[RISCV][InsertVSETVLI] Support constant VLs larger than immediate encoding

[reames]

Summary

The immediate field on the vsetivli is fairly limited. For larger vectors, we end up having to materialize a constant in a register. We hadn't plumbed the infrastructure to treat such materialized constants as constants for purpose of vsetvli elimination.

I only bothered to handle LI. We could extend this to LUI sequences, but well, 2048 elements is probably enough for all practical fixed length vector codegen. :)

The test delta does point out a related problem. At LMUL8, we see increased register allocation pressure, and we should probably either a) address register allocation remat, or b) be less aggressive about eliminating vsetvlis at high lmul. Note that high LMUL code is not generated much by default.

[luke]

Makes sense to me. As you say LMUL=8 is pretty rare, and in those cases I presume we would want to be using scalable vectors anyway

llvm/lib/Target/RISCV/RISCVInsertVSETVLI.cpp

375

Duplicate ;

[craig.topper]

llvm/test/CodeGen/RISCV/rvv/fixed-vectors-masked-gather.ll

12509

Is this instruction dead? Do we need to try to delete these li instructions or run DCE after vsetvli?

[reames]

Address style comment

[reames]

llvm/test/CodeGen/RISCV/rvv/fixed-vectors-masked-gather.ll

12509

It looks like it yes. We already kill other vsetvlis with register VLs, so it's possible we can already see this elsewhere. Note sure about that to be honest; I can't quickly come up with a test case where we would.

If you okay with it, I'd like to address this in a following patch.

[craig.topper]

LGTM

llvm/test/CodeGen/RISCV/rvv/fixed-vectors-masked-gather.ll

12509

We can address it as a follow up.

llvm/lib/Target/RISCV/RISCVInsertVSETVLI.cpp

public:
   unsigned getAVLImm() const {
     assert(hasAVLImm());
     return AVLImm;
   }
 
   unsigned getSEW() const { return SEW; }
   RISCVII::VLMUL getVLMUL() const { return VLMul; }
 
-  bool hasNonZeroAVL() const {
+  bool hasNonZeroAVL(const MachineRegisterInfo &MRI) const {
     if (hasAVLImm())
       return getAVLImm() > 0;
-    if (hasAVLReg())
-      return getAVLReg() == RISCV::X0;
+    if (hasAVLReg()) {
+      if (getAVLReg() == RISCV::X0)
+        return true;
+      if (MachineInstr *MI = MRI.getVRegDef(getAVLReg());
+          MI && MI->getOpcode() == RISCV::ADDI &&
+          MI->getOperand(1).isReg() && MI->getOperand(2).isImm() &&
+          MI->getOperand(1).getReg() == RISCV::X0 &&
+          MI->getOperand(2).getImm() != 0)
+        return true;
+      return false;;
+    }
     return false;
   }
 
-  bool hasEquallyZeroAVL(const VSETVLIInfo &Other) const {
+  bool hasEquallyZeroAVL(const VSETVLIInfo &Other,
+                         const MachineRegisterInfo &MRI) const {
     if (hasSameAVL(Other))
       return true;
-    return (hasNonZeroAVL() && Other.hasNonZeroAVL());
+    return (hasNonZeroAVL(MRI) && Other.hasNonZeroAVL(MRI));
   }
 
   bool hasSameAVL(const VSETVLIInfo &Other) const {
     if (hasAVLReg() && Other.hasAVLReg())
       return getAVLReg() == Other.getAVLReg();
 
     if (hasAVLImm() && Other.hasAVLImm())
       return getAVLImm() == Other.getAVLImm();
public:
   bool hasCompatibleVTYPE(const DemandedFields &Used,
                           const VSETVLIInfo &Require) const {
     return areCompatibleVTYPEs(encodeVTYPE(), Require.encodeVTYPE(), Used);
   }
 
   // Determine whether the vector instructions requirements represented by
   // Require are compatible with the previous vsetvli instruction represented
   // by this.  MI is the instruction whose requirements we're considering.
-  bool isCompatible(const DemandedFields &Used, const VSETVLIInfo &Require) const {
+  bool isCompatible(const DemandedFields &Used, const VSETVLIInfo &Require,
+                    const MachineRegisterInfo &MRI) const {
     assert(isValid() && Require.isValid() &&
            "Can't compare invalid VSETVLIInfos");
     assert(!Require.SEWLMULRatioOnly &&
            "Expected a valid VTYPE for instruction!");
     // Nothing is compatible with Unknown.
     if (isUnknown() || Require.isUnknown())
       return false;
 
     // If only our VLMAX ratio is valid, then this isn't compatible.
     if (SEWLMULRatioOnly)
       return false;
 
     // If the instruction doesn't need an AVLReg and the SEW matches, consider
     // it compatible.
     if (Require.hasAVLReg() && Require.AVLReg == RISCV::NoRegister)
       if (SEW == Require.SEW)
         return true;
 
     if (Used.VLAny && !hasSameAVL(Require))
       return false;
 
-    if (Used.VLZeroness && !hasEquallyZeroAVL(Require))
+    if (Used.VLZeroness && !hasEquallyZeroAVL(Require, MRI))
       return false;
 
     return areCompatibleVTYPEs(encodeVTYPE(), Require.encodeVTYPE(), Used);
   }
 
   bool operator==(const VSETVLIInfo &Other) const {
     // Uninitialized is only equal to another Uninitialized.
     if (!isValid())
if (isVSlideInstr(MI) && Require.hasAVLImm() && Require.getAVLImm() == 1 &&
     if (VRegDef && VRegDef->isImplicitDef()) {
       Used.VLAny = false;
       Used.VLZeroness = true;
       Used.LMUL = false;
       Used.TailPolicy = false;
     }
   }
 
-  if (CurInfo.isCompatible(Used, Require))
+  if (CurInfo.isCompatible(Used, Require, *MRI))
     return false;
 
   // We didn't find a compatible value. If our AVL is a virtual register,
   // it might be defined by a VSET(I)VLI. If it has the same VLMAX we need
   // and the last VL/VTYPE we observed is the same, we don't need a
   // VSETVLI here.
   if (Require.hasAVLReg() && Require.getAVLReg().isVirtual() &&
       CurInfo.hasCompatibleVTYPE(Used, Require)) {
void RISCVInsertVSETVLI::transferBefore(VSETVLIInfo &Info, const MachineInstr &MI) {
   // VL > 0. We can discard the user requested AVL and just use the last
   // one if we can prove it equally zero.  This removes a vsetvli entirely
   // if the types match or allows use of cheaper avl preserving variant
   // if VLMAX doesn't change.  If VLMAX might change, we couldn't use
   // the 'vsetvli x0, x0, vtype" variant, so we avoid the transform to
   // prevent extending live range of an avl register operand.
   // TODO: We can probably relax this for immediates.
   if (isScalarMoveInstr(MI) && PrevInfo.isValid() &&
-      PrevInfo.hasEquallyZeroAVL(Info) &&
+      PrevInfo.hasEquallyZeroAVL(Info, *MRI) &&
       Info.hasSameVLMAX(PrevInfo)) {
     if (PrevInfo.hasAVLImm())
       Info.setAVLImm(PrevInfo.getAVLImm());
     else
       Info.setAVLReg(PrevInfo.getAVLReg());
     return;
   }
 

llvm/test/CodeGen/RISCV/rvv/fixed-vectors-fp-interleave.ll

 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
-; RUN: llc -mtriple=riscv32 -mattr=+v,+zfh,+experimental-zvfh -riscv-v-vector-bits-min=128 -verify-machineinstrs < %s | FileCheck %s --check-prefixes=CHECK,V128,RV32-V128
-; RUN: llc -mtriple=riscv64 -mattr=+v,+zfh,+experimental-zvfh -riscv-v-vector-bits-min=128 -verify-machineinstrs < %s | FileCheck %s --check-prefixes=CHECK,V128,RV64-V128
-; RUN: llc -mtriple=riscv32 -mattr=+v,+zfh,+experimental-zvfh -riscv-v-vector-bits-min=512 -verify-machineinstrs < %s | FileCheck %s --check-prefixes=CHECK,V512,RV32-V512
-; RUN: llc -mtriple=riscv64 -mattr=+v,+zfh,+experimental-zvfh -riscv-v-vector-bits-min=512 -verify-machineinstrs < %s | FileCheck %s --check-prefixes=CHECK,V512,RV64-V512
+; RUN: llc -mtriple=riscv32 -mattr=+v,+zfh,+experimental-zvfh,+m -riscv-v-vector-bits-min=128 -verify-machineinstrs < %s | FileCheck %s --check-prefixes=CHECK,V128,RV32-V128
+; RUN: llc -mtriple=riscv64 -mattr=+v,+zfh,+experimental-zvfh,+m -riscv-v-vector-bits-min=128 -verify-machineinstrs < %s | FileCheck %s --check-prefixes=CHECK,V128,RV64-V128
+; RUN: llc -mtriple=riscv32 -mattr=+v,+zfh,+experimental-zvfh,+m -riscv-v-vector-bits-min=512 -verify-machineinstrs < %s | FileCheck %s --check-prefixes=CHECK,V512,RV32-V512
+; RUN: llc -mtriple=riscv64 -mattr=+v,+zfh,+experimental-zvfh,+m -riscv-v-vector-bits-min=512 -verify-machineinstrs < %s | FileCheck %s --check-prefixes=CHECK,V512,RV64-V512
 
 ; Test optimizing interleaves to widening arithmetic.
 
