEXT:    vsext.vf4 v12, v9
+; CHECK-NEXT:    vsetvli zero, a1, e64, m2, ta, ma
+; CHECK-NEXT:    vsll.vv v8, v10, v12, v0.t
+; CHECK-NEXT:    ret
+;
+; CHECK-ZVBB-LABEL: vwsll_vx_i16_nxv2i64_sext:
+; CHECK-ZVBB:       # %bb.0:
+; CHECK-ZVBB-NEXT:    vsetvli zero, a1, e32, m1, ta, ma
+; CHECK-ZVBB-NEXT:    vwsll.vx v10, v8, a0, v0.t
+; CHECK-ZVBB-NEXT:    vmv2r.v v8, v10
+; CHECK-ZVBB-NEXT:    ret
+  %head = insertelement <vscale x 2 x i16> poison, i16 %b, i32 0
+  %splat = shufflevector <vscale x 2 x i16> %head, <vscale x 2 x i16> poison, <vscale x 2 x i32> zeroinitializer
+  %x = zext <vscale x 2 x i32> %a to <vscale x 2 x i64>
+  %y = sext <vscale x 2 x i16> %splat to <vscale x 2 x i64>
+  %z = call <vscale x 2 x i64> @llvm.vp.shl.nxv2i64(<vscale x 2 x i64> %x, <vscale x 2 x i64> %y, <vscale x 2 x i1> %m, i32 %vl)
+  ret <vscale x 2 x i64> %z
+}
+
+define <vscale x 2 x i64> @vwsll_vx_i16_nxv2i64_zext(<vscale x 2 x i32> %a, i16 %b, <vscale x 2 x i1> %m, i32 zeroext %vl) {
+; CHECK-LABEL: vwsll_vx_i16_nxv2i64_zext:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a2, zero, e16, mf2, ta, ma
+; CHECK-NEXT:    vmv.v.x v9, a0
+; CHECK-NEXT:    vsetvli zero, zero, e64, m2, ta, ma
+; CHECK-NEXT:    vzext.vf2 v10, v8
+; CHECK-NEXT:    vzext.vf4 v12, v9
+; CHECK-NEXT:    vsetvli zero, a1, e64, m2, ta, ma
+; CHECK-NEXT:    vsll.vv v8, v10, v12, v0.t
+; CHECK-NEXT:    ret
+;
+; CHECK-ZVBB-LABEL: vwsll_vx_i16_nxv2i64_zext:
+; CHECK-ZVBB:       # %bb.0:
+; CHECK-ZVBB-NEXT:    vsetvli zero, a1, e32, m1, ta, ma
+; CHECK-ZVBB-NEXT:    vwsll.vx v10, v8, a0, v0.t
+; CHECK-ZVBB-NEXT:    vmv2r.v v8, v10
+; CHECK-ZVBB-NEXT:    ret
+  %head = insertelement <vscale x 2 x i16> poison, i16 %b, i32 0
+  %splat = shufflevector <vscale x 2 x i16> %head, <vscale x 2 x i16> poison, <vscale x 2 x i32> zeroinitializer
+  %x = zext <vscale x 2 x i32> %a to <vscale x 2 x i64>
+  %y = zext <vscale x 2 x i16> %splat to <vscale x 2 x i64>
+  %z = call <vscale x 2 x i64> @llvm.vp.shl.nxv2i64(<vscale x 2 x i64> %x, <vscale x 2 x i64> %y, <vscale x 2 x i1> %m, i32 %vl)
+  ret <vscale x 2 x i64> %z
+}
+
+define <vscale x 2 x i64> @vwsll_vx_i8_nxv2i64_sext(<vscale x 2 x i32> %a, i8 %b, <vscale x 2 x i1> %m, i32 zeroext %vl) {
+; CHECK-LABEL: vwsll_vx_i8_nxv2i64_sext:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a2, zero, e8, mf4, ta, ma
+; CHECK-NEXT:    vmv.v.x v9, a0
+; CHECK-NEXT:    vsetvli zero, zero, e64, m2, ta, ma
+; CHECK-NEXT:    vzext.vf2 v10, v8
+; CHECK-NEXT:    vsext.vf8 v12, v9
+; CHECK-NEXT:    vsetvli zero, a1, e64, m2, ta, ma
+; CHECK-NEXT:    vsll.vv v8, v10, v12, v0.t
+; CHECK-NEXT:    ret
+;
+; CHECK-ZVBB-LABEL: vwsll_vx_i8_nxv2i64_sext:
+; CHECK-ZVBB:       # %bb.0:
+; CHECK-ZVBB-NEXT:    vsetvli zero, a1, e32, m1, ta, ma
+; CHECK-ZVBB-NEXT:    vwsll.vx v10, v8, a0, v0.t
+; CHECK-ZVBB-NEXT:    vmv2r.v v8, v10
+; CHECK-ZVBB-NEXT:    ret
+  %head = insertelement <vscale x 2 x i8> poison, i8 %b, i32 0
+  %splat = shufflevector <vscale x 2 x i8> %head, <vscale x 2 x i8> poison, <vscale x 2 x i32> zeroinitializer
+  %x = zext <vscale x 2 x i32> %a to <vscale x 2 x i64>
+  %y = sext <vscale x 2 x i8> %splat to <vscale x 2 x i64>
+  %z = call <vscale x 2 x i64> @llvm.vp.shl.nxv2i64(<vscale x 2 x i64> %x, <vscale x 2 x i64> %y, <vscale x 2 x i1> %m, i32 %vl)
+  ret <vscale x 2 x i64> %z
+}
+
+define <vscale x 2 x i64> @vwsll_vx_i8_nxv2i64_zext(<vscale x 2 x i32> %a, i8 %b, <vscale x 2 x i1> %m, i32 zeroext %vl) {
+; CHECK-LABEL: vwsll_vx_i8_nxv2i64_zext:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a2, zero, e8, mf4, ta, ma
