diff --git a/README.md b/README.md
index ad990691c54dcc..70fcf0b2d4eba2 100644
--- a/README.md
+++ b/README.md
@@ -39,6 +39,10 @@ Any feature not listed below but present in the specification should be consider
       - (Done) `vand.{vv,vx,vi}`
       - (Done) `vor.{vv,vx,vi}`
       - (Done) `vxor.{vv,vx,vi}`
+    - (Done) `12.5. Vector Single-Width Bit Shift Instructions`
+      - (Done) `vsll.{vv,vx,vi}`
+      - (Done) `vsrl.{vv,vx,vi}`
+      - (Done) `vsra.{vv,vx,vi}`
 - (WIP) Clang intrinsics related to the `XTHeadVector` extension:
   - (WIP) `6. Configuration-Setting and Utility`
     - (Done) `6.1. Set vl and vtype`
diff --git a/llvm/include/llvm/IR/IntrinsicsRISCVXTHeadV.td b/llvm/include/llvm/IR/IntrinsicsRISCVXTHeadV.td
index 91a1a4b9d52048..5ea855df71a04c 100644
--- a/llvm/include/llvm/IR/IntrinsicsRISCVXTHeadV.td
+++ b/llvm/include/llvm/IR/IntrinsicsRISCVXTHeadV.td
@@ -586,6 +586,18 @@ let TargetPrefix = "riscv" in {
     let VLOperand = 4;
   }
 
+  // For destination vector type is the same as
+  // first source vector (with mask but no policy).
+  // The second source operand must match the destination type or be an XLen scalar.
+  // Input: (maskedoff, vector_in, vector_in/scalar_in, mask, vl)
+  class XVBinaryAAShiftMasked
+       : DefaultAttrsIntrinsic<[llvm_anyvector_ty],
+                   [LLVMMatchType<0>, LLVMMatchType<0>, llvm_any_ty,
+                    LLVMScalarOrSameVectorWidth<0, llvm_i1_ty>, llvm_anyint_ty],
+                   [IntrNoMem]>, RISCVVIntrinsic {
+    let VLOperand = 4;
+  }
+
   multiclass XVBinaryAAX {
     def "int_riscv_" # NAME : RISCVBinaryAAXUnMasked;
     def "int_riscv_" # NAME # "_mask" : XVBinaryAAXMasked;
@@ -595,6 +607,11 @@ let TargetPrefix = "riscv" in {
     def "int_riscv_" # NAME : RISCVBinaryABXUnMasked;
     def "int_riscv_" # NAME # "_mask" : XVBinaryABXMasked;
   }
+
+  multiclass XVBinaryAAShift {
+    def "int_riscv_" # NAME : RISCVBinaryAAShiftUnMasked;
+    def "int_riscv_" # NAME # "_mask" : XVBinaryAAShiftMasked;
+  }
 }
 
 let TargetPrefix = "riscv" in {
@@ -624,6 +641,11 @@ let TargetPrefix = "riscv" in {
   defm th_vand : XVBinaryAAX;
   defm th_vor : XVBinaryAAX;
   defm th_vxor : XVBinaryAAX;
+
+  // 12.5. Vector Single-Width Bit Shift Instructions
+  defm th_vsll : XVBinaryAAShift;
+  defm th_vsrl : XVBinaryAAShift;
+  defm th_vsra : XVBinaryAAShift;
 } // TargetPrefix = "riscv"
 
 let TargetPrefix = "riscv" in {
diff --git a/llvm/lib/Target/RISCV/RISCVInstrInfoXTHeadVPseudos.td b/llvm/lib/Target/RISCV/RISCVInstrInfoXTHeadVPseudos.td
index bd06cc79150a2e..4ddca578683a99 100644
--- a/llvm/lib/Target/RISCV/RISCVInstrInfoXTHeadVPseudos.td
+++ b/llvm/lib/Target/RISCV/RISCVInstrInfoXTHeadVPseudos.td
@@ -1789,6 +1789,24 @@ multiclass XVPseudoVCALUM_VM_XM<string Constraint> {
   }
 }
 
+multiclass XVPseudoVSHT_VV_VX_VI<Operand ImmType = simm5, string Constraint = ""> {
+  foreach m = MxListXTHeadV in {
+    defvar mx = m.MX;
+    defvar WriteVShiftV_MX = !cast<SchedWrite>("WriteVShiftV_" # mx);
+    defvar WriteVShiftX_MX = !cast<SchedWrite>("WriteVShiftX_" # mx);
+    defvar WriteVShiftI_MX = !cast<SchedWrite>("WriteVShiftI_" # mx);
+    defvar ReadVShiftV_MX = !cast<SchedRead>("ReadVShiftV_" # mx);
+    defvar ReadVShiftX_MX = !cast<SchedRead>("ReadVShiftX_" # mx);
+
+    defm "" : XVPseudoBinaryV_VV<m, Constraint>,
+              Sched<[WriteVShiftV_MX, ReadVShiftV_MX, ReadVShiftV_MX, ReadVMask]>;
+    defm "" : XVPseudoBinaryV_VX<m, Constraint>,
+              Sched<[WriteVShiftX_MX, ReadVShiftV_MX, ReadVShiftX_MX, ReadVMask]>;
+    defm "" : XVPseudoBinaryV_VI<ImmType, m, Constraint>,
+              Sched<[WriteVShiftI_MX, ReadVShiftV_MX, ReadVMask]>;
+  }
+}
+
 //===----------------------------------------------------------------------===//
 // Helpers to define the intrinsic patterns for the XTHeadVector extension.
 //===----------------------------------------------------------------------===//
@@ -2207,6 +2225,48 @@ let Predicates = [HasVendorXTHeadV] in {
   defm : XVPatBinaryV_VV_VX_VI<"int_riscv_th_vxor", "PseudoTH_VXOR", AllIntegerXVectors>;
 }
 
+//===----------------------------------------------------------------------===//
+// 12.5. Vector Single-Width Bit Shift Instructions
+//===----------------------------------------------------------------------===//
+let Predicates = [HasVendorXTHeadV] in {
+  defm PseudoTH_VSLL : XVPseudoVSHT_VV_VX_VI<uimm5>;
+  defm PseudoTH_VSRL : XVPseudoVSHT_VV_VX_VI<uimm5>;
+  defm PseudoTH_VSRA : XVPseudoVSHT_VV_VX_VI<uimm5>;
+} // Predicates = [HasVendorXTHeadV]
+
+let Predicates = [HasVendorXTHeadV] in {
+  defm : XVPatBinaryV_VV_VX_VI<"int_riscv_th_vsll", "PseudoTH_VSLL", AllIntegerXVectors,
+                               uimm5>;
+  defm : XVPatBinaryV_VV_VX_VI<"int_riscv_th_vsrl", "PseudoTH_VSRL", AllIntegerXVectors,
+                               uimm5>;
+  defm : XVPatBinaryV_VV_VX_VI<"int_riscv_th_vsra", "PseudoTH_VSRA", AllIntegerXVectors,
+                               uimm5>;
+
+  foreach vti = AllIntegerXVectors in {
+    // Emit shift by 1 as an add since it might be faster.
+    let Predicates = GetXVTypePredicates<vti>.Predicates in {
+      def : Pat<(vti.Vector (int_riscv_th_vsll (vti.Vector undef),
+                                               (vti.Vector vti.RegClass:$rs1),
+                                               (XLenVT 1), VLOpFrag)),
+                (!cast<Instruction>("PseudoTH_VADD_VV_"#vti.LMul.MX)
+                  (vti.Vector (IMPLICIT_DEF)), vti.RegClass:$rs1,
+                  vti.RegClass:$rs1, GPR:$vl, vti.Log2SEW)>;
+      def : Pat<(vti.Vector (int_riscv_th_vsll_mask (vti.Vector vti.RegClass:$merge),
+                                                    (vti.Vector vti.RegClass:$rs1),
+                                                    (XLenVT 1),
+                                                    (vti.Mask V0),
+                                                    VLOpFrag)),
+                (!cast<Instruction>("PseudoTH_VADD_VV_"#vti.LMul.MX#"_MASK")
+                                                            vti.RegClass:$merge,
+                                                            vti.RegClass:$rs1,
+                                                            vti.RegClass:$rs1,
+                                                            (vti.Mask V0),
+                                                            GPR:$vl,
+                                                            vti.Log2SEW)>;
+    }
+  }
+} // Predicates = [HasVendorXTHeadV]
+
 //===----------------------------------------------------------------------===//
 // 12.14. Vector Integer Merge and Move Instructions
 //===----------------------------------------------------------------------===//
diff --git a/llvm/test/CodeGen/RISCV/rvv0p71/vsll.ll b/llvm/test/CodeGen/RISCV/rvv0p71/vsll.ll
new file mode 100644
index 00000000000000..7ac6ea68b347cb
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv0p71/vsll.ll
@@ -0,0 +1,3150 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+xtheadvector \
+; RUN:   -verify-machineinstrs | FileCheck %s
+; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+xtheadvector \
+; RUN:   -verify-machineinstrs | FileCheck %s
+
+declare <vscale x 8 x i8> @llvm.riscv.th.vsll.nxv8i8.nxv8i8(
+  <vscale x 8 x i8>,
+  <vscale x 8 x i8>,
+  <vscale x 8 x i8>,
+  iXLen);
+
+define <vscale x 8 x i8> @intrinsic_vsll_vv_nxv8i8_nxv8i8_nxv8i8(<vscale x 8 x i8> %0, <vscale x 8 x i8> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsll_vv_nxv8i8_nxv8i8_nxv8i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e8, m1, d1
+; CHECK-NEXT:    th.vsll.vv v8, v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i8> @llvm.riscv.th.vsll.nxv8i8.nxv8i8(
+    <vscale x 8 x i8> undef,
+    <vscale x 8 x i8> %0,
+    <vscale x 8 x i8> %1,
+    iXLen %2)
+
+  ret <vscale x 8 x i8> %a
+}
+
+declare <vscale x 8 x i8> @llvm.riscv.th.vsll.mask.nxv8i8.nxv8i8(
+  <vscale x 8 x i8>,
+  <vscale x 8 x i8>,
+  <vscale x 8 x i8>,
+  <vscale x 8 x i1>,
+  iXLen);
+
+define <vscale x 8 x i8> @intrinsic_vsll_mask_vv_nxv8i8_nxv8i8_nxv8i8(<vscale x 8 x i8> %0, <vscale x 8 x i8> %1, <vscale x 8 x i8> %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsll_mask_vv_nxv8i8_nxv8i8_nxv8i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e8, m1, d1
+; CHECK-NEXT:    th.vsll.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i8> @llvm.riscv.th.vsll.mask.nxv8i8.nxv8i8(
+    <vscale x 8 x i8> %0,
+    <vscale x 8 x i8> %1,
+    <vscale x 8 x i8> %2,
+    <vscale x 8 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 8 x i8> %a
+}
+
+declare <vscale x 16 x i8> @llvm.riscv.th.vsll.nxv16i8.nxv16i8(
+  <vscale x 16 x i8>,
+  <vscale x 16 x i8>,
+  <vscale x 16 x i8>,
+  iXLen);
+
+define <vscale x 16 x i8> @intrinsic_vsll_vv_nxv16i8_nxv16i8_nxv16i8(<vscale x 16 x i8> %0, <vscale x 16 x i8> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsll_vv_nxv16i8_nxv16i8_nxv16i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e8, m2, d1
+; CHECK-NEXT:    th.vsll.vv v8, v8, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i8> @llvm.riscv.th.vsll.nxv16i8.nxv16i8(
+    <vscale x 16 x i8> undef,
+    <vscale x 16 x i8> %0,
+    <vscale x 16 x i8> %1,
+    iXLen %2)
+
+  ret <vscale x 16 x i8> %a
+}
+
+declare <vscale x 16 x i8> @llvm.riscv.th.vsll.mask.nxv16i8.nxv16i8(
+  <vscale x 16 x i8>,
+  <vscale x 16 x i8>,
+  <vscale x 16 x i8>,
+  <vscale x 16 x i1>,
+  iXLen);
+
+define <vscale x 16 x i8> @intrinsic_vsll_mask_vv_nxv16i8_nxv16i8_nxv16i8(<vscale x 16 x i8> %0, <vscale x 16 x i8> %1, <vscale x 16 x i8> %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsll_mask_vv_nxv16i8_nxv16i8_nxv16i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e8, m2, d1
+; CHECK-NEXT:    th.vsll.vv v8, v10, v12, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i8> @llvm.riscv.th.vsll.mask.nxv16i8.nxv16i8(
+    <vscale x 16 x i8> %0,
+    <vscale x 16 x i8> %1,
+    <vscale x 16 x i8> %2,
+    <vscale x 16 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 16 x i8> %a
+}
+
+declare <vscale x 32 x i8> @llvm.riscv.th.vsll.nxv32i8.nxv32i8(
+  <vscale x 32 x i8>,
+  <vscale x 32 x i8>,
+  <vscale x 32 x i8>,
+  iXLen);
+
+define <vscale x 32 x i8> @intrinsic_vsll_vv_nxv32i8_nxv32i8_nxv32i8(<vscale x 32 x i8> %0, <vscale x 32 x i8> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsll_vv_nxv32i8_nxv32i8_nxv32i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e8, m4, d1
+; CHECK-NEXT:    th.vsll.vv v8, v8, v12
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x i8> @llvm.riscv.th.vsll.nxv32i8.nxv32i8(
+    <vscale x 32 x i8> undef,
+    <vscale x 32 x i8> %0,
+    <vscale x 32 x i8> %1,
+    iXLen %2)
+
+  ret <vscale x 32 x i8> %a
+}
+
+declare <vscale x 32 x i8> @llvm.riscv.th.vsll.mask.nxv32i8.nxv32i8(
+  <vscale x 32 x i8>,
+  <vscale x 32 x i8>,
+  <vscale x 32 x i8>,
+  <vscale x 32 x i1>,
+  iXLen);
+
+define <vscale x 32 x i8> @intrinsic_vsll_mask_vv_nxv32i8_nxv32i8_nxv32i8(<vscale x 32 x i8> %0, <vscale x 32 x i8> %1, <vscale x 32 x i8> %2, <vscale x 32 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsll_mask_vv_nxv32i8_nxv32i8_nxv32i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e8, m4, d1
+; CHECK-NEXT:    th.vsll.vv v8, v12, v16, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x i8> @llvm.riscv.th.vsll.mask.nxv32i8.nxv32i8(
+    <vscale x 32 x i8> %0,
+    <vscale x 32 x i8> %1,
+    <vscale x 32 x i8> %2,
+    <vscale x 32 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 32 x i8> %a
+}
+
+declare <vscale x 64 x i8> @llvm.riscv.th.vsll.nxv64i8.nxv64i8(
+  <vscale x 64 x i8>,
+  <vscale x 64 x i8>,
+  <vscale x 64 x i8>,
+  iXLen);
+
+define <vscale x 64 x i8> @intrinsic_vsll_vv_nxv64i8_nxv64i8_nxv64i8(<vscale x 64 x i8> %0, <vscale x 64 x i8> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsll_vv_nxv64i8_nxv64i8_nxv64i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e8, m8, d1
+; CHECK-NEXT:    th.vsll.vv v8, v8, v16
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 64 x i8> @llvm.riscv.th.vsll.nxv64i8.nxv64i8(
+    <vscale x 64 x i8> undef,
+    <vscale x 64 x i8> %0,
+    <vscale x 64 x i8> %1,
+    iXLen %2)
+
+  ret <vscale x 64 x i8> %a
+}
+
+declare <vscale x 64 x i8> @llvm.riscv.th.vsll.mask.nxv64i8.nxv64i8(
+  <vscale x 64 x i8>,
+  <vscale x 64 x i8>,
+  <vscale x 64 x i8>,
+  <vscale x 64 x i1>,
+  iXLen);
+
+define <vscale x 64 x i8> @intrinsic_vsll_mask_vv_nxv64i8_nxv64i8_nxv64i8(<vscale x 64 x i8> %0, <vscale x 64 x i8> %1, <vscale x 64 x i8> %2, <vscale x 64 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsll_mask_vv_nxv64i8_nxv64i8_nxv64i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m8, d1
+; CHECK-NEXT:    th.vle.v v24, (a0)
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a0, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a0, a2
+; CHECK-NEXT:    th.vsetvli zero, a1, e8, m8, d1
+; CHECK-NEXT:    th.vsll.vv v8, v16, v24, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 64 x i8> @llvm.riscv.th.vsll.mask.nxv64i8.nxv64i8(
+    <vscale x 64 x i8> %0,
+    <vscale x 64 x i8> %1,
+    <vscale x 64 x i8> %2,
+    <vscale x 64 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 64 x i8> %a
+}
+
+declare <vscale x 4 x i16> @llvm.riscv.th.vsll.nxv4i16.nxv4i16(
+  <vscale x 4 x i16>,
+  <vscale x 4 x i16>,
+  <vscale x 4 x i16>,
+  iXLen);
+
+define <vscale x 4 x i16> @intrinsic_vsll_vv_nxv4i16_nxv4i16_nxv4i16(<vscale x 4 x i16> %0, <vscale x 4 x i16> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsll_vv_nxv4i16_nxv4i16_nxv4i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e16, m1, d1
+; CHECK-NEXT:    th.vsll.vv v8, v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i16> @llvm.riscv.th.vsll.nxv4i16.nxv4i16(
+    <vscale x 4 x i16> undef,
+    <vscale x 4 x i16> %0,
+    <vscale x 4 x i16> %1,
+    iXLen %2)
+
+  ret <vscale x 4 x i16> %a
+}
+
+declare <vscale x 4 x i16> @llvm.riscv.th.vsll.mask.nxv4i16.nxv4i16(
+  <vscale x 4 x i16>,
+  <vscale x 4 x i16>,
+  <vscale x 4 x i16>,
+  <vscale x 4 x i1>,
+  iXLen);
+
+define <vscale x 4 x i16> @intrinsic_vsll_mask_vv_nxv4i16_nxv4i16_nxv4i16(<vscale x 4 x i16> %0, <vscale x 4 x i16> %1, <vscale x 4 x i16> %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsll_mask_vv_nxv4i16_nxv4i16_nxv4i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e16, m1, d1
+; CHECK-NEXT:    th.vsll.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i16> @llvm.riscv.th.vsll.mask.nxv4i16.nxv4i16(
+    <vscale x 4 x i16> %0,
+    <vscale x 4 x i16> %1,
+    <vscale x 4 x i16> %2,
+    <vscale x 4 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 4 x i16> %a
+}
+
+declare <vscale x 8 x i16> @llvm.riscv.th.vsll.nxv8i16.nxv8i16(
+  <vscale x 8 x i16>,
+  <vscale x 8 x i16>,
+  <vscale x 8 x i16>,
+  iXLen);
+
+define <vscale x 8 x i16> @intrinsic_vsll_vv_nxv8i16_nxv8i16_nxv8i16(<vscale x 8 x i16> %0, <vscale x 8 x i16> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsll_vv_nxv8i16_nxv8i16_nxv8i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e16, m2, d1
+; CHECK-NEXT:    th.vsll.vv v8, v8, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i16> @llvm.riscv.th.vsll.nxv8i16.nxv8i16(
+    <vscale x 8 x i16> undef,
+    <vscale x 8 x i16> %0,
+    <vscale x 8 x i16> %1,
+    iXLen %2)
+
+  ret <vscale x 8 x i16> %a
+}
+
+declare <vscale x 8 x i16> @llvm.riscv.th.vsll.mask.nxv8i16.nxv8i16(
+  <vscale x 8 x i16>,
+  <vscale x 8 x i16>,
+  <vscale x 8 x i16>,
+  <vscale x 8 x i1>,
+  iXLen);
+
+define <vscale x 8 x i16> @intrinsic_vsll_mask_vv_nxv8i16_nxv8i16_nxv8i16(<vscale x 8 x i16> %0, <vscale x 8 x i16> %1, <vscale x 8 x i16> %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsll_mask_vv_nxv8i16_nxv8i16_nxv8i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e16, m2, d1
+; CHECK-NEXT:    th.vsll.vv v8, v10, v12, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i16> @llvm.riscv.th.vsll.mask.nxv8i16.nxv8i16(
+    <vscale x 8 x i16> %0,
+    <vscale x 8 x i16> %1,
+    <vscale x 8 x i16> %2,
+    <vscale x 8 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 8 x i16> %a
+}
+
+declare <vscale x 16 x i16> @llvm.riscv.th.vsll.nxv16i16.nxv16i16(
+  <vscale x 16 x i16>,
+  <vscale x 16 x i16>,
+  <vscale x 16 x i16>,
+  iXLen);
+
+define <vscale x 16 x i16> @intrinsic_vsll_vv_nxv16i16_nxv16i16_nxv16i16(<vscale x 16 x i16> %0, <vscale x 16 x i16> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsll_vv_nxv16i16_nxv16i16_nxv16i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e16, m4, d1
+; CHECK-NEXT:    th.vsll.vv v8, v8, v12
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i16> @llvm.riscv.th.vsll.nxv16i16.nxv16i16(
+    <vscale x 16 x i16> undef,
+    <vscale x 16 x i16> %0,
+    <vscale x 16 x i16> %1,
+    iXLen %2)
+
+  ret <vscale x 16 x i16> %a
+}
+
+declare <vscale x 16 x i16> @llvm.riscv.th.vsll.mask.nxv16i16.nxv16i16(
+  <vscale x 16 x i16>,
+  <vscale x 16 x i16>,
+  <vscale x 16 x i16>,
+  <vscale x 16 x i1>,
+  iXLen);
+
+define <vscale x 16 x i16> @intrinsic_vsll_mask_vv_nxv16i16_nxv16i16_nxv16i16(<vscale x 16 x i16> %0, <vscale x 16 x i16> %1, <vscale x 16 x i16> %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsll_mask_vv_nxv16i16_nxv16i16_nxv16i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e16, m4, d1
+; CHECK-NEXT:    th.vsll.vv v8, v12, v16, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i16> @llvm.riscv.th.vsll.mask.nxv16i16.nxv16i16(
+    <vscale x 16 x i16> %0,
+    <vscale x 16 x i16> %1,
+    <vscale x 16 x i16> %2,
+    <vscale x 16 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 16 x i16> %a
+}
+
+declare <vscale x 32 x i16> @llvm.riscv.th.vsll.nxv32i16.nxv32i16(
+  <vscale x 32 x i16>,
+  <vscale x 32 x i16>,
+  <vscale x 32 x i16>,
+  iXLen);
+
+define <vscale x 32 x i16> @intrinsic_vsll_vv_nxv32i16_nxv32i16_nxv32i16(<vscale x 32 x i16> %0, <vscale x 32 x i16> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsll_vv_nxv32i16_nxv32i16_nxv32i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e16, m8, d1
+; CHECK-NEXT:    th.vsll.vv v8, v8, v16
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x i16> @llvm.riscv.th.vsll.nxv32i16.nxv32i16(
+    <vscale x 32 x i16> undef,
+    <vscale x 32 x i16> %0,
+    <vscale x 32 x i16> %1,
+    iXLen %2)
+
+  ret <vscale x 32 x i16> %a
+}
+
+declare <vscale x 32 x i16> @llvm.riscv.th.vsll.mask.nxv32i16.nxv32i16(
+  <vscale x 32 x i16>,
+  <vscale x 32 x i16>,
+  <vscale x 32 x i16>,
+  <vscale x 32 x i1>,
+  iXLen);
+
+define <vscale x 32 x i16> @intrinsic_vsll_mask_vv_nxv32i16_nxv32i16_nxv32i16(<vscale x 32 x i16> %0, <vscale x 32 x i16> %1, <vscale x 32 x i16> %2, <vscale x 32 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsll_mask_vv_nxv32i16_nxv32i16_nxv32i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e16, m8, d1
+; CHECK-NEXT:    th.vle.v v24, (a0)
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a0, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a0, a2
+; CHECK-NEXT:    th.vsetvli zero, a1, e16, m8, d1
+; CHECK-NEXT:    th.vsll.vv v8, v16, v24, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x i16> @llvm.riscv.th.vsll.mask.nxv32i16.nxv32i16(
+    <vscale x 32 x i16> %0,
+    <vscale x 32 x i16> %1,
+    <vscale x 32 x i16> %2,
+    <vscale x 32 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 32 x i16> %a
+}
+
+declare <vscale x 2 x i32> @llvm.riscv.th.vsll.nxv2i32.nxv2i32(
+  <vscale x 2 x i32>,
+  <vscale x 2 x i32>,
+  <vscale x 2 x i32>,
+  iXLen);
+
+define <vscale x 2 x i32> @intrinsic_vsll_vv_nxv2i32_nxv2i32_nxv2i32(<vscale x 2 x i32> %0, <vscale x 2 x i32> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsll_vv_nxv2i32_nxv2i32_nxv2i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e32, m1, d1
+; CHECK-NEXT:    th.vsll.vv v8, v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i32> @llvm.riscv.th.vsll.nxv2i32.nxv2i32(
+    <vscale x 2 x i32> undef,
+    <vscale x 2 x i32> %0,
+    <vscale x 2 x i32> %1,
+    iXLen %2)
+
+  ret <vscale x 2 x i32> %a
+}
+
+declare <vscale x 2 x i32> @llvm.riscv.th.vsll.mask.nxv2i32.nxv2i32(
+  <vscale x 2 x i32>,
+  <vscale x 2 x i32>,
+  <vscale x 2 x i32>,
+  <vscale x 2 x i1>,
+  iXLen);
+
+define <vscale x 2 x i32> @intrinsic_vsll_mask_vv_nxv2i32_nxv2i32_nxv2i32(<vscale x 2 x i32> %0, <vscale x 2 x i32> %1, <vscale x 2 x i32> %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsll_mask_vv_nxv2i32_nxv2i32_nxv2i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e32, m1, d1
+; CHECK-NEXT:    th.vsll.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i32> @llvm.riscv.th.vsll.mask.nxv2i32.nxv2i32(
+    <vscale x 2 x i32> %0,
+    <vscale x 2 x i32> %1,
+    <vscale x 2 x i32> %2,
+    <vscale x 2 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 2 x i32> %a
+}
+
+declare <vscale x 4 x i32> @llvm.riscv.th.vsll.nxv4i32.nxv4i32(
+  <vscale x 4 x i32>,
+  <vscale x 4 x i32>,
+  <vscale x 4 x i32>,
+  iXLen);
+
+define <vscale x 4 x i32> @intrinsic_vsll_vv_nxv4i32_nxv4i32_nxv4i32(<vscale x 4 x i32> %0, <vscale x 4 x i32> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsll_vv_nxv4i32_nxv4i32_nxv4i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e32, m2, d1
+; CHECK-NEXT:    th.vsll.vv v8, v8, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i32> @llvm.riscv.th.vsll.nxv4i32.nxv4i32(
+    <vscale x 4 x i32> undef,
+    <vscale x 4 x i32> %0,
+    <vscale x 4 x i32> %1,
+    iXLen %2)
+
+  ret <vscale x 4 x i32> %a
+}
+
+declare <vscale x 4 x i32> @llvm.riscv.th.vsll.mask.nxv4i32.nxv4i32(
+  <vscale x 4 x i32>,
+  <vscale x 4 x i32>,
+  <vscale x 4 x i32>,
+  <vscale x 4 x i1>,
+  iXLen);
+
+define <vscale x 4 x i32> @intrinsic_vsll_mask_vv_nxv4i32_nxv4i32_nxv4i32(<vscale x 4 x i32> %0, <vscale x 4 x i32> %1, <vscale x 4 x i32> %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsll_mask_vv_nxv4i32_nxv4i32_nxv4i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e32, m2, d1
+; CHECK-NEXT:    th.vsll.vv v8, v10, v12, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i32> @llvm.riscv.th.vsll.mask.nxv4i32.nxv4i32(
+    <vscale x 4 x i32> %0,
+    <vscale x 4 x i32> %1,
+    <vscale x 4 x i32> %2,
+    <vscale x 4 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 4 x i32> %a
+}
+
+declare <vscale x 8 x i32> @llvm.riscv.th.vsll.nxv8i32.nxv8i32(
+  <vscale x 8 x i32>,
+  <vscale x 8 x i32>,
+  <vscale x 8 x i32>,
+  iXLen);
+
+define <vscale x 8 x i32> @intrinsic_vsll_vv_nxv8i32_nxv8i32_nxv8i32(<vscale x 8 x i32> %0, <vscale x 8 x i32> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsll_vv_nxv8i32_nxv8i32_nxv8i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e32, m4, d1
+; CHECK-NEXT:    th.vsll.vv v8, v8, v12
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i32> @llvm.riscv.th.vsll.nxv8i32.nxv8i32(
+    <vscale x 8 x i32> undef,
+    <vscale x 8 x i32> %0,
+    <vscale x 8 x i32> %1,
+    iXLen %2)
+
+  ret <vscale x 8 x i32> %a
+}
+
+declare <vscale x 8 x i32> @llvm.riscv.th.vsll.mask.nxv8i32.nxv8i32(
+  <vscale x 8 x i32>,
+  <vscale x 8 x i32>,
+  <vscale x 8 x i32>,
+  <vscale x 8 x i1>,
+  iXLen);
+
+define <vscale x 8 x i32> @intrinsic_vsll_mask_vv_nxv8i32_nxv8i32_nxv8i32(<vscale x 8 x i32> %0, <vscale x 8 x i32> %1, <vscale x 8 x i32> %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsll_mask_vv_nxv8i32_nxv8i32_nxv8i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e32, m4, d1
+; CHECK-NEXT:    th.vsll.vv v8, v12, v16, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i32> @llvm.riscv.th.vsll.mask.nxv8i32.nxv8i32(
+    <vscale x 8 x i32> %0,
+    <vscale x 8 x i32> %1,
+    <vscale x 8 x i32> %2,
+    <vscale x 8 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 8 x i32> %a
+}
+
+declare <vscale x 16 x i32> @llvm.riscv.th.vsll.nxv16i32.nxv16i32(
+  <vscale x 16 x i32>,
+  <vscale x 16 x i32>,
+  <vscale x 16 x i32>,
+  iXLen);
+
+define <vscale x 16 x i32> @intrinsic_vsll_vv_nxv16i32_nxv16i32_nxv16i32(<vscale x 16 x i32> %0, <vscale x 16 x i32> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsll_vv_nxv16i32_nxv16i32_nxv16i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e32, m8, d1
+; CHECK-NEXT:    th.vsll.vv v8, v8, v16
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i32> @llvm.riscv.th.vsll.nxv16i32.nxv16i32(
+    <vscale x 16 x i32> undef,
+    <vscale x 16 x i32> %0,
+    <vscale x 16 x i32> %1,
+    iXLen %2)
+
+  ret <vscale x 16 x i32> %a