 define <4 x half> @interleave_v2f16(<2 x half> %x, <2 x half> %y) {
 ; CHECK-LABEL: interleave_v2f16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetivli zero, 2, e16, mf4, ta, ma
 ; CHECK-NEXT:    vwaddu.vv v10, v8, v9

 }
 
 define <64 x float> @interleave_v32f32(<32 x float> %x, <32 x float> %y) {
 ; RV32-V128-LABEL: interleave_v32f32:
 ; RV32-V128:       # %bb.0:
 ; RV32-V128-NEXT:    addi sp, sp, -16
 ; RV32-V128-NEXT:    .cfi_def_cfa_offset 16
 ; RV32-V128-NEXT:    csrr a0, vlenb
-; RV32-V128-NEXT:    slli a0, a0, 4
+; RV32-V128-NEXT:    li a1, 24
+; RV32-V128-NEXT:    mul a0, a0, a1
 ; RV32-V128-NEXT:    sub sp, sp, a0
-; RV32-V128-NEXT:    .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x10, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 16 * vlenb
+; RV32-V128-NEXT:    .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x18, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 24 * vlenb
+; RV32-V128-NEXT:    csrr a0, vlenb
+; RV32-V128-NEXT:    slli a0, a0, 3
+; RV32-V128-NEXT:    add a0, sp, a0
+; RV32-V128-NEXT:    addi a0, a0, 16
+; RV32-V128-NEXT:    vs8r.v v16, (a0) # Unknown-size Folded Spill
 ; RV32-V128-NEXT:    lui a0, %hi(.LCPI10_0)
 ; RV32-V128-NEXT:    addi a0, a0, %lo(.LCPI10_0)
 ; RV32-V128-NEXT:    li a1, 32
-; RV32-V128-NEXT:    vsetvli zero, a1, e32, m8, ta, ma
-; RV32-V128-NEXT:    vle32.v v0, (a0)
-; RV32-V128-NEXT:    vmv8r.v v24, v8
-; RV32-V128-NEXT:    vrgather.vv v8, v24, v0
-; RV32-V128-NEXT:    addi a0, sp, 16
-; RV32-V128-NEXT:    vs8r.v v24, (a0) # Unknown-size Folded Spill
+; RV32-V128-NEXT:    vsetvli zero, a1, e32, m8, ta, mu
+; RV32-V128-NEXT:    vle32.v v24, (a0)
 ; RV32-V128-NEXT:    lui a0, %hi(.LCPI10_1)
 ; RV32-V128-NEXT:    addi a0, a0, %lo(.LCPI10_1)
-; RV32-V128-NEXT:    vle32.v v24, (a0)
+; RV32-V128-NEXT:    vle32.v v16, (a0)
 ; RV32-V128-NEXT:    csrr a0, vlenb
-; RV32-V128-NEXT:    slli a0, a0, 3
+; RV32-V128-NEXT:    slli a0, a0, 4
 ; RV32-V128-NEXT:    add a0, sp, a0
 ; RV32-V128-NEXT:    addi a0, a0, 16
-; RV32-V128-NEXT:    vs8r.v v24, (a0) # Unknown-size Folded Spill
+; RV32-V128-NEXT:    vs8r.v v16, (a0) # Unknown-size Folded Spill
 ; RV32-V128-NEXT:    lui a0, 699051
 ; RV32-V128-NEXT:    addi a0, a0, -1366
-; RV32-V128-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
 ; RV32-V128-NEXT:    vmv.s.x v0, a0
-; RV32-V128-NEXT:    vsetvli zero, a1, e32, m8, ta, mu
+; RV32-V128-NEXT:    vrgather.vv v16, v8, v24
+; RV32-V128-NEXT:    addi a0, sp, 16
+; RV32-V128-NEXT:    vs8r.v v8, (a0) # Unknown-size Folded Spill
 ; RV32-V128-NEXT:    csrr a0, vlenb
-; RV32-V128-NEXT:    slli a0, a0, 3
+; RV32-V128-NEXT:    slli a0, a0, 4
 ; RV32-V128-NEXT:    add a0, sp, a0
 ; RV32-V128-NEXT:    addi a0, a0, 16
 ; RV32-V128-NEXT:    vl8r.v v24, (a0) # Unknown-size Folded Reload
-; RV32-V128-NEXT:    vrgather.vv v8, v16, v24, v0.t
-; RV32-V128-NEXT:    vmv.v.v v24, v8
+; RV32-V128-NEXT:    csrr a0, vlenb
+; RV32-V128-NEXT:    slli a0, a0, 3
+; RV32-V128-NEXT:    add a0, sp, a0
+; RV32-V128-NEXT:    addi a0, a0, 16
+; RV32-V128-NEXT:    vl8r.v v8, (a0) # Unknown-size Folded Reload
+; RV32-V128-NEXT:    vrgather.vv v16, v8, v24, v0.t
 ; RV32-V128-NEXT:    vsetivli zero, 16, e32, m4, ta, ma
+; RV32-V128-NEXT:    vmv4r.v v24, v8
 ; RV32-V128-NEXT:    addi a0, sp, 16
 ; RV32-V128-NEXT:    vl8r.v v8, (a0) # Unknown-size Folded Reload
-; RV32-V128-NEXT:    vwaddu.vv v0, v8, v16
+; RV32-V128-NEXT:    vwaddu.vv v0, v8, v24
 ; RV32-V128-NEXT:    li a0, -1
-; RV32-V128-NEXT:    vwmaccu.vx v0, a0, v16
+; RV32-V128-NEXT:    vwmaccu.vx v0, a0, v24
 ; RV32-V128-NEXT:    vmv8r.v v8, v0
-; RV32-V128-NEXT:    vmv8r.v v16, v24
 ; RV32-V128-NEXT:    csrr a0, vlenb
-; RV32-V128-NEXT:    slli a0, a0, 4
+; RV32-V128-NEXT:    li a1, 24
+; RV32-V128-NEXT:    mul a0, a0, a1
 ; RV32-V128-NEXT:    add sp, sp, a0
 ; RV32-V128-NEXT:    addi sp, sp, 16
 ; RV32-V128-NEXT:    ret
 ;
 ; RV64-V128-LABEL: interleave_v32f32:
 ; RV64-V128:       # %bb.0:
 ; RV64-V128-NEXT:    addi sp, sp, -16
 ; RV64-V128-NEXT:    .cfi_def_cfa_offset 16
 ; RV64-V128-NEXT:    csrr a0, vlenb
-; RV64-V128-NEXT:    slli a0, a0, 4
+; RV64-V128-NEXT:    li a1, 24
+; RV64-V128-NEXT:    mul a0, a0, a1
 ; RV64-V128-NEXT:    sub sp, sp, a0
-; RV64-V128-NEXT:    .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x10, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 16 * vlenb
+; RV64-V128-NEXT:    .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x18, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 24 * vlenb
+; RV64-V128-NEXT:    csrr a0, vlenb
+; RV64-V128-NEXT:    slli a0, a0, 3
+; RV64-V128-NEXT:    add a0, sp, a0
+; RV64-V128-NEXT:    addi a0, a0, 16
+; RV64-V128-NEXT:    vs8r.v v16, (a0) # Unknown-size Folded Spill
 ; RV64-V128-NEXT:    lui a0, %hi(.LCPI10_0)
 ; RV64-V128-NEXT:    addi a0, a0, %lo(.LCPI10_0)
 ; RV64-V128-NEXT:    li a1, 32
-; RV64-V128-NEXT:    vsetvli zero, a1, e32, m8, ta, ma
-; RV64-V128-NEXT:    vle32.v v0, (a0)
-; RV64-V128-NEXT:    vmv8r.v v24, v8
-; RV64-V128-NEXT:    vrgather.vv v8, v24, v0
-; RV64-V128-NEXT:    addi a0, sp, 16
-; RV64-V128-NEXT:    vs8r.v v24, (a0) # Unknown-size Folded Spill
+; RV64-V128-NEXT:    vsetvli zero, a1, e32, m8, ta, mu
+; RV64-V128-NEXT:    vle32.v v24, (a0)
 ; RV64-V128-NEXT:    lui a0, %hi(.LCPI10_1)
 ; RV64-V128-NEXT:    addi a0, a0, %lo(.LCPI10_1)
-; RV64-V128-NEXT:    vle32.v v24, (a0)
+; RV64-V128-NEXT:    vle32.v v16, (a0)
 ; RV64-V128-NEXT:    csrr a0, vlenb
-; RV64-V128-NEXT:    slli a0, a0, 3
+; RV64-V128-NEXT:    slli a0, a0, 4
 ; RV64-V128-NEXT:    add a0, sp, a0
 ; RV64-V128-NEXT:    addi a0, a0, 16
-; RV64-V128-NEXT:    vs8r.v v24, (a0) # Unknown-size Folded Spill
+; RV64-V128-NEXT:    vs8r.v v16, (a0) # Unknown-size Folded Spill
 ; RV64-V128-NEXT:    lui a0, 699051
 ; RV64-V128-NEXT:    addiw a0, a0, -1366
-; RV64-V128-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
 ; RV64-V128-NEXT:    vmv.s.x v0, a0
-; RV64-V128-NEXT:    vsetvli zero, a1, e32, m8, ta, mu
+; RV64-V128-NEXT:    vrgather.vv v16, v8, v24
+; RV64-V128-NEXT:    addi a0, sp, 16
+; RV64-V128-NEXT:    vs8r.v v8, (a0) # Unknown-size Folded Spill
 ; RV64-V128-NEXT:    csrr a0, vlenb
-; RV64-V128-NEXT:    slli a0, a0, 3
+; RV64-V128-NEXT:    slli a0, a0, 4
 ; RV64-V128-NEXT:    add a0, sp, a0
 ; RV64-V128-NEXT:    addi a0, a0, 16
 ; RV64-V128-NEXT:    vl8r.v v24, (a0) # Unknown-size Folded Reload
-; RV64-V128-NEXT:    vrgather.vv v8, v16, v24, v0.t
-; RV64-V128-NEXT:    vmv.v.v v24, v8
+; RV64-V128-NEXT:    csrr a0, vlenb
+; RV64-V128-NEXT:    slli a0, a0, 3
+; RV64-V128-NEXT:    add a0, sp, a0
+; RV64-V128-NEXT:    addi a0, a0, 16
+; RV64-V128-NEXT:    vl8r.v v8, (a0) # Unknown-size Folded Reload
+; RV64-V128-NEXT:    vrgather.vv v16, v8, v24, v0.t
 ; RV64-V128-NEXT:    vsetivli zero, 16, e32, m4, ta, ma
+; RV64-V128-NEXT:    vmv4r.v v24, v8
 ; RV64-V128-NEXT:    addi a0, sp, 16
 ; RV64-V128-NEXT:    vl8r.v v8, (a0) # Unknown-size Folded Reload
-; RV64-V128-NEXT:    vwaddu.vv v0, v8, v16
+; RV64-V128-NEXT:    vwaddu.vv v0, v8, v24
 ; RV64-V128-NEXT:    li a0, -1
-; RV64-V128-NEXT:    vwmaccu.vx v0, a0, v16
+; RV64-V128-NEXT:    vwmaccu.vx v0, a0, v24
 ; RV64-V128-NEXT:    vmv8r.v v8, v0
-; RV64-V128-NEXT:    vmv8r.v v16, v24
 ; RV64-V128-NEXT:    csrr a0, vlenb
-; RV64-V128-NEXT:    slli a0, a0, 4
+; RV64-V128-NEXT:    li a1, 24
+; RV64-V128-NEXT:    mul a0, a0, a1
 ; RV64-V128-NEXT:    add sp, sp, a0
 ; RV64-V128-NEXT:    addi sp, sp, 16
 ; RV64-V128-NEXT:    ret
 ;
 ; V512-LABEL: interleave_v32f32:
 ; V512:       # %bb.0:
 ; V512-NEXT:    li a0, 32
 ; V512-NEXT:    vsetvli zero, a0, e32, m2, ta, ma