+; CHECK-NEXT:    vmv.v.x v9, a0
+; CHECK-NEXT:    vsetvli zero, zero, e64, m2, ta, ma
+; CHECK-NEXT:    vzext.vf2 v10, v8
+; CHECK-NEXT:    vzext.vf8 v12, v9
+; CHECK-NEXT:    vsetvli zero, a1, e64, m2, ta, ma
+; CHECK-NEXT:    vsll.vv v8, v10, v12, v0.t
+; CHECK-NEXT:    ret
+;
+; CHECK-ZVBB-LABEL: vwsll_vx_i8_nxv2i64_zext:
+; CHECK-ZVBB:       # %bb.0:
+; CHECK-ZVBB-NEXT:    vsetvli zero, a1, e32, m1, ta, ma
+; CHECK-ZVBB-NEXT:    vwsll.vx v10, v8, a0, v0.t
+; CHECK-ZVBB-NEXT:    vmv2r.v v8, v10
+; CHECK-ZVBB-NEXT:    ret
+  %head = insertelement <vscale x 2 x i8> poison, i8 %b, i32 0
+  %splat = shufflevector <vscale x 2 x i8> %head, <vscale x 2 x i8> poison, <vscale x 2 x i32> zeroinitializer
+  %x = zext <vscale x 2 x i32> %a to <vscale x 2 x i64>
+  %y = zext <vscale x 2 x i8> %splat to <vscale x 2 x i64>
+  %z = call <vscale x 2 x i64> @llvm.vp.shl.nxv2i64(<vscale x 2 x i64> %x, <vscale x 2 x i64> %y, <vscale x 2 x i1> %m, i32 %vl)
+  ret <vscale x 2 x i64> %z
+}
+
+define <vscale x 2 x i64> @vwsll_vi_nxv2i64(<vscale x 2 x i32> %a, <vscale x 2 x i1> %m, i32 zeroext %vl) {
+; CHECK-LABEL: vwsll_vi_nxv2i64:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a1, zero, e64, m2, ta, ma
+; CHECK-NEXT:    vzext.vf2 v10, v8
+; CHECK-NEXT:    vsetvli zero, a0, e64, m2, ta, ma
+; CHECK-NEXT:    vsll.vi v8, v10, 2, v0.t
+; CHECK-NEXT:    ret
+;
+; CHECK-ZVBB-LABEL: vwsll_vi_nxv2i64:
+; CHECK-ZVBB:       # %bb.0:
+; CHECK-ZVBB-NEXT:    vsetvli zero, a0, e32, m1, ta, ma
+; CHECK-ZVBB-NEXT:    vwsll.vi v10, v8, 2, v0.t
+; CHECK-ZVBB-NEXT:    vmv2r.v v8, v10
+; CHECK-ZVBB-NEXT:    ret
+  %x = zext <vscale x 2 x i32> %a to <vscale x 2 x i64>
+  %z = call <vscale x 2 x i64> @llvm.vp.shl.nxv2i64(<vscale x 2 x i64> %x, <vscale x 2 x i64> shufflevector(<vscale x 2 x i64> insertelement(<vscale x 2 x i64> poison, i64 2, i32 0), <vscale x 2 x i64> poison, <vscale x 2 x i32> zeroinitializer), <vscale x 2 x i1> %m, i32 %vl)
+  ret <vscale x 2 x i64> %z
+}
+
+; ==============================================================================
+; i16 -> i32
+; ==============================================================================
+
+declare <vscale x 4 x i32> @llvm.vp.shl.nxv4i32(<vscale x 4 x i32>, <vscale x 4 x i32>, <vscale x 4 x i1>, i32)
+
+define <vscale x 4 x i32> @vwsll_vv_nxv4i32_sext(<vscale x 4 x i16> %a, <vscale x 4 x i16> %b, <vscale x 4 x i1> %m, i32 zeroext %vl) {
+; CHECK-LABEL: vwsll_vv_nxv4i32_sext:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a1, zero, e32, m2, ta, ma
+; CHECK-NEXT:    vzext.vf2 v10, v8
+; CHECK-NEXT:    vsext.vf2 v12, v9
+; CHECK-NEXT:    vsetvli zero, a0, e32, m2, ta, ma
+; CHECK-NEXT:    vsll.vv v8, v10, v12, v0.t
+; CHECK-NEXT:    ret
+;
+; CHECK-ZVBB-LABEL: vwsll_vv_nxv4i32_sext:
+; CHECK-ZVBB:       # %bb.0:
+; CHECK-ZVBB-NEXT:    vsetvli zero, a0, e16, m1, ta, ma
+; CHECK-ZVBB-NEXT:    vwsll.vv v10, v8, v9, v0.t
+; CHECK-ZVBB-NEXT:    vmv2r.v v8, v10
+; CHECK-ZVBB-NEXT:    ret
+  %x = zext <vscale x 4 x i16> %a to <vscale x 4 x i32>
+  %y = sext <vscale x 4 x i16> %b to <vscale x 4 x i32>
+  %z = call <vscale x 4 x i32> @llvm.vp.shl.nxv4i32(<vscale x 4 x i32> %x, <vscale x 4 x i32> %y, <vscale x 4 x i1> %m, i32 %vl)
+  ret <vscale x 4 x i32> %z
+}
+
+define <vscale x 4 x i32> @vwsll_vv_nxv4i32_zext(<vscale x 4 x i16> %a, <vscale x 4 x i16> %b, <vscale x 4 x i1> %m, i32 zeroext %vl) {
+; CHECK-LABEL: vwsll_vv_nxv4i32_zext:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a1, zero, e32, m2, ta, ma
+; CHECK-NEXT:    vzext.vf2 v10, v8