+}
+
+declare <vscale x 16 x i32> @llvm.riscv.th.vsll.mask.nxv16i32.nxv16i32(
+  <vscale x 16 x i32>,
+  <vscale x 16 x i32>,
+  <vscale x 16 x i32>,
+  <vscale x 16 x i1>,
+  iXLen);
+
+define <vscale x 16 x i32> @intrinsic_vsll_mask_vv_nxv16i32_nxv16i32_nxv16i32(<vscale x 16 x i32> %0, <vscale x 16 x i32> %1, <vscale x 16 x i32> %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsll_mask_vv_nxv16i32_nxv16i32_nxv16i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e32, m8, d1
+; CHECK-NEXT:    th.vle.v v24, (a0)
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a0, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a0, a2
+; CHECK-NEXT:    th.vsetvli zero, a1, e32, m8, d1
+; CHECK-NEXT:    th.vsll.vv v8, v16, v24, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i32> @llvm.riscv.th.vsll.mask.nxv16i32.nxv16i32(
+    <vscale x 16 x i32> %0,
+    <vscale x 16 x i32> %1,
+    <vscale x 16 x i32> %2,
+    <vscale x 16 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 16 x i32> %a
+}
+
+declare <vscale x 1 x i64> @llvm.riscv.th.vsll.nxv1i64.nxv1i64(
+  <vscale x 1 x i64>,
+  <vscale x 1 x i64>,
+  <vscale x 1 x i64>,
+  iXLen);
+
+define <vscale x 1 x i64> @intrinsic_vsll_vv_nxv1i64_nxv1i64_nxv1i64(<vscale x 1 x i64> %0, <vscale x 1 x i64> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsll_vv_nxv1i64_nxv1i64_nxv1i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e64, m1, d1
+; CHECK-NEXT:    th.vsll.vv v8, v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x i64> @llvm.riscv.th.vsll.nxv1i64.nxv1i64(
+    <vscale x 1 x i64> undef,
+    <vscale x 1 x i64> %0,
+    <vscale x 1 x i64> %1,
+    iXLen %2)
+
+  ret <vscale x 1 x i64> %a
+}
+
+declare <vscale x 1 x i64> @llvm.riscv.th.vsll.mask.nxv1i64.nxv1i64(
+  <vscale x 1 x i64>,
+  <vscale x 1 x i64>,
+  <vscale x 1 x i64>,
+  <vscale x 1 x i1>,
+  iXLen);
+
+define <vscale x 1 x i64> @intrinsic_vsll_mask_vv_nxv1i64_nxv1i64_nxv1i64(<vscale x 1 x i64> %0, <vscale x 1 x i64> %1, <vscale x 1 x i64> %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsll_mask_vv_nxv1i64_nxv1i64_nxv1i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e64, m1, d1
+; CHECK-NEXT:    th.vsll.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x i64> @llvm.riscv.th.vsll.mask.nxv1i64.nxv1i64(
+    <vscale x 1 x i64> %0,
+    <vscale x 1 x i64> %1,
+    <vscale x 1 x i64> %2,
+    <vscale x 1 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 1 x i64> %a
+}
+
+declare <vscale x 2 x i64> @llvm.riscv.th.vsll.nxv2i64.nxv2i64(
+  <vscale x 2 x i64>,
+  <vscale x 2 x i64>,
+  <vscale x 2 x i64>,
+  iXLen);
+
+define <vscale x 2 x i64> @intrinsic_vsll_vv_nxv2i64_nxv2i64_nxv2i64(<vscale x 2 x i64> %0, <vscale x 2 x i64> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsll_vv_nxv2i64_nxv2i64_nxv2i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e64, m2, d1
+; CHECK-NEXT:    th.vsll.vv v8, v8, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i64> @llvm.riscv.th.vsll.nxv2i64.nxv2i64(
+    <vscale x 2 x i64> undef,
+    <vscale x 2 x i64> %0,
+    <vscale x 2 x i64> %1,
+    iXLen %2)
+
+  ret <vscale x 2 x i64> %a
+}
+
+declare <vscale x 2 x i64> @llvm.riscv.th.vsll.mask.nxv2i64.nxv2i64(
+  <vscale x 2 x i64>,
+  <vscale x 2 x i64>,
+  <vscale x 2 x i64>,
+  <vscale x 2 x i1>,
+  iXLen);
+
+define <vscale x 2 x i64> @intrinsic_vsll_mask_vv_nxv2i64_nxv2i64_nxv2i64(<vscale x 2 x i64> %0, <vscale x 2 x i64> %1, <vscale x 2 x i64> %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsll_mask_vv_nxv2i64_nxv2i64_nxv2i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e64, m2, d1
+; CHECK-NEXT:    th.vsll.vv v8, v10, v12, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i64> @llvm.riscv.th.vsll.mask.nxv2i64.nxv2i64(
+    <vscale x 2 x i64> %0,
+    <vscale x 2 x i64> %1,
+    <vscale x 2 x i64> %2,
+    <vscale x 2 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 2 x i64> %a
+}
+
+declare <vscale x 4 x i64> @llvm.riscv.th.vsll.nxv4i64.nxv4i64(
+  <vscale x 4 x i64>,
+  <vscale x 4 x i64>,
+  <vscale x 4 x i64>,
+  iXLen);
+
+define <vscale x 4 x i64> @intrinsic_vsll_vv_nxv4i64_nxv4i64_nxv4i64(<vscale x 4 x i64> %0, <vscale x 4 x i64> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsll_vv_nxv4i64_nxv4i64_nxv4i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e64, m4, d1
+; CHECK-NEXT:    th.vsll.vv v8, v8, v12
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i64> @llvm.riscv.th.vsll.nxv4i64.nxv4i64(
+    <vscale x 4 x i64> undef,
+    <vscale x 4 x i64> %0,
+    <vscale x 4 x i64> %1,
+    iXLen %2)
+
+  ret <vscale x 4 x i64> %a
+}
+
+declare <vscale x 4 x i64> @llvm.riscv.th.vsll.mask.nxv4i64.nxv4i64(
+  <vscale x 4 x i64>,
+  <vscale x 4 x i64>,
+  <vscale x 4 x i64>,
+  <vscale x 4 x i1>,
+  iXLen);
+
+define <vscale x 4 x i64> @intrinsic_vsll_mask_vv_nxv4i64_nxv4i64_nxv4i64(<vscale x 4 x i64> %0, <vscale x 4 x i64> %1, <vscale x 4 x i64> %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsll_mask_vv_nxv4i64_nxv4i64_nxv4i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e64, m4, d1
+; CHECK-NEXT:    th.vsll.vv v8, v12, v16, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i64> @llvm.riscv.th.vsll.mask.nxv4i64.nxv4i64(
+    <vscale x 4 x i64> %0,
+    <vscale x 4 x i64> %1,
+    <vscale x 4 x i64> %2,
+    <vscale x 4 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 4 x i64> %a
+}
+
+declare <vscale x 8 x i64> @llvm.riscv.th.vsll.nxv8i64.nxv8i64(
+  <vscale x 8 x i64>,
+  <vscale x 8 x i64>,
+  <vscale x 8 x i64>,
+  iXLen);
+
+define <vscale x 8 x i64> @intrinsic_vsll_vv_nxv8i64_nxv8i64_nxv8i64(<vscale x 8 x i64> %0, <vscale x 8 x i64> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsll_vv_nxv8i64_nxv8i64_nxv8i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e64, m8, d1
+; CHECK-NEXT:    th.vsll.vv v8, v8, v16
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i64> @llvm.riscv.th.vsll.nxv8i64.nxv8i64(
+    <vscale x 8 x i64> undef,
+    <vscale x 8 x i64> %0,
+    <vscale x 8 x i64> %1,
+    iXLen %2)
+
+  ret <vscale x 8 x i64> %a
+}
+
+declare <vscale x 8 x i64> @llvm.riscv.th.vsll.mask.nxv8i64.nxv8i64(
+  <vscale x 8 x i64>,
+  <vscale x 8 x i64>,
+  <vscale x 8 x i64>,
+  <vscale x 8 x i1>,
+  iXLen);
+
+define <vscale x 8 x i64> @intrinsic_vsll_mask_vv_nxv8i64_nxv8i64_nxv8i64(<vscale x 8 x i64> %0, <vscale x 8 x i64> %1, <vscale x 8 x i64> %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsll_mask_vv_nxv8i64_nxv8i64_nxv8i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e64, m8, d1
+; CHECK-NEXT:    th.vle.v v24, (a0)
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a0, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a0, a2
+; CHECK-NEXT:    th.vsetvli zero, a1, e64, m8, d1
+; CHECK-NEXT:    th.vsll.vv v8, v16, v24, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i64> @llvm.riscv.th.vsll.mask.nxv8i64.nxv8i64(
+    <vscale x 8 x i64> %0,
+    <vscale x 8 x i64> %1,
+    <vscale x 8 x i64> %2,
+    <vscale x 8 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 8 x i64> %a
+}
+
+declare <vscale x 8 x i8> @llvm.riscv.th.vsll.nxv8i8(
+  <vscale x 8 x i8>,
+  <vscale x 8 x i8>,
+  iXLen,
+  iXLen);
+
+define <vscale x 8 x i8> @intrinsic_vsll_vx_nxv8i8_nxv8i8(<vscale x 8 x i8> %0, iXLen %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsll_vx_nxv8i8_nxv8i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e8, m1, d1
+; CHECK-NEXT:    th.vsll.vx v8, v8, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i8> @llvm.riscv.th.vsll.nxv8i8(
+    <vscale x 8 x i8> undef,
+    <vscale x 8 x i8> %0,
+    iXLen %1,
+    iXLen %2)
+
+  ret <vscale x 8 x i8> %a
+}
+
+declare <vscale x 8 x i8> @llvm.riscv.th.vsll.mask.nxv8i8(
+  <vscale x 8 x i8>,
+  <vscale x 8 x i8>,
+  iXLen,
+  <vscale x 8 x i1>,
+  iXLen);
+
+define <vscale x 8 x i8> @intrinsic_vsll_mask_vx_nxv8i8_nxv8i8(<vscale x 8 x i8> %0, <vscale x 8 x i8> %1, iXLen %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsll_mask_vx_nxv8i8_nxv8i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e8, m1, d1
+; CHECK-NEXT:    th.vsll.vx v8, v9, a0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i8> @llvm.riscv.th.vsll.mask.nxv8i8(
+    <vscale x 8 x i8> %0,
+    <vscale x 8 x i8> %1,
+    iXLen %2,
+    <vscale x 8 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 8 x i8> %a
+}
+
+declare <vscale x 16 x i8> @llvm.riscv.th.vsll.nxv16i8(
+  <vscale x 16 x i8>,
+  <vscale x 16 x i8>,
+  iXLen,
+  iXLen);
+
+define <vscale x 16 x i8> @intrinsic_vsll_vx_nxv16i8_nxv16i8(<vscale x 16 x i8> %0, iXLen %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsll_vx_nxv16i8_nxv16i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e8, m2, d1
+; CHECK-NEXT:    th.vsll.vx v8, v8, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i8> @llvm.riscv.th.vsll.nxv16i8(
+    <vscale x 16 x i8> undef,
+    <vscale x 16 x i8> %0,
+    iXLen %1,
+    iXLen %2)
+
+  ret <vscale x 16 x i8> %a
+}
+
+declare <vscale x 16 x i8> @llvm.riscv.th.vsll.mask.nxv16i8(
+  <vscale x 16 x i8>,
+  <vscale x 16 x i8>,
+  iXLen,
+  <vscale x 16 x i1>,
+  iXLen);
+
+define <vscale x 16 x i8> @intrinsic_vsll_mask_vx_nxv16i8_nxv16i8(<vscale x 16 x i8> %0, <vscale x 16 x i8> %1, iXLen %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsll_mask_vx_nxv16i8_nxv16i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e8, m2, d1
+; CHECK-NEXT:    th.vsll.vx v8, v10, a0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i8> @llvm.riscv.th.vsll.mask.nxv16i8(
+    <vscale x 16 x i8> %0,
+    <vscale x 16 x i8> %1,
+    iXLen %2,
+    <vscale x 16 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 16 x i8> %a
+}
+
+declare <vscale x 32 x i8> @llvm.riscv.th.vsll.nxv32i8(
+  <vscale x 32 x i8>,
+  <vscale x 32 x i8>,
+  iXLen,
+  iXLen);
+
+define <vscale x 32 x i8> @intrinsic_vsll_vx_nxv32i8_nxv32i8(<vscale x 32 x i8> %0, iXLen %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsll_vx_nxv32i8_nxv32i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e8, m4, d1
+; CHECK-NEXT:    th.vsll.vx v8, v8, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x i8> @llvm.riscv.th.vsll.nxv32i8(
+    <vscale x 32 x i8> undef,
+    <vscale x 32 x i8> %0,
+    iXLen %1,
+    iXLen %2)
+
+  ret <vscale x 32 x i8> %a
+}
+
+declare <vscale x 32 x i8> @llvm.riscv.th.vsll.mask.nxv32i8(
+  <vscale x 32 x i8>,
+  <vscale x 32 x i8>,
+  iXLen,
+  <vscale x 32 x i1>,
+  iXLen);
+
+define <vscale x 32 x i8> @intrinsic_vsll_mask_vx_nxv32i8_nxv32i8(<vscale x 32 x i8> %0, <vscale x 32 x i8> %1, iXLen %2, <vscale x 32 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsll_mask_vx_nxv32i8_nxv32i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e8, m4, d1
+; CHECK-NEXT:    th.vsll.vx v8, v12, a0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x i8> @llvm.riscv.th.vsll.mask.nxv32i8(
+    <vscale x 32 x i8> %0,
+    <vscale x 32 x i8> %1,
+    iXLen %2,
+    <vscale x 32 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 32 x i8> %a
+}
+
+declare <vscale x 64 x i8> @llvm.riscv.th.vsll.nxv64i8(
+  <vscale x 64 x i8>,
+  <vscale x 64 x i8>,
+  iXLen,
+  iXLen);
+
+define <vscale x 64 x i8> @intrinsic_vsll_vx_nxv64i8_nxv64i8(<vscale x 64 x i8> %0, iXLen %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsll_vx_nxv64i8_nxv64i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e8, m8, d1
+; CHECK-NEXT:    th.vsll.vx v8, v8, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 64 x i8> @llvm.riscv.th.vsll.nxv64i8(
+    <vscale x 64 x i8> undef,
+    <vscale x 64 x i8> %0,
+    iXLen %1,
+    iXLen %2)
+
+  ret <vscale x 64 x i8> %a
+}
+
+declare <vscale x 64 x i8> @llvm.riscv.th.vsll.mask.nxv64i8(
+  <vscale x 64 x i8>,
+  <vscale x 64 x i8>,
+  iXLen,
+  <vscale x 64 x i1>,
+  iXLen);
+
+define <vscale x 64 x i8> @intrinsic_vsll_mask_vx_nxv64i8_nxv64i8(<vscale x 64 x i8> %0, <vscale x 64 x i8> %1, iXLen %2, <vscale x 64 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsll_mask_vx_nxv64i8_nxv64i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e8, m8, d1
+; CHECK-NEXT:    th.vsll.vx v8, v16, a0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 64 x i8> @llvm.riscv.th.vsll.mask.nxv64i8(
+    <vscale x 64 x i8> %0,
+    <vscale x 64 x i8> %1,
+    iXLen %2,
+    <vscale x 64 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 64 x i8> %a
+}
+
+declare <vscale x 4 x i16> @llvm.riscv.th.vsll.nxv4i16(
+  <vscale x 4 x i16>,
+  <vscale x 4 x i16>,
+  iXLen,
+  iXLen);
+
+define <vscale x 4 x i16> @intrinsic_vsll_vx_nxv4i16_nxv4i16(<vscale x 4 x i16> %0, iXLen %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsll_vx_nxv4i16_nxv4i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e16, m1, d1
+; CHECK-NEXT:    th.vsll.vx v8, v8, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i16> @llvm.riscv.th.vsll.nxv4i16(
+    <vscale x 4 x i16> undef,
+    <vscale x 4 x i16> %0,
+    iXLen %1,
+    iXLen %2)
+
+  ret <vscale x 4 x i16> %a
+}
+
+declare <vscale x 4 x i16> @llvm.riscv.th.vsll.mask.nxv4i16(
+  <vscale x 4 x i16>,
+  <vscale x 4 x i16>,
+  iXLen,
+  <vscale x 4 x i1>,
+  iXLen);
+
+define <vscale x 4 x i16> @intrinsic_vsll_mask_vx_nxv4i16_nxv4i16(<vscale x 4 x i16> %0, <vscale x 4 x i16> %1, iXLen %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsll_mask_vx_nxv4i16_nxv4i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e16, m1, d1
+; CHECK-NEXT:    th.vsll.vx v8, v9, a0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i16> @llvm.riscv.th.vsll.mask.nxv4i16(
+    <vscale x 4 x i16> %0,
+    <vscale x 4 x i16> %1,
+    iXLen %2,
+    <vscale x 4 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 4 x i16> %a
+}
+
+declare <vscale x 8 x i16> @llvm.riscv.th.vsll.nxv8i16(
+  <vscale x 8 x i16>,
+  <vscale x 8 x i16>,
+  iXLen,
+  iXLen);
+
+define <vscale x 8 x i16> @intrinsic_vsll_vx_nxv8i16_nxv8i16(<vscale x 8 x i16> %0, iXLen %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsll_vx_nxv8i16_nxv8i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e16, m2, d1
+; CHECK-NEXT:    th.vsll.vx v8, v8, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i16> @llvm.riscv.th.vsll.nxv8i16(
+    <vscale x 8 x i16> undef,
+    <vscale x 8 x i16> %0,
+    iXLen %1,
+    iXLen %2)
+
+  ret <vscale x 8 x i16> %a
+}
+
+declare <vscale x 8 x i16> @llvm.riscv.th.vsll.mask.nxv8i16(
+  <vscale x 8 x i16>,
+  <vscale x 8 x i16>,
+  iXLen,
+  <vscale x 8 x i1>,
+  iXLen);
+
+define <vscale x 8 x i16> @intrinsic_vsll_mask_vx_nxv8i16_nxv8i16(<vscale x 8 x i16> %0, <vscale x 8 x i16> %1, iXLen %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsll_mask_vx_nxv8i16_nxv8i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e16, m2, d1
+; CHECK-NEXT:    th.vsll.vx v8, v10, a0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i16> @llvm.riscv.th.vsll.mask.nxv8i16(
+    <vscale x 8 x i16> %0,
+    <vscale x 8 x i16> %1,
+    iXLen %2,
+    <vscale x 8 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 8 x i16> %a
+}
+
+declare <vscale x 16 x i16> @llvm.riscv.th.vsll.nxv16i16(
+  <vscale x 16 x i16>,
+  <vscale x 16 x i16>,
+  iXLen,
+  iXLen);
+
+define <vscale x 16 x i16> @intrinsic_vsll_vx_nxv16i16_nxv16i16(<vscale x 16 x i16> %0, iXLen %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsll_vx_nxv16i16_nxv16i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e16, m4, d1
+; CHECK-NEXT:    th.vsll.vx v8, v8, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i16> @llvm.riscv.th.vsll.nxv16i16(
+    <vscale x 16 x i16> undef,
+    <vscale x 16 x i16> %0,
+    iXLen %1,
+    iXLen %2)
+
+  ret <vscale x 16 x i16> %a
+}
+
+declare <vscale x 16 x i16> @llvm.riscv.th.vsll.mask.nxv16i16(
+  <vscale x 16 x i16>,
+  <vscale x 16 x i16>,
+  iXLen,
+  <vscale x 16 x i1>,
+  iXLen);
+
+define <vscale x 16 x i16> @intrinsic_vsll_mask_vx_nxv16i16_nxv16i16(<vscale x 16 x i16> %0, <vscale x 16 x i16> %1, iXLen %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsll_mask_vx_nxv16i16_nxv16i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e16, m4, d1
+; CHECK-NEXT:    th.vsll.vx v8, v12, a0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i16> @llvm.riscv.th.vsll.mask.nxv16i16(
+    <vscale x 16 x i16> %0,
+    <vscale x 16 x i16> %1,
+    iXLen %2,
+    <vscale x 16 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 16 x i16> %a
+}
+
+declare <vscale x 32 x i16> @llvm.riscv.th.vsll.nxv32i16(
+  <vscale x 32 x i16>,
+  <vscale x 32 x i16>,
+  iXLen,
+  iXLen);
+
+define <vscale x 32 x i16> @intrinsic_vsll_vx_nxv32i16_nxv32i16(<vscale x 32 x i16> %0, iXLen %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsll_vx_nxv32i16_nxv32i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e16, m8, d1
+; CHECK-NEXT:    th.vsll.vx v8, v8, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x i16> @llvm.riscv.th.vsll.nxv32i16(
+    <vscale x 32 x i16> undef,
+    <vscale x 32 x i16> %0,
+    iXLen %1,
+    iXLen %2)
+
+  ret <vscale x 32 x i16> %a
+}
+
+declare <vscale x 32 x i16> @llvm.riscv.th.vsll.mask.nxv32i16(
+  <vscale x 32 x i16>,
+  <vscale x 32 x i16>,
+  iXLen,
+  <vscale x 32 x i1>,
+  iXLen);
+
+define <vscale x 32 x i16> @intrinsic_vsll_mask_vx_nxv32i16_nxv32i16(<vscale x 32 x i16> %0, <vscale x 32 x i16> %1, iXLen %2, <vscale x 32 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsll_mask_vx_nxv32i16_nxv32i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e16, m8, d1
+; CHECK-NEXT:    th.vsll.vx v8, v16, a0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x i16> @llvm.riscv.th.vsll.mask.nxv32i16(
+    <vscale x 32 x i16> %0,
+    <vscale x 32 x i16> %1,
+    iXLen %2,
+    <vscale x 32 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 32 x i16> %a
+}
+
+declare <vscale x 2 x i32> @llvm.riscv.th.vsll.nxv2i32(
+  <vscale x 2 x i32>,
+  <vscale x 2 x i32>,
+  iXLen,
+  iXLen);
+
+define <vscale x 2 x i32> @intrinsic_vsll_vx_nxv2i32_nxv2i32(<vscale x 2 x i32> %0, iXLen %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsll_vx_nxv2i32_nxv2i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e32, m1, d1
+; CHECK-NEXT:    th.vsll.vx v8, v8, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i32> @llvm.riscv.th.vsll.nxv2i32(
+    <vscale x 2 x i32> undef,
+    <vscale x 2 x i32> %0,
+    iXLen %1,
+    iXLen %2)
+
+  ret <vscale x 2 x i32> %a
+}
+
+declare <vscale x 2 x i32> @llvm.riscv.th.vsll.mask.nxv2i32(
+  <vscale x 2 x i32>,
+  <vscale x 2 x i32>,
+  iXLen,
+  <vscale x 2 x i1>,
+  iXLen);
+
+define <vscale x 2 x i32> @intrinsic_vsll_mask_vx_nxv2i32_nxv2i32(<vscale x 2 x i32> %0, <vscale x 2 x i32> %1, iXLen %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsll_mask_vx_nxv2i32_nxv2i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e32, m1, d1
+; CHECK-NEXT:    th.vsll.vx v8, v9, a0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i32> @llvm.riscv.th.vsll.mask.nxv2i32(
+    <vscale x 2 x i32> %0,
+    <vscale x 2 x i32> %1,
+    iXLen %2,
+    <vscale x 2 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 2 x i32> %a
+}
+
+declare <vscale x 4 x i32> @llvm.riscv.th.vsll.nxv4i32(
+  <vscale x 4 x i32>,
+  <vscale x 4 x i32>,
+  iXLen,
+  iXLen);
+
+define <vscale x 4 x i32> @intrinsic_vsll_vx_nxv4i32_nxv4i32(<vscale x 4 x i32> %0, iXLen %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsll_vx_nxv4i32_nxv4i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e32, m2, d1
+; CHECK-NEXT:    th.vsll.vx v8, v8, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i32> @llvm.riscv.th.vsll.nxv4i32(
+    <vscale x 4 x i32> undef,
+    <vscale x 4 x i32> %0,
+    iXLen %1,
+    iXLen %2)
+
+  ret <vscale x 4 x i32> %a
+}
+
+declare <vscale x 4 x i32> @llvm.riscv.th.vsll.mask.nxv4i32(
+  <vscale x 4 x i32>,
+  <vscale x 4 x i32>,
+  iXLen,
+  <vscale x 4 x i1>,
+  iXLen);
+
+define <vscale x 4 x i32> @intrinsic_vsll_mask_vx_nxv4i32_nxv4i32(<vscale x 4 x i32> %0, <vscale x 4 x i32> %1, iXLen %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsll_mask_vx_nxv4i32_nxv4i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e32, m2, d1
+; CHECK-NEXT:    th.vsll.vx v8, v10, a0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i32> @llvm.riscv.th.vsll.mask.nxv4i32(
+    <vscale x 4 x i32> %0,
+    <vscale x 4 x i32> %1,
+    iXLen %2,
+    <vscale x 4 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 4 x i32> %a
+}
+
+declare <vscale x 8 x i32> @llvm.riscv.th.vsll.nxv8i32(
+  <vscale x 8 x i32>,
+  <vscale x 8 x i32>,
+  iXLen,
+  iXLen);
+
+define <vscale x 8 x i32> @intrinsic_vsll_vx_nxv8i32_nxv8i32(<vscale x 8 x i32> %0, iXLen %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsll_vx_nxv8i32_nxv8i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e32, m4, d1
+; CHECK-NEXT:    th.vsll.vx v8, v8, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i32> @llvm.riscv.th.vsll.nxv8i32(
+    <vscale x 8 x i32> undef,
+    <vscale x 8 x i32> %0,
+    iXLen %1,
+    iXLen %2)
+
+  ret <vscale x 8 x i32> %a
+}
+
+declare <vscale x 8 x i32> @llvm.riscv.th.vsll.mask.nxv8i32(
+  <vscale x 8 x i32>,
+  <vscale x 8 x i32>,
+  iXLen,
+  <vscale x 8 x i1>,
+  iXLen);
+
+define <vscale x 8 x i32> @intrinsic_vsll_mask_vx_nxv8i32_nxv8i32(<vscale x 8 x i32> %0, <vscale x 8 x i32> %1, iXLen %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsll_mask_vx_nxv8i32_nxv8i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e32, m4, d1
+; CHECK-NEXT:    th.vsll.vx v8, v12, a0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i32> @llvm.riscv.th.vsll.mask.nxv8i32(
+    <vscale x 8 x i32> %0,
+    <vscale x 8 x i32> %1,
+    iXLen %2,
+    <vscale x 8 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 8 x i32> %a
+}
+
+declare <vscale x 16 x i32> @llvm.riscv.th.vsll.nxv16i32(
+  <vscale x 16 x i32>,
+  <vscale x 16 x i32>,
+  iXLen,
+  iXLen);
+
+define <vscale x 16 x i32> @intrinsic_vsll_vx_nxv16i32_nxv16i32(<vscale x 16 x i32> %0, iXLen %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsll_vx_nxv16i32_nxv16i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e32, m8, d1
+; CHECK-NEXT:    th.vsll.vx v8, v8, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i32> @llvm.riscv.th.vsll.nxv16i32(
+    <vscale x 16 x i32> undef,
+    <vscale x 16 x i32> %0,
+    iXLen %1,
+    iXLen %2)
+
+  ret <vscale x 16 x i32> %a
+}
+
+declare <vscale x 16 x i32> @llvm.riscv.th.vsll.mask.nxv16i32(
+  <vscale x 16 x i32>,
+  <vscale x 16 x i32>,
+  iXLen,
+  <vscale x 16 x i1>,
+  iXLen);
+
+define <vscale x 16 x i32> @intrinsic_vsll_mask_vx_nxv16i32_nxv16i32(<vscale x 16 x i32> %0, <vscale x 16 x i32> %1, iXLen %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsll_mask_vx_nxv16i32_nxv16i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e32, m8, d1
+; CHECK-NEXT:    th.vsll.vx v8, v16, a0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i32> @llvm.riscv.th.vsll.mask.nxv16i32(
+    <vscale x 16 x i32> %0,
+    <vscale x 16 x i32> %1,
+    iXLen %2,
+    <vscale x 16 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 16 x i32> %a
+}
+
+declare <vscale x 1 x i64> @llvm.riscv.th.vsll.nxv1i64(
+  <vscale x 1 x i64>,
+  <vscale x 1 x i64>,
+  iXLen,
+  iXLen);
+
+define <vscale x 1 x i64> @intrinsic_vsll_vx_nxv1i64_nxv1i64(<vscale x 1 x i64> %0, iXLen %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsll_vx_nxv1i64_nxv1i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e64, m1, d1
+; CHECK-NEXT:    th.vsll.vx v8, v8, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x i64> @llvm.riscv.th.vsll.nxv1i64(
+    <vscale x 1 x i64> undef,
+    <vscale x 1 x i64> %0,
+    iXLen %1,
+    iXLen %2)
+
+  ret <vscale x 1 x i64> %a
+}
+
+declare <vscale x 1 x i64> @llvm.riscv.th.vsll.mask.nxv1i64(
+  <vscale x 1 x i64>,
+  <vscale x 1 x i64>,
+  iXLen,
+  <vscale x 1 x i1>,
+  iXLen);
+
+define <vscale x 1 x i64> @intrinsic_vsll_mask_vx_nxv1i64_nxv1i64(<vscale x 1 x i64> %0, <vscale x 1 x i64> %1, iXLen %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsll_mask_vx_nxv1i64_nxv1i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e64, m1, d1
+; CHECK-NEXT:    th.vsll.vx v8, v9, a0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x i64> @llvm.riscv.th.vsll.mask.nxv1i64(
+    <vscale x 1 x i64> %0,
+    <vscale x 1 x i64> %1,
+    iXLen %2,
+    <vscale x 1 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 1 x i64> %a
+}
+
+declare <vscale x 2 x i64> @llvm.riscv.th.vsll.nxv2i64(
+  <vscale x 2 x i64>,
+  <vscale x 2 x i64>,
+  iXLen,
+  iXLen);
+