llvm/test/CodeGen/RISCV/rvv/fixed-vectors-int-interleave.ll

 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
-; RUN: llc -mtriple=riscv32 -mattr=+v -riscv-v-vector-bits-min=128 -verify-machineinstrs < %s | FileCheck %s --check-prefixes=CHECK,V128,RV32-V128
-; RUN: llc -mtriple=riscv64 -mattr=+v -riscv-v-vector-bits-min=128 -verify-machineinstrs < %s | FileCheck %s --check-prefixes=CHECK,V128,RV64-V128
-; RUN: llc -mtriple=riscv32 -mattr=+v -riscv-v-vector-bits-min=512 -verify-machineinstrs < %s | FileCheck %s --check-prefixes=CHECK,V512,RV32-V512
-; RUN: llc -mtriple=riscv64 -mattr=+v -riscv-v-vector-bits-min=512 -verify-machineinstrs < %s | FileCheck %s --check-prefixes=CHECK,V512,RV64-V512
+; RUN: llc -mtriple=riscv32 -mattr=+v,+m -riscv-v-vector-bits-min=128 -verify-machineinstrs < %s | FileCheck %s --check-prefixes=CHECK,V128,RV32-V128
+; RUN: llc -mtriple=riscv64 -mattr=+v,+m -riscv-v-vector-bits-min=128 -verify-machineinstrs < %s | FileCheck %s --check-prefixes=CHECK,V128,RV64-V128
+; RUN: llc -mtriple=riscv32 -mattr=+v,+m -riscv-v-vector-bits-min=512 -verify-machineinstrs < %s | FileCheck %s --check-prefixes=CHECK,V512,RV32-V512
+; RUN: llc -mtriple=riscv64 -mattr=+v,+m -riscv-v-vector-bits-min=512 -verify-machineinstrs < %s | FileCheck %s --check-prefixes=CHECK,V512,RV64-V512
 
 ; Test optimizing interleaves to widening arithmetic.
 
 define <4 x i8> @interleave_v2i8(<2 x i8> %x, <2 x i8> %y) {
 ; CHECK-LABEL: interleave_v2i8:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetivli zero, 2, e8, mf8, ta, ma
 ; CHECK-NEXT:    vwaddu.vv v10, v8, v9