+; CHECK-NEXT:    vzext.vf2 v12, v9
+; CHECK-NEXT:    vsetvli zero, a0, e32, m2, ta, ma
+; CHECK-NEXT:    vsll.vv v8, v10, v12, v0.t
+; CHECK-NEXT:    ret
+;
+; CHECK-ZVBB-LABEL: vwsll_vv_nxv4i32_zext:
+; CHECK-ZVBB:       # %bb.0:
+; CHECK-ZVBB-NEXT:    vsetvli zero, a0, e16, m1, ta, ma
+; CHECK-ZVBB-NEXT:    vwsll.vv v10, v8, v9, v0.t
+; CHECK-ZVBB-NEXT:    vmv2r.v v8, v10
+; CHECK-ZVBB-NEXT:    ret
+  %x = zext <vscale x 4 x i16> %a to <vscale x 4 x i32>
+  %y = zext <vscale x 4 x i16> %b to <vscale x 4 x i32>
+  %z = call <vscale x 4 x i32> @llvm.vp.shl.nxv4i32(<vscale x 4 x i32> %x, <vscale x 4 x i32> %y, <vscale x 4 x i1> %m, i32 %vl)
+  ret <vscale x 4 x i32> %z
+}
+
+define <vscale x 4 x i32> @vwsll_vx_i64_nxv4i32(<vscale x 4 x i16> %a, i64 %b, <vscale x 4 x i1> %m, i32 zeroext %vl) {
+; CHECK-RV32-LABEL: vwsll_vx_i64_nxv4i32:
+; CHECK-RV32:       # %bb.0:
+; CHECK-RV32-NEXT:    addi sp, sp, -16
+; CHECK-RV32-NEXT:    .cfi_def_cfa_offset 16
+; CHECK-RV32-NEXT:    sw a1, 12(sp)
+; CHECK-RV32-NEXT:    sw a0, 8(sp)
+; CHECK-RV32-NEXT:    addi a0, sp, 8
+; CHECK-RV32-NEXT:    vsetvli a1, zero, e32, m2, ta, ma
+; CHECK-RV32-NEXT:    vlse64.v v12, (a0), zero
+; CHECK-RV32-NEXT:    vzext.vf2 v10, v8
+; CHECK-RV32-NEXT:    vnsrl.wi v8, v12, 0
+; CHECK-RV32-NEXT:    vsetvli zero, a2, e32, m2, ta, ma
+; CHECK-RV32-NEXT:    vsll.vv v8, v10, v8, v0.t
+; CHECK-RV32-NEXT:    addi sp, sp, 16
+; CHECK-RV32-NEXT:    ret
+;
+; CHECK-RV64-LABEL: vwsll_vx_i64_nxv4i32:
+; CHECK-RV64:       # %bb.0:
+; CHECK-RV64-NEXT:    vsetvli a2, zero, e64, m4, ta, ma
+; CHECK-RV64-NEXT:    vmv.v.x v12, a0
+; CHECK-RV64-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
+; CHECK-RV64-NEXT:    vzext.vf2 v10, v8
+; CHECK-RV64-NEXT:    vnsrl.wi v8, v12, 0
+; CHECK-RV64-NEXT:    vsetvli zero, a1, e32, m2, ta, ma
+; CHECK-RV64-NEXT:    vsll.vv v8, v10, v8, v0.t
+; CHECK-RV64-NEXT:    ret
+;
+; CHECK-ZVBB32-LABEL: vwsll_vx_i64_nxv4i32:
+; CHECK-ZVBB32:       # %bb.0:
+; CHECK-ZVBB32-NEXT:    addi sp, sp, -16
+; CHECK-ZVBB32-NEXT:    .cfi_def_cfa_offset 16
+; CHECK-ZVBB32-NEXT:    sw a1, 12(sp)
+; CHECK-ZVBB32-NEXT:    sw a0, 8(sp)
+; CHECK-ZVBB32-NEXT:    addi a0, sp, 8
+; CHECK-ZVBB32-NEXT:    vsetvli a1, zero, e32, m2, ta, ma
+; CHECK-ZVBB32-NEXT:    vlse64.v v12, (a0), zero
+; CHECK-ZVBB32-NEXT:    vzext.vf2 v10, v8
+; CHECK-ZVBB32-NEXT:    vnsrl.wi v8, v12, 0
+; CHECK-ZVBB32-NEXT:    vsetvli zero, a2, e32, m2, ta, ma
+; CHECK-ZVBB32-NEXT:    vsll.vv v8, v10, v8, v0.t
+; CHECK-ZVBB32-NEXT:    addi sp, sp, 16
+; CHECK-ZVBB32-NEXT:    ret
+;
+; CHECK-ZVBB64-LABEL: vwsll_vx_i64_nxv4i32:
+; CHECK-ZVBB64:       # %bb.0:
+; CHECK-ZVBB64-NEXT:    vsetvli zero, a1, e16, m1, ta, ma
+; CHECK-ZVBB64-NEXT:    vwsll.vx v10, v8, a0, v0.t
+; CHECK-ZVBB64-NEXT:    vmv2r.v v8, v10
+; CHECK-ZVBB64-NEXT:    ret
+  %head = insertelement <vscale x 4 x i64> poison, i64 %b, i32 0
+  %splat = shufflevector <vscale x 4 x i64> %head, <vscale x 4 x i64> poison, <vscale x 4 x i32> zeroinitializer
+  %x = zext <vscale x 4 x i16> %a to <vscale x 4 x i32>
+  %y = trunc <vscale x 4 x i64> %splat to <vscale x 4 x i32>
+  %z = call <vscale x 4 x i32> @llvm.vp.shl.nxv4i32(<vscale x 4 x i32> %x, <vscale x 4 x i32> %y, <vscale x 4 x i1> %m, i32 %vl)
+  ret <vscale x 4 x i32> %z
+}
+
+define <vscale x 4 x i32> @vwsll_vx_i32_nxv4i32(<vscale x 4 x i16> %a, i32 %b, <vscale x 4 x i1> %m, i32 zeroext %vl) {
+; CHECK-LABEL: vwsll_vx_i32_nxv4i32:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a2, zero, e32, m2, ta, ma
+; CHECK-NEXT:    vzext.vf2 v10, v8