+define <vscale x 2 x i64> @intrinsic_vsll_vx_nxv2i64_nxv2i64(<vscale x 2 x i64> %0, iXLen %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsll_vx_nxv2i64_nxv2i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e64, m2, d1
+; CHECK-NEXT:    th.vsll.vx v8, v8, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i64> @llvm.riscv.th.vsll.nxv2i64(
+    <vscale x 2 x i64> undef,
+    <vscale x 2 x i64> %0,
+    iXLen %1,
+    iXLen %2)
+
+  ret <vscale x 2 x i64> %a
+}
+
+declare <vscale x 2 x i64> @llvm.riscv.th.vsll.mask.nxv2i64(
+  <vscale x 2 x i64>,
+  <vscale x 2 x i64>,
+  iXLen,
+  <vscale x 2 x i1>,
+  iXLen);
+
+define <vscale x 2 x i64> @intrinsic_vsll_mask_vx_nxv2i64_nxv2i64(<vscale x 2 x i64> %0, <vscale x 2 x i64> %1, iXLen %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsll_mask_vx_nxv2i64_nxv2i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e64, m2, d1
+; CHECK-NEXT:    th.vsll.vx v8, v10, a0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i64> @llvm.riscv.th.vsll.mask.nxv2i64(
+    <vscale x 2 x i64> %0,
+    <vscale x 2 x i64> %1,
+    iXLen %2,
+    <vscale x 2 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 2 x i64> %a
+}
+
+declare <vscale x 4 x i64> @llvm.riscv.th.vsll.nxv4i64(
+  <vscale x 4 x i64>,
+  <vscale x 4 x i64>,
+  iXLen,
+  iXLen);
+
+define <vscale x 4 x i64> @intrinsic_vsll_vx_nxv4i64_nxv4i64(<vscale x 4 x i64> %0, iXLen %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsll_vx_nxv4i64_nxv4i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e64, m4, d1
+; CHECK-NEXT:    th.vsll.vx v8, v8, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i64> @llvm.riscv.th.vsll.nxv4i64(
+    <vscale x 4 x i64> undef,
+    <vscale x 4 x i64> %0,
+    iXLen %1,
+    iXLen %2)
+
+  ret <vscale x 4 x i64> %a
+}
+
+declare <vscale x 4 x i64> @llvm.riscv.th.vsll.mask.nxv4i64(
+  <vscale x 4 x i64>,
+  <vscale x 4 x i64>,
+  iXLen,
+  <vscale x 4 x i1>,
+  iXLen);
+
+define <vscale x 4 x i64> @intrinsic_vsll_mask_vx_nxv4i64_nxv4i64(<vscale x 4 x i64> %0, <vscale x 4 x i64> %1, iXLen %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsll_mask_vx_nxv4i64_nxv4i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e64, m4, d1
+; CHECK-NEXT:    th.vsll.vx v8, v12, a0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i64> @llvm.riscv.th.vsll.mask.nxv4i64(
+    <vscale x 4 x i64> %0,
+    <vscale x 4 x i64> %1,
+    iXLen %2,
+    <vscale x 4 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 4 x i64> %a
+}
+
+declare <vscale x 8 x i64> @llvm.riscv.th.vsll.nxv8i64(
+  <vscale x 8 x i64>,
+  <vscale x 8 x i64>,
+  iXLen,
+  iXLen);
+
+define <vscale x 8 x i64> @intrinsic_vsll_vx_nxv8i64_nxv8i64(<vscale x 8 x i64> %0, iXLen %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsll_vx_nxv8i64_nxv8i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e64, m8, d1
+; CHECK-NEXT:    th.vsll.vx v8, v8, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i64> @llvm.riscv.th.vsll.nxv8i64(
+    <vscale x 8 x i64> undef,
+    <vscale x 8 x i64> %0,
+    iXLen %1,
+    iXLen %2)
+
+  ret <vscale x 8 x i64> %a
+}
+
+declare <vscale x 8 x i64> @llvm.riscv.th.vsll.mask.nxv8i64(
+  <vscale x 8 x i64>,
+  <vscale x 8 x i64>,
+  iXLen,
+  <vscale x 8 x i1>,
+  iXLen);
+
+define <vscale x 8 x i64> @intrinsic_vsll_mask_vx_nxv8i64_nxv8i64(<vscale x 8 x i64> %0, <vscale x 8 x i64> %1, iXLen %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsll_mask_vx_nxv8i64_nxv8i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e64, m8, d1
+; CHECK-NEXT:    th.vsll.vx v8, v16, a0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i64> @llvm.riscv.th.vsll.mask.nxv8i64(
+    <vscale x 8 x i64> %0,
+    <vscale x 8 x i64> %1,
+    iXLen %2,
+    <vscale x 8 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 8 x i64> %a
+}
+
+define <vscale x 8 x i8> @intrinsic_vsll_vi_nxv8i8_nxv8i8_i8(<vscale x 8 x i8> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vsll_vi_nxv8i8_nxv8i8_i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e8, m1, d1
+; CHECK-NEXT:    th.vsll.vi v8, v8, 9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i8> @llvm.riscv.th.vsll.nxv8i8(
+    <vscale x 8 x i8> undef,
+    <vscale x 8 x i8> %0,
+    iXLen 9,
+    iXLen %1)
+
+  ret <vscale x 8 x i8> %a
+}
+
+define <vscale x 8 x i8> @intrinsic_vsll_mask_vi_nxv8i8_nxv8i8_i8(<vscale x 8 x i8> %0, <vscale x 8 x i8> %1, <vscale x 8 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vsll_mask_vi_nxv8i8_nxv8i8_i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e8, m1, d1
+; CHECK-NEXT:    th.vsll.vi v8, v9, 9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i8> @llvm.riscv.th.vsll.mask.nxv8i8(
+    <vscale x 8 x i8> %0,
+    <vscale x 8 x i8> %1,
+    iXLen 9,
+    <vscale x 8 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 8 x i8> %a
+}
+
+define <vscale x 8 x i8> @intrinsic_vsll_1_nxv8i8_nxv8i8_i8(<vscale x 8 x i8> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vsll_1_nxv8i8_nxv8i8_i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e8, m1, d1
+; CHECK-NEXT:    th.vadd.vv v8, v8, v8
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i8> @llvm.riscv.th.vsll.nxv8i8(
+    <vscale x 8 x i8> undef,
+    <vscale x 8 x i8> %0,
+    iXLen 1,
+    iXLen %1)
+
+  ret <vscale x 8 x i8> %a
+}
+
+define <vscale x 8 x i8> @intrinsic_vsll_mask_1_nxv8i8_nxv8i8_i8(<vscale x 8 x i8> %0, <vscale x 8 x i8> %1, <vscale x 8 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vsll_mask_1_nxv8i8_nxv8i8_i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e8, m1, d1
+; CHECK-NEXT:    th.vadd.vv v8, v9, v9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i8> @llvm.riscv.th.vsll.mask.nxv8i8(
+    <vscale x 8 x i8> %0,
+    <vscale x 8 x i8> %1,
+    iXLen 1,
+    <vscale x 8 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 8 x i8> %a
+}
+
+define <vscale x 16 x i8> @intrinsic_vsll_vi_nxv16i8_nxv16i8_i8(<vscale x 16 x i8> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vsll_vi_nxv16i8_nxv16i8_i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e8, m2, d1
+; CHECK-NEXT:    th.vsll.vi v8, v8, 9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i8> @llvm.riscv.th.vsll.nxv16i8(
+    <vscale x 16 x i8> undef,
+    <vscale x 16 x i8> %0,
+    iXLen 9,
+    iXLen %1)
+
+  ret <vscale x 16 x i8> %a
+}
+
+define <vscale x 16 x i8> @intrinsic_vsll_mask_vi_nxv16i8_nxv16i8_i8(<vscale x 16 x i8> %0, <vscale x 16 x i8> %1, <vscale x 16 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vsll_mask_vi_nxv16i8_nxv16i8_i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e8, m2, d1
+; CHECK-NEXT:    th.vsll.vi v8, v10, 9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i8> @llvm.riscv.th.vsll.mask.nxv16i8(
+    <vscale x 16 x i8> %0,
+    <vscale x 16 x i8> %1,
+    iXLen 9,
+    <vscale x 16 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 16 x i8> %a
+}
+
+define <vscale x 32 x i8> @intrinsic_vsll_vi_nxv32i8_nxv32i8_i8(<vscale x 32 x i8> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vsll_vi_nxv32i8_nxv32i8_i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e8, m4, d1
+; CHECK-NEXT:    th.vsll.vi v8, v8, 9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x i8> @llvm.riscv.th.vsll.nxv32i8(
+    <vscale x 32 x i8> undef,
+    <vscale x 32 x i8> %0,
+    iXLen 9,
+    iXLen %1)
+
+  ret <vscale x 32 x i8> %a
+}
+
+define <vscale x 32 x i8> @intrinsic_vsll_mask_vi_nxv32i8_nxv32i8_i8(<vscale x 32 x i8> %0, <vscale x 32 x i8> %1, <vscale x 32 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vsll_mask_vi_nxv32i8_nxv32i8_i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e8, m4, d1
+; CHECK-NEXT:    th.vsll.vi v8, v12, 9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x i8> @llvm.riscv.th.vsll.mask.nxv32i8(
+    <vscale x 32 x i8> %0,
+    <vscale x 32 x i8> %1,
+    iXLen 9,
+    <vscale x 32 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 32 x i8> %a
+}
+
+define <vscale x 64 x i8> @intrinsic_vsll_vi_nxv64i8_nxv64i8_i8(<vscale x 64 x i8> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vsll_vi_nxv64i8_nxv64i8_i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e8, m8, d1
+; CHECK-NEXT:    th.vsll.vi v8, v8, 9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 64 x i8> @llvm.riscv.th.vsll.nxv64i8(
+    <vscale x 64 x i8> undef,
+    <vscale x 64 x i8> %0,
+    iXLen 9,
+    iXLen %1)
+
+  ret <vscale x 64 x i8> %a
+}
+
+define <vscale x 64 x i8> @intrinsic_vsll_mask_vi_nxv64i8_nxv64i8_i8(<vscale x 64 x i8> %0, <vscale x 64 x i8> %1, <vscale x 64 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vsll_mask_vi_nxv64i8_nxv64i8_i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e8, m8, d1
+; CHECK-NEXT:    th.vsll.vi v8, v16, 9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 64 x i8> @llvm.riscv.th.vsll.mask.nxv64i8(
+    <vscale x 64 x i8> %0,
+    <vscale x 64 x i8> %1,
+    iXLen 9,
+    <vscale x 64 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 64 x i8> %a
+}
+
+define <vscale x 4 x i16> @intrinsic_vsll_vi_nxv4i16_nxv4i16_i16(<vscale x 4 x i16> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vsll_vi_nxv4i16_nxv4i16_i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e16, m1, d1
+; CHECK-NEXT:    th.vsll.vi v8, v8, 9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i16> @llvm.riscv.th.vsll.nxv4i16(
+    <vscale x 4 x i16> undef,
+    <vscale x 4 x i16> %0,
+    iXLen 9,
+    iXLen %1)
+
+  ret <vscale x 4 x i16> %a
+}
+
+define <vscale x 4 x i16> @intrinsic_vsll_mask_vi_nxv4i16_nxv4i16_i16(<vscale x 4 x i16> %0, <vscale x 4 x i16> %1, <vscale x 4 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vsll_mask_vi_nxv4i16_nxv4i16_i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e16, m1, d1
+; CHECK-NEXT:    th.vsll.vi v8, v9, 9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i16> @llvm.riscv.th.vsll.mask.nxv4i16(
+    <vscale x 4 x i16> %0,
+    <vscale x 4 x i16> %1,
+    iXLen 9,
+    <vscale x 4 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 4 x i16> %a
+}
+
+define <vscale x 8 x i16> @intrinsic_vsll_vi_nxv8i16_nxv8i16_i16(<vscale x 8 x i16> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vsll_vi_nxv8i16_nxv8i16_i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e16, m2, d1
+; CHECK-NEXT:    th.vsll.vi v8, v8, 9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i16> @llvm.riscv.th.vsll.nxv8i16(
+    <vscale x 8 x i16> undef,
+    <vscale x 8 x i16> %0,
+    iXLen 9,
+    iXLen %1)
+
+  ret <vscale x 8 x i16> %a
+}
+
+define <vscale x 8 x i16> @intrinsic_vsll_mask_vi_nxv8i16_nxv8i16_i16(<vscale x 8 x i16> %0, <vscale x 8 x i16> %1, <vscale x 8 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vsll_mask_vi_nxv8i16_nxv8i16_i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e16, m2, d1
+; CHECK-NEXT:    th.vsll.vi v8, v10, 9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i16> @llvm.riscv.th.vsll.mask.nxv8i16(
+    <vscale x 8 x i16> %0,
+    <vscale x 8 x i16> %1,
+    iXLen 9,
+    <vscale x 8 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 8 x i16> %a
+}
+
+define <vscale x 16 x i16> @intrinsic_vsll_vi_nxv16i16_nxv16i16_i16(<vscale x 16 x i16> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vsll_vi_nxv16i16_nxv16i16_i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e16, m4, d1
+; CHECK-NEXT:    th.vsll.vi v8, v8, 9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i16> @llvm.riscv.th.vsll.nxv16i16(
+    <vscale x 16 x i16> undef,
+    <vscale x 16 x i16> %0,
+    iXLen 9,
+    iXLen %1)
+
+  ret <vscale x 16 x i16> %a
+}
+
+define <vscale x 16 x i16> @intrinsic_vsll_mask_vi_nxv16i16_nxv16i16_i16(<vscale x 16 x i16> %0, <vscale x 16 x i16> %1, <vscale x 16 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vsll_mask_vi_nxv16i16_nxv16i16_i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e16, m4, d1
+; CHECK-NEXT:    th.vsll.vi v8, v12, 9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i16> @llvm.riscv.th.vsll.mask.nxv16i16(
+    <vscale x 16 x i16> %0,
+    <vscale x 16 x i16> %1,
+    iXLen 9,
+    <vscale x 16 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 16 x i16> %a
+}
+
+define <vscale x 32 x i16> @intrinsic_vsll_vi_nxv32i16_nxv32i16_i16(<vscale x 32 x i16> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vsll_vi_nxv32i16_nxv32i16_i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e16, m8, d1
+; CHECK-NEXT:    th.vsll.vi v8, v8, 9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x i16> @llvm.riscv.th.vsll.nxv32i16(
+    <vscale x 32 x i16> undef,
+    <vscale x 32 x i16> %0,
+    iXLen 9,
+    iXLen %1)
+
+  ret <vscale x 32 x i16> %a
+}
+
+define <vscale x 32 x i16> @intrinsic_vsll_mask_vi_nxv32i16_nxv32i16_i16(<vscale x 32 x i16> %0, <vscale x 32 x i16> %1, <vscale x 32 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vsll_mask_vi_nxv32i16_nxv32i16_i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e16, m8, d1
+; CHECK-NEXT:    th.vsll.vi v8, v16, 9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x i16> @llvm.riscv.th.vsll.mask.nxv32i16(
+    <vscale x 32 x i16> %0,
+    <vscale x 32 x i16> %1,
+    iXLen 9,
+    <vscale x 32 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 32 x i16> %a
+}
+
+define <vscale x 2 x i32> @intrinsic_vsll_vi_nxv2i32_nxv2i32_i32(<vscale x 2 x i32> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vsll_vi_nxv2i32_nxv2i32_i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e32, m1, d1
+; CHECK-NEXT:    th.vsll.vi v8, v8, 9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i32> @llvm.riscv.th.vsll.nxv2i32(
+    <vscale x 2 x i32> undef,
+    <vscale x 2 x i32> %0,
+    iXLen 9,
+    iXLen %1)
+
+  ret <vscale x 2 x i32> %a
+}
+
+define <vscale x 2 x i32> @intrinsic_vsll_mask_vi_nxv2i32_nxv2i32_i32(<vscale x 2 x i32> %0, <vscale x 2 x i32> %1, <vscale x 2 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vsll_mask_vi_nxv2i32_nxv2i32_i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e32, m1, d1
+; CHECK-NEXT:    th.vsll.vi v8, v9, 9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i32> @llvm.riscv.th.vsll.mask.nxv2i32(
+    <vscale x 2 x i32> %0,
+    <vscale x 2 x i32> %1,
+    iXLen 9,
+    <vscale x 2 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 2 x i32> %a
+}
+
+define <vscale x 4 x i32> @intrinsic_vsll_vi_nxv4i32_nxv4i32_i32(<vscale x 4 x i32> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vsll_vi_nxv4i32_nxv4i32_i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e32, m2, d1
+; CHECK-NEXT:    th.vsll.vi v8, v8, 9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i32> @llvm.riscv.th.vsll.nxv4i32(
+    <vscale x 4 x i32> undef,
+    <vscale x 4 x i32> %0,
+    iXLen 9,
+    iXLen %1)
+
+  ret <vscale x 4 x i32> %a
+}
+
+define <vscale x 4 x i32> @intrinsic_vsll_mask_vi_nxv4i32_nxv4i32_i32(<vscale x 4 x i32> %0, <vscale x 4 x i32> %1, <vscale x 4 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vsll_mask_vi_nxv4i32_nxv4i32_i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e32, m2, d1
+; CHECK-NEXT:    th.vsll.vi v8, v10, 9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i32> @llvm.riscv.th.vsll.mask.nxv4i32(
+    <vscale x 4 x i32> %0,
+    <vscale x 4 x i32> %1,
+    iXLen 9,
+    <vscale x 4 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 4 x i32> %a
+}
+
+define <vscale x 8 x i32> @intrinsic_vsll_vi_nxv8i32_nxv8i32_i32(<vscale x 8 x i32> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vsll_vi_nxv8i32_nxv8i32_i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e32, m4, d1
+; CHECK-NEXT:    th.vsll.vi v8, v8, 9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i32> @llvm.riscv.th.vsll.nxv8i32(
+    <vscale x 8 x i32> undef,
+    <vscale x 8 x i32> %0,
+    iXLen 9,
+    iXLen %1)
+
+  ret <vscale x 8 x i32> %a
+}
+
+define <vscale x 8 x i32> @intrinsic_vsll_mask_vi_nxv8i32_nxv8i32_i32(<vscale x 8 x i32> %0, <vscale x 8 x i32> %1, <vscale x 8 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vsll_mask_vi_nxv8i32_nxv8i32_i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e32, m4, d1
+; CHECK-NEXT:    th.vsll.vi v8, v12, 9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i32> @llvm.riscv.th.vsll.mask.nxv8i32(
+    <vscale x 8 x i32> %0,
+    <vscale x 8 x i32> %1,
+    iXLen 9,
+    <vscale x 8 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 8 x i32> %a
+}
+
+define <vscale x 16 x i32> @intrinsic_vsll_vi_nxv16i32_nxv16i32_i32(<vscale x 16 x i32> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vsll_vi_nxv16i32_nxv16i32_i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e32, m8, d1
+; CHECK-NEXT:    th.vsll.vi v8, v8, 9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i32> @llvm.riscv.th.vsll.nxv16i32(
+    <vscale x 16 x i32> undef,
+    <vscale x 16 x i32> %0,
+    iXLen 9,
+    iXLen %1)
+
+  ret <vscale x 16 x i32> %a
+}
+
+define <vscale x 16 x i32> @intrinsic_vsll_mask_vi_nxv16i32_nxv16i32_i32(<vscale x 16 x i32> %0, <vscale x 16 x i32> %1, <vscale x 16 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vsll_mask_vi_nxv16i32_nxv16i32_i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e32, m8, d1
+; CHECK-NEXT:    th.vsll.vi v8, v16, 9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i32> @llvm.riscv.th.vsll.mask.nxv16i32(
+    <vscale x 16 x i32> %0,
+    <vscale x 16 x i32> %1,
+    iXLen 9,
+    <vscale x 16 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 16 x i32> %a
+}
+
+define <vscale x 1 x i64> @intrinsic_vsll_vi_nxv1i64_nxv1i64_i64(<vscale x 1 x i64> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vsll_vi_nxv1i64_nxv1i64_i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e64, m1, d1
+; CHECK-NEXT:    th.vsll.vi v8, v8, 9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x i64> @llvm.riscv.th.vsll.nxv1i64(
+    <vscale x 1 x i64> undef,
+    <vscale x 1 x i64> %0,
+    iXLen 9,
+    iXLen %1)
+
+  ret <vscale x 1 x i64> %a
+}
+
+define <vscale x 1 x i64> @intrinsic_vsll_mask_vi_nxv1i64_nxv1i64_i64(<vscale x 1 x i64> %0, <vscale x 1 x i64> %1, <vscale x 1 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vsll_mask_vi_nxv1i64_nxv1i64_i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e64, m1, d1
+; CHECK-NEXT:    th.vsll.vi v8, v9, 9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x i64> @llvm.riscv.th.vsll.mask.nxv1i64(
+    <vscale x 1 x i64> %0,
+    <vscale x 1 x i64> %1,
+    iXLen 9,
+    <vscale x 1 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 1 x i64> %a
+}
+
+define <vscale x 2 x i64> @intrinsic_vsll_vi_nxv2i64_nxv2i64_i64(<vscale x 2 x i64> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vsll_vi_nxv2i64_nxv2i64_i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e64, m2, d1
+; CHECK-NEXT:    th.vsll.vi v8, v8, 9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i64> @llvm.riscv.th.vsll.nxv2i64(
+    <vscale x 2 x i64> undef,
+    <vscale x 2 x i64> %0,
+    iXLen 9,
+    iXLen %1)
+
+  ret <vscale x 2 x i64> %a
+}
+
+define <vscale x 2 x i64> @intrinsic_vsll_mask_vi_nxv2i64_nxv2i64_i64(<vscale x 2 x i64> %0, <vscale x 2 x i64> %1, <vscale x 2 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vsll_mask_vi_nxv2i64_nxv2i64_i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e64, m2, d1
+; CHECK-NEXT:    th.vsll.vi v8, v10, 9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i64> @llvm.riscv.th.vsll.mask.nxv2i64(
+    <vscale x 2 x i64> %0,
+    <vscale x 2 x i64> %1,
+    iXLen 9,
+    <vscale x 2 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 2 x i64> %a
+}
+
+define <vscale x 4 x i64> @intrinsic_vsll_vi_nxv4i64_nxv4i64_i64(<vscale x 4 x i64> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vsll_vi_nxv4i64_nxv4i64_i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e64, m4, d1
+; CHECK-NEXT:    th.vsll.vi v8, v8, 9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i64> @llvm.riscv.th.vsll.nxv4i64(
+    <vscale x 4 x i64> undef,
+    <vscale x 4 x i64> %0,
+    iXLen 9,
+    iXLen %1)
+
+  ret <vscale x 4 x i64> %a
+}
+
+define <vscale x 4 x i64> @intrinsic_vsll_mask_vi_nxv4i64_nxv4i64_i64(<vscale x 4 x i64> %0, <vscale x 4 x i64> %1, <vscale x 4 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vsll_mask_vi_nxv4i64_nxv4i64_i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e64, m4, d1
+; CHECK-NEXT:    th.vsll.vi v8, v12, 9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i64> @llvm.riscv.th.vsll.mask.nxv4i64(
+    <vscale x 4 x i64> %0,
+    <vscale x 4 x i64> %1,
+    iXLen 9,
+    <vscale x 4 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 4 x i64> %a
+}
+
+define <vscale x 8 x i64> @intrinsic_vsll_vi_nxv8i64_nxv8i64_i64(<vscale x 8 x i64> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vsll_vi_nxv8i64_nxv8i64_i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e64, m8, d1
+; CHECK-NEXT:    th.vsll.vi v8, v8, 9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i64> @llvm.riscv.th.vsll.nxv8i64(
+    <vscale x 8 x i64> undef,
+    <vscale x 8 x i64> %0,
+    iXLen 9,
+    iXLen %1)
+
+  ret <vscale x 8 x i64> %a
+}
+
+define <vscale x 8 x i64> @intrinsic_vsll_mask_vi_nxv8i64_nxv8i64_i64(<vscale x 8 x i64> %0, <vscale x 8 x i64> %1, <vscale x 8 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vsll_mask_vi_nxv8i64_nxv8i64_i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e64, m8, d1
+; CHECK-NEXT:    th.vsll.vi v8, v16, 9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i64> @llvm.riscv.th.vsll.mask.nxv8i64(
+    <vscale x 8 x i64> %0,
+    <vscale x 8 x i64> %1,
+    iXLen 9,
+    <vscale x 8 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 8 x i64> %a
+}
diff --git a/llvm/test/CodeGen/RISCV/rvv0p71/vsra.ll b/llvm/test/CodeGen/RISCV/rvv0p71/vsra.ll
new file mode 100644
index 00000000000000..55c446b95c5dcf
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv0p71/vsra.ll
@@ -0,0 +1,3097 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+xtheadvector \
+; RUN:   -verify-machineinstrs | FileCheck %s
+; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+xtheadvector \
+; RUN:   -verify-machineinstrs | FileCheck %s
+
+declare <vscale x 8 x i8> @llvm.riscv.th.vsra.nxv8i8.nxv8i8(
+  <vscale x 8 x i8>,
+  <vscale x 8 x i8>,
+  <vscale x 8 x i8>,
+  iXLen);
+
+define <vscale x 8 x i8> @intrinsic_vsra_vv_nxv8i8_nxv8i8_nxv8i8(<vscale x 8 x i8> %0, <vscale x 8 x i8> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsra_vv_nxv8i8_nxv8i8_nxv8i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e8, m1, d1
+; CHECK-NEXT:    th.vsra.vv v8, v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i8> @llvm.riscv.th.vsra.nxv8i8.nxv8i8(
+    <vscale x 8 x i8> undef,
+    <vscale x 8 x i8> %0,
+    <vscale x 8 x i8> %1,
+    iXLen %2)
+
+  ret <vscale x 8 x i8> %a
+}
+
+declare <vscale x 8 x i8> @llvm.riscv.th.vsra.mask.nxv8i8.nxv8i8(
+  <vscale x 8 x i8>,
+  <vscale x 8 x i8>,
+  <vscale x 8 x i8>,
+  <vscale x 8 x i1>,
+  iXLen);
+
+define <vscale x 8 x i8> @intrinsic_vsra_mask_vv_nxv8i8_nxv8i8_nxv8i8(<vscale x 8 x i8> %0, <vscale x 8 x i8> %1, <vscale x 8 x i8> %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsra_mask_vv_nxv8i8_nxv8i8_nxv8i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e8, m1, d1
+; CHECK-NEXT:    th.vsra.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i8> @llvm.riscv.th.vsra.mask.nxv8i8.nxv8i8(
+    <vscale x 8 x i8> %0,
+    <vscale x 8 x i8> %1,
+    <vscale x 8 x i8> %2,
+    <vscale x 8 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 8 x i8> %a
+}
+
+declare <vscale x 16 x i8> @llvm.riscv.th.vsra.nxv16i8.nxv16i8(
+  <vscale x 16 x i8>,
+  <vscale x 16 x i8>,
+  <vscale x 16 x i8>,
+  iXLen);
+
+define <vscale x 16 x i8> @intrinsic_vsra_vv_nxv16i8_nxv16i8_nxv16i8(<vscale x 16 x i8> %0, <vscale x 16 x i8> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsra_vv_nxv16i8_nxv16i8_nxv16i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e8, m2, d1
+; CHECK-NEXT:    th.vsra.vv v8, v8, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i8> @llvm.riscv.th.vsra.nxv16i8.nxv16i8(
+    <vscale x 16 x i8> undef,
+    <vscale x 16 x i8> %0,
+    <vscale x 16 x i8> %1,
+    iXLen %2)
+
+  ret <vscale x 16 x i8> %a
+}
+
+declare <vscale x 16 x i8> @llvm.riscv.th.vsra.mask.nxv16i8.nxv16i8(
+  <vscale x 16 x i8>,
+  <vscale x 16 x i8>,
+  <vscale x 16 x i8>,
+  <vscale x 16 x i1>,
+  iXLen);
+
+define <vscale x 16 x i8> @intrinsic_vsra_mask_vv_nxv16i8_nxv16i8_nxv16i8(<vscale x 16 x i8> %0, <vscale x 16 x i8> %1, <vscale x 16 x i8> %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsra_mask_vv_nxv16i8_nxv16i8_nxv16i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e8, m2, d1