 }
 
 define <64 x i32> @interleave_v32i32(<32 x i32> %x, <32 x i32> %y) {
 ; RV32-V128-LABEL: interleave_v32i32:
 ; RV32-V128:       # %bb.0:
 ; RV32-V128-NEXT:    addi sp, sp, -16
 ; RV32-V128-NEXT:    .cfi_def_cfa_offset 16
 ; RV32-V128-NEXT:    csrr a0, vlenb
-; RV32-V128-NEXT:    slli a0, a0, 4
+; RV32-V128-NEXT:    li a1, 24
+; RV32-V128-NEXT:    mul a0, a0, a1
 ; RV32-V128-NEXT:    sub sp, sp, a0
-; RV32-V128-NEXT:    .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x10, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 16 * vlenb
+; RV32-V128-NEXT:    .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x18, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 24 * vlenb
+; RV32-V128-NEXT:    csrr a0, vlenb
+; RV32-V128-NEXT:    slli a0, a0, 3
+; RV32-V128-NEXT:    add a0, sp, a0
+; RV32-V128-NEXT:    addi a0, a0, 16
+; RV32-V128-NEXT:    vs8r.v v16, (a0) # Unknown-size Folded Spill
 ; RV32-V128-NEXT:    lui a0, %hi(.LCPI17_0)
 ; RV32-V128-NEXT:    addi a0, a0, %lo(.LCPI17_0)
 ; RV32-V128-NEXT:    li a1, 32
-; RV32-V128-NEXT:    vsetvli zero, a1, e32, m8, ta, ma
-; RV32-V128-NEXT:    vle32.v v0, (a0)
-; RV32-V128-NEXT:    vmv8r.v v24, v8
-; RV32-V128-NEXT:    vrgather.vv v8, v24, v0
-; RV32-V128-NEXT:    addi a0, sp, 16
-; RV32-V128-NEXT:    vs8r.v v24, (a0) # Unknown-size Folded Spill
+; RV32-V128-NEXT:    vsetvli zero, a1, e32, m8, ta, mu
+; RV32-V128-NEXT:    vle32.v v24, (a0)
 ; RV32-V128-NEXT:    lui a0, %hi(.LCPI17_1)
 ; RV32-V128-NEXT:    addi a0, a0, %lo(.LCPI17_1)
-; RV32-V128-NEXT:    vle32.v v24, (a0)
+; RV32-V128-NEXT:    vle32.v v16, (a0)
 ; RV32-V128-NEXT:    csrr a0, vlenb
-; RV32-V128-NEXT:    slli a0, a0, 3
+; RV32-V128-NEXT:    slli a0, a0, 4
 ; RV32-V128-NEXT:    add a0, sp, a0
 ; RV32-V128-NEXT:    addi a0, a0, 16
-; RV32-V128-NEXT:    vs8r.v v24, (a0) # Unknown-size Folded Spill
+; RV32-V128-NEXT:    vs8r.v v16, (a0) # Unknown-size Folded Spill
 ; RV32-V128-NEXT:    lui a0, 699051
 ; RV32-V128-NEXT:    addi a0, a0, -1366
-; RV32-V128-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
 ; RV32-V128-NEXT:    vmv.s.x v0, a0
-; RV32-V128-NEXT:    vsetvli zero, a1, e32, m8, ta, mu
+; RV32-V128-NEXT:    vrgather.vv v16, v8, v24
+; RV32-V128-NEXT:    addi a0, sp, 16
+; RV32-V128-NEXT:    vs8r.v v8, (a0) # Unknown-size Folded Spill
 ; RV32-V128-NEXT:    csrr a0, vlenb
-; RV32-V128-NEXT:    slli a0, a0, 3
+; RV32-V128-NEXT:    slli a0, a0, 4
 ; RV32-V128-NEXT:    add a0, sp, a0
 ; RV32-V128-NEXT:    addi a0, a0, 16
 ; RV32-V128-NEXT:    vl8r.v v24, (a0) # Unknown-size Folded Reload
-; RV32-V128-NEXT:    vrgather.vv v8, v16, v24, v0.t
-; RV32-V128-NEXT:    vmv.v.v v24, v8
+; RV32-V128-NEXT:    csrr a0, vlenb
+; RV32-V128-NEXT:    slli a0, a0, 3
+; RV32-V128-NEXT:    add a0, sp, a0
+; RV32-V128-NEXT:    addi a0, a0, 16
+; RV32-V128-NEXT:    vl8r.v v8, (a0) # Unknown-size Folded Reload
+; RV32-V128-NEXT:    vrgather.vv v16, v8, v24, v0.t
 ; RV32-V128-NEXT:    vsetivli zero, 16, e32, m4, ta, ma
+; RV32-V128-NEXT:    vmv4r.v v24, v8
 ; RV32-V128-NEXT:    addi a0, sp, 16
 ; RV32-V128-NEXT:    vl8r.v v8, (a0) # Unknown-size Folded Reload
-; RV32-V128-NEXT:    vwaddu.vv v0, v8, v16
+; RV32-V128-NEXT:    vwaddu.vv v0, v8, v24
 ; RV32-V128-NEXT:    li a0, -1
-; RV32-V128-NEXT:    vwmaccu.vx v0, a0, v16
+; RV32-V128-NEXT:    vwmaccu.vx v0, a0, v24
 ; RV32-V128-NEXT:    vmv8r.v v8, v0
-; RV32-V128-NEXT:    vmv8r.v v16, v24
 ; RV32-V128-NEXT:    csrr a0, vlenb
-; RV32-V128-NEXT:    slli a0, a0, 4
+; RV32-V128-NEXT:    li a1, 24
+; RV32-V128-NEXT:    mul a0, a0, a1
 ; RV32-V128-NEXT:    add sp, sp, a0
 ; RV32-V128-NEXT:    addi sp, sp, 16
 ; RV32-V128-NEXT:    ret
 ;
 ; RV64-V128-LABEL: interleave_v32i32:
 ; RV64-V128:       # %bb.0:
 ; RV64-V128-NEXT:    addi sp, sp, -16
 ; RV64-V128-NEXT:    .cfi_def_cfa_offset 16
 ; RV64-V128-NEXT:    csrr a0, vlenb
-; RV64-V128-NEXT:    slli a0, a0, 4
+; RV64-V128-NEXT:    li a1, 24
+; RV64-V128-NEXT:    mul a0, a0, a1
 ; RV64-V128-NEXT:    sub sp, sp, a0
-; RV64-V128-NEXT:    .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x10, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 16 * vlenb
+; RV64-V128-NEXT:    .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x18, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 24 * vlenb
+; RV64-V128-NEXT:    csrr a0, vlenb
+; RV64-V128-NEXT:    slli a0, a0, 3
+; RV64-V128-NEXT:    add a0, sp, a0
+; RV64-V128-NEXT:    addi a0, a0, 16
+; RV64-V128-NEXT:    vs8r.v v16, (a0) # Unknown-size Folded Spill
 ; RV64-V128-NEXT:    lui a0, %hi(.LCPI17_0)
 ; RV64-V128-NEXT:    addi a0, a0, %lo(.LCPI17_0)
 ; RV64-V128-NEXT:    li a1, 32
-; RV64-V128-NEXT:    vsetvli zero, a1, e32, m8, ta, ma
-; RV64-V128-NEXT:    vle32.v v0, (a0)
-; RV64-V128-NEXT:    vmv8r.v v24, v8
-; RV64-V128-NEXT:    vrgather.vv v8, v24, v0
-; RV64-V128-NEXT:    addi a0, sp, 16
-; RV64-V128-NEXT:    vs8r.v v24, (a0) # Unknown-size Folded Spill
+; RV64-V128-NEXT:    vsetvli zero, a1, e32, m8, ta, mu
+; RV64-V128-NEXT:    vle32.v v24, (a0)
 ; RV64-V128-NEXT:    lui a0, %hi(.LCPI17_1)
 ; RV64-V128-NEXT:    addi a0, a0, %lo(.LCPI17_1)
-; RV64-V128-NEXT:    vle32.v v24, (a0)
+; RV64-V128-NEXT:    vle32.v v16, (a0)
 ; RV64-V128-NEXT:    csrr a0, vlenb
-; RV64-V128-NEXT:    slli a0, a0, 3
+; RV64-V128-NEXT:    slli a0, a0, 4
 ; RV64-V128-NEXT:    add a0, sp, a0
 ; RV64-V128-NEXT:    addi a0, a0, 16
-; RV64-V128-NEXT:    vs8r.v v24, (a0) # Unknown-size Folded Spill
+; RV64-V128-NEXT:    vs8r.v v16, (a0) # Unknown-size Folded Spill
 ; RV64-V128-NEXT:    lui a0, 699051
 ; RV64-V128-NEXT:    addiw a0, a0, -1366
-; RV64-V128-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
 ; RV64-V128-NEXT:    vmv.s.x v0, a0
-; RV64-V128-NEXT:    vsetvli zero, a1, e32, m8, ta, mu
+; RV64-V128-NEXT:    vrgather.vv v16, v8, v24
+; RV64-V128-NEXT:    addi a0, sp, 16
+; RV64-V128-NEXT:    vs8r.v v8, (a0) # Unknown-size Folded Spill
 ; RV64-V128-NEXT:    csrr a0, vlenb
-; RV64-V128-NEXT:    slli a0, a0, 3
+; RV64-V128-NEXT:    slli a0, a0, 4
 ; RV64-V128-NEXT:    add a0, sp, a0
 ; RV64-V128-NEXT:    addi a0, a0, 16
 ; RV64-V128-NEXT:    vl8r.v v24, (a0) # Unknown-size Folded Reload
-; RV64-V128-NEXT:    vrgather.vv v8, v16, v24, v0.t
-; RV64-V128-NEXT:    vmv.v.v v24, v8
+; RV64-V128-NEXT:    csrr a0, vlenb
+; RV64-V128-NEXT:    slli a0, a0, 3
+; RV64-V128-NEXT:    add a0, sp, a0
+; RV64-V128-NEXT:    addi a0, a0, 16
+; RV64-V128-NEXT:    vl8r.v v8, (a0) # Unknown-size Folded Reload
+; RV64-V128-NEXT:    vrgather.vv v16, v8, v24, v0.t
 ; RV64-V128-NEXT:    vsetivli zero, 16, e32, m4, ta, ma
+; RV64-V128-NEXT:    vmv4r.v v24, v8
 ; RV64-V128-NEXT:    addi a0, sp, 16
 ; RV64-V128-NEXT:    vl8r.v v8, (a0) # Unknown-size Folded Reload
-; RV64-V128-NEXT:    vwaddu.vv v0, v8, v16
+; RV64-V128-NEXT:    vwaddu.vv v0, v8, v24
 ; RV64-V128-NEXT:    li a0, -1
-; RV64-V128-NEXT:    vwmaccu.vx v0, a0, v16
+; RV64-V128-NEXT:    vwmaccu.vx v0, a0, v24
 ; RV64-V128-NEXT:    vmv8r.v v8, v0
-; RV64-V128-NEXT:    vmv8r.v v16, v24
 ; RV64-V128-NEXT:    csrr a0, vlenb
-; RV64-V128-NEXT:    slli a0, a0, 4
+; RV64-V128-NEXT:    li a1, 24
+; RV64-V128-NEXT:    mul a0, a0, a1
 ; RV64-V128-NEXT:    add sp, sp, a0
 ; RV64-V128-NEXT:    addi sp, sp, 16
 ; RV64-V128-NEXT:    ret
 ;
 ; V512-LABEL: interleave_v32i32:
 ; V512:       # %bb.0:
 ; V512-NEXT:    li a0, 32
 ; V512-NEXT:    vsetvli zero, a0, e32, m2, ta, ma