+; CHECK-NEXT:    vsetvli zero, a1, e32, m2, ta, ma
+; CHECK-NEXT:    vsll.vx v8, v10, a0, v0.t
+; CHECK-NEXT:    ret
+;
+; CHECK-ZVBB-LABEL: vwsll_vx_i32_nxv4i32:
+; CHECK-ZVBB:       # %bb.0:
+; CHECK-ZVBB-NEXT:    vsetvli zero, a1, e16, m1, ta, ma
+; CHECK-ZVBB-NEXT:    vwsll.vx v10, v8, a0, v0.t
+; CHECK-ZVBB-NEXT:    vmv2r.v v8, v10
+; CHECK-ZVBB-NEXT:    ret
+  %head = insertelement <vscale x 4 x i32> poison, i32 %b, i32 0
+  %splat = shufflevector <vscale x 4 x i32> %head, <vscale x 4 x i32> poison, <vscale x 4 x i32> zeroinitializer
+  %x = zext <vscale x 4 x i16> %a to <vscale x 4 x i32>
+  %z = call <vscale x 4 x i32> @llvm.vp.shl.nxv4i32(<vscale x 4 x i32> %x, <vscale x 4 x i32> %splat, <vscale x 4 x i1> %m, i32 %vl)
+  ret <vscale x 4 x i32> %z
+}
+
+define <vscale x 4 x i32> @vwsll_vx_i16_nxv4i32_sext(<vscale x 4 x i16> %a, i16 %b, <vscale x 4 x i1> %m, i32 zeroext %vl) {
+; CHECK-LABEL: vwsll_vx_i16_nxv4i32_sext:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a2, zero, e16, m1, ta, ma
+; CHECK-NEXT:    vmv.v.x v9, a0
+; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
+; CHECK-NEXT:    vzext.vf2 v10, v8
+; CHECK-NEXT:    vsext.vf2 v12, v9
+; CHECK-NEXT:    vsetvli zero, a1, e32, m2, ta, ma
+; CHECK-NEXT:    vsll.vv v8, v10, v12, v0.t
+; CHECK-NEXT:    ret
+;
+; CHECK-ZVBB-LABEL: vwsll_vx_i16_nxv4i32_sext:
+; CHECK-ZVBB:       # %bb.0:
+; CHECK-ZVBB-NEXT:    vsetvli zero, a1, e16, m1, ta, ma
+; CHECK-ZVBB-NEXT:    vwsll.vx v10, v8, a0, v0.t
+; CHECK-ZVBB-NEXT:    vmv2r.v v8, v10
+; CHECK-ZVBB-NEXT:    ret
+  %head = insertelement <vscale x 4 x i16> poison, i16 %b, i32 0
+  %splat = shufflevector <vscale x 4 x i16> %head, <vscale x 4 x i16> poison, <vscale x 4 x i32> zeroinitializer
+  %x = zext <vscale x 4 x i16> %a to <vscale x 4 x i32>
+  %y = sext <vscale x 4 x i16> %splat to <vscale x 4 x i32>
+  %z = call <vscale x 4 x i32> @llvm.vp.shl.nxv4i32(<vscale x 4 x i32> %x, <vscale x 4 x i32> %y, <vscale x 4 x i1> %m, i32 %vl)
+  ret <vscale x 4 x i32> %z
+}
+
+define <vscale x 4 x i32> @vwsll_vx_i16_nxv4i32_zext(<vscale x 4 x i16> %a, i16 %b, <vscale x 4 x i1> %m, i32 zeroext %vl) {
+; CHECK-LABEL: vwsll_vx_i16_nxv4i32_zext:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a2, zero, e16, m1, ta, ma
+; CHECK-NEXT:    vmv.v.x v9, a0
+; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
+; CHECK-NEXT:    vzext.vf2 v10, v8
+; CHECK-NEXT:    vzext.vf2 v12, v9
+; CHECK-NEXT:    vsetvli zero, a1, e32, m2, ta, ma
+; CHECK-NEXT:    vsll.vv v8, v10, v12, v0.t
+; CHECK-NEXT:    ret
+;
+; CHECK-ZVBB-LABEL: vwsll_vx_i16_nxv4i32_zext:
+; CHECK-ZVBB:       # %bb.0:
+; CHECK-ZVBB-NEXT:    vsetvli zero, a1, e16, m1, ta, ma
+; CHECK-ZVBB-NEXT:    vwsll.vx v10, v8, a0, v0.t
+; CHECK-ZVBB-NEXT:    vmv2r.v v8, v10
+; CHECK-ZVBB-NEXT:    ret
+  %head = insertelement <vscale x 4 x i16> poison, i16 %b, i32 0
+  %splat = shufflevector <vscale x 4 x i16> %head, <vscale x 4 x i16> poison, <vscale x 4 x i32> zeroinitializer
+  %x = zext <vscale x 4 x i16> %a to <vscale x 4 x i32>
+  %y = zext <vscale x 4 x i16> %splat to <vscale x 4 x i32>
+  %z = call <vscale x 4 x i32> @llvm.vp.shl.nxv4i32(<vscale x 4 x i32> %x, <vscale x 4 x i32> %y, <vscale x 4 x i1> %m, i32 %vl)
+  ret <vscale x 4 x i32> %z
+}
+
+define <vscale x 4 x i32> @vwsll_vx_i8_nxv4i32_sext(<vscale x 4 x i16> %a, i8 %b, <vscale x 4 x i1> %m, i32 zeroext %vl) {
+; CHECK-LABEL: vwsll_vx_i8_nxv4i32_sext:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a2, zero, e8, mf2, ta, ma
+; CHECK-NEXT:    vmv.v.x v9, a0
+; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