+; CHECK-NEXT:    th.vsra.vv v8, v10, v12, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i8> @llvm.riscv.th.vsra.mask.nxv16i8.nxv16i8(
+    <vscale x 16 x i8> %0,
+    <vscale x 16 x i8> %1,
+    <vscale x 16 x i8> %2,
+    <vscale x 16 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 16 x i8> %a
+}
+
+declare <vscale x 32 x i8> @llvm.riscv.th.vsra.nxv32i8.nxv32i8(
+  <vscale x 32 x i8>,
+  <vscale x 32 x i8>,
+  <vscale x 32 x i8>,
+  iXLen);
+
+define <vscale x 32 x i8> @intrinsic_vsra_vv_nxv32i8_nxv32i8_nxv32i8(<vscale x 32 x i8> %0, <vscale x 32 x i8> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsra_vv_nxv32i8_nxv32i8_nxv32i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e8, m4, d1
+; CHECK-NEXT:    th.vsra.vv v8, v8, v12
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x i8> @llvm.riscv.th.vsra.nxv32i8.nxv32i8(
+    <vscale x 32 x i8> undef,
+    <vscale x 32 x i8> %0,
+    <vscale x 32 x i8> %1,
+    iXLen %2)
+
+  ret <vscale x 32 x i8> %a
+}
+
+declare <vscale x 32 x i8> @llvm.riscv.th.vsra.mask.nxv32i8.nxv32i8(
+  <vscale x 32 x i8>,
+  <vscale x 32 x i8>,
+  <vscale x 32 x i8>,
+  <vscale x 32 x i1>,
+  iXLen);
+
+define <vscale x 32 x i8> @intrinsic_vsra_mask_vv_nxv32i8_nxv32i8_nxv32i8(<vscale x 32 x i8> %0, <vscale x 32 x i8> %1, <vscale x 32 x i8> %2, <vscale x 32 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsra_mask_vv_nxv32i8_nxv32i8_nxv32i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e8, m4, d1
+; CHECK-NEXT:    th.vsra.vv v8, v12, v16, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x i8> @llvm.riscv.th.vsra.mask.nxv32i8.nxv32i8(
+    <vscale x 32 x i8> %0,
+    <vscale x 32 x i8> %1,
+    <vscale x 32 x i8> %2,
+    <vscale x 32 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 32 x i8> %a
+}
+
+declare <vscale x 64 x i8> @llvm.riscv.th.vsra.nxv64i8.nxv64i8(
+  <vscale x 64 x i8>,
+  <vscale x 64 x i8>,
+  <vscale x 64 x i8>,
+  iXLen);
+
+define <vscale x 64 x i8> @intrinsic_vsra_vv_nxv64i8_nxv64i8_nxv64i8(<vscale x 64 x i8> %0, <vscale x 64 x i8> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsra_vv_nxv64i8_nxv64i8_nxv64i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e8, m8, d1
+; CHECK-NEXT:    th.vsra.vv v8, v8, v16
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 64 x i8> @llvm.riscv.th.vsra.nxv64i8.nxv64i8(
+    <vscale x 64 x i8> undef,
+    <vscale x 64 x i8> %0,
+    <vscale x 64 x i8> %1,
+    iXLen %2)
+
+  ret <vscale x 64 x i8> %a
+}
+
+declare <vscale x 64 x i8> @llvm.riscv.th.vsra.mask.nxv64i8.nxv64i8(
+  <vscale x 64 x i8>,
+  <vscale x 64 x i8>,
+  <vscale x 64 x i8>,
+  <vscale x 64 x i1>,
+  iXLen);
+
+define <vscale x 64 x i8> @intrinsic_vsra_mask_vv_nxv64i8_nxv64i8_nxv64i8(<vscale x 64 x i8> %0, <vscale x 64 x i8> %1, <vscale x 64 x i8> %2, <vscale x 64 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsra_mask_vv_nxv64i8_nxv64i8_nxv64i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m8, d1
+; CHECK-NEXT:    th.vle.v v24, (a0)
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a0, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a0, a2
+; CHECK-NEXT:    th.vsetvli zero, a1, e8, m8, d1
+; CHECK-NEXT:    th.vsra.vv v8, v16, v24, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 64 x i8> @llvm.riscv.th.vsra.mask.nxv64i8.nxv64i8(
+    <vscale x 64 x i8> %0,
+    <vscale x 64 x i8> %1,
+    <vscale x 64 x i8> %2,
+    <vscale x 64 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 64 x i8> %a
+}
+
+declare <vscale x 4 x i16> @llvm.riscv.th.vsra.nxv4i16.nxv4i16(
+  <vscale x 4 x i16>,
+  <vscale x 4 x i16>,
+  <vscale x 4 x i16>,
+  iXLen);
+
+define <vscale x 4 x i16> @intrinsic_vsra_vv_nxv4i16_nxv4i16_nxv4i16(<vscale x 4 x i16> %0, <vscale x 4 x i16> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsra_vv_nxv4i16_nxv4i16_nxv4i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e16, m1, d1
+; CHECK-NEXT:    th.vsra.vv v8, v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i16> @llvm.riscv.th.vsra.nxv4i16.nxv4i16(
+    <vscale x 4 x i16> undef,
+    <vscale x 4 x i16> %0,
+    <vscale x 4 x i16> %1,
+    iXLen %2)
+
+  ret <vscale x 4 x i16> %a
+}
+
+declare <vscale x 4 x i16> @llvm.riscv.th.vsra.mask.nxv4i16.nxv4i16(
+  <vscale x 4 x i16>,
+  <vscale x 4 x i16>,
+  <vscale x 4 x i16>,
+  <vscale x 4 x i1>,
+  iXLen);
+
+define <vscale x 4 x i16> @intrinsic_vsra_mask_vv_nxv4i16_nxv4i16_nxv4i16(<vscale x 4 x i16> %0, <vscale x 4 x i16> %1, <vscale x 4 x i16> %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsra_mask_vv_nxv4i16_nxv4i16_nxv4i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e16, m1, d1
+; CHECK-NEXT:    th.vsra.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i16> @llvm.riscv.th.vsra.mask.nxv4i16.nxv4i16(
+    <vscale x 4 x i16> %0,
+    <vscale x 4 x i16> %1,
+    <vscale x 4 x i16> %2,
+    <vscale x 4 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 4 x i16> %a
+}
+
+declare <vscale x 8 x i16> @llvm.riscv.th.vsra.nxv8i16.nxv8i16(
+  <vscale x 8 x i16>,
+  <vscale x 8 x i16>,
+  <vscale x 8 x i16>,
+  iXLen);
+
+define <vscale x 8 x i16> @intrinsic_vsra_vv_nxv8i16_nxv8i16_nxv8i16(<vscale x 8 x i16> %0, <vscale x 8 x i16> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsra_vv_nxv8i16_nxv8i16_nxv8i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e16, m2, d1
+; CHECK-NEXT:    th.vsra.vv v8, v8, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i16> @llvm.riscv.th.vsra.nxv8i16.nxv8i16(
+    <vscale x 8 x i16> undef,
+    <vscale x 8 x i16> %0,
+    <vscale x 8 x i16> %1,
+    iXLen %2)
+
+  ret <vscale x 8 x i16> %a
+}
+
+declare <vscale x 8 x i16> @llvm.riscv.th.vsra.mask.nxv8i16.nxv8i16(
+  <vscale x 8 x i16>,
+  <vscale x 8 x i16>,
+  <vscale x 8 x i16>,
+  <vscale x 8 x i1>,
+  iXLen);
+
+define <vscale x 8 x i16> @intrinsic_vsra_mask_vv_nxv8i16_nxv8i16_nxv8i16(<vscale x 8 x i16> %0, <vscale x 8 x i16> %1, <vscale x 8 x i16> %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsra_mask_vv_nxv8i16_nxv8i16_nxv8i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e16, m2, d1
+; CHECK-NEXT:    th.vsra.vv v8, v10, v12, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i16> @llvm.riscv.th.vsra.mask.nxv8i16.nxv8i16(
+    <vscale x 8 x i16> %0,
+    <vscale x 8 x i16> %1,
+    <vscale x 8 x i16> %2,
+    <vscale x 8 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 8 x i16> %a
+}
+
+declare <vscale x 16 x i16> @llvm.riscv.th.vsra.nxv16i16.nxv16i16(
+  <vscale x 16 x i16>,
+  <vscale x 16 x i16>,
+  <vscale x 16 x i16>,
+  iXLen);
+
+define <vscale x 16 x i16> @intrinsic_vsra_vv_nxv16i16_nxv16i16_nxv16i16(<vscale x 16 x i16> %0, <vscale x 16 x i16> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsra_vv_nxv16i16_nxv16i16_nxv16i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e16, m4, d1
+; CHECK-NEXT:    th.vsra.vv v8, v8, v12
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i16> @llvm.riscv.th.vsra.nxv16i16.nxv16i16(
+    <vscale x 16 x i16> undef,
+    <vscale x 16 x i16> %0,
+    <vscale x 16 x i16> %1,
+    iXLen %2)
+
+  ret <vscale x 16 x i16> %a
+}
+
+declare <vscale x 16 x i16> @llvm.riscv.th.vsra.mask.nxv16i16.nxv16i16(
+  <vscale x 16 x i16>,
+  <vscale x 16 x i16>,
+  <vscale x 16 x i16>,
+  <vscale x 16 x i1>,
+  iXLen);
+
+define <vscale x 16 x i16> @intrinsic_vsra_mask_vv_nxv16i16_nxv16i16_nxv16i16(<vscale x 16 x i16> %0, <vscale x 16 x i16> %1, <vscale x 16 x i16> %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsra_mask_vv_nxv16i16_nxv16i16_nxv16i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e16, m4, d1
+; CHECK-NEXT:    th.vsra.vv v8, v12, v16, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i16> @llvm.riscv.th.vsra.mask.nxv16i16.nxv16i16(
+    <vscale x 16 x i16> %0,
+    <vscale x 16 x i16> %1,
+    <vscale x 16 x i16> %2,
+    <vscale x 16 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 16 x i16> %a
+}
+
+declare <vscale x 32 x i16> @llvm.riscv.th.vsra.nxv32i16.nxv32i16(
+  <vscale x 32 x i16>,
+  <vscale x 32 x i16>,
+  <vscale x 32 x i16>,
+  iXLen);
+
+define <vscale x 32 x i16> @intrinsic_vsra_vv_nxv32i16_nxv32i16_nxv32i16(<vscale x 32 x i16> %0, <vscale x 32 x i16> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsra_vv_nxv32i16_nxv32i16_nxv32i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e16, m8, d1
+; CHECK-NEXT:    th.vsra.vv v8, v8, v16
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x i16> @llvm.riscv.th.vsra.nxv32i16.nxv32i16(
+    <vscale x 32 x i16> undef,
+    <vscale x 32 x i16> %0,
+    <vscale x 32 x i16> %1,
+    iXLen %2)
+
+  ret <vscale x 32 x i16> %a
+}
+
+declare <vscale x 32 x i16> @llvm.riscv.th.vsra.mask.nxv32i16.nxv32i16(
+  <vscale x 32 x i16>,
+  <vscale x 32 x i16>,
+  <vscale x 32 x i16>,
+  <vscale x 32 x i1>,
+  iXLen);
+
+define <vscale x 32 x i16> @intrinsic_vsra_mask_vv_nxv32i16_nxv32i16_nxv32i16(<vscale x 32 x i16> %0, <vscale x 32 x i16> %1, <vscale x 32 x i16> %2, <vscale x 32 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsra_mask_vv_nxv32i16_nxv32i16_nxv32i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e16, m8, d1
+; CHECK-NEXT:    th.vle.v v24, (a0)
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a0, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a0, a2
+; CHECK-NEXT:    th.vsetvli zero, a1, e16, m8, d1
+; CHECK-NEXT:    th.vsra.vv v8, v16, v24, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x i16> @llvm.riscv.th.vsra.mask.nxv32i16.nxv32i16(
+    <vscale x 32 x i16> %0,
+    <vscale x 32 x i16> %1,
+    <vscale x 32 x i16> %2,
+    <vscale x 32 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 32 x i16> %a
+}
+
+declare <vscale x 2 x i32> @llvm.riscv.th.vsra.nxv2i32.nxv2i32(
+  <vscale x 2 x i32>,
+  <vscale x 2 x i32>,
+  <vscale x 2 x i32>,
+  iXLen);
+
+define <vscale x 2 x i32> @intrinsic_vsra_vv_nxv2i32_nxv2i32_nxv2i32(<vscale x 2 x i32> %0, <vscale x 2 x i32> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsra_vv_nxv2i32_nxv2i32_nxv2i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e32, m1, d1
+; CHECK-NEXT:    th.vsra.vv v8, v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i32> @llvm.riscv.th.vsra.nxv2i32.nxv2i32(
+    <vscale x 2 x i32> undef,
+    <vscale x 2 x i32> %0,
+    <vscale x 2 x i32> %1,
+    iXLen %2)
+
+  ret <vscale x 2 x i32> %a
+}
+
+declare <vscale x 2 x i32> @llvm.riscv.th.vsra.mask.nxv2i32.nxv2i32(
+  <vscale x 2 x i32>,
+  <vscale x 2 x i32>,
+  <vscale x 2 x i32>,
+  <vscale x 2 x i1>,
+  iXLen);
+
+define <vscale x 2 x i32> @intrinsic_vsra_mask_vv_nxv2i32_nxv2i32_nxv2i32(<vscale x 2 x i32> %0, <vscale x 2 x i32> %1, <vscale x 2 x i32> %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsra_mask_vv_nxv2i32_nxv2i32_nxv2i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e32, m1, d1
+; CHECK-NEXT:    th.vsra.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i32> @llvm.riscv.th.vsra.mask.nxv2i32.nxv2i32(
+    <vscale x 2 x i32> %0,
+    <vscale x 2 x i32> %1,
+    <vscale x 2 x i32> %2,
+    <vscale x 2 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 2 x i32> %a
+}
+
+declare <vscale x 4 x i32> @llvm.riscv.th.vsra.nxv4i32.nxv4i32(
+  <vscale x 4 x i32>,
+  <vscale x 4 x i32>,
+  <vscale x 4 x i32>,
+  iXLen);
+
+define <vscale x 4 x i32> @intrinsic_vsra_vv_nxv4i32_nxv4i32_nxv4i32(<vscale x 4 x i32> %0, <vscale x 4 x i32> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsra_vv_nxv4i32_nxv4i32_nxv4i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e32, m2, d1
+; CHECK-NEXT:    th.vsra.vv v8, v8, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i32> @llvm.riscv.th.vsra.nxv4i32.nxv4i32(
+    <vscale x 4 x i32> undef,
+    <vscale x 4 x i32> %0,
+    <vscale x 4 x i32> %1,
+    iXLen %2)
+
+  ret <vscale x 4 x i32> %a
+}
+
+declare <vscale x 4 x i32> @llvm.riscv.th.vsra.mask.nxv4i32.nxv4i32(
+  <vscale x 4 x i32>,
+  <vscale x 4 x i32>,
+  <vscale x 4 x i32>,
+  <vscale x 4 x i1>,
+  iXLen);
+
+define <vscale x 4 x i32> @intrinsic_vsra_mask_vv_nxv4i32_nxv4i32_nxv4i32(<vscale x 4 x i32> %0, <vscale x 4 x i32> %1, <vscale x 4 x i32> %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsra_mask_vv_nxv4i32_nxv4i32_nxv4i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e32, m2, d1
+; CHECK-NEXT:    th.vsra.vv v8, v10, v12, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i32> @llvm.riscv.th.vsra.mask.nxv4i32.nxv4i32(
+    <vscale x 4 x i32> %0,
+    <vscale x 4 x i32> %1,
+    <vscale x 4 x i32> %2,
+    <vscale x 4 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 4 x i32> %a
+}
+
+declare <vscale x 8 x i32> @llvm.riscv.th.vsra.nxv8i32.nxv8i32(
+  <vscale x 8 x i32>,
+  <vscale x 8 x i32>,
+  <vscale x 8 x i32>,
+  iXLen);
+
+define <vscale x 8 x i32> @intrinsic_vsra_vv_nxv8i32_nxv8i32_nxv8i32(<vscale x 8 x i32> %0, <vscale x 8 x i32> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsra_vv_nxv8i32_nxv8i32_nxv8i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e32, m4, d1
+; CHECK-NEXT:    th.vsra.vv v8, v8, v12
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i32> @llvm.riscv.th.vsra.nxv8i32.nxv8i32(
+    <vscale x 8 x i32> undef,
+    <vscale x 8 x i32> %0,
+    <vscale x 8 x i32> %1,
+    iXLen %2)
+
+  ret <vscale x 8 x i32> %a
+}
+
+declare <vscale x 8 x i32> @llvm.riscv.th.vsra.mask.nxv8i32.nxv8i32(
+  <vscale x 8 x i32>,
+  <vscale x 8 x i32>,
+  <vscale x 8 x i32>,
+  <vscale x 8 x i1>,
+  iXLen);
+
+define <vscale x 8 x i32> @intrinsic_vsra_mask_vv_nxv8i32_nxv8i32_nxv8i32(<vscale x 8 x i32> %0, <vscale x 8 x i32> %1, <vscale x 8 x i32> %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsra_mask_vv_nxv8i32_nxv8i32_nxv8i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e32, m4, d1
+; CHECK-NEXT:    th.vsra.vv v8, v12, v16, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i32> @llvm.riscv.th.vsra.mask.nxv8i32.nxv8i32(
+    <vscale x 8 x i32> %0,
+    <vscale x 8 x i32> %1,
+    <vscale x 8 x i32> %2,
+    <vscale x 8 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 8 x i32> %a
+}
+
+declare <vscale x 16 x i32> @llvm.riscv.th.vsra.nxv16i32.nxv16i32(
+  <vscale x 16 x i32>,
+  <vscale x 16 x i32>,
+  <vscale x 16 x i32>,
+  iXLen);
+
+define <vscale x 16 x i32> @intrinsic_vsra_vv_nxv16i32_nxv16i32_nxv16i32(<vscale x 16 x i32> %0, <vscale x 16 x i32> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsra_vv_nxv16i32_nxv16i32_nxv16i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e32, m8, d1
+; CHECK-NEXT:    th.vsra.vv v8, v8, v16
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i32> @llvm.riscv.th.vsra.nxv16i32.nxv16i32(
+    <vscale x 16 x i32> undef,
+    <vscale x 16 x i32> %0,
+    <vscale x 16 x i32> %1,
+    iXLen %2)
+
+  ret <vscale x 16 x i32> %a
+}
+
+declare <vscale x 16 x i32> @llvm.riscv.th.vsra.mask.nxv16i32.nxv16i32(
+  <vscale x 16 x i32>,
+  <vscale x 16 x i32>,
+  <vscale x 16 x i32>,
+  <vscale x 16 x i1>,
+  iXLen);
+
+define <vscale x 16 x i32> @intrinsic_vsra_mask_vv_nxv16i32_nxv16i32_nxv16i32(<vscale x 16 x i32> %0, <vscale x 16 x i32> %1, <vscale x 16 x i32> %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsra_mask_vv_nxv16i32_nxv16i32_nxv16i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e32, m8, d1
+; CHECK-NEXT:    th.vle.v v24, (a0)
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a0, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a0, a2
+; CHECK-NEXT:    th.vsetvli zero, a1, e32, m8, d1
+; CHECK-NEXT:    th.vsra.vv v8, v16, v24, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i32> @llvm.riscv.th.vsra.mask.nxv16i32.nxv16i32(
+    <vscale x 16 x i32> %0,
+    <vscale x 16 x i32> %1,
+    <vscale x 16 x i32> %2,
+    <vscale x 16 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 16 x i32> %a
+}
+
+declare <vscale x 1 x i64> @llvm.riscv.th.vsra.nxv1i64.nxv1i64(
+  <vscale x 1 x i64>,
+  <vscale x 1 x i64>,
+  <vscale x 1 x i64>,
+  iXLen);
+
+define <vscale x 1 x i64> @intrinsic_vsra_vv_nxv1i64_nxv1i64_nxv1i64(<vscale x 1 x i64> %0, <vscale x 1 x i64> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsra_vv_nxv1i64_nxv1i64_nxv1i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e64, m1, d1
+; CHECK-NEXT:    th.vsra.vv v8, v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x i64> @llvm.riscv.th.vsra.nxv1i64.nxv1i64(
+    <vscale x 1 x i64> undef,
+    <vscale x 1 x i64> %0,
+    <vscale x 1 x i64> %1,
+    iXLen %2)
+
+  ret <vscale x 1 x i64> %a
+}
+
+declare <vscale x 1 x i64> @llvm.riscv.th.vsra.mask.nxv1i64.nxv1i64(
+  <vscale x 1 x i64>,
+  <vscale x 1 x i64>,
+  <vscale x 1 x i64>,
+  <vscale x 1 x i1>,
+  iXLen);
+
+define <vscale x 1 x i64> @intrinsic_vsra_mask_vv_nxv1i64_nxv1i64_nxv1i64(<vscale x 1 x i64> %0, <vscale x 1 x i64> %1, <vscale x 1 x i64> %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsra_mask_vv_nxv1i64_nxv1i64_nxv1i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e64, m1, d1
+; CHECK-NEXT:    th.vsra.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x i64> @llvm.riscv.th.vsra.mask.nxv1i64.nxv1i64(
+    <vscale x 1 x i64> %0,
+    <vscale x 1 x i64> %1,
+    <vscale x 1 x i64> %2,
+    <vscale x 1 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 1 x i64> %a
+}
+
+declare <vscale x 2 x i64> @llvm.riscv.th.vsra.nxv2i64.nxv2i64(
+  <vscale x 2 x i64>,
+  <vscale x 2 x i64>,
+  <vscale x 2 x i64>,
+  iXLen);
+
+define <vscale x 2 x i64> @intrinsic_vsra_vv_nxv2i64_nxv2i64_nxv2i64(<vscale x 2 x i64> %0, <vscale x 2 x i64> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsra_vv_nxv2i64_nxv2i64_nxv2i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e64, m2, d1
+; CHECK-NEXT:    th.vsra.vv v8, v8, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i64> @llvm.riscv.th.vsra.nxv2i64.nxv2i64(
+    <vscale x 2 x i64> undef,
+    <vscale x 2 x i64> %0,
+    <vscale x 2 x i64> %1,
+    iXLen %2)
+
+  ret <vscale x 2 x i64> %a
+}
+
+declare <vscale x 2 x i64> @llvm.riscv.th.vsra.mask.nxv2i64.nxv2i64(
+  <vscale x 2 x i64>,
+  <vscale x 2 x i64>,
+  <vscale x 2 x i64>,
+  <vscale x 2 x i1>,
+  iXLen);
+
+define <vscale x 2 x i64> @intrinsic_vsra_mask_vv_nxv2i64_nxv2i64_nxv2i64(<vscale x 2 x i64> %0, <vscale x 2 x i64> %1, <vscale x 2 x i64> %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsra_mask_vv_nxv2i64_nxv2i64_nxv2i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e64, m2, d1
+; CHECK-NEXT:    th.vsra.vv v8, v10, v12, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i64> @llvm.riscv.th.vsra.mask.nxv2i64.nxv2i64(
+    <vscale x 2 x i64> %0,
+    <vscale x 2 x i64> %1,
+    <vscale x 2 x i64> %2,
+    <vscale x 2 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 2 x i64> %a
+}
+
+declare <vscale x 4 x i64> @llvm.riscv.th.vsra.nxv4i64.nxv4i64(
+  <vscale x 4 x i64>,
+  <vscale x 4 x i64>,
+  <vscale x 4 x i64>,
+  iXLen);
+
+define <vscale x 4 x i64> @intrinsic_vsra_vv_nxv4i64_nxv4i64_nxv4i64(<vscale x 4 x i64> %0, <vscale x 4 x i64> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsra_vv_nxv4i64_nxv4i64_nxv4i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e64, m4, d1
+; CHECK-NEXT:    th.vsra.vv v8, v8, v12
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i64> @llvm.riscv.th.vsra.nxv4i64.nxv4i64(
+    <vscale x 4 x i64> undef,
+    <vscale x 4 x i64> %0,
+    <vscale x 4 x i64> %1,
+    iXLen %2)
+
+  ret <vscale x 4 x i64> %a
+}
+
+declare <vscale x 4 x i64> @llvm.riscv.th.vsra.mask.nxv4i64.nxv4i64(
+  <vscale x 4 x i64>,
+  <vscale x 4 x i64>,
+  <vscale x 4 x i64>,
+  <vscale x 4 x i1>,
+  iXLen);
+
+define <vscale x 4 x i64> @intrinsic_vsra_mask_vv_nxv4i64_nxv4i64_nxv4i64(<vscale x 4 x i64> %0, <vscale x 4 x i64> %1, <vscale x 4 x i64> %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsra_mask_vv_nxv4i64_nxv4i64_nxv4i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e64, m4, d1
+; CHECK-NEXT:    th.vsra.vv v8, v12, v16, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i64> @llvm.riscv.th.vsra.mask.nxv4i64.nxv4i64(
+    <vscale x 4 x i64> %0,
+    <vscale x 4 x i64> %1,
+    <vscale x 4 x i64> %2,
+    <vscale x 4 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 4 x i64> %a
+}
+
+declare <vscale x 8 x i64> @llvm.riscv.th.vsra.nxv8i64.nxv8i64(
+  <vscale x 8 x i64>,
+  <vscale x 8 x i64>,
+  <vscale x 8 x i64>,
+  iXLen);
+
+define <vscale x 8 x i64> @intrinsic_vsra_vv_nxv8i64_nxv8i64_nxv8i64(<vscale x 8 x i64> %0, <vscale x 8 x i64> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsra_vv_nxv8i64_nxv8i64_nxv8i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e64, m8, d1
+; CHECK-NEXT:    th.vsra.vv v8, v8, v16
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i64> @llvm.riscv.th.vsra.nxv8i64.nxv8i64(
+    <vscale x 8 x i64> undef,
+    <vscale x 8 x i64> %0,
+    <vscale x 8 x i64> %1,
+    iXLen %2)
+
+  ret <vscale x 8 x i64> %a
+}
+
+declare <vscale x 8 x i64> @llvm.riscv.th.vsra.mask.nxv8i64.nxv8i64(
+  <vscale x 8 x i64>,
+  <vscale x 8 x i64>,
+  <vscale x 8 x i64>,
+  <vscale x 8 x i1>,
+  iXLen);
+
+define <vscale x 8 x i64> @intrinsic_vsra_mask_vv_nxv8i64_nxv8i64_nxv8i64(<vscale x 8 x i64> %0, <vscale x 8 x i64> %1, <vscale x 8 x i64> %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsra_mask_vv_nxv8i64_nxv8i64_nxv8i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e64, m8, d1
+; CHECK-NEXT:    th.vle.v v24, (a0)
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a0, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a0, a2
+; CHECK-NEXT:    th.vsetvli zero, a1, e64, m8, d1
+; CHECK-NEXT:    th.vsra.vv v8, v16, v24, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i64> @llvm.riscv.th.vsra.mask.nxv8i64.nxv8i64(
+    <vscale x 8 x i64> %0,
+    <vscale x 8 x i64> %1,
+    <vscale x 8 x i64> %2,
+    <vscale x 8 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 8 x i64> %a
+}
+
+declare <vscale x 8 x i8> @llvm.riscv.th.vsra.nxv8i8(
+  <vscale x 8 x i8>,
+  <vscale x 8 x i8>,
+  iXLen,
+  iXLen);
+
+define <vscale x 8 x i8> @intrinsic_vsra_vx_nxv8i8_nxv8i8(<vscale x 8 x i8> %0, iXLen %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsra_vx_nxv8i8_nxv8i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e8, m1, d1
+; CHECK-NEXT:    th.vsra.vx v8, v8, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i8> @llvm.riscv.th.vsra.nxv8i8(
+    <vscale x 8 x i8> undef,
+    <vscale x 8 x i8> %0,
+    iXLen %1,
+    iXLen %2)
+
+  ret <vscale x 8 x i8> %a
+}
+
+declare <vscale x 8 x i8> @llvm.riscv.th.vsra.mask.nxv8i8(
+  <vscale x 8 x i8>,
+  <vscale x 8 x i8>,
+  iXLen,
+  <vscale x 8 x i1>,
+  iXLen);
+
+define <vscale x 8 x i8> @intrinsic_vsra_mask_vx_nxv8i8_nxv8i8(<vscale x 8 x i8> %0, <vscale x 8 x i8> %1, iXLen %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsra_mask_vx_nxv8i8_nxv8i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e8, m1, d1
+; CHECK-NEXT:    th.vsra.vx v8, v9, a0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i8> @llvm.riscv.th.vsra.mask.nxv8i8(
+    <vscale x 8 x i8> %0,
+    <vscale x 8 x i8> %1,
+    iXLen %2,
+    <vscale x 8 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 8 x i8> %a
+}
+
+declare <vscale x 16 x i8> @llvm.riscv.th.vsra.nxv16i8(
+  <vscale x 16 x i8>,
+  <vscale x 16 x i8>,
+  iXLen,
+  iXLen);
+
+define <vscale x 16 x i8> @intrinsic_vsra_vx_nxv16i8_nxv16i8(<vscale x 16 x i8> %0, iXLen %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsra_vx_nxv16i8_nxv16i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e8, m2, d1
+; CHECK-NEXT:    th.vsra.vx v8, v8, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i8> @llvm.riscv.th.vsra.nxv16i8(
+    <vscale x 16 x i8> undef,
+    <vscale x 16 x i8> %0,
+    iXLen %1,
+    iXLen %2)
+
+  ret <vscale x 16 x i8> %a
+}
+
+declare <vscale x 16 x i8> @llvm.riscv.th.vsra.mask.nxv16i8(
+  <vscale x 16 x i8>,
+  <vscale x 16 x i8>,
+  iXLen,
+  <vscale x 16 x i1>,
+  iXLen);
+
+define <vscale x 16 x i8> @intrinsic_vsra_mask_vx_nxv16i8_nxv16i8(<vscale x 16 x i8> %0, <vscale x 16 x i8> %1, iXLen %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsra_mask_vx_nxv16i8_nxv16i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e8, m2, d1
+; CHECK-NEXT:    th.vsra.vx v8, v10, a0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i8> @llvm.riscv.th.vsra.mask.nxv16i8(
+    <vscale x 16 x i8> %0,
+    <vscale x 16 x i8> %1,
+    iXLen %2,
+    <vscale x 16 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 16 x i8> %a
+}
+
+declare <vscale x 32 x i8> @llvm.riscv.th.vsra.nxv32i8(
+  <vscale x 32 x i8>,
+  <vscale x 32 x i8>,