llvm/test/CodeGen/RISCV/rvv/fixed-vectors-interleaved-access.ll

 
 ; LMUL * NF is > 8 here and so shouldn't be lowered to a vlseg
 define {<8 x i64>, <8 x i64>, <8 x i64>, <8 x i64>, <8 x i64>, <8 x i64>} @load_factor6_too_big(ptr %ptr) {
 ; RV32-LABEL: load_factor6_too_big:
 ; RV32:       # %bb.0:
 ; RV32-NEXT:    addi sp, sp, -16
 ; RV32-NEXT:    .cfi_def_cfa_offset 16
 ; RV32-NEXT:    csrr a2, vlenb
-; RV32-NEXT:    li a3, 78
+; RV32-NEXT:    li a3, 82
 ; RV32-NEXT:    mul a2, a2, a3
 ; RV32-NEXT:    sub sp, sp, a2
-; RV32-NEXT:    .cfi_escape 0x0f, 0x0e, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0xce, 0x00, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 78 * vlenb
+; RV32-NEXT:    .cfi_escape 0x0f, 0x0e, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0xd2, 0x00, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 82 * vlenb
 ; RV32-NEXT:    addi a3, a1, 256
 ; RV32-NEXT:    li a2, 32
 ; RV32-NEXT:    vsetvli zero, a2, e32, m8, ta, ma
 ; RV32-NEXT:    vle32.v v24, (a3)
 ; RV32-NEXT:    vsetivli zero, 16, e32, m4, ta, ma
 ; RV32-NEXT:    vid.v v16
 ; RV32-NEXT:    csrr a3, vlenb
 ; RV32-NEXT:    li a4, 37
 ; RV32-NEXT:    mul a3, a3, a4
 ; RV32-NEXT:    add a3, sp, a3
 ; RV32-NEXT:    addi a3, a3, 16
 ; RV32-NEXT:    vs4r.v v16, (a3) # Unknown-size Folded Spill
 ; RV32-NEXT:    vadd.vi v8, v16, -4
 ; RV32-NEXT:    csrr a3, vlenb
-; RV32-NEXT:    li a4, 13
-; RV32-NEXT:    mul a3, a3, a4
+; RV32-NEXT:    slli a4, a3, 4
+; RV32-NEXT:    add a3, a4, a3
 ; RV32-NEXT:    add a3, sp, a3
 ; RV32-NEXT:    addi a3, a3, 16
 ; RV32-NEXT:    vs4r.v v8, (a3) # Unknown-size Folded Spill
-; RV32-NEXT:    vrgather.vv v4, v24, v8
+; RV32-NEXT:    vrgather.vv v12, v24, v8
 ; RV32-NEXT:    csrr a3, vlenb
-; RV32-NEXT:    li a4, 41
+; RV32-NEXT:    li a4, 45
 ; RV32-NEXT:    mul a3, a3, a4
 ; RV32-NEXT:    add a3, sp, a3
 ; RV32-NEXT:    addi a3, a3, 16
-; RV32-NEXT:    vs4r.v v4, (a3) # Unknown-size Folded Spill
+; RV32-NEXT:    vs4r.v v12, (a3) # Unknown-size Folded Spill
 ; RV32-NEXT:    vadd.vi v8, v16, -10
 ; RV32-NEXT:    lui a3, 12
 ; RV32-NEXT:    vmv.s.x v0, a3
 ; RV32-NEXT:    csrr a3, vlenb
-; RV32-NEXT:    li a4, 29
-; RV32-NEXT:    mul a3, a3, a4
+; RV32-NEXT:    slli a4, a3, 5
+; RV32-NEXT:    add a3, a4, a3
 ; RV32-NEXT:    add a3, sp, a3
 ; RV32-NEXT:    addi a3, a3, 16
 ; RV32-NEXT:    vs1r.v v0, (a3) # Unknown-size Folded Spill
 ; RV32-NEXT:    vsetivli zero, 16, e32, m8, ta, ma
 ; RV32-NEXT:    vslidedown.vi v16, v24, 16
 ; RV32-NEXT:    csrr a3, vlenb
-; RV32-NEXT:    li a4, 53
+; RV32-NEXT:    li a4, 57
 ; RV32-NEXT:    mul a3, a3, a4
 ; RV32-NEXT:    add a3, sp, a3
 ; RV32-NEXT:    addi a3, a3, 16
 ; RV32-NEXT:    vs8r.v v24, (a3) # Unknown-size Folded Spill
 ; RV32-NEXT:    vsetivli zero, 16, e32, m4, ta, mu
-; RV32-NEXT:    vrgather.vv v4, v16, v8, v0.t
-; RV32-NEXT:    csrr a3, vlenb
-; RV32-NEXT:    li a4, 61
-; RV32-NEXT:    mul a3, a3, a4
-; RV32-NEXT:    add a3, sp, a3
-; RV32-NEXT:    addi a3, a3, 16
-; RV32-NEXT:    vs8r.v v16, (a3) # Unknown-size Folded Spill
-; RV32-NEXT:    lui a3, %hi(.LCPI6_0)
-; RV32-NEXT:    addi a3, a3, %lo(.LCPI6_0)
-; RV32-NEXT:    vsetvli zero, a2, e32, m8, ta, ma
-; RV32-NEXT:    vle32.v v24, (a3)
-; RV32-NEXT:    vle32.v v16, (a1)
+; RV32-NEXT:    vrgather.vv v12, v16, v8, v0.t
 ; RV32-NEXT:    csrr a3, vlenb
-; RV32-NEXT:    li a4, 69
+; RV32-NEXT:    li a4, 41
 ; RV32-NEXT:    mul a3, a3, a4
 ; RV32-NEXT:    add a3, sp, a3
 ; RV32-NEXT:    addi a3, a3, 16
-; RV32-NEXT:    vs8r.v v16, (a3) # Unknown-size Folded Spill
-; RV32-NEXT:    addi a1, a1, 128
-; RV32-NEXT:    vrgather.vv v8, v16, v24
-; RV32-NEXT:    lui a3, %hi(.LCPI6_1)
-; RV32-NEXT:    addi a3, a3, %lo(.LCPI6_1)
-; RV32-NEXT:    lui a4, 1
-; RV32-NEXT:    addi a4, a4, -64
-; RV32-NEXT:    vle32.v v16, (a3)
+; RV32-NEXT:    vs4r.v v12, (a3) # Unknown-size Folded Spill
 ; RV32-NEXT:    csrr a3, vlenb
-; RV32-NEXT:    li a5, 21
-; RV32-NEXT:    mul a3, a3, a5
+; RV32-NEXT:    slli a4, a3, 6
+; RV32-NEXT:    add a3, a4, a3
 ; RV32-NEXT:    add a3, sp, a3
 ; RV32-NEXT:    addi a3, a3, 16
 ; RV32-NEXT:    vs8r.v v16, (a3) # Unknown-size Folded Spill
+; RV32-NEXT:    addi a3, a1, 128
+; RV32-NEXT:    lui a4, %hi(.LCPI6_0)
+; RV32-NEXT:    addi a4, a4, %lo(.LCPI6_0)
+; RV32-NEXT:    vsetvli zero, a2, e32, m8, ta, mu
+; RV32-NEXT:    lui a5, %hi(.LCPI6_1)
+; RV32-NEXT:    addi a5, a5, %lo(.LCPI6_1)
+; RV32-NEXT:    lui a6, 1
+; RV32-NEXT:    vle32.v v8, (a4)
+; RV32-NEXT:    csrr a4, vlenb
+; RV32-NEXT:    slli a7, a4, 3
+; RV32-NEXT:    add a4, a7, a4
+; RV32-NEXT:    add a4, sp, a4
+; RV32-NEXT:    addi a4, a4, 16
+; RV32-NEXT:    vs8r.v v8, (a4) # Unknown-size Folded Spill
 ; RV32-NEXT:    vle32.v v16, (a1)
 ; RV32-NEXT:    csrr a1, vlenb
-; RV32-NEXT:    li a3, 45
-; RV32-NEXT:    mul a1, a1, a3
+; RV32-NEXT:    li a4, 73
+; RV32-NEXT:    mul a1, a1, a4
 ; RV32-NEXT:    add a1, sp, a1
 ; RV32-NEXT:    addi a1, a1, 16
 ; RV32-NEXT:    vs8r.v v16, (a1) # Unknown-size Folded Spill
-; RV32-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
-; RV32-NEXT:    vmv.s.x v0, a4
+; RV32-NEXT:    vle32.v v8, (a5)
 ; RV32-NEXT:    csrr a1, vlenb
-; RV32-NEXT:    slli a3, a1, 4
+; RV32-NEXT:    li a4, 25
+; RV32-NEXT:    mul a1, a1, a4
+; RV32-NEXT:    add a1, sp, a1
+; RV32-NEXT:    addi a1, a1, 16
+; RV32-NEXT:    vs8r.v v8, (a1) # Unknown-size Folded Spill
+; RV32-NEXT:    vle32.v v8, (a3)
+; RV32-NEXT:    csrr a1, vlenb
+; RV32-NEXT:    li a3, 49
+; RV32-NEXT:    mul a1, a1, a3
+; RV32-NEXT:    add a1, sp, a1
+; RV32-NEXT:    addi a1, a1, 16
+; RV32-NEXT:    vs8r.v v8, (a1) # Unknown-size Folded Spill
+; RV32-NEXT:    addi a1, a6, -64
+; RV32-NEXT:    vmv.s.x v24, a1
+; RV32-NEXT:    csrr a1, vlenb
+; RV32-NEXT:    slli a1, a1, 3
+; RV32-NEXT:    add a1, sp, a1
+; RV32-NEXT:    addi a1, a1, 16
+; RV32-NEXT:    vs1r.v v24, (a1) # Unknown-size Folded Spill
+; RV32-NEXT:    csrr a1, vlenb
+; RV32-NEXT:    slli a3, a1, 3
 ; RV32-NEXT:    add a1, a3, a1
 ; RV32-NEXT:    add a1, sp, a1