+; CHECK-NEXT:    vzext.vf2 v10, v8
+; CHECK-NEXT:    vsext.vf4 v12, v9
+; CHECK-NEXT:    vsetvli zero, a1, e32, m2, ta, ma
+; CHECK-NEXT:    vsll.vv v8, v10, v12, v0.t
+; CHECK-NEXT:    ret
+;
+; CHECK-ZVBB-LABEL: vwsll_vx_i8_nxv4i32_sext:
+; CHECK-ZVBB:       # %bb.0:
+; CHECK-ZVBB-NEXT:    vsetvli zero, a1, e16, m1, ta, ma
+; CHECK-ZVBB-NEXT:    vwsll.vx v10, v8, a0, v0.t
+; CHECK-ZVBB-NEXT:    vmv2r.v v8, v10
+; CHECK-ZVBB-NEXT:    ret
+  %head = insertelement <vscale x 4 x i8> poison, i8 %b, i32 0
+  %splat = shufflevector <vscale x 4 x i8> %head, <vscale x 4 x i8> poison, <vscale x 4 x i32> zeroinitializer
+  %x = zext <vscale x 4 x i16> %a to <vscale x 4 x i32>
+  %y = sext <vscale x 4 x i8> %splat to <vscale x 4 x i32>
+  %z = call <vscale x 4 x i32> @llvm.vp.shl.nxv4i32(<vscale x 4 x i32> %x, <vscale x 4 x i32> %y, <vscale x 4 x i1> %m, i32 %vl)
+  ret <vscale x 4 x i32> %z
+}
+
+define <vscale x 4 x i32> @vwsll_vx_i8_nxv4i32_zext(<vscale x 4 x i16> %a, i8 %b, <vscale x 4 x i1> %m, i32 zeroext %vl) {
+; CHECK-LABEL: vwsll_vx_i8_nxv4i32_zext:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a2, zero, e8, mf2, ta, ma
+; CHECK-NEXT:    vmv.v.x v9, a0
+; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
+; CHECK-NEXT:    vzext.vf2 v10, v8
+; CHECK-NEXT:    vzext.vf4 v12, v9
+; CHECK-NEXT:    vsetvli zero, a1, e32, m2, ta, ma
+; CHECK-NEXT:    vsll.vv v8, v10, v12, v0.t
+; CHECK-NEXT:    ret
+;
+; CHECK-ZVBB-LABEL: vwsll_vx_i8_nxv4i32_zext:
+; CHECK-ZVBB:       # %bb.0:
+; CHECK-ZVBB-NEXT:    vsetvli zero, a1, e16, m1, ta, ma
+; CHECK-ZVBB-NEXT:    vwsll.vx v10, v8, a0, v0.t
+; CHECK-ZVBB-NEXT:    vmv2r.v v8, v10
+; CHECK-ZVBB-NEXT:    ret
+  %head = insertelement <vscale x 4 x i8> poison, i8 %b, i32 0
+  %splat = shufflevector <vscale x 4 x i8> %head, <vscale x 4 x i8> poison, <vscale x 4 x i32> zeroinitializer
+  %x = zext <vscale x 4 x i16> %a to <vscale x 4 x i32>
+  %y = zext <vscale x 4 x i8> %splat to <vscale x 4 x i32>
+  %z = call <vscale x 4 x i32> @llvm.vp.shl.nxv4i32(<vscale x 4 x i32> %x, <vscale x 4 x i32> %y, <vscale x 4 x i1> %m, i32 %vl)
+  ret <vscale x 4 x i32> %z
+}
+
+define <vscale x 4 x i32> @vwsll_vi_nxv4i32(<vscale x 4 x i16> %a, <vscale x 4 x i1> %m, i32 zeroext %vl) {
+; CHECK-LABEL: vwsll_vi_nxv4i32:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a1, zero, e32, m2, ta, ma
+; CHECK-NEXT:    vzext.vf2 v10, v8
+; CHECK-NEXT:    vsetvli zero, a0, e32, m2, ta, ma
+; CHECK-NEXT:    vsll.vi v8, v10, 2, v0.t
+; CHECK-NEXT:    ret
+;
+; CHECK-ZVBB-LABEL: vwsll_vi_nxv4i32:
+; CHECK-ZVBB:       # %bb.0:
+; CHECK-ZVBB-NEXT:    vsetvli zero, a0, e16, m1, ta, ma
+; CHECK-ZVBB-NEXT:    vwsll.vi v10, v8, 2, v0.t
+; CHECK-ZVBB-NEXT:    vmv2r.v v8, v10
+; CHECK-ZVBB-NEXT:    ret
+  %x = zext <vscale x 4 x i16> %a to <vscale x 4 x i32>
+  %z = call <vscale x 4 x i32> @llvm.vp.shl.nxv4i32(<vscale x 4 x i32> %x, <vscale x 4 x i32> shufflevector(<vscale x 4 x i32> insertelement(<vscale x 4 x i32> poison, i32 2, i32 0), <vscale x 4 x i32> poison, <vscale x 4 x i32> zeroinitializer), <vscale x 4 x i1> %m, i32 %vl)
+  ret <vscale x 4 x i32> %z
+}
+
+
+; ==============================================================================
+; i8 -> i16
+; ==============================================================================
+
+declare <vscale x 8 x i16> @llvm.vp.shl.nxv8i16(<vscale x 8 x i16>, <vscale x 8 x i16>, <vscale x 8 x i1>, i32)
+
+define <vscale x 8 x i16> @vwsll_vv_nxv8i16_sext(<vscale x 8 x i8> %a, <vscale x 8 x i8> %b, <vscale x 8 x i1> %m, i32 zeroext %vl) {