+  iXLen,
+  iXLen);
+
+define <vscale x 32 x i8> @intrinsic_vsra_vx_nxv32i8_nxv32i8(<vscale x 32 x i8> %0, iXLen %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsra_vx_nxv32i8_nxv32i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e8, m4, d1
+; CHECK-NEXT:    th.vsra.vx v8, v8, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x i8> @llvm.riscv.th.vsra.nxv32i8(
+    <vscale x 32 x i8> undef,
+    <vscale x 32 x i8> %0,
+    iXLen %1,
+    iXLen %2)
+
+  ret <vscale x 32 x i8> %a
+}
+
+declare <vscale x 32 x i8> @llvm.riscv.th.vsra.mask.nxv32i8(
+  <vscale x 32 x i8>,
+  <vscale x 32 x i8>,
+  iXLen,
+  <vscale x 32 x i1>,
+  iXLen);
+
+define <vscale x 32 x i8> @intrinsic_vsra_mask_vx_nxv32i8_nxv32i8(<vscale x 32 x i8> %0, <vscale x 32 x i8> %1, iXLen %2, <vscale x 32 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsra_mask_vx_nxv32i8_nxv32i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e8, m4, d1
+; CHECK-NEXT:    th.vsra.vx v8, v12, a0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x i8> @llvm.riscv.th.vsra.mask.nxv32i8(
+    <vscale x 32 x i8> %0,
+    <vscale x 32 x i8> %1,
+    iXLen %2,
+    <vscale x 32 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 32 x i8> %a
+}
+
+declare <vscale x 64 x i8> @llvm.riscv.th.vsra.nxv64i8(
+  <vscale x 64 x i8>,
+  <vscale x 64 x i8>,
+  iXLen,
+  iXLen);
+
+define <vscale x 64 x i8> @intrinsic_vsra_vx_nxv64i8_nxv64i8(<vscale x 64 x i8> %0, iXLen %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsra_vx_nxv64i8_nxv64i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e8, m8, d1
+; CHECK-NEXT:    th.vsra.vx v8, v8, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 64 x i8> @llvm.riscv.th.vsra.nxv64i8(
+    <vscale x 64 x i8> undef,
+    <vscale x 64 x i8> %0,
+    iXLen %1,
+    iXLen %2)
+
+  ret <vscale x 64 x i8> %a
+}
+
+declare <vscale x 64 x i8> @llvm.riscv.th.vsra.mask.nxv64i8(
+  <vscale x 64 x i8>,
+  <vscale x 64 x i8>,
+  iXLen,
+  <vscale x 64 x i1>,
+  iXLen);
+
+define <vscale x 64 x i8> @intrinsic_vsra_mask_vx_nxv64i8_nxv64i8(<vscale x 64 x i8> %0, <vscale x 64 x i8> %1, iXLen %2, <vscale x 64 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsra_mask_vx_nxv64i8_nxv64i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e8, m8, d1
+; CHECK-NEXT:    th.vsra.vx v8, v16, a0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 64 x i8> @llvm.riscv.th.vsra.mask.nxv64i8(
+    <vscale x 64 x i8> %0,
+    <vscale x 64 x i8> %1,
+    iXLen %2,
+    <vscale x 64 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 64 x i8> %a
+}
+
+declare <vscale x 4 x i16> @llvm.riscv.th.vsra.nxv4i16(
+  <vscale x 4 x i16>,
+  <vscale x 4 x i16>,
+  iXLen,
+  iXLen);
+
+define <vscale x 4 x i16> @intrinsic_vsra_vx_nxv4i16_nxv4i16(<vscale x 4 x i16> %0, iXLen %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsra_vx_nxv4i16_nxv4i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e16, m1, d1
+; CHECK-NEXT:    th.vsra.vx v8, v8, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i16> @llvm.riscv.th.vsra.nxv4i16(
+    <vscale x 4 x i16> undef,
+    <vscale x 4 x i16> %0,
+    iXLen %1,
+    iXLen %2)
+
+  ret <vscale x 4 x i16> %a
+}
+
+declare <vscale x 4 x i16> @llvm.riscv.th.vsra.mask.nxv4i16(
+  <vscale x 4 x i16>,
+  <vscale x 4 x i16>,
+  iXLen,
+  <vscale x 4 x i1>,
+  iXLen);
+
+define <vscale x 4 x i16> @intrinsic_vsra_mask_vx_nxv4i16_nxv4i16(<vscale x 4 x i16> %0, <vscale x 4 x i16> %1, iXLen %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsra_mask_vx_nxv4i16_nxv4i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e16, m1, d1
+; CHECK-NEXT:    th.vsra.vx v8, v9, a0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i16> @llvm.riscv.th.vsra.mask.nxv4i16(
+    <vscale x 4 x i16> %0,
+    <vscale x 4 x i16> %1,
+    iXLen %2,
+    <vscale x 4 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 4 x i16> %a
+}
+
+declare <vscale x 8 x i16> @llvm.riscv.th.vsra.nxv8i16(
+  <vscale x 8 x i16>,
+  <vscale x 8 x i16>,
+  iXLen,
+  iXLen);
+
+define <vscale x 8 x i16> @intrinsic_vsra_vx_nxv8i16_nxv8i16(<vscale x 8 x i16> %0, iXLen %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsra_vx_nxv8i16_nxv8i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e16, m2, d1
+; CHECK-NEXT:    th.vsra.vx v8, v8, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i16> @llvm.riscv.th.vsra.nxv8i16(
+    <vscale x 8 x i16> undef,
+    <vscale x 8 x i16> %0,
+    iXLen %1,
+    iXLen %2)
+
+  ret <vscale x 8 x i16> %a
+}
+
+declare <vscale x 8 x i16> @llvm.riscv.th.vsra.mask.nxv8i16(
+  <vscale x 8 x i16>,
+  <vscale x 8 x i16>,
+  iXLen,
+  <vscale x 8 x i1>,
+  iXLen);
+
+define <vscale x 8 x i16> @intrinsic_vsra_mask_vx_nxv8i16_nxv8i16(<vscale x 8 x i16> %0, <vscale x 8 x i16> %1, iXLen %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsra_mask_vx_nxv8i16_nxv8i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e16, m2, d1
+; CHECK-NEXT:    th.vsra.vx v8, v10, a0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i16> @llvm.riscv.th.vsra.mask.nxv8i16(
+    <vscale x 8 x i16> %0,
+    <vscale x 8 x i16> %1,
+    iXLen %2,
+    <vscale x 8 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 8 x i16> %a
+}
+
+declare <vscale x 16 x i16> @llvm.riscv.th.vsra.nxv16i16(
+  <vscale x 16 x i16>,
+  <vscale x 16 x i16>,
+  iXLen,
+  iXLen);
+
+define <vscale x 16 x i16> @intrinsic_vsra_vx_nxv16i16_nxv16i16(<vscale x 16 x i16> %0, iXLen %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsra_vx_nxv16i16_nxv16i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e16, m4, d1
+; CHECK-NEXT:    th.vsra.vx v8, v8, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i16> @llvm.riscv.th.vsra.nxv16i16(
+    <vscale x 16 x i16> undef,
+    <vscale x 16 x i16> %0,
+    iXLen %1,
+    iXLen %2)
+
+  ret <vscale x 16 x i16> %a
+}
+
+declare <vscale x 16 x i16> @llvm.riscv.th.vsra.mask.nxv16i16(
+  <vscale x 16 x i16>,
+  <vscale x 16 x i16>,
+  iXLen,
+  <vscale x 16 x i1>,
+  iXLen);
+
+define <vscale x 16 x i16> @intrinsic_vsra_mask_vx_nxv16i16_nxv16i16(<vscale x 16 x i16> %0, <vscale x 16 x i16> %1, iXLen %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsra_mask_vx_nxv16i16_nxv16i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e16, m4, d1
+; CHECK-NEXT:    th.vsra.vx v8, v12, a0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i16> @llvm.riscv.th.vsra.mask.nxv16i16(
+    <vscale x 16 x i16> %0,
+    <vscale x 16 x i16> %1,
+    iXLen %2,
+    <vscale x 16 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 16 x i16> %a
+}
+
+declare <vscale x 32 x i16> @llvm.riscv.th.vsra.nxv32i16(
+  <vscale x 32 x i16>,
+  <vscale x 32 x i16>,
+  iXLen,
+  iXLen);
+
+define <vscale x 32 x i16> @intrinsic_vsra_vx_nxv32i16_nxv32i16(<vscale x 32 x i16> %0, iXLen %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsra_vx_nxv32i16_nxv32i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e16, m8, d1
+; CHECK-NEXT:    th.vsra.vx v8, v8, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x i16> @llvm.riscv.th.vsra.nxv32i16(
+    <vscale x 32 x i16> undef,
+    <vscale x 32 x i16> %0,
+    iXLen %1,
+    iXLen %2)
+
+  ret <vscale x 32 x i16> %a
+}
+
+declare <vscale x 32 x i16> @llvm.riscv.th.vsra.mask.nxv32i16(
+  <vscale x 32 x i16>,
+  <vscale x 32 x i16>,
+  iXLen,
+  <vscale x 32 x i1>,
+  iXLen);
+
+define <vscale x 32 x i16> @intrinsic_vsra_mask_vx_nxv32i16_nxv32i16(<vscale x 32 x i16> %0, <vscale x 32 x i16> %1, iXLen %2, <vscale x 32 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsra_mask_vx_nxv32i16_nxv32i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e16, m8, d1
+; CHECK-NEXT:    th.vsra.vx v8, v16, a0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x i16> @llvm.riscv.th.vsra.mask.nxv32i16(
+    <vscale x 32 x i16> %0,
+    <vscale x 32 x i16> %1,
+    iXLen %2,
+    <vscale x 32 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 32 x i16> %a
+}
+
+declare <vscale x 2 x i32> @llvm.riscv.th.vsra.nxv2i32(
+  <vscale x 2 x i32>,
+  <vscale x 2 x i32>,
+  iXLen,
+  iXLen);
+
+define <vscale x 2 x i32> @intrinsic_vsra_vx_nxv2i32_nxv2i32(<vscale x 2 x i32> %0, iXLen %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsra_vx_nxv2i32_nxv2i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e32, m1, d1
+; CHECK-NEXT:    th.vsra.vx v8, v8, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i32> @llvm.riscv.th.vsra.nxv2i32(
+    <vscale x 2 x i32> undef,
+    <vscale x 2 x i32> %0,
+    iXLen %1,
+    iXLen %2)
+
+  ret <vscale x 2 x i32> %a
+}
+
+declare <vscale x 2 x i32> @llvm.riscv.th.vsra.mask.nxv2i32(
+  <vscale x 2 x i32>,
+  <vscale x 2 x i32>,
+  iXLen,
+  <vscale x 2 x i1>,
+  iXLen);
+
+define <vscale x 2 x i32> @intrinsic_vsra_mask_vx_nxv2i32_nxv2i32(<vscale x 2 x i32> %0, <vscale x 2 x i32> %1, iXLen %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsra_mask_vx_nxv2i32_nxv2i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e32, m1, d1
+; CHECK-NEXT:    th.vsra.vx v8, v9, a0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i32> @llvm.riscv.th.vsra.mask.nxv2i32(
+    <vscale x 2 x i32> %0,
+    <vscale x 2 x i32> %1,
+    iXLen %2,
+    <vscale x 2 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 2 x i32> %a
+}
+
+declare <vscale x 4 x i32> @llvm.riscv.th.vsra.nxv4i32(
+  <vscale x 4 x i32>,
+  <vscale x 4 x i32>,
+  iXLen,
+  iXLen);
+
+define <vscale x 4 x i32> @intrinsic_vsra_vx_nxv4i32_nxv4i32(<vscale x 4 x i32> %0, iXLen %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsra_vx_nxv4i32_nxv4i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e32, m2, d1
+; CHECK-NEXT:    th.vsra.vx v8, v8, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i32> @llvm.riscv.th.vsra.nxv4i32(
+    <vscale x 4 x i32> undef,
+    <vscale x 4 x i32> %0,
+    iXLen %1,
+    iXLen %2)
+
+  ret <vscale x 4 x i32> %a
+}
+
+declare <vscale x 4 x i32> @llvm.riscv.th.vsra.mask.nxv4i32(
+  <vscale x 4 x i32>,
+  <vscale x 4 x i32>,
+  iXLen,
+  <vscale x 4 x i1>,
+  iXLen);
+
+define <vscale x 4 x i32> @intrinsic_vsra_mask_vx_nxv4i32_nxv4i32(<vscale x 4 x i32> %0, <vscale x 4 x i32> %1, iXLen %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsra_mask_vx_nxv4i32_nxv4i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e32, m2, d1
+; CHECK-NEXT:    th.vsra.vx v8, v10, a0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i32> @llvm.riscv.th.vsra.mask.nxv4i32(
+    <vscale x 4 x i32> %0,
+    <vscale x 4 x i32> %1,
+    iXLen %2,
+    <vscale x 4 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 4 x i32> %a
+}
+
+declare <vscale x 8 x i32> @llvm.riscv.th.vsra.nxv8i32(
+  <vscale x 8 x i32>,
+  <vscale x 8 x i32>,
+  iXLen,
+  iXLen);
+
+define <vscale x 8 x i32> @intrinsic_vsra_vx_nxv8i32_nxv8i32(<vscale x 8 x i32> %0, iXLen %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsra_vx_nxv8i32_nxv8i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e32, m4, d1
+; CHECK-NEXT:    th.vsra.vx v8, v8, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i32> @llvm.riscv.th.vsra.nxv8i32(
+    <vscale x 8 x i32> undef,
+    <vscale x 8 x i32> %0,
+    iXLen %1,
+    iXLen %2)
+
+  ret <vscale x 8 x i32> %a
+}
+
+declare <vscale x 8 x i32> @llvm.riscv.th.vsra.mask.nxv8i32(
+  <vscale x 8 x i32>,
+  <vscale x 8 x i32>,
+  iXLen,
+  <vscale x 8 x i1>,
+  iXLen);
+
+define <vscale x 8 x i32> @intrinsic_vsra_mask_vx_nxv8i32_nxv8i32(<vscale x 8 x i32> %0, <vscale x 8 x i32> %1, iXLen %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsra_mask_vx_nxv8i32_nxv8i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e32, m4, d1
+; CHECK-NEXT:    th.vsra.vx v8, v12, a0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i32> @llvm.riscv.th.vsra.mask.nxv8i32(
+    <vscale x 8 x i32> %0,
+    <vscale x 8 x i32> %1,
+    iXLen %2,
+    <vscale x 8 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 8 x i32> %a
+}
+
+declare <vscale x 16 x i32> @llvm.riscv.th.vsra.nxv16i32(
+  <vscale x 16 x i32>,
+  <vscale x 16 x i32>,
+  iXLen,
+  iXLen);
+
+define <vscale x 16 x i32> @intrinsic_vsra_vx_nxv16i32_nxv16i32(<vscale x 16 x i32> %0, iXLen %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsra_vx_nxv16i32_nxv16i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e32, m8, d1
+; CHECK-NEXT:    th.vsra.vx v8, v8, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i32> @llvm.riscv.th.vsra.nxv16i32(
+    <vscale x 16 x i32> undef,
+    <vscale x 16 x i32> %0,
+    iXLen %1,
+    iXLen %2)
+
+  ret <vscale x 16 x i32> %a
+}
+
+declare <vscale x 16 x i32> @llvm.riscv.th.vsra.mask.nxv16i32(
+  <vscale x 16 x i32>,
+  <vscale x 16 x i32>,
+  iXLen,
+  <vscale x 16 x i1>,
+  iXLen);
+
+define <vscale x 16 x i32> @intrinsic_vsra_mask_vx_nxv16i32_nxv16i32(<vscale x 16 x i32> %0, <vscale x 16 x i32> %1, iXLen %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsra_mask_vx_nxv16i32_nxv16i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e32, m8, d1
+; CHECK-NEXT:    th.vsra.vx v8, v16, a0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i32> @llvm.riscv.th.vsra.mask.nxv16i32(
+    <vscale x 16 x i32> %0,
+    <vscale x 16 x i32> %1,
+    iXLen %2,
+    <vscale x 16 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 16 x i32> %a
+}
+
+declare <vscale x 1 x i64> @llvm.riscv.th.vsra.nxv1i64(
+  <vscale x 1 x i64>,
+  <vscale x 1 x i64>,
+  iXLen,
+  iXLen);
+
+define <vscale x 1 x i64> @intrinsic_vsra_vx_nxv1i64_nxv1i64(<vscale x 1 x i64> %0, iXLen %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsra_vx_nxv1i64_nxv1i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e64, m1, d1
+; CHECK-NEXT:    th.vsra.vx v8, v8, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x i64> @llvm.riscv.th.vsra.nxv1i64(
+    <vscale x 1 x i64> undef,
+    <vscale x 1 x i64> %0,
+    iXLen %1,
+    iXLen %2)
+
+  ret <vscale x 1 x i64> %a
+}
+
+declare <vscale x 1 x i64> @llvm.riscv.th.vsra.mask.nxv1i64(
+  <vscale x 1 x i64>,
+  <vscale x 1 x i64>,
+  iXLen,
+  <vscale x 1 x i1>,
+  iXLen);
+
+define <vscale x 1 x i64> @intrinsic_vsra_mask_vx_nxv1i64_nxv1i64(<vscale x 1 x i64> %0, <vscale x 1 x i64> %1, iXLen %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsra_mask_vx_nxv1i64_nxv1i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e64, m1, d1
+; CHECK-NEXT:    th.vsra.vx v8, v9, a0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x i64> @llvm.riscv.th.vsra.mask.nxv1i64(
+    <vscale x 1 x i64> %0,
+    <vscale x 1 x i64> %1,
+    iXLen %2,
+    <vscale x 1 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 1 x i64> %a
+}
+
+declare <vscale x 2 x i64> @llvm.riscv.th.vsra.nxv2i64(
+  <vscale x 2 x i64>,
+  <vscale x 2 x i64>,
+  iXLen,
+  iXLen);
+
+define <vscale x 2 x i64> @intrinsic_vsra_vx_nxv2i64_nxv2i64(<vscale x 2 x i64> %0, iXLen %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsra_vx_nxv2i64_nxv2i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e64, m2, d1
+; CHECK-NEXT:    th.vsra.vx v8, v8, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i64> @llvm.riscv.th.vsra.nxv2i64(
+    <vscale x 2 x i64> undef,
+    <vscale x 2 x i64> %0,
+    iXLen %1,
+    iXLen %2)
+
+  ret <vscale x 2 x i64> %a
+}
+
+declare <vscale x 2 x i64> @llvm.riscv.th.vsra.mask.nxv2i64(
+  <vscale x 2 x i64>,
+  <vscale x 2 x i64>,
+  iXLen,
+  <vscale x 2 x i1>,
+  iXLen);
+
+define <vscale x 2 x i64> @intrinsic_vsra_mask_vx_nxv2i64_nxv2i64(<vscale x 2 x i64> %0, <vscale x 2 x i64> %1, iXLen %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsra_mask_vx_nxv2i64_nxv2i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e64, m2, d1
+; CHECK-NEXT:    th.vsra.vx v8, v10, a0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i64> @llvm.riscv.th.vsra.mask.nxv2i64(
+    <vscale x 2 x i64> %0,
+    <vscale x 2 x i64> %1,
+    iXLen %2,
+    <vscale x 2 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 2 x i64> %a
+}
+
+declare <vscale x 4 x i64> @llvm.riscv.th.vsra.nxv4i64(
+  <vscale x 4 x i64>,
+  <vscale x 4 x i64>,
+  iXLen,
+  iXLen);
+
+define <vscale x 4 x i64> @intrinsic_vsra_vx_nxv4i64_nxv4i64(<vscale x 4 x i64> %0, iXLen %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsra_vx_nxv4i64_nxv4i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e64, m4, d1
+; CHECK-NEXT:    th.vsra.vx v8, v8, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i64> @llvm.riscv.th.vsra.nxv4i64(
+    <vscale x 4 x i64> undef,
+    <vscale x 4 x i64> %0,
+    iXLen %1,
+    iXLen %2)
+
+  ret <vscale x 4 x i64> %a
+}
+
+declare <vscale x 4 x i64> @llvm.riscv.th.vsra.mask.nxv4i64(
+  <vscale x 4 x i64>,
+  <vscale x 4 x i64>,
+  iXLen,
+  <vscale x 4 x i1>,
+  iXLen);
+
+define <vscale x 4 x i64> @intrinsic_vsra_mask_vx_nxv4i64_nxv4i64(<vscale x 4 x i64> %0, <vscale x 4 x i64> %1, iXLen %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsra_mask_vx_nxv4i64_nxv4i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e64, m4, d1
+; CHECK-NEXT:    th.vsra.vx v8, v12, a0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i64> @llvm.riscv.th.vsra.mask.nxv4i64(
+    <vscale x 4 x i64> %0,
+    <vscale x 4 x i64> %1,
+    iXLen %2,
+    <vscale x 4 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 4 x i64> %a
+}
+
+declare <vscale x 8 x i64> @llvm.riscv.th.vsra.nxv8i64(
+  <vscale x 8 x i64>,
+  <vscale x 8 x i64>,
+  iXLen,
+  iXLen);
+
+define <vscale x 8 x i64> @intrinsic_vsra_vx_nxv8i64_nxv8i64(<vscale x 8 x i64> %0, iXLen %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsra_vx_nxv8i64_nxv8i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e64, m8, d1
+; CHECK-NEXT:    th.vsra.vx v8, v8, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i64> @llvm.riscv.th.vsra.nxv8i64(
+    <vscale x 8 x i64> undef,
+    <vscale x 8 x i64> %0,
+    iXLen %1,
+    iXLen %2)
+
+  ret <vscale x 8 x i64> %a
+}
+
+declare <vscale x 8 x i64> @llvm.riscv.th.vsra.mask.nxv8i64(
+  <vscale x 8 x i64>,
+  <vscale x 8 x i64>,
+  iXLen,
+  <vscale x 8 x i1>,
+  iXLen);
+
+define <vscale x 8 x i64> @intrinsic_vsra_mask_vx_nxv8i64_nxv8i64(<vscale x 8 x i64> %0, <vscale x 8 x i64> %1, iXLen %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsra_mask_vx_nxv8i64_nxv8i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e64, m8, d1
+; CHECK-NEXT:    th.vsra.vx v8, v16, a0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i64> @llvm.riscv.th.vsra.mask.nxv8i64(
+    <vscale x 8 x i64> %0,
+    <vscale x 8 x i64> %1,
+    iXLen %2,
+    <vscale x 8 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 8 x i64> %a
+}
+
+define <vscale x 8 x i8> @intrinsic_vsra_vi_nxv8i8_nxv8i8_i8(<vscale x 8 x i8> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vsra_vi_nxv8i8_nxv8i8_i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e8, m1, d1
+; CHECK-NEXT:    th.vsra.vi v8, v8, 9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i8> @llvm.riscv.th.vsra.nxv8i8(
+    <vscale x 8 x i8> undef,
+    <vscale x 8 x i8> %0,
+    iXLen 9,
+    iXLen %1)
+
+  ret <vscale x 8 x i8> %a
+}
+
+define <vscale x 8 x i8> @intrinsic_vsra_mask_vi_nxv8i8_nxv8i8_i8(<vscale x 8 x i8> %0, <vscale x 8 x i8> %1, <vscale x 8 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vsra_mask_vi_nxv8i8_nxv8i8_i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e8, m1, d1
+; CHECK-NEXT:    th.vsra.vi v8, v9, 9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i8> @llvm.riscv.th.vsra.mask.nxv8i8(
+    <vscale x 8 x i8> %0,
+    <vscale x 8 x i8> %1,
+    iXLen 9,
+    <vscale x 8 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 8 x i8> %a
+}
+
+define <vscale x 16 x i8> @intrinsic_vsra_vi_nxv16i8_nxv16i8_i8(<vscale x 16 x i8> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vsra_vi_nxv16i8_nxv16i8_i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e8, m2, d1
+; CHECK-NEXT:    th.vsra.vi v8, v8, 9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i8> @llvm.riscv.th.vsra.nxv16i8(
+    <vscale x 16 x i8> undef,
+    <vscale x 16 x i8> %0,
+    iXLen 9,
+    iXLen %1)
+
+  ret <vscale x 16 x i8> %a
+}
+
+define <vscale x 16 x i8> @intrinsic_vsra_mask_vi_nxv16i8_nxv16i8_i8(<vscale x 16 x i8> %0, <vscale x 16 x i8> %1, <vscale x 16 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vsra_mask_vi_nxv16i8_nxv16i8_i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e8, m2, d1
+; CHECK-NEXT:    th.vsra.vi v8, v10, 9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i8> @llvm.riscv.th.vsra.mask.nxv16i8(
+    <vscale x 16 x i8> %0,
+    <vscale x 16 x i8> %1,
+    iXLen 9,
+    <vscale x 16 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 16 x i8> %a
+}
+
+define <vscale x 32 x i8> @intrinsic_vsra_vi_nxv32i8_nxv32i8_i8(<vscale x 32 x i8> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vsra_vi_nxv32i8_nxv32i8_i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e8, m4, d1
+; CHECK-NEXT:    th.vsra.vi v8, v8, 9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x i8> @llvm.riscv.th.vsra.nxv32i8(
+    <vscale x 32 x i8> undef,
+    <vscale x 32 x i8> %0,
+    iXLen 9,
+    iXLen %1)
+
+  ret <vscale x 32 x i8> %a
+}
+
+define <vscale x 32 x i8> @intrinsic_vsra_mask_vi_nxv32i8_nxv32i8_i8(<vscale x 32 x i8> %0, <vscale x 32 x i8> %1, <vscale x 32 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vsra_mask_vi_nxv32i8_nxv32i8_i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e8, m4, d1
+; CHECK-NEXT:    th.vsra.vi v8, v12, 9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x i8> @llvm.riscv.th.vsra.mask.nxv32i8(
+    <vscale x 32 x i8> %0,
+    <vscale x 32 x i8> %1,
+    iXLen 9,
+    <vscale x 32 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 32 x i8> %a
+}
+
+define <vscale x 64 x i8> @intrinsic_vsra_vi_nxv64i8_nxv64i8_i8(<vscale x 64 x i8> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vsra_vi_nxv64i8_nxv64i8_i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e8, m8, d1
+; CHECK-NEXT:    th.vsra.vi v8, v8, 9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 64 x i8> @llvm.riscv.th.vsra.nxv64i8(
+    <vscale x 64 x i8> undef,
+    <vscale x 64 x i8> %0,
+    iXLen 9,
+    iXLen %1)
+
+  ret <vscale x 64 x i8> %a
+}
+
+define <vscale x 64 x i8> @intrinsic_vsra_mask_vi_nxv64i8_nxv64i8_i8(<vscale x 64 x i8> %0, <vscale x 64 x i8> %1, <vscale x 64 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vsra_mask_vi_nxv64i8_nxv64i8_i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e8, m8, d1
+; CHECK-NEXT:    th.vsra.vi v8, v16, 9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 64 x i8> @llvm.riscv.th.vsra.mask.nxv64i8(
+    <vscale x 64 x i8> %0,
+    <vscale x 64 x i8> %1,
+    iXLen 9,
+    <vscale x 64 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 64 x i8> %a
+}
+
+define <vscale x 4 x i16> @intrinsic_vsra_vi_nxv4i16_nxv4i16_i16(<vscale x 4 x i16> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vsra_vi_nxv4i16_nxv4i16_i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e16, m1, d1
+; CHECK-NEXT:    th.vsra.vi v8, v8, 9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i16> @llvm.riscv.th.vsra.nxv4i16(
+    <vscale x 4 x i16> undef,
+    <vscale x 4 x i16> %0,
+    iXLen 9,
+    iXLen %1)
+
+  ret <vscale x 4 x i16> %a
+}
+
+define <vscale x 4 x i16> @intrinsic_vsra_mask_vi_nxv4i16_nxv4i16_i16(<vscale x 4 x i16> %0, <vscale x 4 x i16> %1, <vscale x 4 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vsra_mask_vi_nxv4i16_nxv4i16_i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e16, m1, d1
+; CHECK-NEXT:    th.vsra.vi v8, v9, 9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i16> @llvm.riscv.th.vsra.mask.nxv4i16(
+    <vscale x 4 x i16> %0,
+    <vscale x 4 x i16> %1,
+    iXLen 9,
+    <vscale x 4 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 4 x i16> %a
+}
+
+define <vscale x 8 x i16> @intrinsic_vsra_vi_nxv8i16_nxv8i16_i16(<vscale x 8 x i16> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vsra_vi_nxv8i16_nxv8i16_i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e16, m2, d1
+; CHECK-NEXT:    th.vsra.vi v8, v8, 9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i16> @llvm.riscv.th.vsra.nxv8i16(
+    <vscale x 8 x i16> undef,
+    <vscale x 8 x i16> %0,
+    iXLen 9,
+    iXLen %1)
+
+  ret <vscale x 8 x i16> %a
+}
+
+define <vscale x 8 x i16> @intrinsic_vsra_mask_vi_nxv8i16_nxv8i16_i16(<vscale x 8 x i16> %0, <vscale x 8 x i16> %1, <vscale x 8 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vsra_mask_vi_nxv8i16_nxv8i16_i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e16, m2, d1