 ; RV32-NEXT:    addi a1, a1, 16
-; RV32-NEXT:    vs1r.v v0, (a1) # Unknown-size Folded Spill
-; RV32-NEXT:    vsetvli zero, a2, e32, m8, ta, mu
+; RV32-NEXT:    vl8r.v v0, (a1) # Unknown-size Folded Reload
+; RV32-NEXT:    vrgather.vv v24, v16, v0
+; RV32-NEXT:    csrr a1, vlenb
+; RV32-NEXT:    slli a1, a1, 3
+; RV32-NEXT:    add a1, sp, a1
+; RV32-NEXT:    addi a1, a1, 16
+; RV32-NEXT:    vl1r.v v2, (a1) # Unknown-size Folded Reload
+; RV32-NEXT:    vmv1r.v v0, v2
 ; RV32-NEXT:    csrr a1, vlenb
-; RV32-NEXT:    li a3, 21
+; RV32-NEXT:    li a3, 25
 ; RV32-NEXT:    mul a1, a1, a3
 ; RV32-NEXT:    add a1, sp, a1
 ; RV32-NEXT:    addi a1, a1, 16
-; RV32-NEXT:    vl8r.v v24, (a1) # Unknown-size Folded Reload
-; RV32-NEXT:    vrgather.vv v8, v16, v24, v0.t
+; RV32-NEXT:    vl8r.v v16, (a1) # Unknown-size Folded Reload
+; RV32-NEXT:    vrgather.vv v24, v8, v16, v0.t
 ; RV32-NEXT:    vsetivli zero, 12, e32, m4, tu, ma
-; RV32-NEXT:    vslideup.vi v4, v8, 0
 ; RV32-NEXT:    csrr a1, vlenb
-; RV32-NEXT:    li a3, 21
+; RV32-NEXT:    li a3, 41
 ; RV32-NEXT:    mul a1, a1, a3
 ; RV32-NEXT:    add a1, sp, a1
 ; RV32-NEXT:    addi a1, a1, 16
-; RV32-NEXT:    vs4r.v v4, (a1) # Unknown-size Folded Spill
+; RV32-NEXT:    vl4r.v v8, (a1) # Unknown-size Folded Reload
+; RV32-NEXT:    vslideup.vi v8, v24, 0
+; RV32-NEXT:    csrr a1, vlenb
+; RV32-NEXT:    li a3, 41
+; RV32-NEXT:    mul a1, a1, a3
+; RV32-NEXT:    add a1, sp, a1
+; RV32-NEXT:    addi a1, a1, 16
+; RV32-NEXT:    vs4r.v v8, (a1) # Unknown-size Folded Spill
 ; RV32-NEXT:    vsetivli zero, 16, e32, m4, ta, mu
 ; RV32-NEXT:    csrr a1, vlenb
 ; RV32-NEXT:    li a3, 37
 ; RV32-NEXT:    mul a1, a1, a3
 ; RV32-NEXT:    add a1, sp, a1
 ; RV32-NEXT:    addi a1, a1, 16
-; RV32-NEXT:    vl4r.v v24, (a1) # Unknown-size Folded Reload
-; RV32-NEXT:    vadd.vi v8, v24, -2
+; RV32-NEXT:    vl4r.v v12, (a1) # Unknown-size Folded Reload
+; RV32-NEXT:    vadd.vi v16, v12, -2
 ; RV32-NEXT:    csrr a1, vlenb
-; RV32-NEXT:    li a3, 53
+; RV32-NEXT:    li a3, 57
 ; RV32-NEXT:    mul a1, a1, a3
 ; RV32-NEXT:    add a1, sp, a1
 ; RV32-NEXT:    addi a1, a1, 16
-; RV32-NEXT:    vl8r.v v16, (a1) # Unknown-size Folded Reload
-; RV32-NEXT:    vrgather.vv v4, v16, v8
-; RV32-NEXT:    vadd.vi v8, v24, -8
+; RV32-NEXT:    vl8r.v v24, (a1) # Unknown-size Folded Reload
+; RV32-NEXT:    vrgather.vv v8, v24, v16
+; RV32-NEXT:    vadd.vi v16, v12, -8
 ; RV32-NEXT:    csrr a1, vlenb
-; RV32-NEXT:    li a3, 29
-; RV32-NEXT:    mul a1, a1, a3
+; RV32-NEXT:    slli a3, a1, 5
+; RV32-NEXT:    add a1, a3, a1
 ; RV32-NEXT:    add a1, sp, a1
 ; RV32-NEXT:    addi a1, a1, 16
 ; RV32-NEXT:    vl1r.v v0, (a1) # Unknown-size Folded Reload
+; RV32-NEXT:    vmv1r.v v1, v0
 ; RV32-NEXT:    csrr a1, vlenb
-; RV32-NEXT:    li a3, 61
-; RV32-NEXT:    mul a1, a1, a3
+; RV32-NEXT:    slli a3, a1, 6
+; RV32-NEXT:    add a1, a3, a1
 ; RV32-NEXT:    add a1, sp, a1
 ; RV32-NEXT:    addi a1, a1, 16
-; RV32-NEXT:    vl8r.v v16, (a1) # Unknown-size Folded Reload
-; RV32-NEXT:    vrgather.vv v4, v16, v8, v0.t
+; RV32-NEXT:    vl8r.v v24, (a1) # Unknown-size Folded Reload
+; RV32-NEXT:    vrgather.vv v8, v24, v16, v0.t
+; RV32-NEXT:    vmv.v.v v4, v8
 ; RV32-NEXT:    lui a1, %hi(.LCPI6_2)
 ; RV32-NEXT:    addi a1, a1, %lo(.LCPI6_2)
 ; RV32-NEXT:    vsetvli zero, a2, e32, m8, ta, mu
 ; RV32-NEXT:    lui a3, %hi(.LCPI6_3)
 ; RV32-NEXT:    addi a3, a3, %lo(.LCPI6_3)
 ; RV32-NEXT:    vle32.v v16, (a1)
 ; RV32-NEXT:    vle32.v v8, (a3)
 ; RV32-NEXT:    csrr a1, vlenb
-; RV32-NEXT:    slli a3, a1, 2
-; RV32-NEXT:    add a1, a3, a1
+; RV32-NEXT:    li a3, 25
+; RV32-NEXT:    mul a1, a1, a3
 ; RV32-NEXT:    add a1, sp, a1
 ; RV32-NEXT:    addi a1, a1, 16
 ; RV32-NEXT:    vs8r.v v8, (a1) # Unknown-size Folded Spill
 ; RV32-NEXT:    csrr a1, vlenb
-; RV32-NEXT:    li a3, 69
+; RV32-NEXT:    li a3, 73
 ; RV32-NEXT:    mul a1, a1, a3
 ; RV32-NEXT:    add a1, sp, a1
 ; RV32-NEXT:    addi a1, a1, 16
-; RV32-NEXT:    vl8r.v v24, (a1) # Unknown-size Folded Reload
-; RV32-NEXT:    vrgather.vv v8, v24, v16
-; RV32-NEXT:    csrr a1, vlenb
-; RV32-NEXT:    slli a3, a1, 4
-; RV32-NEXT:    add a1, a3, a1
-; RV32-NEXT:    add a1, sp, a1
-; RV32-NEXT:    addi a1, a1, 16
-; RV32-NEXT:    vl1r.v v0, (a1) # Unknown-size Folded Reload
+; RV32-NEXT:    vl8r.v v8, (a1) # Unknown-size Folded Reload
+; RV32-NEXT:    vrgather.vv v24, v8, v16
+; RV32-NEXT:    vmv1r.v v0, v2
 ; RV32-NEXT:    csrr a1, vlenb
-; RV32-NEXT:    li a3, 45
+; RV32-NEXT:    li a3, 49
 ; RV32-NEXT:    mul a1, a1, a3
 ; RV32-NEXT:    add a1, sp, a1
 ; RV32-NEXT:    addi a1, a1, 16
-; RV32-NEXT:    vl8r.v v24, (a1) # Unknown-size Folded Reload
+; RV32-NEXT:    vl8r.v v8, (a1) # Unknown-size Folded Reload
 ; RV32-NEXT:    csrr a1, vlenb
-; RV32-NEXT:    slli a3, a1, 2
-; RV32-NEXT:    add a1, a3, a1
+; RV32-NEXT:    li a3, 25
+; RV32-NEXT:    mul a1, a1, a3
 ; RV32-NEXT:    add a1, sp, a1
 ; RV32-NEXT:    addi a1, a1, 16
 ; RV32-NEXT:    vl8r.v v16, (a1) # Unknown-size Folded Reload
-; RV32-NEXT:    vrgather.vv v8, v24, v16, v0.t
+; RV32-NEXT:    vrgather.vv v24, v8, v16, v0.t
 ; RV32-NEXT:    vsetivli zero, 12, e32, m4, tu, ma
-; RV32-NEXT:    vslideup.vi v4, v8, 0
+; RV32-NEXT:    vslideup.vi v4, v24, 0
 ; RV32-NEXT:    csrr a1, vlenb
-; RV32-NEXT:    slli a3, a1, 4
-; RV32-NEXT:    add a1, a3, a1
+; RV32-NEXT:    li a3, 25
+; RV32-NEXT:    mul a1, a1, a3
 ; RV32-NEXT:    add a1, sp, a1
 ; RV32-NEXT:    addi a1, a1, 16
 ; RV32-NEXT:    vs4r.v v4, (a1) # Unknown-size Folded Spill
 ; RV32-NEXT:    lui a1, %hi(.LCPI6_4)
 ; RV32-NEXT:    addi a1, a1, %lo(.LCPI6_4)
 ; RV32-NEXT:    vsetivli zero, 16, e32, m4, ta, mu
 ; RV32-NEXT:    vle32.v v8, (a1)
 ; RV32-NEXT:    csrr a1, vlenb
-; RV32-NEXT:    li a3, 53
+; RV32-NEXT:    li a3, 57
 ; RV32-NEXT:    mul a1, a1, a3
 ; RV32-NEXT:    add a1, sp, a1
 ; RV32-NEXT:    addi a1, a1, 16
-; RV32-NEXT:    vl8r.v v0, (a1) # Unknown-size Folded Reload
-; RV32-NEXT:    vrgather.vv v12, v0, v8