+; CHECK-LABEL: vwsll_vv_nxv8i16_sext:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a1, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vzext.vf2 v10, v8
+; CHECK-NEXT:    vsext.vf2 v12, v9
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
+; CHECK-NEXT:    vsll.vv v8, v10, v12, v0.t
+; CHECK-NEXT:    ret
+;
+; CHECK-ZVBB-LABEL: vwsll_vv_nxv8i16_sext:
+; CHECK-ZVBB:       # %bb.0:
+; CHECK-ZVBB-NEXT:    vsetvli zero, a0, e8, m1, ta, ma
+; CHECK-ZVBB-NEXT:    vwsll.vv v10, v8, v9, v0.t
+; CHECK-ZVBB-NEXT:    vmv2r.v v8, v10
+; CHECK-ZVBB-NEXT:    ret
+  %x = zext <vscale x 8 x i8> %a to <vscale x 8 x i16>
+  %y = sext <vscale x 8 x i8> %b to <vscale x 8 x i16>
+  %z = call <vscale x 8 x i16> @llvm.vp.shl.nxv8i16(<vscale x 8 x i16> %x, <vscale x 8 x i16> %y, <vscale x 8 x i1> %m, i32 %vl)
+  ret <vscale x 8 x i16> %z
+}
+
+define <vscale x 8 x i16> @vwsll_vv_nxv8i16_zext(<vscale x 8 x i8> %a, <vscale x 8 x i8> %b, <vscale x 8 x i1> %m, i32 zeroext %vl) {
+; CHECK-LABEL: vwsll_vv_nxv8i16_zext:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a1, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vzext.vf2 v10, v8
+; CHECK-NEXT:    vzext.vf2 v12, v9
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
+; CHECK-NEXT:    vsll.vv v8, v10, v12, v0.t
+; CHECK-NEXT:    ret
+;
+; CHECK-ZVBB-LABEL: vwsll_vv_nxv8i16_zext:
+; CHECK-ZVBB:       # %bb.0:
+; CHECK-ZVBB-NEXT:    vsetvli zero, a0, e8, m1, ta, ma
+; CHECK-ZVBB-NEXT:    vwsll.vv v10, v8, v9, v0.t
+; CHECK-ZVBB-NEXT:    vmv2r.v v8, v10
+; CHECK-ZVBB-NEXT:    ret
+  %x = zext <vscale x 8 x i8> %a to <vscale x 8 x i16>
+  %y = zext <vscale x 8 x i8> %b to <vscale x 8 x i16>
+  %z = call <vscale x 8 x i16> @llvm.vp.shl.nxv8i16(<vscale x 8 x i16> %x, <vscale x 8 x i16> %y, <vscale x 8 x i1> %m, i32 %vl)
+  ret <vscale x 8 x i16> %z
+}
+
+define <vscale x 8 x i16> @vwsll_vx_i64_nxv8i16(<vscale x 8 x i8> %a, i64 %b, <vscale x 8 x i1> %m, i32 zeroext %vl) {
+; CHECK-RV32-LABEL: vwsll_vx_i64_nxv8i16:
+; CHECK-RV32:       # %bb.0:
+; CHECK-RV32-NEXT:    addi sp, sp, -16
+; CHECK-RV32-NEXT:    .cfi_def_cfa_offset 16
+; CHECK-RV32-NEXT:    sw a1, 12(sp)
+; CHECK-RV32-NEXT:    sw a0, 8(sp)
+; CHECK-RV32-NEXT:    addi a0, sp, 8
+; CHECK-RV32-NEXT:    vsetvli a1, zero, e16, m2, ta, ma
+; CHECK-RV32-NEXT:    vlse64.v v16, (a0), zero
+; CHECK-RV32-NEXT:    vzext.vf2 v10, v8
+; CHECK-RV32-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
+; CHECK-RV32-NEXT:    vnsrl.wi v12, v16, 0
+; CHECK-RV32-NEXT:    vsetvli zero, zero, e16, m2, ta, ma
+; CHECK-RV32-NEXT:    vnsrl.wi v8, v12, 0
+; CHECK-RV32-NEXT:    vsetvli zero, a2, e16, m2, ta, ma
+; CHECK-RV32-NEXT:    vsll.vv v8, v10, v8, v0.t
+; CHECK-RV32-NEXT:    addi sp, sp, 16
+; CHECK-RV32-NEXT:    ret
+;
+; CHECK-RV64-LABEL: vwsll_vx_i64_nxv8i16:
+; CHECK-RV64:       # %bb.0:
+; CHECK-RV64-NEXT:    vsetvli a2, zero, e64, m8, ta, ma
+; CHECK-RV64-NEXT:    vmv.v.x v16, a0
+; CHECK-RV64-NEXT:    vsetvli zero, zero, e16, m2, ta, ma
+; CHECK-RV64-NEXT:    vzext.vf2 v10, v8
+; CHECK-RV64-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
+; CHECK-RV64-NEXT:    vnsrl.wi v12, v16, 0
+; CHECK-RV64-NEXT:    vsetvli zero, zero, e16, m2, ta, ma
+; CHECK-RV64-NEXT:    vnsrl.wi v8, v12, 0
+; CHECK-RV64-NEXT:    vsetvli zero, a1, e16, m2, ta, ma
+; CHECK-RV64-NEXT:    vsll.vv v8, v10, v8, v0.t
+; CHECK-RV64-NEXT:    ret
+;
+; CHECK-ZVBB32-LABEL: vwsll_vx_i64_nxv8i16:
+; CHECK-ZVBB32:       # %bb.0:
+; CHECK-ZVBB32-NEXT:    addi sp, sp, -16
+; CHECK-ZVBB32-NEXT:    .cfi_def_cfa_offset 16
+; CHECK-ZVBB32-NEXT:    sw a1, 12(sp)