+; CHECK-NEXT:    th.vsra.vi v8, v10, 9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i16> @llvm.riscv.th.vsra.mask.nxv8i16(
+    <vscale x 8 x i16> %0,
+    <vscale x 8 x i16> %1,
+    iXLen 9,
+    <vscale x 8 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 8 x i16> %a
+}
+
+define <vscale x 16 x i16> @intrinsic_vsra_vi_nxv16i16_nxv16i16_i16(<vscale x 16 x i16> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vsra_vi_nxv16i16_nxv16i16_i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e16, m4, d1
+; CHECK-NEXT:    th.vsra.vi v8, v8, 9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i16> @llvm.riscv.th.vsra.nxv16i16(
+    <vscale x 16 x i16> undef,
+    <vscale x 16 x i16> %0,
+    iXLen 9,
+    iXLen %1)
+
+  ret <vscale x 16 x i16> %a
+}
+
+define <vscale x 16 x i16> @intrinsic_vsra_mask_vi_nxv16i16_nxv16i16_i16(<vscale x 16 x i16> %0, <vscale x 16 x i16> %1, <vscale x 16 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vsra_mask_vi_nxv16i16_nxv16i16_i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e16, m4, d1
+; CHECK-NEXT:    th.vsra.vi v8, v12, 9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i16> @llvm.riscv.th.vsra.mask.nxv16i16(
+    <vscale x 16 x i16> %0,
+    <vscale x 16 x i16> %1,
+    iXLen 9,
+    <vscale x 16 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 16 x i16> %a
+}
+
+define <vscale x 32 x i16> @intrinsic_vsra_vi_nxv32i16_nxv32i16_i16(<vscale x 32 x i16> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vsra_vi_nxv32i16_nxv32i16_i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e16, m8, d1
+; CHECK-NEXT:    th.vsra.vi v8, v8, 9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x i16> @llvm.riscv.th.vsra.nxv32i16(
+    <vscale x 32 x i16> undef,
+    <vscale x 32 x i16> %0,
+    iXLen 9,
+    iXLen %1)
+
+  ret <vscale x 32 x i16> %a
+}
+
+define <vscale x 32 x i16> @intrinsic_vsra_mask_vi_nxv32i16_nxv32i16_i16(<vscale x 32 x i16> %0, <vscale x 32 x i16> %1, <vscale x 32 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vsra_mask_vi_nxv32i16_nxv32i16_i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e16, m8, d1
+; CHECK-NEXT:    th.vsra.vi v8, v16, 9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x i16> @llvm.riscv.th.vsra.mask.nxv32i16(
+    <vscale x 32 x i16> %0,
+    <vscale x 32 x i16> %1,
+    iXLen 9,
+    <vscale x 32 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 32 x i16> %a
+}
+
+define <vscale x 2 x i32> @intrinsic_vsra_vi_nxv2i32_nxv2i32_i32(<vscale x 2 x i32> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vsra_vi_nxv2i32_nxv2i32_i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e32, m1, d1
+; CHECK-NEXT:    th.vsra.vi v8, v8, 9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i32> @llvm.riscv.th.vsra.nxv2i32(
+    <vscale x 2 x i32> undef,
+    <vscale x 2 x i32> %0,
+    iXLen 9,
+    iXLen %1)
+
+  ret <vscale x 2 x i32> %a
+}
+
+define <vscale x 2 x i32> @intrinsic_vsra_mask_vi_nxv2i32_nxv2i32_i32(<vscale x 2 x i32> %0, <vscale x 2 x i32> %1, <vscale x 2 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vsra_mask_vi_nxv2i32_nxv2i32_i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e32, m1, d1
+; CHECK-NEXT:    th.vsra.vi v8, v9, 9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i32> @llvm.riscv.th.vsra.mask.nxv2i32(
+    <vscale x 2 x i32> %0,
+    <vscale x 2 x i32> %1,
+    iXLen 9,
+    <vscale x 2 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 2 x i32> %a
+}
+
+define <vscale x 4 x i32> @intrinsic_vsra_vi_nxv4i32_nxv4i32_i32(<vscale x 4 x i32> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vsra_vi_nxv4i32_nxv4i32_i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e32, m2, d1
+; CHECK-NEXT:    th.vsra.vi v8, v8, 9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i32> @llvm.riscv.th.vsra.nxv4i32(
+    <vscale x 4 x i32> undef,
+    <vscale x 4 x i32> %0,
+    iXLen 9,
+    iXLen %1)
+
+  ret <vscale x 4 x i32> %a
+}
+
+define <vscale x 4 x i32> @intrinsic_vsra_mask_vi_nxv4i32_nxv4i32_i32(<vscale x 4 x i32> %0, <vscale x 4 x i32> %1, <vscale x 4 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vsra_mask_vi_nxv4i32_nxv4i32_i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e32, m2, d1
+; CHECK-NEXT:    th.vsra.vi v8, v10, 9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i32> @llvm.riscv.th.vsra.mask.nxv4i32(
+    <vscale x 4 x i32> %0,
+    <vscale x 4 x i32> %1,
+    iXLen 9,
+    <vscale x 4 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 4 x i32> %a
+}
+
+define <vscale x 8 x i32> @intrinsic_vsra_vi_nxv8i32_nxv8i32_i32(<vscale x 8 x i32> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vsra_vi_nxv8i32_nxv8i32_i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e32, m4, d1
+; CHECK-NEXT:    th.vsra.vi v8, v8, 9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i32> @llvm.riscv.th.vsra.nxv8i32(
+    <vscale x 8 x i32> undef,
+    <vscale x 8 x i32> %0,
+    iXLen 9,
+    iXLen %1)
+
+  ret <vscale x 8 x i32> %a
+}
+
+define <vscale x 8 x i32> @intrinsic_vsra_mask_vi_nxv8i32_nxv8i32_i32(<vscale x 8 x i32> %0, <vscale x 8 x i32> %1, <vscale x 8 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vsra_mask_vi_nxv8i32_nxv8i32_i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e32, m4, d1
+; CHECK-NEXT:    th.vsra.vi v8, v12, 9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i32> @llvm.riscv.th.vsra.mask.nxv8i32(
+    <vscale x 8 x i32> %0,
+    <vscale x 8 x i32> %1,
+    iXLen 9,
+    <vscale x 8 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 8 x i32> %a
+}
+
+define <vscale x 16 x i32> @intrinsic_vsra_vi_nxv16i32_nxv16i32_i32(<vscale x 16 x i32> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vsra_vi_nxv16i32_nxv16i32_i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e32, m8, d1
+; CHECK-NEXT:    th.vsra.vi v8, v8, 9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i32> @llvm.riscv.th.vsra.nxv16i32(
+    <vscale x 16 x i32> undef,
+    <vscale x 16 x i32> %0,
+    iXLen 9,
+    iXLen %1)
+
+  ret <vscale x 16 x i32> %a
+}
+
+define <vscale x 16 x i32> @intrinsic_vsra_mask_vi_nxv16i32_nxv16i32_i32(<vscale x 16 x i32> %0, <vscale x 16 x i32> %1, <vscale x 16 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vsra_mask_vi_nxv16i32_nxv16i32_i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e32, m8, d1
+; CHECK-NEXT:    th.vsra.vi v8, v16, 9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i32> @llvm.riscv.th.vsra.mask.nxv16i32(
+    <vscale x 16 x i32> %0,
+    <vscale x 16 x i32> %1,
+    iXLen 9,
+    <vscale x 16 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 16 x i32> %a
+}
+
+define <vscale x 1 x i64> @intrinsic_vsra_vi_nxv1i64_nxv1i64_i64(<vscale x 1 x i64> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vsra_vi_nxv1i64_nxv1i64_i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e64, m1, d1
+; CHECK-NEXT:    th.vsra.vi v8, v8, 9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x i64> @llvm.riscv.th.vsra.nxv1i64(
+    <vscale x 1 x i64> undef,
+    <vscale x 1 x i64> %0,
+    iXLen 9,
+    iXLen %1)
+
+  ret <vscale x 1 x i64> %a
+}
+
+define <vscale x 1 x i64> @intrinsic_vsra_mask_vi_nxv1i64_nxv1i64_i64(<vscale x 1 x i64> %0, <vscale x 1 x i64> %1, <vscale x 1 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vsra_mask_vi_nxv1i64_nxv1i64_i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e64, m1, d1
+; CHECK-NEXT:    th.vsra.vi v8, v9, 9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x i64> @llvm.riscv.th.vsra.mask.nxv1i64(
+    <vscale x 1 x i64> %0,
+    <vscale x 1 x i64> %1,
+    iXLen 9,
+    <vscale x 1 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 1 x i64> %a
+}
+
+define <vscale x 2 x i64> @intrinsic_vsra_vi_nxv2i64_nxv2i64_i64(<vscale x 2 x i64> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vsra_vi_nxv2i64_nxv2i64_i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e64, m2, d1
+; CHECK-NEXT:    th.vsra.vi v8, v8, 9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i64> @llvm.riscv.th.vsra.nxv2i64(
+    <vscale x 2 x i64> undef,
+    <vscale x 2 x i64> %0,
+    iXLen 9,
+    iXLen %1)
+
+  ret <vscale x 2 x i64> %a
+}
+
+define <vscale x 2 x i64> @intrinsic_vsra_mask_vi_nxv2i64_nxv2i64_i64(<vscale x 2 x i64> %0, <vscale x 2 x i64> %1, <vscale x 2 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vsra_mask_vi_nxv2i64_nxv2i64_i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e64, m2, d1
+; CHECK-NEXT:    th.vsra.vi v8, v10, 9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i64> @llvm.riscv.th.vsra.mask.nxv2i64(
+    <vscale x 2 x i64> %0,
+    <vscale x 2 x i64> %1,
+    iXLen 9,
+    <vscale x 2 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 2 x i64> %a
+}
+
+define <vscale x 4 x i64> @intrinsic_vsra_vi_nxv4i64_nxv4i64_i64(<vscale x 4 x i64> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vsra_vi_nxv4i64_nxv4i64_i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e64, m4, d1
+; CHECK-NEXT:    th.vsra.vi v8, v8, 9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i64> @llvm.riscv.th.vsra.nxv4i64(
+    <vscale x 4 x i64> undef,
+    <vscale x 4 x i64> %0,
+    iXLen 9,
+    iXLen %1)
+
+  ret <vscale x 4 x i64> %a
+}
+
+define <vscale x 4 x i64> @intrinsic_vsra_mask_vi_nxv4i64_nxv4i64_i64(<vscale x 4 x i64> %0, <vscale x 4 x i64> %1, <vscale x 4 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vsra_mask_vi_nxv4i64_nxv4i64_i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e64, m4, d1
+; CHECK-NEXT:    th.vsra.vi v8, v12, 9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i64> @llvm.riscv.th.vsra.mask.nxv4i64(
+    <vscale x 4 x i64> %0,
+    <vscale x 4 x i64> %1,
+    iXLen 9,
+    <vscale x 4 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 4 x i64> %a
+}
+
+define <vscale x 8 x i64> @intrinsic_vsra_vi_nxv8i64_nxv8i64_i64(<vscale x 8 x i64> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vsra_vi_nxv8i64_nxv8i64_i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e64, m8, d1
+; CHECK-NEXT:    th.vsra.vi v8, v8, 9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i64> @llvm.riscv.th.vsra.nxv8i64(
+    <vscale x 8 x i64> undef,
+    <vscale x 8 x i64> %0,
+    iXLen 9,
+    iXLen %1)
+
+  ret <vscale x 8 x i64> %a
+}
+
+define <vscale x 8 x i64> @intrinsic_vsra_mask_vi_nxv8i64_nxv8i64_i64(<vscale x 8 x i64> %0, <vscale x 8 x i64> %1, <vscale x 8 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vsra_mask_vi_nxv8i64_nxv8i64_i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e64, m8, d1
+; CHECK-NEXT:    th.vsra.vi v8, v16, 9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i64> @llvm.riscv.th.vsra.mask.nxv8i64(
+    <vscale x 8 x i64> %0,
+    <vscale x 8 x i64> %1,
+    iXLen 9,
+    <vscale x 8 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 8 x i64> %a
+}
diff --git a/llvm/test/CodeGen/RISCV/rvv0p71/vsrl.ll b/llvm/test/CodeGen/RISCV/rvv0p71/vsrl.ll
new file mode 100644
index 00000000000000..0a15bb8fc4b8cc
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv0p71/vsrl.ll
@@ -0,0 +1,3097 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+xtheadvector \
+; RUN:   -verify-machineinstrs | FileCheck %s
+; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+xtheadvector \
+; RUN:   -verify-machineinstrs | FileCheck %s
+
+declare <vscale x 8 x i8> @llvm.riscv.th.vsrl.nxv8i8.nxv8i8(
+  <vscale x 8 x i8>,
+  <vscale x 8 x i8>,
+  <vscale x 8 x i8>,
+  iXLen);
+
+define <vscale x 8 x i8> @intrinsic_vsrl_vv_nxv8i8_nxv8i8_nxv8i8(<vscale x 8 x i8> %0, <vscale x 8 x i8> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_vv_nxv8i8_nxv8i8_nxv8i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e8, m1, d1
+; CHECK-NEXT:    th.vsrl.vv v8, v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i8> @llvm.riscv.th.vsrl.nxv8i8.nxv8i8(
+    <vscale x 8 x i8> undef,
+    <vscale x 8 x i8> %0,
+    <vscale x 8 x i8> %1,
+    iXLen %2)
+
+  ret <vscale x 8 x i8> %a
+}
+
+declare <vscale x 8 x i8> @llvm.riscv.th.vsrl.mask.nxv8i8.nxv8i8(
+  <vscale x 8 x i8>,
+  <vscale x 8 x i8>,
+  <vscale x 8 x i8>,
+  <vscale x 8 x i1>,
+  iXLen);
+
+define <vscale x 8 x i8> @intrinsic_vsrl_mask_vv_nxv8i8_nxv8i8_nxv8i8(<vscale x 8 x i8> %0, <vscale x 8 x i8> %1, <vscale x 8 x i8> %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_mask_vv_nxv8i8_nxv8i8_nxv8i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e8, m1, d1
+; CHECK-NEXT:    th.vsrl.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i8> @llvm.riscv.th.vsrl.mask.nxv8i8.nxv8i8(
+    <vscale x 8 x i8> %0,
+    <vscale x 8 x i8> %1,
+    <vscale x 8 x i8> %2,
+    <vscale x 8 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 8 x i8> %a
+}
+
+declare <vscale x 16 x i8> @llvm.riscv.th.vsrl.nxv16i8.nxv16i8(
+  <vscale x 16 x i8>,
+  <vscale x 16 x i8>,
+  <vscale x 16 x i8>,
+  iXLen);
+
+define <vscale x 16 x i8> @intrinsic_vsrl_vv_nxv16i8_nxv16i8_nxv16i8(<vscale x 16 x i8> %0, <vscale x 16 x i8> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_vv_nxv16i8_nxv16i8_nxv16i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e8, m2, d1
+; CHECK-NEXT:    th.vsrl.vv v8, v8, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i8> @llvm.riscv.th.vsrl.nxv16i8.nxv16i8(
+    <vscale x 16 x i8> undef,
+    <vscale x 16 x i8> %0,
+    <vscale x 16 x i8> %1,
+    iXLen %2)
+
+  ret <vscale x 16 x i8> %a
+}
+
+declare <vscale x 16 x i8> @llvm.riscv.th.vsrl.mask.nxv16i8.nxv16i8(
+  <vscale x 16 x i8>,
+  <vscale x 16 x i8>,
+  <vscale x 16 x i8>,
+  <vscale x 16 x i1>,
+  iXLen);
+
+define <vscale x 16 x i8> @intrinsic_vsrl_mask_vv_nxv16i8_nxv16i8_nxv16i8(<vscale x 16 x i8> %0, <vscale x 16 x i8> %1, <vscale x 16 x i8> %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_mask_vv_nxv16i8_nxv16i8_nxv16i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e8, m2, d1
+; CHECK-NEXT:    th.vsrl.vv v8, v10, v12, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i8> @llvm.riscv.th.vsrl.mask.nxv16i8.nxv16i8(
+    <vscale x 16 x i8> %0,
+    <vscale x 16 x i8> %1,
+    <vscale x 16 x i8> %2,
+    <vscale x 16 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 16 x i8> %a
+}
+
+declare <vscale x 32 x i8> @llvm.riscv.th.vsrl.nxv32i8.nxv32i8(
+  <vscale x 32 x i8>,
+  <vscale x 32 x i8>,
+  <vscale x 32 x i8>,
+  iXLen);
+
+define <vscale x 32 x i8> @intrinsic_vsrl_vv_nxv32i8_nxv32i8_nxv32i8(<vscale x 32 x i8> %0, <vscale x 32 x i8> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_vv_nxv32i8_nxv32i8_nxv32i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e8, m4, d1
+; CHECK-NEXT:    th.vsrl.vv v8, v8, v12
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x i8> @llvm.riscv.th.vsrl.nxv32i8.nxv32i8(
+    <vscale x 32 x i8> undef,
+    <vscale x 32 x i8> %0,
+    <vscale x 32 x i8> %1,
+    iXLen %2)
+
+  ret <vscale x 32 x i8> %a
+}
+
+declare <vscale x 32 x i8> @llvm.riscv.th.vsrl.mask.nxv32i8.nxv32i8(
+  <vscale x 32 x i8>,
+  <vscale x 32 x i8>,
+  <vscale x 32 x i8>,
+  <vscale x 32 x i1>,
+  iXLen);
+
+define <vscale x 32 x i8> @intrinsic_vsrl_mask_vv_nxv32i8_nxv32i8_nxv32i8(<vscale x 32 x i8> %0, <vscale x 32 x i8> %1, <vscale x 32 x i8> %2, <vscale x 32 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_mask_vv_nxv32i8_nxv32i8_nxv32i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e8, m4, d1
+; CHECK-NEXT:    th.vsrl.vv v8, v12, v16, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x i8> @llvm.riscv.th.vsrl.mask.nxv32i8.nxv32i8(
+    <vscale x 32 x i8> %0,
+    <vscale x 32 x i8> %1,
+    <vscale x 32 x i8> %2,
+    <vscale x 32 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 32 x i8> %a
+}
+
+declare <vscale x 64 x i8> @llvm.riscv.th.vsrl.nxv64i8.nxv64i8(
+  <vscale x 64 x i8>,
+  <vscale x 64 x i8>,
+  <vscale x 64 x i8>,
+  iXLen);
+
+define <vscale x 64 x i8> @intrinsic_vsrl_vv_nxv64i8_nxv64i8_nxv64i8(<vscale x 64 x i8> %0, <vscale x 64 x i8> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_vv_nxv64i8_nxv64i8_nxv64i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e8, m8, d1
+; CHECK-NEXT:    th.vsrl.vv v8, v8, v16
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 64 x i8> @llvm.riscv.th.vsrl.nxv64i8.nxv64i8(
+    <vscale x 64 x i8> undef,
+    <vscale x 64 x i8> %0,
+    <vscale x 64 x i8> %1,
+    iXLen %2)
+
+  ret <vscale x 64 x i8> %a
+}
+
+declare <vscale x 64 x i8> @llvm.riscv.th.vsrl.mask.nxv64i8.nxv64i8(
+  <vscale x 64 x i8>,
+  <vscale x 64 x i8>,
+  <vscale x 64 x i8>,
+  <vscale x 64 x i1>,
+  iXLen);
+
+define <vscale x 64 x i8> @intrinsic_vsrl_mask_vv_nxv64i8_nxv64i8_nxv64i8(<vscale x 64 x i8> %0, <vscale x 64 x i8> %1, <vscale x 64 x i8> %2, <vscale x 64 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_mask_vv_nxv64i8_nxv64i8_nxv64i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m8, d1
+; CHECK-NEXT:    th.vle.v v24, (a0)
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a0, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a0, a2
+; CHECK-NEXT:    th.vsetvli zero, a1, e8, m8, d1
+; CHECK-NEXT:    th.vsrl.vv v8, v16, v24, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 64 x i8> @llvm.riscv.th.vsrl.mask.nxv64i8.nxv64i8(
+    <vscale x 64 x i8> %0,
+    <vscale x 64 x i8> %1,
+    <vscale x 64 x i8> %2,
+    <vscale x 64 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 64 x i8> %a
+}
+
+declare <vscale x 4 x i16> @llvm.riscv.th.vsrl.nxv4i16.nxv4i16(
+  <vscale x 4 x i16>,
+  <vscale x 4 x i16>,
+  <vscale x 4 x i16>,
+  iXLen);
+
+define <vscale x 4 x i16> @intrinsic_vsrl_vv_nxv4i16_nxv4i16_nxv4i16(<vscale x 4 x i16> %0, <vscale x 4 x i16> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_vv_nxv4i16_nxv4i16_nxv4i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e16, m1, d1
+; CHECK-NEXT:    th.vsrl.vv v8, v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i16> @llvm.riscv.th.vsrl.nxv4i16.nxv4i16(
+    <vscale x 4 x i16> undef,
+    <vscale x 4 x i16> %0,
+    <vscale x 4 x i16> %1,
+    iXLen %2)
+
+  ret <vscale x 4 x i16> %a
+}
+
+declare <vscale x 4 x i16> @llvm.riscv.th.vsrl.mask.nxv4i16.nxv4i16(
+  <vscale x 4 x i16>,
+  <vscale x 4 x i16>,
+  <vscale x 4 x i16>,
+  <vscale x 4 x i1>,
+  iXLen);
+
+define <vscale x 4 x i16> @intrinsic_vsrl_mask_vv_nxv4i16_nxv4i16_nxv4i16(<vscale x 4 x i16> %0, <vscale x 4 x i16> %1, <vscale x 4 x i16> %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_mask_vv_nxv4i16_nxv4i16_nxv4i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e16, m1, d1
+; CHECK-NEXT:    th.vsrl.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i16> @llvm.riscv.th.vsrl.mask.nxv4i16.nxv4i16(
+    <vscale x 4 x i16> %0,
+    <vscale x 4 x i16> %1,
+    <vscale x 4 x i16> %2,
+    <vscale x 4 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 4 x i16> %a
+}
+
+declare <vscale x 8 x i16> @llvm.riscv.th.vsrl.nxv8i16.nxv8i16(
+  <vscale x 8 x i16>,
+  <vscale x 8 x i16>,
+  <vscale x 8 x i16>,
+  iXLen);
+
+define <vscale x 8 x i16> @intrinsic_vsrl_vv_nxv8i16_nxv8i16_nxv8i16(<vscale x 8 x i16> %0, <vscale x 8 x i16> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_vv_nxv8i16_nxv8i16_nxv8i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e16, m2, d1
+; CHECK-NEXT:    th.vsrl.vv v8, v8, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i16> @llvm.riscv.th.vsrl.nxv8i16.nxv8i16(
+    <vscale x 8 x i16> undef,
+    <vscale x 8 x i16> %0,
+    <vscale x 8 x i16> %1,
+    iXLen %2)
+
+  ret <vscale x 8 x i16> %a
+}
+
+declare <vscale x 8 x i16> @llvm.riscv.th.vsrl.mask.nxv8i16.nxv8i16(
+  <vscale x 8 x i16>,
+  <vscale x 8 x i16>,
+  <vscale x 8 x i16>,
+  <vscale x 8 x i1>,
+  iXLen);
+
+define <vscale x 8 x i16> @intrinsic_vsrl_mask_vv_nxv8i16_nxv8i16_nxv8i16(<vscale x 8 x i16> %0, <vscale x 8 x i16> %1, <vscale x 8 x i16> %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_mask_vv_nxv8i16_nxv8i16_nxv8i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e16, m2, d1
+; CHECK-NEXT:    th.vsrl.vv v8, v10, v12, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i16> @llvm.riscv.th.vsrl.mask.nxv8i16.nxv8i16(
+    <vscale x 8 x i16> %0,
+    <vscale x 8 x i16> %1,
+    <vscale x 8 x i16> %2,
+    <vscale x 8 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 8 x i16> %a
+}
+
+declare <vscale x 16 x i16> @llvm.riscv.th.vsrl.nxv16i16.nxv16i16(
+  <vscale x 16 x i16>,
+  <vscale x 16 x i16>,
+  <vscale x 16 x i16>,
+  iXLen);
+
+define <vscale x 16 x i16> @intrinsic_vsrl_vv_nxv16i16_nxv16i16_nxv16i16(<vscale x 16 x i16> %0, <vscale x 16 x i16> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_vv_nxv16i16_nxv16i16_nxv16i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e16, m4, d1
+; CHECK-NEXT:    th.vsrl.vv v8, v8, v12
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i16> @llvm.riscv.th.vsrl.nxv16i16.nxv16i16(
+    <vscale x 16 x i16> undef,
+    <vscale x 16 x i16> %0,
+    <vscale x 16 x i16> %1,
+    iXLen %2)
+
+  ret <vscale x 16 x i16> %a
+}
+
+declare <vscale x 16 x i16> @llvm.riscv.th.vsrl.mask.nxv16i16.nxv16i16(
+  <vscale x 16 x i16>,
+  <vscale x 16 x i16>,
+  <vscale x 16 x i16>,
+  <vscale x 16 x i1>,
+  iXLen);
+
+define <vscale x 16 x i16> @intrinsic_vsrl_mask_vv_nxv16i16_nxv16i16_nxv16i16(<vscale x 16 x i16> %0, <vscale x 16 x i16> %1, <vscale x 16 x i16> %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_mask_vv_nxv16i16_nxv16i16_nxv16i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e16, m4, d1
+; CHECK-NEXT:    th.vsrl.vv v8, v12, v16, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i16> @llvm.riscv.th.vsrl.mask.nxv16i16.nxv16i16(
+    <vscale x 16 x i16> %0,
+    <vscale x 16 x i16> %1,
+    <vscale x 16 x i16> %2,
+    <vscale x 16 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 16 x i16> %a
+}
+
+declare <vscale x 32 x i16> @llvm.riscv.th.vsrl.nxv32i16.nxv32i16(
+  <vscale x 32 x i16>,
+  <vscale x 32 x i16>,
+  <vscale x 32 x i16>,
+  iXLen);
+
+define <vscale x 32 x i16> @intrinsic_vsrl_vv_nxv32i16_nxv32i16_nxv32i16(<vscale x 32 x i16> %0, <vscale x 32 x i16> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_vv_nxv32i16_nxv32i16_nxv32i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e16, m8, d1
+; CHECK-NEXT:    th.vsrl.vv v8, v8, v16
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x i16> @llvm.riscv.th.vsrl.nxv32i16.nxv32i16(
+    <vscale x 32 x i16> undef,
+    <vscale x 32 x i16> %0,
+    <vscale x 32 x i16> %1,
+    iXLen %2)
+
+  ret <vscale x 32 x i16> %a
+}
+
+declare <vscale x 32 x i16> @llvm.riscv.th.vsrl.mask.nxv32i16.nxv32i16(
+  <vscale x 32 x i16>,
+  <vscale x 32 x i16>,
+  <vscale x 32 x i16>,
+  <vscale x 32 x i1>,
+  iXLen);
+
+define <vscale x 32 x i16> @intrinsic_vsrl_mask_vv_nxv32i16_nxv32i16_nxv32i16(<vscale x 32 x i16> %0, <vscale x 32 x i16> %1, <vscale x 32 x i16> %2, <vscale x 32 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_mask_vv_nxv32i16_nxv32i16_nxv32i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e16, m8, d1
+; CHECK-NEXT:    th.vle.v v24, (a0)
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a0, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a0, a2
+; CHECK-NEXT:    th.vsetvli zero, a1, e16, m8, d1
+; CHECK-NEXT:    th.vsrl.vv v8, v16, v24, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x i16> @llvm.riscv.th.vsrl.mask.nxv32i16.nxv32i16(
+    <vscale x 32 x i16> %0,
+    <vscale x 32 x i16> %1,
+    <vscale x 32 x i16> %2,
+    <vscale x 32 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 32 x i16> %a
+}
+
+declare <vscale x 2 x i32> @llvm.riscv.th.vsrl.nxv2i32.nxv2i32(
+  <vscale x 2 x i32>,
+  <vscale x 2 x i32>,
+  <vscale x 2 x i32>,
+  iXLen);
+
+define <vscale x 2 x i32> @intrinsic_vsrl_vv_nxv2i32_nxv2i32_nxv2i32(<vscale x 2 x i32> %0, <vscale x 2 x i32> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_vv_nxv2i32_nxv2i32_nxv2i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e32, m1, d1
+; CHECK-NEXT:    th.vsrl.vv v8, v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i32> @llvm.riscv.th.vsrl.nxv2i32.nxv2i32(
+    <vscale x 2 x i32> undef,
+    <vscale x 2 x i32> %0,
+    <vscale x 2 x i32> %1,
+    iXLen %2)
+
+  ret <vscale x 2 x i32> %a
+}
+
+declare <vscale x 2 x i32> @llvm.riscv.th.vsrl.mask.nxv2i32.nxv2i32(
+  <vscale x 2 x i32>,
+  <vscale x 2 x i32>,
+  <vscale x 2 x i32>,
+  <vscale x 2 x i1>,
+  iXLen);
+
+define <vscale x 2 x i32> @intrinsic_vsrl_mask_vv_nxv2i32_nxv2i32_nxv2i32(<vscale x 2 x i32> %0, <vscale x 2 x i32> %1, <vscale x 2 x i32> %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_mask_vv_nxv2i32_nxv2i32_nxv2i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e32, m1, d1