+; RV32-NEXT:    vl8r.v v16, (a1) # Unknown-size Folded Reload
+; RV32-NEXT:    vrgather.vv v12, v16, v8
 ; RV32-NEXT:    csrr a1, vlenb
 ; RV32-NEXT:    li a3, 37
 ; RV32-NEXT:    mul a1, a1, a3
 ; RV32-NEXT:    add a1, sp, a1
 ; RV32-NEXT:    addi a1, a1, 16
 ; RV32-NEXT:    vl4r.v v8, (a1) # Unknown-size Folded Reload
 ; RV32-NEXT:    vadd.vi v8, v8, -6
 ; RV32-NEXT:    csrr a1, vlenb
+; RV32-NEXT:    slli a3, a1, 3
+; RV32-NEXT:    add a1, a3, a1
 ; RV32-NEXT:    add a1, sp, a1
 ; RV32-NEXT:    addi a1, a1, 16
 ; RV32-NEXT:    vs4r.v v8, (a1) # Unknown-size Folded Spill
+; RV32-NEXT:    vmv1r.v v0, v1
 ; RV32-NEXT:    csrr a1, vlenb
-; RV32-NEXT:    li a3, 29
-; RV32-NEXT:    mul a1, a1, a3
+; RV32-NEXT:    slli a3, a1, 6
+; RV32-NEXT:    add a1, a3, a1
 ; RV32-NEXT:    add a1, sp, a1
 ; RV32-NEXT:    addi a1, a1, 16
-; RV32-NEXT:    vl1r.v v1, (a1) # Unknown-size Folded Reload
-; RV32-NEXT:    vmv1r.v v0, v1
+; RV32-NEXT:    vl8r.v v16, (a1) # Unknown-size Folded Reload
+; RV32-NEXT:    vrgather.vv v12, v16, v8, v0.t
 ; RV32-NEXT:    csrr a1, vlenb
-; RV32-NEXT:    li a3, 61
+; RV32-NEXT:    li a3, 37
 ; RV32-NEXT:    mul a1, a1, a3
 ; RV32-NEXT:    add a1, sp, a1
 ; RV32-NEXT:    addi a1, a1, 16
-; RV32-NEXT:    vl8r.v v16, (a1) # Unknown-size Folded Reload
-; RV32-NEXT:    vrgather.vv v12, v16, v8, v0.t
-; RV32-NEXT:    vmv.v.v v4, v12
+; RV32-NEXT:    vs4r.v v12, (a1) # Unknown-size Folded Spill
 ; RV32-NEXT:    lui a1, %hi(.LCPI6_5)
 ; RV32-NEXT:    addi a1, a1, %lo(.LCPI6_5)
-; RV32-NEXT:    vsetvli zero, a2, e32, m8, ta, ma
+; RV32-NEXT:    vsetvli zero, a2, e32, m8, ta, mu
+; RV32-NEXT:    lui a3, %hi(.LCPI6_6)
+; RV32-NEXT:    addi a3, a3, %lo(.LCPI6_6)
 ; RV32-NEXT:    vle32.v v16, (a1)
+; RV32-NEXT:    vle32.v v8, (a3)
+; RV32-NEXT:    addi a1, sp, 16
+; RV32-NEXT:    vs8r.v v8, (a1) # Unknown-size Folded Spill
+; RV32-NEXT:    li a1, 960
+; RV32-NEXT:    vmv.s.x v0, a1
+; RV32-NEXT:    csrr a1, vlenb
+; RV32-NEXT:    slli a1, a1, 3
+; RV32-NEXT:    add a1, sp, a1
+; RV32-NEXT:    addi a1, a1, 16
+; RV32-NEXT:    vs1r.v v0, (a1) # Unknown-size Folded Spill
 ; RV32-NEXT:    csrr a1, vlenb
-; RV32-NEXT:    li a3, 69
+; RV32-NEXT:    li a3, 73
 ; RV32-NEXT:    mul a1, a1, a3
 ; RV32-NEXT:    add a1, sp, a1
 ; RV32-NEXT:    addi a1, a1, 16
 ; RV32-NEXT:    vl8r.v v24, (a1) # Unknown-size Folded Reload
 ; RV32-NEXT:    vrgather.vv v8, v24, v16
-; RV32-NEXT:    lui a1, %hi(.LCPI6_6)
-; RV32-NEXT:    addi a1, a1, %lo(.LCPI6_6)
-; RV32-NEXT:    li a3, 960
-; RV32-NEXT:    vle32.v v24, (a1)
-; RV32-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
-; RV32-NEXT:    vmv.s.x v0, a3
-; RV32-NEXT:    addi a1, sp, 16
-; RV32-NEXT:    vs1r.v v0, (a1) # Unknown-size Folded Spill
-; RV32-NEXT:    vsetvli zero, a2, e32, m8, ta, mu
 ; RV32-NEXT:    csrr a1, vlenb
-; RV32-NEXT:    li a3, 45
+; RV32-NEXT:    li a3, 49
 ; RV32-NEXT:    mul a1, a1, a3
 ; RV32-NEXT:    add a1, sp, a1
 ; RV32-NEXT:    addi a1, a1, 16
 ; RV32-NEXT:    vl8r.v v16, (a1) # Unknown-size Folded Reload
+; RV32-NEXT:    addi a1, sp, 16
+; RV32-NEXT:    vl8r.v v24, (a1) # Unknown-size Folded Reload
 ; RV32-NEXT:    vrgather.vv v8, v16, v24, v0.t
 ; RV32-NEXT:    vsetivli zero, 10, e32, m4, tu, ma
-; RV32-NEXT:    vslideup.vi v4, v8, 0
 ; RV32-NEXT:    csrr a1, vlenb
-; RV32-NEXT:    slli a3, a1, 2
-; RV32-NEXT:    add a1, a3, a1
+; RV32-NEXT:    li a3, 37
+; RV32-NEXT:    mul a1, a1, a3
 ; RV32-NEXT:    add a1, sp, a1
 ; RV32-NEXT:    addi a1, a1, 16
-; RV32-NEXT:    vs4r.v v4, (a1) # Unknown-size Folded Spill
+; RV32-NEXT:    vl4r.v v12, (a1) # Unknown-size Folded Reload
+; RV32-NEXT:    vslideup.vi v12, v8, 0
+; RV32-NEXT:    csrr a1, vlenb
+; RV32-NEXT:    li a3, 37
+; RV32-NEXT:    mul a1, a1, a3
+; RV32-NEXT:    add a1, sp, a1
+; RV32-NEXT:    addi a1, a1, 16
+; RV32-NEXT:    vs4r.v v12, (a1) # Unknown-size Folded Spill
 ; RV32-NEXT:    lui a1, %hi(.LCPI6_7)
 ; RV32-NEXT:    addi a1, a1, %lo(.LCPI6_7)
 ; RV32-NEXT:    vsetivli zero, 16, e32, m4, ta, mu
 ; RV32-NEXT:    vle32.v v8, (a1)
 ; RV32-NEXT:    csrr a1, vlenb
-; RV32-NEXT:    li a3, 53
+; RV32-NEXT:    li a3, 57
 ; RV32-NEXT:    mul a1, a1, a3
 ; RV32-NEXT:    add a1, sp, a1
 ; RV32-NEXT:    addi a1, a1, 16
 ; RV32-NEXT:    vl8r.v v24, (a1) # Unknown-size Folded Reload
 ; RV32-NEXT:    vrgather.vv v12, v24, v8
-; RV32-NEXT:    vmv1r.v v0, v1
 ; RV32-NEXT:    csrr a1, vlenb
-; RV32-NEXT:    li a3, 13
-; RV32-NEXT:    mul a1, a1, a3
+; RV32-NEXT:    slli a3, a1, 5
+; RV32-NEXT:    add a1, a3, a1
+; RV32-NEXT:    add a1, sp, a1
+; RV32-NEXT:    addi a1, a1, 16
+; RV32-NEXT:    vl1r.v v0, (a1) # Unknown-size Folded Reload
+; RV32-NEXT:    csrr a1, vlenb
+; RV32-NEXT:    slli a3, a1, 4
+; RV32-NEXT:    add a1, a3, a1
 ; RV32-NEXT:    add a1, sp, a1
 ; RV32-NEXT:    addi a1, a1, 16
 ; RV32-NEXT:    vl4r.v v8, (a1) # Unknown-size Folded Reload
 ; RV32-NEXT:    csrr a1, vlenb
-; RV32-NEXT:    li a3, 61
-; RV32-NEXT:    mul a1, a1, a3
+; RV32-NEXT:    slli a3, a1, 6
+; RV32-NEXT:    add a1, a3, a1
 ; RV32-NEXT:    add a1, sp, a1
 ; RV32-NEXT:    addi a1, a1, 16
 ; RV32-NEXT:    vl8r.v v24, (a1) # Unknown-size Folded Reload
 ; RV32-NEXT:    vrgather.vv v12, v24, v8, v0.t
 ; RV32-NEXT:    csrr a1, vlenb
-; RV32-NEXT:    li a3, 37
-; RV32-NEXT:    mul a1, a1, a3
+; RV32-NEXT:    slli a3, a1, 5
+; RV32-NEXT:    add a1, a3, a1
 ; RV32-NEXT:    add a1, sp, a1
 ; RV32-NEXT:    addi a1, a1, 16
 ; RV32-NEXT:    vs4r.v v12, (a1) # Unknown-size Folded Spill
 ; RV32-NEXT:    lui a1, %hi(.LCPI6_8)
 ; RV32-NEXT:    addi a1, a1, %lo(.LCPI6_8)
 ; RV32-NEXT:    vsetvli zero, a2, e32, m8, ta, mu
 ; RV32-NEXT:    lui a3, %hi(.LCPI6_9)
 ; RV32-NEXT:    addi a3, a3, %lo(.LCPI6_9)
 ; RV32-NEXT:    vle32.v v24, (a1)
 ; RV32-NEXT:    vle32.v v8, (a3)
 ; RV32-NEXT:    csrr a1, vlenb
-; RV32-NEXT:    li a3, 29
-; RV32-NEXT:    mul a1, a1, a3
+; RV32-NEXT:    slli a3, a1, 4
+; RV32-NEXT:    add a1, a3, a1
 ; RV32-NEXT:    add a1, sp, a1
 ; RV32-NEXT:    addi a1, a1, 16
 ; RV32-NEXT:    vs8r.v v8, (a1) # Unknown-size Folded Spill