+; CHECK-ZVBB32-NEXT:    sw a0, 8(sp)
+; CHECK-ZVBB32-NEXT:    addi a0, sp, 8
+; CHECK-ZVBB32-NEXT:    vsetvli a1, zero, e16, m2, ta, ma
+; CHECK-ZVBB32-NEXT:    vlse64.v v16, (a0), zero
+; CHECK-ZVBB32-NEXT:    vzext.vf2 v10, v8
+; CHECK-ZVBB32-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
+; CHECK-ZVBB32-NEXT:    vnsrl.wi v12, v16, 0
+; CHECK-ZVBB32-NEXT:    vsetvli zero, zero, e16, m2, ta, ma
+; CHECK-ZVBB32-NEXT:    vnsrl.wi v8, v12, 0
+; CHECK-ZVBB32-NEXT:    vsetvli zero, a2, e16, m2, ta, ma
+; CHECK-ZVBB32-NEXT:    vsll.vv v8, v10, v8, v0.t
+; CHECK-ZVBB32-NEXT:    addi sp, sp, 16
+; CHECK-ZVBB32-NEXT:    ret
+;
+; CHECK-ZVBB64-LABEL: vwsll_vx_i64_nxv8i16:
+; CHECK-ZVBB64:       # %bb.0:
+; CHECK-ZVBB64-NEXT:    vsetvli zero, a1, e8, m1, ta, ma
+; CHECK-ZVBB64-NEXT:    vwsll.vx v10, v8, a0, v0.t
+; CHECK-ZVBB64-NEXT:    vmv2r.v v8, v10
+; CHECK-ZVBB64-NEXT:    ret
+  %head = insertelement <vscale x 4 x i64> poison, i64 %b, i32 0
+  %splat = shufflevector <vscale x 4 x i64> %head, <vscale x 4 x i64> poison, <vscale x 8 x i32> zeroinitializer
+  %x = zext <vscale x 8 x i8> %a to <vscale x 8 x i16>
+  %y = trunc <vscale x 8 x i64> %splat to <vscale x 8 x i16>
+  %z = call <vscale x 8 x i16> @llvm.vp.shl.nxv8i16(<vscale x 8 x i16> %x, <vscale x 8 x i16> %y, <vscale x 8 x i1> %m, i32 %vl)
+  ret <vscale x 8 x i16> %z
+}
+
+define <vscale x 8 x i16> @vwsll_vx_i32_nxv8i16(<vscale x 8 x i8> %a, i32 %b, <vscale x 8 x i1> %m, i32 zeroext %vl) {
+; CHECK-LABEL: vwsll_vx_i32_nxv8i16:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a2, zero, e32, m4, ta, ma
+; CHECK-NEXT:    vmv.v.x v12, a0
+; CHECK-NEXT:    vsetvli zero, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vzext.vf2 v10, v8
+; CHECK-NEXT:    vnsrl.wi v8, v12, 0
+; CHECK-NEXT:    vsetvli zero, a1, e16, m2, ta, ma
+; CHECK-NEXT:    vsll.vv v8, v10, v8, v0.t
+; CHECK-NEXT:    ret
+;
+; CHECK-ZVBB-LABEL: vwsll_vx_i32_nxv8i16:
+; CHECK-ZVBB:       # %bb.0:
+; CHECK-ZVBB-NEXT:    vsetvli zero, a1, e8, m1, ta, ma
+; CHECK-ZVBB-NEXT:    vwsll.vx v10, v8, a0, v0.t
+; CHECK-ZVBB-NEXT:    vmv2r.v v8, v10
+; CHECK-ZVBB-NEXT:    ret
+  %head = insertelement <vscale x 8 x i32> poison, i32 %b, i32 0
+  %splat = shufflevector <vscale x 8 x i32> %head, <vscale x 8 x i32> poison, <vscale x 8 x i32> zeroinitializer
+  %x = zext <vscale x 8 x i8> %a to <vscale x 8 x i16>
+  %y = trunc <vscale x 8 x i32> %splat to <vscale x 8 x i16>
+  %z = call <vscale x 8 x i16> @llvm.vp.shl.nxv8i16(<vscale x 8 x i16> %x, <vscale x 8 x i16> %y, <vscale x 8 x i1> %m, i32 %vl)
+  ret <vscale x 8 x i16> %z
+}
+
+define <vscale x 8 x i16> @vwsll_vx_i16_nxv8i16(<vscale x 8 x i8> %a, i16 %b, <vscale x 8 x i1> %m, i32 zeroext %vl) {
+; CHECK-LABEL: vwsll_vx_i16_nxv8i16:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a2, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vzext.vf2 v10, v8
+; CHECK-NEXT:    vsetvli zero, a1, e16, m2, ta, ma
+; CHECK-NEXT:    vsll.vx v8, v10, a0, v0.t
+; CHECK-NEXT:    ret
+;
+; CHECK-ZVBB-LABEL: vwsll_vx_i16_nxv8i16:
+; CHECK-ZVBB:       # %bb.0:
+; CHECK-ZVBB-NEXT:    vsetvli zero, a1, e8, m1, ta, ma
+; CHECK-ZVBB-NEXT:    vwsll.vx v10, v8, a0, v0.t
+; CHECK-ZVBB-NEXT:    vmv2r.v v8, v10
+; CHECK-ZVBB-NEXT:    ret
+  %head = insertelement <vscale x 8 x i16> poison, i16 %b, i32 0
+  %splat = shufflevector <vscale x 8 x i16> %head, <vscale x 8 x i16> poison, <vscale x 8 x i32> zeroinitializer
+  %x = zext <vscale x 8 x i8> %a to <vscale x 8 x i16>
+  %z = call <vscale x 8 x i16> @llvm.vp.shl.nxv8i16(<vscale x 8 x i16> %x, <vscale x 8 x i16> %splat, <vscale x 8 x i1> %m, i32 %vl)