+; CHECK-NEXT:    th.vsrl.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i32> @llvm.riscv.th.vsrl.mask.nxv2i32.nxv2i32(
+    <vscale x 2 x i32> %0,
+    <vscale x 2 x i32> %1,
+    <vscale x 2 x i32> %2,
+    <vscale x 2 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 2 x i32> %a
+}
+
+declare <vscale x 4 x i32> @llvm.riscv.th.vsrl.nxv4i32.nxv4i32(
+  <vscale x 4 x i32>,
+  <vscale x 4 x i32>,
+  <vscale x 4 x i32>,
+  iXLen);
+
+define <vscale x 4 x i32> @intrinsic_vsrl_vv_nxv4i32_nxv4i32_nxv4i32(<vscale x 4 x i32> %0, <vscale x 4 x i32> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_vv_nxv4i32_nxv4i32_nxv4i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e32, m2, d1
+; CHECK-NEXT:    th.vsrl.vv v8, v8, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i32> @llvm.riscv.th.vsrl.nxv4i32.nxv4i32(
+    <vscale x 4 x i32> undef,
+    <vscale x 4 x i32> %0,
+    <vscale x 4 x i32> %1,
+    iXLen %2)
+
+  ret <vscale x 4 x i32> %a
+}
+
+declare <vscale x 4 x i32> @llvm.riscv.th.vsrl.mask.nxv4i32.nxv4i32(
+  <vscale x 4 x i32>,
+  <vscale x 4 x i32>,
+  <vscale x 4 x i32>,
+  <vscale x 4 x i1>,
+  iXLen);
+
+define <vscale x 4 x i32> @intrinsic_vsrl_mask_vv_nxv4i32_nxv4i32_nxv4i32(<vscale x 4 x i32> %0, <vscale x 4 x i32> %1, <vscale x 4 x i32> %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_mask_vv_nxv4i32_nxv4i32_nxv4i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e32, m2, d1
+; CHECK-NEXT:    th.vsrl.vv v8, v10, v12, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i32> @llvm.riscv.th.vsrl.mask.nxv4i32.nxv4i32(
+    <vscale x 4 x i32> %0,
+    <vscale x 4 x i32> %1,
+    <vscale x 4 x i32> %2,
+    <vscale x 4 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 4 x i32> %a
+}
+
+declare <vscale x 8 x i32> @llvm.riscv.th.vsrl.nxv8i32.nxv8i32(
+  <vscale x 8 x i32>,
+  <vscale x 8 x i32>,
+  <vscale x 8 x i32>,
+  iXLen);
+
+define <vscale x 8 x i32> @intrinsic_vsrl_vv_nxv8i32_nxv8i32_nxv8i32(<vscale x 8 x i32> %0, <vscale x 8 x i32> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_vv_nxv8i32_nxv8i32_nxv8i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e32, m4, d1
+; CHECK-NEXT:    th.vsrl.vv v8, v8, v12
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i32> @llvm.riscv.th.vsrl.nxv8i32.nxv8i32(
+    <vscale x 8 x i32> undef,
+    <vscale x 8 x i32> %0,
+    <vscale x 8 x i32> %1,
+    iXLen %2)
+
+  ret <vscale x 8 x i32> %a
+}
+
+declare <vscale x 8 x i32> @llvm.riscv.th.vsrl.mask.nxv8i32.nxv8i32(
+  <vscale x 8 x i32>,
+  <vscale x 8 x i32>,
+  <vscale x 8 x i32>,
+  <vscale x 8 x i1>,
+  iXLen);
+
+define <vscale x 8 x i32> @intrinsic_vsrl_mask_vv_nxv8i32_nxv8i32_nxv8i32(<vscale x 8 x i32> %0, <vscale x 8 x i32> %1, <vscale x 8 x i32> %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_mask_vv_nxv8i32_nxv8i32_nxv8i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e32, m4, d1
+; CHECK-NEXT:    th.vsrl.vv v8, v12, v16, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i32> @llvm.riscv.th.vsrl.mask.nxv8i32.nxv8i32(
+    <vscale x 8 x i32> %0,
+    <vscale x 8 x i32> %1,
+    <vscale x 8 x i32> %2,
+    <vscale x 8 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 8 x i32> %a
+}
+
+declare <vscale x 16 x i32> @llvm.riscv.th.vsrl.nxv16i32.nxv16i32(
+  <vscale x 16 x i32>,
+  <vscale x 16 x i32>,
+  <vscale x 16 x i32>,
+  iXLen);
+
+define <vscale x 16 x i32> @intrinsic_vsrl_vv_nxv16i32_nxv16i32_nxv16i32(<vscale x 16 x i32> %0, <vscale x 16 x i32> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_vv_nxv16i32_nxv16i32_nxv16i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e32, m8, d1
+; CHECK-NEXT:    th.vsrl.vv v8, v8, v16
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i32> @llvm.riscv.th.vsrl.nxv16i32.nxv16i32(
+    <vscale x 16 x i32> undef,
+    <vscale x 16 x i32> %0,
+    <vscale x 16 x i32> %1,
+    iXLen %2)
+
+  ret <vscale x 16 x i32> %a
+}
+
+declare <vscale x 16 x i32> @llvm.riscv.th.vsrl.mask.nxv16i32.nxv16i32(
+  <vscale x 16 x i32>,
+  <vscale x 16 x i32>,
+  <vscale x 16 x i32>,
+  <vscale x 16 x i1>,
+  iXLen);
+
+define <vscale x 16 x i32> @intrinsic_vsrl_mask_vv_nxv16i32_nxv16i32_nxv16i32(<vscale x 16 x i32> %0, <vscale x 16 x i32> %1, <vscale x 16 x i32> %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_mask_vv_nxv16i32_nxv16i32_nxv16i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e32, m8, d1
+; CHECK-NEXT:    th.vle.v v24, (a0)
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a0, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a0, a2
+; CHECK-NEXT:    th.vsetvli zero, a1, e32, m8, d1
+; CHECK-NEXT:    th.vsrl.vv v8, v16, v24, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i32> @llvm.riscv.th.vsrl.mask.nxv16i32.nxv16i32(
+    <vscale x 16 x i32> %0,
+    <vscale x 16 x i32> %1,
+    <vscale x 16 x i32> %2,
+    <vscale x 16 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 16 x i32> %a
+}
+
+declare <vscale x 1 x i64> @llvm.riscv.th.vsrl.nxv1i64.nxv1i64(
+  <vscale x 1 x i64>,
+  <vscale x 1 x i64>,
+  <vscale x 1 x i64>,
+  iXLen);
+
+define <vscale x 1 x i64> @intrinsic_vsrl_vv_nxv1i64_nxv1i64_nxv1i64(<vscale x 1 x i64> %0, <vscale x 1 x i64> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_vv_nxv1i64_nxv1i64_nxv1i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e64, m1, d1
+; CHECK-NEXT:    th.vsrl.vv v8, v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x i64> @llvm.riscv.th.vsrl.nxv1i64.nxv1i64(
+    <vscale x 1 x i64> undef,
+    <vscale x 1 x i64> %0,
+    <vscale x 1 x i64> %1,
+    iXLen %2)
+
+  ret <vscale x 1 x i64> %a
+}
+
+declare <vscale x 1 x i64> @llvm.riscv.th.vsrl.mask.nxv1i64.nxv1i64(
+  <vscale x 1 x i64>,
+  <vscale x 1 x i64>,
+  <vscale x 1 x i64>,
+  <vscale x 1 x i1>,
+  iXLen);
+
+define <vscale x 1 x i64> @intrinsic_vsrl_mask_vv_nxv1i64_nxv1i64_nxv1i64(<vscale x 1 x i64> %0, <vscale x 1 x i64> %1, <vscale x 1 x i64> %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_mask_vv_nxv1i64_nxv1i64_nxv1i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e64, m1, d1
+; CHECK-NEXT:    th.vsrl.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x i64> @llvm.riscv.th.vsrl.mask.nxv1i64.nxv1i64(
+    <vscale x 1 x i64> %0,
+    <vscale x 1 x i64> %1,
+    <vscale x 1 x i64> %2,
+    <vscale x 1 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 1 x i64> %a
+}
+
+declare <vscale x 2 x i64> @llvm.riscv.th.vsrl.nxv2i64.nxv2i64(
+  <vscale x 2 x i64>,
+  <vscale x 2 x i64>,
+  <vscale x 2 x i64>,
+  iXLen);
+
+define <vscale x 2 x i64> @intrinsic_vsrl_vv_nxv2i64_nxv2i64_nxv2i64(<vscale x 2 x i64> %0, <vscale x 2 x i64> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_vv_nxv2i64_nxv2i64_nxv2i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e64, m2, d1
+; CHECK-NEXT:    th.vsrl.vv v8, v8, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i64> @llvm.riscv.th.vsrl.nxv2i64.nxv2i64(
+    <vscale x 2 x i64> undef,
+    <vscale x 2 x i64> %0,
+    <vscale x 2 x i64> %1,
+    iXLen %2)
+
+  ret <vscale x 2 x i64> %a
+}
+
+declare <vscale x 2 x i64> @llvm.riscv.th.vsrl.mask.nxv2i64.nxv2i64(
+  <vscale x 2 x i64>,
+  <vscale x 2 x i64>,
+  <vscale x 2 x i64>,
+  <vscale x 2 x i1>,
+  iXLen);
+
+define <vscale x 2 x i64> @intrinsic_vsrl_mask_vv_nxv2i64_nxv2i64_nxv2i64(<vscale x 2 x i64> %0, <vscale x 2 x i64> %1, <vscale x 2 x i64> %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_mask_vv_nxv2i64_nxv2i64_nxv2i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e64, m2, d1
+; CHECK-NEXT:    th.vsrl.vv v8, v10, v12, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i64> @llvm.riscv.th.vsrl.mask.nxv2i64.nxv2i64(
+    <vscale x 2 x i64> %0,
+    <vscale x 2 x i64> %1,
+    <vscale x 2 x i64> %2,
+    <vscale x 2 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 2 x i64> %a
+}
+
+declare <vscale x 4 x i64> @llvm.riscv.th.vsrl.nxv4i64.nxv4i64(
+  <vscale x 4 x i64>,
+  <vscale x 4 x i64>,
+  <vscale x 4 x i64>,
+  iXLen);
+
+define <vscale x 4 x i64> @intrinsic_vsrl_vv_nxv4i64_nxv4i64_nxv4i64(<vscale x 4 x i64> %0, <vscale x 4 x i64> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_vv_nxv4i64_nxv4i64_nxv4i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e64, m4, d1
+; CHECK-NEXT:    th.vsrl.vv v8, v8, v12
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i64> @llvm.riscv.th.vsrl.nxv4i64.nxv4i64(
+    <vscale x 4 x i64> undef,
+    <vscale x 4 x i64> %0,
+    <vscale x 4 x i64> %1,
+    iXLen %2)
+
+  ret <vscale x 4 x i64> %a
+}
+
+declare <vscale x 4 x i64> @llvm.riscv.th.vsrl.mask.nxv4i64.nxv4i64(
+  <vscale x 4 x i64>,
+  <vscale x 4 x i64>,
+  <vscale x 4 x i64>,
+  <vscale x 4 x i1>,
+  iXLen);
+
+define <vscale x 4 x i64> @intrinsic_vsrl_mask_vv_nxv4i64_nxv4i64_nxv4i64(<vscale x 4 x i64> %0, <vscale x 4 x i64> %1, <vscale x 4 x i64> %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_mask_vv_nxv4i64_nxv4i64_nxv4i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e64, m4, d1
+; CHECK-NEXT:    th.vsrl.vv v8, v12, v16, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i64> @llvm.riscv.th.vsrl.mask.nxv4i64.nxv4i64(
+    <vscale x 4 x i64> %0,
+    <vscale x 4 x i64> %1,
+    <vscale x 4 x i64> %2,
+    <vscale x 4 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 4 x i64> %a
+}
+
+declare <vscale x 8 x i64> @llvm.riscv.th.vsrl.nxv8i64.nxv8i64(
+  <vscale x 8 x i64>,
+  <vscale x 8 x i64>,
+  <vscale x 8 x i64>,
+  iXLen);
+
+define <vscale x 8 x i64> @intrinsic_vsrl_vv_nxv8i64_nxv8i64_nxv8i64(<vscale x 8 x i64> %0, <vscale x 8 x i64> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_vv_nxv8i64_nxv8i64_nxv8i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e64, m8, d1
+; CHECK-NEXT:    th.vsrl.vv v8, v8, v16
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i64> @llvm.riscv.th.vsrl.nxv8i64.nxv8i64(
+    <vscale x 8 x i64> undef,
+    <vscale x 8 x i64> %0,
+    <vscale x 8 x i64> %1,
+    iXLen %2)
+
+  ret <vscale x 8 x i64> %a
+}
+
+declare <vscale x 8 x i64> @llvm.riscv.th.vsrl.mask.nxv8i64.nxv8i64(
+  <vscale x 8 x i64>,
+  <vscale x 8 x i64>,
+  <vscale x 8 x i64>,
+  <vscale x 8 x i1>,
+  iXLen);
+
+define <vscale x 8 x i64> @intrinsic_vsrl_mask_vv_nxv8i64_nxv8i64_nxv8i64(<vscale x 8 x i64> %0, <vscale x 8 x i64> %1, <vscale x 8 x i64> %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_mask_vv_nxv8i64_nxv8i64_nxv8i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e64, m8, d1
+; CHECK-NEXT:    th.vle.v v24, (a0)
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a0, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a0, a2
+; CHECK-NEXT:    th.vsetvli zero, a1, e64, m8, d1
+; CHECK-NEXT:    th.vsrl.vv v8, v16, v24, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i64> @llvm.riscv.th.vsrl.mask.nxv8i64.nxv8i64(
+    <vscale x 8 x i64> %0,
+    <vscale x 8 x i64> %1,
+    <vscale x 8 x i64> %2,
+    <vscale x 8 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 8 x i64> %a
+}
+
+declare <vscale x 8 x i8> @llvm.riscv.th.vsrl.nxv8i8(
+  <vscale x 8 x i8>,
+  <vscale x 8 x i8>,
+  iXLen,
+  iXLen);
+
+define <vscale x 8 x i8> @intrinsic_vsrl_vx_nxv8i8_nxv8i8(<vscale x 8 x i8> %0, iXLen %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_vx_nxv8i8_nxv8i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e8, m1, d1
+; CHECK-NEXT:    th.vsrl.vx v8, v8, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i8> @llvm.riscv.th.vsrl.nxv8i8(
+    <vscale x 8 x i8> undef,
+    <vscale x 8 x i8> %0,
+    iXLen %1,
+    iXLen %2)
+
+  ret <vscale x 8 x i8> %a
+}
+
+declare <vscale x 8 x i8> @llvm.riscv.th.vsrl.mask.nxv8i8(
+  <vscale x 8 x i8>,
+  <vscale x 8 x i8>,
+  iXLen,
+  <vscale x 8 x i1>,
+  iXLen);
+
+define <vscale x 8 x i8> @intrinsic_vsrl_mask_vx_nxv8i8_nxv8i8(<vscale x 8 x i8> %0, <vscale x 8 x i8> %1, iXLen %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_mask_vx_nxv8i8_nxv8i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e8, m1, d1
+; CHECK-NEXT:    th.vsrl.vx v8, v9, a0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i8> @llvm.riscv.th.vsrl.mask.nxv8i8(
+    <vscale x 8 x i8> %0,
+    <vscale x 8 x i8> %1,
+    iXLen %2,
+    <vscale x 8 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 8 x i8> %a
+}
+
+declare <vscale x 16 x i8> @llvm.riscv.th.vsrl.nxv16i8(
+  <vscale x 16 x i8>,
+  <vscale x 16 x i8>,
+  iXLen,
+  iXLen);
+
+define <vscale x 16 x i8> @intrinsic_vsrl_vx_nxv16i8_nxv16i8(<vscale x 16 x i8> %0, iXLen %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_vx_nxv16i8_nxv16i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e8, m2, d1
+; CHECK-NEXT:    th.vsrl.vx v8, v8, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i8> @llvm.riscv.th.vsrl.nxv16i8(
+    <vscale x 16 x i8> undef,
+    <vscale x 16 x i8> %0,
+    iXLen %1,
+    iXLen %2)
+
+  ret <vscale x 16 x i8> %a
+}
+
+declare <vscale x 16 x i8> @llvm.riscv.th.vsrl.mask.nxv16i8(
+  <vscale x 16 x i8>,
+  <vscale x 16 x i8>,
+  iXLen,
+  <vscale x 16 x i1>,
+  iXLen);
+
+define <vscale x 16 x i8> @intrinsic_vsrl_mask_vx_nxv16i8_nxv16i8(<vscale x 16 x i8> %0, <vscale x 16 x i8> %1, iXLen %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_mask_vx_nxv16i8_nxv16i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e8, m2, d1
+; CHECK-NEXT:    th.vsrl.vx v8, v10, a0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i8> @llvm.riscv.th.vsrl.mask.nxv16i8(
+    <vscale x 16 x i8> %0,
+    <vscale x 16 x i8> %1,
+    iXLen %2,
+    <vscale x 16 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 16 x i8> %a
+}
+
+declare <vscale x 32 x i8> @llvm.riscv.th.vsrl.nxv32i8(
+  <vscale x 32 x i8>,
+  <vscale x 32 x i8>,
+  iXLen,
+  iXLen);
+
+define <vscale x 32 x i8> @intrinsic_vsrl_vx_nxv32i8_nxv32i8(<vscale x 32 x i8> %0, iXLen %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_vx_nxv32i8_nxv32i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e8, m4, d1
+; CHECK-NEXT:    th.vsrl.vx v8, v8, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x i8> @llvm.riscv.th.vsrl.nxv32i8(
+    <vscale x 32 x i8> undef,
+    <vscale x 32 x i8> %0,
+    iXLen %1,
+    iXLen %2)
+
+  ret <vscale x 32 x i8> %a
+}
+
+declare <vscale x 32 x i8> @llvm.riscv.th.vsrl.mask.nxv32i8(
+  <vscale x 32 x i8>,
+  <vscale x 32 x i8>,
+  iXLen,
+  <vscale x 32 x i1>,
+  iXLen);
+
+define <vscale x 32 x i8> @intrinsic_vsrl_mask_vx_nxv32i8_nxv32i8(<vscale x 32 x i8> %0, <vscale x 32 x i8> %1, iXLen %2, <vscale x 32 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_mask_vx_nxv32i8_nxv32i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e8, m4, d1
+; CHECK-NEXT:    th.vsrl.vx v8, v12, a0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x i8> @llvm.riscv.th.vsrl.mask.nxv32i8(
+    <vscale x 32 x i8> %0,
+    <vscale x 32 x i8> %1,
+    iXLen %2,
+    <vscale x 32 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 32 x i8> %a
+}
+
+declare <vscale x 64 x i8> @llvm.riscv.th.vsrl.nxv64i8(
+  <vscale x 64 x i8>,
+  <vscale x 64 x i8>,
+  iXLen,
+  iXLen);
+
+define <vscale x 64 x i8> @intrinsic_vsrl_vx_nxv64i8_nxv64i8(<vscale x 64 x i8> %0, iXLen %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_vx_nxv64i8_nxv64i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e8, m8, d1
+; CHECK-NEXT:    th.vsrl.vx v8, v8, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 64 x i8> @llvm.riscv.th.vsrl.nxv64i8(
+    <vscale x 64 x i8> undef,
+    <vscale x 64 x i8> %0,
+    iXLen %1,
+    iXLen %2)
+
+  ret <vscale x 64 x i8> %a
+}
+
+declare <vscale x 64 x i8> @llvm.riscv.th.vsrl.mask.nxv64i8(
+  <vscale x 64 x i8>,
+  <vscale x 64 x i8>,
+  iXLen,
+  <vscale x 64 x i1>,
+  iXLen);
+
+define <vscale x 64 x i8> @intrinsic_vsrl_mask_vx_nxv64i8_nxv64i8(<vscale x 64 x i8> %0, <vscale x 64 x i8> %1, iXLen %2, <vscale x 64 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_mask_vx_nxv64i8_nxv64i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e8, m8, d1
+; CHECK-NEXT:    th.vsrl.vx v8, v16, a0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 64 x i8> @llvm.riscv.th.vsrl.mask.nxv64i8(
+    <vscale x 64 x i8> %0,
+    <vscale x 64 x i8> %1,
+    iXLen %2,
+    <vscale x 64 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 64 x i8> %a
+}
+
+declare <vscale x 4 x i16> @llvm.riscv.th.vsrl.nxv4i16(
+  <vscale x 4 x i16>,
+  <vscale x 4 x i16>,
+  iXLen,
+  iXLen);
+
+define <vscale x 4 x i16> @intrinsic_vsrl_vx_nxv4i16_nxv4i16(<vscale x 4 x i16> %0, iXLen %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_vx_nxv4i16_nxv4i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e16, m1, d1
+; CHECK-NEXT:    th.vsrl.vx v8, v8, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i16> @llvm.riscv.th.vsrl.nxv4i16(
+    <vscale x 4 x i16> undef,
+    <vscale x 4 x i16> %0,
+    iXLen %1,
+    iXLen %2)
+
+  ret <vscale x 4 x i16> %a
+}
+
+declare <vscale x 4 x i16> @llvm.riscv.th.vsrl.mask.nxv4i16(
+  <vscale x 4 x i16>,
+  <vscale x 4 x i16>,
+  iXLen,
+  <vscale x 4 x i1>,
+  iXLen);
+
+define <vscale x 4 x i16> @intrinsic_vsrl_mask_vx_nxv4i16_nxv4i16(<vscale x 4 x i16> %0, <vscale x 4 x i16> %1, iXLen %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_mask_vx_nxv4i16_nxv4i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e16, m1, d1
+; CHECK-NEXT:    th.vsrl.vx v8, v9, a0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i16> @llvm.riscv.th.vsrl.mask.nxv4i16(
+    <vscale x 4 x i16> %0,
+    <vscale x 4 x i16> %1,
+    iXLen %2,
+    <vscale x 4 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 4 x i16> %a
+}
+
+declare <vscale x 8 x i16> @llvm.riscv.th.vsrl.nxv8i16(
+  <vscale x 8 x i16>,
+  <vscale x 8 x i16>,
+  iXLen,
+  iXLen);
+
+define <vscale x 8 x i16> @intrinsic_vsrl_vx_nxv8i16_nxv8i16(<vscale x 8 x i16> %0, iXLen %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_vx_nxv8i16_nxv8i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e16, m2, d1
+; CHECK-NEXT:    th.vsrl.vx v8, v8, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i16> @llvm.riscv.th.vsrl.nxv8i16(
+    <vscale x 8 x i16> undef,
+    <vscale x 8 x i16> %0,
+    iXLen %1,
+    iXLen %2)
+
+  ret <vscale x 8 x i16> %a
+}
+
+declare <vscale x 8 x i16> @llvm.riscv.th.vsrl.mask.nxv8i16(
+  <vscale x 8 x i16>,
+  <vscale x 8 x i16>,
+  iXLen,
+  <vscale x 8 x i1>,
+  iXLen);
+
+define <vscale x 8 x i16> @intrinsic_vsrl_mask_vx_nxv8i16_nxv8i16(<vscale x 8 x i16> %0, <vscale x 8 x i16> %1, iXLen %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_mask_vx_nxv8i16_nxv8i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e16, m2, d1
+; CHECK-NEXT:    th.vsrl.vx v8, v10, a0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i16> @llvm.riscv.th.vsrl.mask.nxv8i16(
+    <vscale x 8 x i16> %0,
+    <vscale x 8 x i16> %1,
+    iXLen %2,
+    <vscale x 8 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 8 x i16> %a
+}
+
+declare <vscale x 16 x i16> @llvm.riscv.th.vsrl.nxv16i16(
+  <vscale x 16 x i16>,
+  <vscale x 16 x i16>,
+  iXLen,
+  iXLen);
+
+define <vscale x 16 x i16> @intrinsic_vsrl_vx_nxv16i16_nxv16i16(<vscale x 16 x i16> %0, iXLen %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_vx_nxv16i16_nxv16i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e16, m4, d1
+; CHECK-NEXT:    th.vsrl.vx v8, v8, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i16> @llvm.riscv.th.vsrl.nxv16i16(
+    <vscale x 16 x i16> undef,
+    <vscale x 16 x i16> %0,
+    iXLen %1,
+    iXLen %2)
+
+  ret <vscale x 16 x i16> %a
+}
+
+declare <vscale x 16 x i16> @llvm.riscv.th.vsrl.mask.nxv16i16(
+  <vscale x 16 x i16>,
+  <vscale x 16 x i16>,
+  iXLen,
+  <vscale x 16 x i1>,
+  iXLen);
+
+define <vscale x 16 x i16> @intrinsic_vsrl_mask_vx_nxv16i16_nxv16i16(<vscale x 16 x i16> %0, <vscale x 16 x i16> %1, iXLen %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_mask_vx_nxv16i16_nxv16i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e16, m4, d1
+; CHECK-NEXT:    th.vsrl.vx v8, v12, a0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i16> @llvm.riscv.th.vsrl.mask.nxv16i16(
+    <vscale x 16 x i16> %0,
+    <vscale x 16 x i16> %1,
+    iXLen %2,
+    <vscale x 16 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 16 x i16> %a
+}
+
+declare <vscale x 32 x i16> @llvm.riscv.th.vsrl.nxv32i16(
+  <vscale x 32 x i16>,
+  <vscale x 32 x i16>,
+  iXLen,
+  iXLen);
+
+define <vscale x 32 x i16> @intrinsic_vsrl_vx_nxv32i16_nxv32i16(<vscale x 32 x i16> %0, iXLen %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_vx_nxv32i16_nxv32i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e16, m8, d1
+; CHECK-NEXT:    th.vsrl.vx v8, v8, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x i16> @llvm.riscv.th.vsrl.nxv32i16(
+    <vscale x 32 x i16> undef,
+    <vscale x 32 x i16> %0,
+    iXLen %1,
+    iXLen %2)
+
+  ret <vscale x 32 x i16> %a
+}
+
+declare <vscale x 32 x i16> @llvm.riscv.th.vsrl.mask.nxv32i16(
+  <vscale x 32 x i16>,
+  <vscale x 32 x i16>,
+  iXLen,
+  <vscale x 32 x i1>,
+  iXLen);
+
+define <vscale x 32 x i16> @intrinsic_vsrl_mask_vx_nxv32i16_nxv32i16(<vscale x 32 x i16> %0, <vscale x 32 x i16> %1, iXLen %2, <vscale x 32 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_mask_vx_nxv32i16_nxv32i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e16, m8, d1
+; CHECK-NEXT:    th.vsrl.vx v8, v16, a0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x i16> @llvm.riscv.th.vsrl.mask.nxv32i16(
+    <vscale x 32 x i16> %0,
+    <vscale x 32 x i16> %1,
+    iXLen %2,
+    <vscale x 32 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 32 x i16> %a
+}
+
+declare <vscale x 2 x i32> @llvm.riscv.th.vsrl.nxv2i32(
+  <vscale x 2 x i32>,
+  <vscale x 2 x i32>,
+  iXLen,
+  iXLen);
+
+define <vscale x 2 x i32> @intrinsic_vsrl_vx_nxv2i32_nxv2i32(<vscale x 2 x i32> %0, iXLen %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_vx_nxv2i32_nxv2i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e32, m1, d1
+; CHECK-NEXT:    th.vsrl.vx v8, v8, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i32> @llvm.riscv.th.vsrl.nxv2i32(
+    <vscale x 2 x i32> undef,
+    <vscale x 2 x i32> %0,
+    iXLen %1,
+    iXLen %2)
+
+  ret <vscale x 2 x i32> %a
+}
+
+declare <vscale x 2 x i32> @llvm.riscv.th.vsrl.mask.nxv2i32(
+  <vscale x 2 x i32>,
+  <vscale x 2 x i32>,
+  iXLen,
+  <vscale x 2 x i1>,
+  iXLen);
+
+define <vscale x 2 x i32> @intrinsic_vsrl_mask_vx_nxv2i32_nxv2i32(<vscale x 2 x i32> %0, <vscale x 2 x i32> %1, iXLen %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_mask_vx_nxv2i32_nxv2i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e32, m1, d1
+; CHECK-NEXT:    th.vsrl.vx v8, v9, a0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i32> @llvm.riscv.th.vsrl.mask.nxv2i32(
+    <vscale x 2 x i32> %0,
+    <vscale x 2 x i32> %1,
+    iXLen %2,
+    <vscale x 2 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 2 x i32> %a
+}
+
+declare <vscale x 4 x i32> @llvm.riscv.th.vsrl.nxv4i32(
+  <vscale x 4 x i32>,
+  <vscale x 4 x i32>,
+  iXLen,
+  iXLen);
+
+define <vscale x 4 x i32> @intrinsic_vsrl_vx_nxv4i32_nxv4i32(<vscale x 4 x i32> %0, iXLen %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_vx_nxv4i32_nxv4i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e32, m2, d1
+; CHECK-NEXT:    th.vsrl.vx v8, v8, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i32> @llvm.riscv.th.vsrl.nxv4i32(
+    <vscale x 4 x i32> undef,
+    <vscale x 4 x i32> %0,
+    iXLen %1,
+    iXLen %2)
+
+  ret <vscale x 4 x i32> %a
+}
+
+declare <vscale x 4 x i32> @llvm.riscv.th.vsrl.mask.nxv4i32(
+  <vscale x 4 x i32>,
+  <vscale x 4 x i32>,
+  iXLen,
+  <vscale x 4 x i1>,
+  iXLen);
+
+define <vscale x 4 x i32> @intrinsic_vsrl_mask_vx_nxv4i32_nxv4i32(<vscale x 4 x i32> %0, <vscale x 4 x i32> %1, iXLen %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_mask_vx_nxv4i32_nxv4i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e32, m2, d1
+; CHECK-NEXT:    th.vsrl.vx v8, v10, a0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i32> @llvm.riscv.th.vsrl.mask.nxv4i32(
+    <vscale x 4 x i32> %0,
+    <vscale x 4 x i32> %1,
+    iXLen %2,
+    <vscale x 4 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 4 x i32> %a
+}
+
+declare <vscale x 8 x i32> @llvm.riscv.th.vsrl.nxv8i32(
+  <vscale x 8 x i32>,
+  <vscale x 8 x i32>,
+  iXLen,
+  iXLen);
+
+define <vscale x 8 x i32> @intrinsic_vsrl_vx_nxv8i32_nxv8i32(<vscale x 8 x i32> %0, iXLen %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_vx_nxv8i32_nxv8i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e32, m4, d1
+; CHECK-NEXT:    th.vsrl.vx v8, v8, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i32> @llvm.riscv.th.vsrl.nxv8i32(
+    <vscale x 8 x i32> undef,
+    <vscale x 8 x i32> %0,
+    iXLen %1,
+    iXLen %2)
+
+  ret <vscale x 8 x i32> %a
+}
+
+declare <vscale x 8 x i32> @llvm.riscv.th.vsrl.mask.nxv8i32(
+  <vscale x 8 x i32>,
+  <vscale x 8 x i32>,
+  iXLen,
+  <vscale x 8 x i1>,
+  iXLen);
+
+define <vscale x 8 x i32> @intrinsic_vsrl_mask_vx_nxv8i32_nxv8i32(<vscale x 8 x i32> %0, <vscale x 8 x i32> %1, iXLen %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_mask_vx_nxv8i32_nxv8i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e32, m4, d1
+; CHECK-NEXT:    th.vsrl.vx v8, v12, a0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i32> @llvm.riscv.th.vsrl.mask.nxv8i32(
+    <vscale x 8 x i32> %0,
+    <vscale x 8 x i32> %1,
+    iXLen %2,
+    <vscale x 8 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 8 x i32> %a
+}
+
+declare <vscale x 16 x i32> @llvm.riscv.th.vsrl.nxv16i32(
+  <vscale x 16 x i32>,
+  <vscale x 16 x i32>,
+  iXLen,
+  iXLen);
+
+define <vscale x 16 x i32> @intrinsic_vsrl_vx_nxv16i32_nxv16i32(<vscale x 16 x i32> %0, iXLen %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_vx_nxv16i32_nxv16i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e32, m8, d1
+; CHECK-NEXT:    th.vsrl.vx v8, v8, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i32> @llvm.riscv.th.vsrl.nxv16i32(
+    <vscale x 16 x i32> undef,
+    <vscale x 16 x i32> %0,
+    iXLen %1,
+    iXLen %2)
+
+  ret <vscale x 16 x i32> %a
+}
+
+declare <vscale x 16 x i32> @llvm.riscv.th.vsrl.mask.nxv16i32(
+  <vscale x 16 x i32>,
+  <vscale x 16 x i32>,
+  iXLen,
+  <vscale x 16 x i1>,
+  iXLen);
+
+define <vscale x 16 x i32> @intrinsic_vsrl_mask_vx_nxv16i32_nxv16i32(<vscale x 16 x i32> %0, <vscale x 16 x i32> %1, iXLen %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_mask_vx_nxv16i32_nxv16i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e32, m8, d1
+; CHECK-NEXT:    th.vsrl.vx v8, v16, a0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i32> @llvm.riscv.th.vsrl.mask.nxv16i32(
+    <vscale x 16 x i32> %0,
+    <vscale x 16 x i32> %1,
+    iXLen %2,
+    <vscale x 16 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 16 x i32> %a
+}
+
+declare <vscale x 1 x i64> @llvm.riscv.th.vsrl.nxv1i64(
+  <vscale x 1 x i64>,
+  <vscale x 1 x i64>,
+  iXLen,
+  iXLen);
+
+define <vscale x 1 x i64> @intrinsic_vsrl_vx_nxv1i64_nxv1i64(<vscale x 1 x i64> %0, iXLen %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_vx_nxv1i64_nxv1i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e64, m1, d1
+; CHECK-NEXT:    th.vsrl.vx v8, v8, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x i64> @llvm.riscv.th.vsrl.nxv1i64(
+    <vscale x 1 x i64> undef,
+    <vscale x 1 x i64> %0,
+    iXLen %1,
+    iXLen %2)
+
+  ret <vscale x 1 x i64> %a
+}
+
+declare <vscale x 1 x i64> @llvm.riscv.th.vsrl.mask.nxv1i64(
+  <vscale x 1 x i64>,
+  <vscale x 1 x i64>,
+  iXLen,
+  <vscale x 1 x i1>,
+  iXLen);
+
+define <vscale x 1 x i64> @intrinsic_vsrl_mask_vx_nxv1i64_nxv1i64(<vscale x 1 x i64> %0, <vscale x 1 x i64> %1, iXLen %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_mask_vx_nxv1i64_nxv1i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e64, m1, d1
+; CHECK-NEXT:    th.vsrl.vx v8, v9, a0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x i64> @llvm.riscv.th.vsrl.mask.nxv1i64(
+    <vscale x 1 x i64> %0,
+    <vscale x 1 x i64> %1,
+    iXLen %2,
+    <vscale x 1 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 1 x i64> %a
+}
+
+declare <vscale x 2 x i64> @llvm.riscv.th.vsrl.nxv2i64(
+  <vscale x 2 x i64>,
+  <vscale x 2 x i64>,
+  iXLen,
+  iXLen);
+
+define <vscale x 2 x i64> @intrinsic_vsrl_vx_nxv2i64_nxv2i64(<vscale x 2 x i64> %0, iXLen %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_vx_nxv2i64_nxv2i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e64, m2, d1
+; CHECK-NEXT:    th.vsrl.vx v8, v8, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i64> @llvm.riscv.th.vsrl.nxv2i64(
+    <vscale x 2 x i64> undef,
+    <vscale x 2 x i64> %0,
+    iXLen %1,
+    iXLen %2)
+
+  ret <vscale x 2 x i64> %a
+}
+
+declare <vscale x 2 x i64> @llvm.riscv.th.vsrl.mask.nxv2i64(
+  <vscale x 2 x i64>,
+  <vscale x 2 x i64>,
+  iXLen,
+  <vscale x 2 x i1>,
+  iXLen);
+
+define <vscale x 2 x i64> @intrinsic_vsrl_mask_vx_nxv2i64_nxv2i64(<vscale x 2 x i64> %0, <vscale x 2 x i64> %1, iXLen %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_mask_vx_nxv2i64_nxv2i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e64, m2, d1
+; CHECK-NEXT:    th.vsrl.vx v8, v10, a0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i64> @llvm.riscv.th.vsrl.mask.nxv2i64(
+    <vscale x 2 x i64> %0,
+    <vscale x 2 x i64> %1,
+    iXLen %2,
+    <vscale x 2 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 2 x i64> %a
+}
+
+declare <vscale x 4 x i64> @llvm.riscv.th.vsrl.nxv4i64(
+  <vscale x 4 x i64>,
+  <vscale x 4 x i64>,
+  iXLen,
+  iXLen);
+
+define <vscale x 4 x i64> @intrinsic_vsrl_vx_nxv4i64_nxv4i64(<vscale x 4 x i64> %0, iXLen %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_vx_nxv4i64_nxv4i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e64, m4, d1
+; CHECK-NEXT:    th.vsrl.vx v8, v8, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i64> @llvm.riscv.th.vsrl.nxv4i64(
+    <vscale x 4 x i64> undef,
+    <vscale x 4 x i64> %0,
+    iXLen %1,
+    iXLen %2)
+
+  ret <vscale x 4 x i64> %a
+}
+
+declare <vscale x 4 x i64> @llvm.riscv.th.vsrl.mask.nxv4i64(
+  <vscale x 4 x i64>,
+  <vscale x 4 x i64>,
+  iXLen,
+  <vscale x 4 x i1>,
+  iXLen);
+
+define <vscale x 4 x i64> @intrinsic_vsrl_mask_vx_nxv4i64_nxv4i64(<vscale x 4 x i64> %0, <vscale x 4 x i64> %1, iXLen %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_mask_vx_nxv4i64_nxv4i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e64, m4, d1
+; CHECK-NEXT:    th.vsrl.vx v8, v12, a0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i64> @llvm.riscv.th.vsrl.mask.nxv4i64(
+    <vscale x 4 x i64> %0,
+    <vscale x 4 x i64> %1,
+    iXLen %2,
+    <vscale x 4 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 4 x i64> %a
+}
+
+declare <vscale x 8 x i64> @llvm.riscv.th.vsrl.nxv8i64(
+  <vscale x 8 x i64>,
+  <vscale x 8 x i64>,
+  iXLen,
+  iXLen);
+
+define <vscale x 8 x i64> @intrinsic_vsrl_vx_nxv8i64_nxv8i64(<vscale x 8 x i64> %0, iXLen %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_vx_nxv8i64_nxv8i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e64, m8, d1
+; CHECK-NEXT:    th.vsrl.vx v8, v8, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i64> @llvm.riscv.th.vsrl.nxv8i64(
+    <vscale x 8 x i64> undef,
+    <vscale x 8 x i64> %0,
+    iXLen %1,
+    iXLen %2)
+
+  ret <vscale x 8 x i64> %a
+}
+
+declare <vscale x 8 x i64> @llvm.riscv.th.vsrl.mask.nxv8i64(
+  <vscale x 8 x i64>,
+  <vscale x 8 x i64>,
+  iXLen,
+  <vscale x 8 x i1>,
+  iXLen);
+
+define <vscale x 8 x i64> @intrinsic_vsrl_mask_vx_nxv8i64_nxv8i64(<vscale x 8 x i64> %0, <vscale x 8 x i64> %1, iXLen %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_mask_vx_nxv8i64_nxv8i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    csrr a2, vl
+; CHECK-NEXT:    csrr a3, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a2, a3
+; CHECK-NEXT:    th.vsetvli zero, a1, e64, m8, d1
+; CHECK-NEXT:    th.vsrl.vx v8, v16, a0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i64> @llvm.riscv.th.vsrl.mask.nxv8i64(
+    <vscale x 8 x i64> %0,
+    <vscale x 8 x i64> %1,
+    iXLen %2,
+    <vscale x 8 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 8 x i64> %a
+}
+
+define <vscale x 8 x i8> @intrinsic_vsrl_vi_nxv8i8_nxv8i8_i8(<vscale x 8 x i8> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_vi_nxv8i8_nxv8i8_i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e8, m1, d1
+; CHECK-NEXT:    th.vsrl.vi v8, v8, 9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i8> @llvm.riscv.th.vsrl.nxv8i8(
+    <vscale x 8 x i8> undef,
+    <vscale x 8 x i8> %0,
+    iXLen 9,
+    iXLen %1)
+
+  ret <vscale x 8 x i8> %a
+}
+
+define <vscale x 8 x i8> @intrinsic_vsrl_mask_vi_nxv8i8_nxv8i8_i8(<vscale x 8 x i8> %0, <vscale x 8 x i8> %1, <vscale x 8 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_mask_vi_nxv8i8_nxv8i8_i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e8, m1, d1
+; CHECK-NEXT:    th.vsrl.vi v8, v9, 9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i8> @llvm.riscv.th.vsrl.mask.nxv8i8(
+    <vscale x 8 x i8> %0,
+    <vscale x 8 x i8> %1,
+    iXLen 9,
+    <vscale x 8 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 8 x i8> %a
+}
+
+define <vscale x 16 x i8> @intrinsic_vsrl_vi_nxv16i8_nxv16i8_i8(<vscale x 16 x i8> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_vi_nxv16i8_nxv16i8_i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e8, m2, d1
+; CHECK-NEXT:    th.vsrl.vi v8, v8, 9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i8> @llvm.riscv.th.vsrl.nxv16i8(
+    <vscale x 16 x i8> undef,
+    <vscale x 16 x i8> %0,
+    iXLen 9,
+    iXLen %1)
+
+  ret <vscale x 16 x i8> %a
+}
+
+define <vscale x 16 x i8> @intrinsic_vsrl_mask_vi_nxv16i8_nxv16i8_i8(<vscale x 16 x i8> %0, <vscale x 16 x i8> %1, <vscale x 16 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_mask_vi_nxv16i8_nxv16i8_i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e8, m2, d1
+; CHECK-NEXT:    th.vsrl.vi v8, v10, 9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i8> @llvm.riscv.th.vsrl.mask.nxv16i8(
+    <vscale x 16 x i8> %0,
+    <vscale x 16 x i8> %1,
+    iXLen 9,
+    <vscale x 16 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 16 x i8> %a
+}
+
+define <vscale x 32 x i8> @intrinsic_vsrl_vi_nxv32i8_nxv32i8_i8(<vscale x 32 x i8> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_vi_nxv32i8_nxv32i8_i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e8, m4, d1
+; CHECK-NEXT:    th.vsrl.vi v8, v8, 9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x i8> @llvm.riscv.th.vsrl.nxv32i8(
+    <vscale x 32 x i8> undef,
+    <vscale x 32 x i8> %0,
+    iXLen 9,
+    iXLen %1)
+
+  ret <vscale x 32 x i8> %a
+}
+
+define <vscale x 32 x i8> @intrinsic_vsrl_mask_vi_nxv32i8_nxv32i8_i8(<vscale x 32 x i8> %0, <vscale x 32 x i8> %1, <vscale x 32 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_mask_vi_nxv32i8_nxv32i8_i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e8, m4, d1
+; CHECK-NEXT:    th.vsrl.vi v8, v12, 9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x i8> @llvm.riscv.th.vsrl.mask.nxv32i8(
+    <vscale x 32 x i8> %0,
+    <vscale x 32 x i8> %1,
+    iXLen 9,
+    <vscale x 32 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 32 x i8> %a
+}
+
+define <vscale x 64 x i8> @intrinsic_vsrl_vi_nxv64i8_nxv64i8_i8(<vscale x 64 x i8> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_vi_nxv64i8_nxv64i8_i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e8, m8, d1
+; CHECK-NEXT:    th.vsrl.vi v8, v8, 9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 64 x i8> @llvm.riscv.th.vsrl.nxv64i8(
+    <vscale x 64 x i8> undef,
+    <vscale x 64 x i8> %0,
+    iXLen 9,
+    iXLen %1)
+
+  ret <vscale x 64 x i8> %a
+}
+
+define <vscale x 64 x i8> @intrinsic_vsrl_mask_vi_nxv64i8_nxv64i8_i8(<vscale x 64 x i8> %0, <vscale x 64 x i8> %1, <vscale x 64 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_mask_vi_nxv64i8_nxv64i8_i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e8, m8, d1
+; CHECK-NEXT:    th.vsrl.vi v8, v16, 9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 64 x i8> @llvm.riscv.th.vsrl.mask.nxv64i8(
+    <vscale x 64 x i8> %0,
+    <vscale x 64 x i8> %1,
+    iXLen 9,
+    <vscale x 64 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 64 x i8> %a
+}
+
+define <vscale x 4 x i16> @intrinsic_vsrl_vi_nxv4i16_nxv4i16_i16(<vscale x 4 x i16> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_vi_nxv4i16_nxv4i16_i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e16, m1, d1
+; CHECK-NEXT:    th.vsrl.vi v8, v8, 9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i16> @llvm.riscv.th.vsrl.nxv4i16(
+    <vscale x 4 x i16> undef,
+    <vscale x 4 x i16> %0,
+    iXLen 9,
+    iXLen %1)
+
+  ret <vscale x 4 x i16> %a
+}
+
+define <vscale x 4 x i16> @intrinsic_vsrl_mask_vi_nxv4i16_nxv4i16_i16(<vscale x 4 x i16> %0, <vscale x 4 x i16> %1, <vscale x 4 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_mask_vi_nxv4i16_nxv4i16_i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e16, m1, d1
+; CHECK-NEXT:    th.vsrl.vi v8, v9, 9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i16> @llvm.riscv.th.vsrl.mask.nxv4i16(
+    <vscale x 4 x i16> %0,
+    <vscale x 4 x i16> %1,
+    iXLen 9,
+    <vscale x 4 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 4 x i16> %a
+}
+
+define <vscale x 8 x i16> @intrinsic_vsrl_vi_nxv8i16_nxv8i16_i16(<vscale x 8 x i16> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_vi_nxv8i16_nxv8i16_i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e16, m2, d1
+; CHECK-NEXT:    th.vsrl.vi v8, v8, 9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i16> @llvm.riscv.th.vsrl.nxv8i16(
+    <vscale x 8 x i16> undef,
+    <vscale x 8 x i16> %0,
+    iXLen 9,
+    iXLen %1)
+
+  ret <vscale x 8 x i16> %a
+}
+
+define <vscale x 8 x i16> @intrinsic_vsrl_mask_vi_nxv8i16_nxv8i16_i16(<vscale x 8 x i16> %0, <vscale x 8 x i16> %1, <vscale x 8 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_mask_vi_nxv8i16_nxv8i16_i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e16, m2, d1
+; CHECK-NEXT:    th.vsrl.vi v8, v10, 9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i16> @llvm.riscv.th.vsrl.mask.nxv8i16(
+    <vscale x 8 x i16> %0,
+    <vscale x 8 x i16> %1,
+    iXLen 9,
+    <vscale x 8 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 8 x i16> %a
+}
+
+define <vscale x 16 x i16> @intrinsic_vsrl_vi_nxv16i16_nxv16i16_i16(<vscale x 16 x i16> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_vi_nxv16i16_nxv16i16_i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e16, m4, d1
+; CHECK-NEXT:    th.vsrl.vi v8, v8, 9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i16> @llvm.riscv.th.vsrl.nxv16i16(
+    <vscale x 16 x i16> undef,
+    <vscale x 16 x i16> %0,
+    iXLen 9,
+    iXLen %1)
+
+  ret <vscale x 16 x i16> %a
+}
+
+define <vscale x 16 x i16> @intrinsic_vsrl_mask_vi_nxv16i16_nxv16i16_i16(<vscale x 16 x i16> %0, <vscale x 16 x i16> %1, <vscale x 16 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_mask_vi_nxv16i16_nxv16i16_i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e16, m4, d1
+; CHECK-NEXT:    th.vsrl.vi v8, v12, 9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i16> @llvm.riscv.th.vsrl.mask.nxv16i16(
+    <vscale x 16 x i16> %0,
+    <vscale x 16 x i16> %1,
+    iXLen 9,
+    <vscale x 16 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 16 x i16> %a
+}
+
+define <vscale x 32 x i16> @intrinsic_vsrl_vi_nxv32i16_nxv32i16_i16(<vscale x 32 x i16> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_vi_nxv32i16_nxv32i16_i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e16, m8, d1
+; CHECK-NEXT:    th.vsrl.vi v8, v8, 9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x i16> @llvm.riscv.th.vsrl.nxv32i16(
+    <vscale x 32 x i16> undef,
+    <vscale x 32 x i16> %0,
+    iXLen 9,
+    iXLen %1)
+
+  ret <vscale x 32 x i16> %a
+}
+
+define <vscale x 32 x i16> @intrinsic_vsrl_mask_vi_nxv32i16_nxv32i16_i16(<vscale x 32 x i16> %0, <vscale x 32 x i16> %1, <vscale x 32 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_mask_vi_nxv32i16_nxv32i16_i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e16, m8, d1
+; CHECK-NEXT:    th.vsrl.vi v8, v16, 9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x i16> @llvm.riscv.th.vsrl.mask.nxv32i16(
+    <vscale x 32 x i16> %0,
+    <vscale x 32 x i16> %1,
+    iXLen 9,
+    <vscale x 32 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 32 x i16> %a
+}
+
+define <vscale x 2 x i32> @intrinsic_vsrl_vi_nxv2i32_nxv2i32_i32(<vscale x 2 x i32> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_vi_nxv2i32_nxv2i32_i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e32, m1, d1
+; CHECK-NEXT:    th.vsrl.vi v8, v8, 9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i32> @llvm.riscv.th.vsrl.nxv2i32(
+    <vscale x 2 x i32> undef,
+    <vscale x 2 x i32> %0,
+    iXLen 9,
+    iXLen %1)
+
+  ret <vscale x 2 x i32> %a
+}
+
+define <vscale x 2 x i32> @intrinsic_vsrl_mask_vi_nxv2i32_nxv2i32_i32(<vscale x 2 x i32> %0, <vscale x 2 x i32> %1, <vscale x 2 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_mask_vi_nxv2i32_nxv2i32_i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e32, m1, d1
+; CHECK-NEXT:    th.vsrl.vi v8, v9, 9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i32> @llvm.riscv.th.vsrl.mask.nxv2i32(
+    <vscale x 2 x i32> %0,
+    <vscale x 2 x i32> %1,
+    iXLen 9,
+    <vscale x 2 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 2 x i32> %a
+}
+
+define <vscale x 4 x i32> @intrinsic_vsrl_vi_nxv4i32_nxv4i32_i32(<vscale x 4 x i32> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_vi_nxv4i32_nxv4i32_i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e32, m2, d1
+; CHECK-NEXT:    th.vsrl.vi v8, v8, 9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i32> @llvm.riscv.th.vsrl.nxv4i32(
+    <vscale x 4 x i32> undef,
+    <vscale x 4 x i32> %0,
+    iXLen 9,
+    iXLen %1)
+
+  ret <vscale x 4 x i32> %a
+}
+
+define <vscale x 4 x i32> @intrinsic_vsrl_mask_vi_nxv4i32_nxv4i32_i32(<vscale x 4 x i32> %0, <vscale x 4 x i32> %1, <vscale x 4 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_mask_vi_nxv4i32_nxv4i32_i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e32, m2, d1
+; CHECK-NEXT:    th.vsrl.vi v8, v10, 9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i32> @llvm.riscv.th.vsrl.mask.nxv4i32(
+    <vscale x 4 x i32> %0,
+    <vscale x 4 x i32> %1,
+    iXLen 9,
+    <vscale x 4 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 4 x i32> %a
+}
+
+define <vscale x 8 x i32> @intrinsic_vsrl_vi_nxv8i32_nxv8i32_i32(<vscale x 8 x i32> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_vi_nxv8i32_nxv8i32_i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e32, m4, d1
+; CHECK-NEXT:    th.vsrl.vi v8, v8, 9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i32> @llvm.riscv.th.vsrl.nxv8i32(
+    <vscale x 8 x i32> undef,
+    <vscale x 8 x i32> %0,
+    iXLen 9,
+    iXLen %1)
+
+  ret <vscale x 8 x i32> %a
+}
+
+define <vscale x 8 x i32> @intrinsic_vsrl_mask_vi_nxv8i32_nxv8i32_i32(<vscale x 8 x i32> %0, <vscale x 8 x i32> %1, <vscale x 8 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_mask_vi_nxv8i32_nxv8i32_i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e32, m4, d1
+; CHECK-NEXT:    th.vsrl.vi v8, v12, 9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i32> @llvm.riscv.th.vsrl.mask.nxv8i32(
+    <vscale x 8 x i32> %0,
+    <vscale x 8 x i32> %1,
+    iXLen 9,
+    <vscale x 8 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 8 x i32> %a
+}
+
+define <vscale x 16 x i32> @intrinsic_vsrl_vi_nxv16i32_nxv16i32_i32(<vscale x 16 x i32> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_vi_nxv16i32_nxv16i32_i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e32, m8, d1
+; CHECK-NEXT:    th.vsrl.vi v8, v8, 9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i32> @llvm.riscv.th.vsrl.nxv16i32(
+    <vscale x 16 x i32> undef,
+    <vscale x 16 x i32> %0,
+    iXLen 9,
+    iXLen %1)
+
+  ret <vscale x 16 x i32> %a
+}
+
+define <vscale x 16 x i32> @intrinsic_vsrl_mask_vi_nxv16i32_nxv16i32_i32(<vscale x 16 x i32> %0, <vscale x 16 x i32> %1, <vscale x 16 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_mask_vi_nxv16i32_nxv16i32_i32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e32, m8, d1
+; CHECK-NEXT:    th.vsrl.vi v8, v16, 9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i32> @llvm.riscv.th.vsrl.mask.nxv16i32(
+    <vscale x 16 x i32> %0,
+    <vscale x 16 x i32> %1,
+    iXLen 9,
+    <vscale x 16 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 16 x i32> %a
+}
+
+define <vscale x 1 x i64> @intrinsic_vsrl_vi_nxv1i64_nxv1i64_i64(<vscale x 1 x i64> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_vi_nxv1i64_nxv1i64_i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e64, m1, d1
+; CHECK-NEXT:    th.vsrl.vi v8, v8, 9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x i64> @llvm.riscv.th.vsrl.nxv1i64(
+    <vscale x 1 x i64> undef,
+    <vscale x 1 x i64> %0,
+    iXLen 9,
+    iXLen %1)
+
+  ret <vscale x 1 x i64> %a
+}
+
+define <vscale x 1 x i64> @intrinsic_vsrl_mask_vi_nxv1i64_nxv1i64_i64(<vscale x 1 x i64> %0, <vscale x 1 x i64> %1, <vscale x 1 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_mask_vi_nxv1i64_nxv1i64_i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e64, m1, d1
+; CHECK-NEXT:    th.vsrl.vi v8, v9, 9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x i64> @llvm.riscv.th.vsrl.mask.nxv1i64(
+    <vscale x 1 x i64> %0,
+    <vscale x 1 x i64> %1,
+    iXLen 9,
+    <vscale x 1 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 1 x i64> %a
+}
+
+define <vscale x 2 x i64> @intrinsic_vsrl_vi_nxv2i64_nxv2i64_i64(<vscale x 2 x i64> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_vi_nxv2i64_nxv2i64_i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e64, m2, d1
+; CHECK-NEXT:    th.vsrl.vi v8, v8, 9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i64> @llvm.riscv.th.vsrl.nxv2i64(
+    <vscale x 2 x i64> undef,
+    <vscale x 2 x i64> %0,
+    iXLen 9,
+    iXLen %1)
+
+  ret <vscale x 2 x i64> %a
+}
+
+define <vscale x 2 x i64> @intrinsic_vsrl_mask_vi_nxv2i64_nxv2i64_i64(<vscale x 2 x i64> %0, <vscale x 2 x i64> %1, <vscale x 2 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_mask_vi_nxv2i64_nxv2i64_i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e64, m2, d1
+; CHECK-NEXT:    th.vsrl.vi v8, v10, 9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i64> @llvm.riscv.th.vsrl.mask.nxv2i64(
+    <vscale x 2 x i64> %0,
+    <vscale x 2 x i64> %1,
+    iXLen 9,
+    <vscale x 2 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 2 x i64> %a
+}
+
+define <vscale x 4 x i64> @intrinsic_vsrl_vi_nxv4i64_nxv4i64_i64(<vscale x 4 x i64> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_vi_nxv4i64_nxv4i64_i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e64, m4, d1
+; CHECK-NEXT:    th.vsrl.vi v8, v8, 9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i64> @llvm.riscv.th.vsrl.nxv4i64(
+    <vscale x 4 x i64> undef,
+    <vscale x 4 x i64> %0,
+    iXLen 9,
+    iXLen %1)
+
+  ret <vscale x 4 x i64> %a
+}
+
+define <vscale x 4 x i64> @intrinsic_vsrl_mask_vi_nxv4i64_nxv4i64_i64(<vscale x 4 x i64> %0, <vscale x 4 x i64> %1, <vscale x 4 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_mask_vi_nxv4i64_nxv4i64_i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e64, m4, d1
+; CHECK-NEXT:    th.vsrl.vi v8, v12, 9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i64> @llvm.riscv.th.vsrl.mask.nxv4i64(
+    <vscale x 4 x i64> %0,
+    <vscale x 4 x i64> %1,
+    iXLen 9,
+    <vscale x 4 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 4 x i64> %a
+}
+
+define <vscale x 8 x i64> @intrinsic_vsrl_vi_nxv8i64_nxv8i64_i64(<vscale x 8 x i64> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_vi_nxv8i64_nxv8i64_i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e64, m8, d1
+; CHECK-NEXT:    th.vsrl.vi v8, v8, 9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i64> @llvm.riscv.th.vsrl.nxv8i64(
+    <vscale x 8 x i64> undef,
+    <vscale x 8 x i64> %0,
+    iXLen 9,
+    iXLen %1)
+
+  ret <vscale x 8 x i64> %a
+}
+
+define <vscale x 8 x i64> @intrinsic_vsrl_mask_vi_nxv8i64_nxv8i64_i64(<vscale x 8 x i64> %0, <vscale x 8 x i64> %1, <vscale x 8 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vsrl_mask_vi_nxv8i64_nxv8i64_i64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    csrr a1, vl
+; CHECK-NEXT:    csrr a2, vtype
+; CHECK-NEXT:    th.vsetvli zero, zero, e8, m1, d1
+; CHECK-NEXT:    th.vsetvl zero, a1, a2
+; CHECK-NEXT:    th.vsetvli zero, a0, e64, m8, d1
+; CHECK-NEXT:    th.vsrl.vi v8, v16, 9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i64> @llvm.riscv.th.vsrl.mask.nxv8i64(
+    <vscale x 8 x i64> %0,
+    <vscale x 8 x i64> %1,
+    iXLen 9,
+    <vscale x 8 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 8 x i64> %a
+}