+  ret <vscale x 8 x i16> %z
+}
+
+define <vscale x 8 x i16> @vwsll_vx_i8_nxv8i16_sext(<vscale x 8 x i8> %a, i8 %b, <vscale x 8 x i1> %m, i32 zeroext %vl) {
+; CHECK-LABEL: vwsll_vx_i8_nxv8i16_sext:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a2, zero, e8, m1, ta, ma
+; CHECK-NEXT:    vmv.v.x v9, a0
+; CHECK-NEXT:    vsetvli zero, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vzext.vf2 v10, v8
+; CHECK-NEXT:    vsext.vf2 v12, v9
+; CHECK-NEXT:    vsetvli zero, a1, e16, m2, ta, ma
+; CHECK-NEXT:    vsll.vv v8, v10, v12, v0.t
+; CHECK-NEXT:    ret
+;
+; CHECK-ZVBB-LABEL: vwsll_vx_i8_nxv8i16_sext:
+; CHECK-ZVBB:       # %bb.0:
+; CHECK-ZVBB-NEXT:    vsetvli zero, a1, e8, m1, ta, ma
+; CHECK-ZVBB-NEXT:    vwsll.vx v10, v8, a0, v0.t
+; CHECK-ZVBB-NEXT:    vmv2r.v v8, v10
+; CHECK-ZVBB-NEXT:    ret
+  %head = insertelement <vscale x 8 x i8> poison, i8 %b, i32 0
+  %splat = shufflevector <vscale x 8 x i8> %head, <vscale x 8 x i8> poison, <vscale x 8 x i32> zeroinitializer
+  %x = zext <vscale x 8 x i8> %a to <vscale x 8 x i16>
+  %y = sext <vscale x 8 x i8> %splat to <vscale x 8 x i16>
+  %z = call <vscale x 8 x i16> @llvm.vp.shl.nxv8i16(<vscale x 8 x i16> %x, <vscale x 8 x i16> %y, <vscale x 8 x i1> %m, i32 %vl)
+  ret <vscale x 8 x i16> %z
+}
+
+define <vscale x 8 x i16> @vwsll_vx_i8_nxv8i16_zext(<vscale x 8 x i8> %a, i8 %b, <vscale x 8 x i1> %m, i32 zeroext %vl) {
+; CHECK-LABEL: vwsll_vx_i8_nxv8i16_zext:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a2, zero, e8, m1, ta, ma
+; CHECK-NEXT:    vmv.v.x v9, a0
+; CHECK-NEXT:    vsetvli zero, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vzext.vf2 v10, v8
+; CHECK-NEXT:    vzext.vf2 v12, v9
+; CHECK-NEXT:    vsetvli zero, a1, e16, m2, ta, ma
+; CHECK-NEXT:    vsll.vv v8, v10, v12, v0.t
+; CHECK-NEXT:    ret
+;
+; CHECK-ZVBB-LABEL: vwsll_vx_i8_nxv8i16_zext:
+; CHECK-ZVBB:       # %bb.0:
+; CHECK-ZVBB-NEXT:    vsetvli zero, a1, e8, m1, ta, ma
+; CHECK-ZVBB-NEXT:    vwsll.vx v10, v8, a0, v0.t
+; CHECK-ZVBB-NEXT:    vmv2r.v v8, v10
+; CHECK-ZVBB-NEXT:    ret
+  %head = insertelement <vscale x 8 x i8> poison, i8 %b, i32 0
+  %splat = shufflevector <vscale x 8 x i8> %head, <vscale x 8 x i8> poison, <vscale x 8 x i32> zeroinitializer
+  %x = zext <vscale x 8 x i8> %a to <vscale x 8 x i16>
+  %y = zext <vscale x 8 x i8> %splat to <vscale x 8 x i16>
+  %z = call <vscale x 8 x i16> @llvm.vp.shl.nxv8i16(<vscale x 8 x i16> %x, <vscale x 8 x i16> %y, <vscale x 8 x i1> %m, i32 %vl)
+  ret <vscale x 8 x i16> %z
+}
+
+define <vscale x 8 x i16> @vwsll_vi_nxv8i16(<vscale x 8 x i8> %a, <vscale x 8 x i1> %m, i32 zeroext %vl) {
+; CHECK-LABEL: vwsll_vi_nxv8i16:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a1, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vzext.vf2 v10, v8
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
+; CHECK-NEXT:    vsll.vi v8, v10, 2, v0.t
+; CHECK-NEXT:    ret
+;
+; CHECK-ZVBB-LABEL: vwsll_vi_nxv8i16:
+; CHECK-ZVBB:       # %bb.0:
+; CHECK-ZVBB-NEXT:    vsetvli zero, a0, e8, m1, ta, ma
+; CHECK-ZVBB-NEXT:    vwsll.vi v10, v8, 2, v0.t
+; CHECK-ZVBB-NEXT:    vmv2r.v v8, v10
+; CHECK-ZVBB-NEXT:    ret
+  %x = zext <vscale x 8 x i8> %a to <vscale x 8 x i16>
+  %z = call <vscale x 8 x i16> @llvm.vp.shl.nxv8i16(<vscale x 8 x i16> %x, <vscale x 8 x i16> shufflevector(<vscale x 8 x i16> insertelement(<vscale x 8 x i16> poison, i16 2, i32 0), <vscale x 8 x i16> poison, <vscale x 8 x i32> zeroinitializer), <vscale x 8 x i1> %m, i32 %vl)
+  ret <vscale x 8 x i16> %z
+}