 ; RV32-NEXT:    csrr a1, vlenb
-; RV32-NEXT:    li a3, 69
+; RV32-NEXT:    li a3, 73
 ; RV32-NEXT:    mul a1, a1, a3
 ; RV32-NEXT:    add a1, sp, a1
 ; RV32-NEXT:    addi a1, a1, 16
 ; RV32-NEXT:    vl8r.v v0, (a1) # Unknown-size Folded Reload
 ; RV32-NEXT:    vrgather.vv v8, v0, v24
-; RV32-NEXT:    addi a1, sp, 16
+; RV32-NEXT:    csrr a1, vlenb
+; RV32-NEXT:    slli a1, a1, 3
+; RV32-NEXT:    add a1, sp, a1
+; RV32-NEXT:    addi a1, a1, 16
 ; RV32-NEXT:    vl1r.v v0, (a1) # Unknown-size Folded Reload
 ; RV32-NEXT:    csrr a1, vlenb
-; RV32-NEXT:    li a3, 29
-; RV32-NEXT:    mul a1, a1, a3
+; RV32-NEXT:    slli a3, a1, 4
+; RV32-NEXT:    add a1, a3, a1
 ; RV32-NEXT:    add a1, sp, a1
 ; RV32-NEXT:    addi a1, a1, 16
 ; RV32-NEXT:    vl8r.v v24, (a1) # Unknown-size Folded Reload
 ; RV32-NEXT:    vrgather.vv v8, v16, v24, v0.t
 ; RV32-NEXT:    vsetivli zero, 10, e32, m4, tu, ma
 ; RV32-NEXT:    csrr a1, vlenb
-; RV32-NEXT:    li a3, 37
-; RV32-NEXT:    mul a1, a1, a3
+; RV32-NEXT:    slli a3, a1, 5
+; RV32-NEXT:    add a1, a3, a1
 ; RV32-NEXT:    add a1, sp, a1
 ; RV32-NEXT:    addi a1, a1, 16
 ; RV32-NEXT:    vl4r.v v12, (a1) # Unknown-size Folded Reload
 ; RV32-NEXT:    vslideup.vi v12, v8, 0
 ; RV32-NEXT:    csrr a1, vlenb
-; RV32-NEXT:    li a3, 37
-; RV32-NEXT:    mul a1, a1, a3
+; RV32-NEXT:    slli a3, a1, 5
+; RV32-NEXT:    add a1, a3, a1
 ; RV32-NEXT:    add a1, sp, a1
 ; RV32-NEXT:    addi a1, a1, 16
 ; RV32-NEXT:    vs4r.v v12, (a1) # Unknown-size Folded Spill
 ; RV32-NEXT:    vsetivli zero, 16, e32, m4, ta, mu
 ; RV32-NEXT:    lui a1, %hi(.LCPI6_10)
 ; RV32-NEXT:    addi a1, a1, %lo(.LCPI6_10)
 ; RV32-NEXT:    vle32.v v8, (a1)
 ; RV32-NEXT:    lui a1, 15
 ; RV32-NEXT:    vmv.s.x v1, a1
 ; RV32-NEXT:    csrr a1, vlenb
-; RV32-NEXT:    li a3, 53
+; RV32-NEXT:    li a3, 57
 ; RV32-NEXT:    mul a1, a1, a3
 ; RV32-NEXT:    add a1, sp, a1
 ; RV32-NEXT:    addi a1, a1, 16
 ; RV32-NEXT:    vl8r.v v16, (a1) # Unknown-size Folded Reload
 ; RV32-NEXT:    csrr a1, vlenb
+; RV32-NEXT:    slli a3, a1, 3
+; RV32-NEXT:    add a1, a3, a1
 ; RV32-NEXT:    add a1, sp, a1
 ; RV32-NEXT:    addi a1, a1, 16
 ; RV32-NEXT:    vl4r.v v12, (a1) # Unknown-size Folded Reload
 ; RV32-NEXT:    vrgather.vv v4, v16, v12
 ; RV32-NEXT:    vmv1r.v v0, v1
 ; RV32-NEXT:    csrr a1, vlenb
-; RV32-NEXT:    li a3, 61
-; RV32-NEXT:    mul a1, a1, a3
+; RV32-NEXT:    slli a3, a1, 6
+; RV32-NEXT:    add a1, a3, a1
 ; RV32-NEXT:    add a1, sp, a1
 ; RV32-NEXT:    addi a1, a1, 16
 ; RV32-NEXT:    vl8r.v v16, (a1) # Unknown-size Folded Reload
 ; RV32-NEXT:    vrgather.vv v4, v16, v8, v0.t
 ; RV32-NEXT:    lui a1, %hi(.LCPI6_11)
 ; RV32-NEXT:    addi a1, a1, %lo(.LCPI6_11)
-; RV32-NEXT:    vsetvli zero, a2, e32, m8, ta, ma
-; RV32-NEXT:    vle32.v v16, (a1)
+; RV32-NEXT:    vsetvli zero, a2, e32, m8, ta, mu
+; RV32-NEXT:    lui a3, %hi(.LCPI6_12)
+; RV32-NEXT:    addi a3, a3, %lo(.LCPI6_12)
+; RV32-NEXT:    vle32.v v24, (a1)
+; RV32-NEXT:    vle32.v v8, (a3)
 ; RV32-NEXT:    csrr a1, vlenb
-; RV32-NEXT:    li a3, 69
+; RV32-NEXT:    li a3, 57
 ; RV32-NEXT:    mul a1, a1, a3
 ; RV32-NEXT:    add a1, sp, a1
 ; RV32-NEXT:    addi a1, a1, 16
-; RV32-NEXT:    vl8r.v v24, (a1) # Unknown-size Folded Reload
-; RV32-NEXT:    vrgather.vv v8, v24, v16
-; RV32-NEXT:    lui a1, %hi(.LCPI6_12)
-; RV32-NEXT:    addi a1, a1, %lo(.LCPI6_12)
-; RV32-NEXT:    li a3, 1008
-; RV32-NEXT:    vle32.v v24, (a1)
-; RV32-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
-; RV32-NEXT:    vmv.s.x v2, a3
-; RV32-NEXT:    vsetvli zero, a2, e32, m8, ta, mu
+; RV32-NEXT:    vs8r.v v8, (a1) # Unknown-size Folded Spill
+; RV32-NEXT:    li a1, 1008
+; RV32-NEXT:    vmv.s.x v2, a1
+; RV32-NEXT:    csrr a1, vlenb
+; RV32-NEXT:    li a3, 73
+; RV32-NEXT:    mul a1, a1, a3
+; RV32-NEXT:    add a1, sp, a1
+; RV32-NEXT:    addi a1, a1, 16
+; RV32-NEXT:    vl8r.v v16, (a1) # Unknown-size Folded Reload
+; RV32-NEXT:    vrgather.vv v8, v16, v24
 ; RV32-NEXT:    vmv1r.v v0, v2
 ; RV32-NEXT:    csrr a1, vlenb
-; RV32-NEXT:    li a3, 45
+; RV32-NEXT:    li a3, 49
 ; RV32-NEXT:    mul a1, a1, a3
 ; RV32-NEXT:    add a1, sp, a1
 ; RV32-NEXT:    addi a1, a1, 16
 ; RV32-NEXT:    vl8r.v v16, (a1) # Unknown-size Folded Reload
+; RV32-NEXT:    csrr a1, vlenb
+; RV32-NEXT:    li a3, 57
+; RV32-NEXT:    mul a1, a1, a3
+; RV32-NEXT:    add a1, sp, a1
+; RV32-NEXT:    addi a1, a1, 16
+; RV32-NEXT:    vl8r.v v24, (a1) # Unknown-size Folded Reload
 ; RV32-NEXT:    vrgather.vv v8, v16, v24, v0.t
 ; RV32-NEXT:    vsetivli zero, 10, e32, m4, tu, ma
 ; RV32-NEXT:    vslideup.vi v4, v8, 0
 ; RV32-NEXT:    lui a1, %hi(.LCPI6_13)
 ; RV32-NEXT:    addi a1, a1, %lo(.LCPI6_13)
 ; RV32-NEXT:    vsetivli zero, 16, e32, m4, ta, mu
 ; RV32-NEXT:    vle32.v v8, (a1)
 ; RV32-NEXT:    vmv1r.v v0, v1
 ; RV32-NEXT:    csrr a1, vlenb
-; RV32-NEXT:    li a3, 41
+; RV32-NEXT:    li a3, 45
 ; RV32-NEXT:    mul a1, a1, a3
 ; RV32-NEXT:    add a1, sp, a1
 ; RV32-NEXT:    addi a1, a1, 16
 ; RV32-NEXT:    vl4r.v v12, (a1) # Unknown-size Folded Reload
 ; RV32-NEXT:    csrr a1, vlenb
-; RV32-NEXT:    li a3, 61
-; RV32-NEXT:    mul a1, a1, a3
+; RV32-NEXT:    slli a3, a1, 6
+; RV32-NEXT:    add a1, a3, a1
 ; RV32-NEXT:    add a1, sp, a1
 ; RV32-NEXT:    addi a1, a1, 16
 ; RV32-NEXT:    vl8r.v v16, (a1) # Unknown-size Folded Reload
 ; RV32-NEXT:    vrgather.vv v12, v16, v8, v0.t
 ; RV32-NEXT:    csrr a1, vlenb
-; RV32-NEXT:    li a3, 41
+; RV32-NEXT:    li a3, 45
 ; RV32-NEXT:    mul a1, a1, a3
 ; RV32-NEXT:    add a1, sp, a1
 ; RV32-NEXT:    addi a1, a1, 16
 ; RV32-NEXT:    vs4r.v v12, (a1) # Unknown-size Folded Spill
 ; RV32-NEXT:    lui a1, %hi(.LCPI6_14)
 ; RV32-NEXT:    addi a1, a1, %lo(.LCPI6_14)
 ; RV32-NEXT:    vsetvli zero, a2, e32, m8, ta, mu
 ; RV32-NEXT:    lui a2, %hi(.LCPI6_15)
 ; RV32-NEXT:    addi a2, a2, %lo(.LCPI6_15)
-; RV32-NEXT:    vle32.v v24, (a1)
+; RV32-NEXT:    vle32.v v16, (a1)
 ; RV32-NEXT:    vle32.v v8, (a2)
 ; RV32-NEXT:    csrr a1, vlenb
-; RV32-NEXT:    li a2, 61
-; RV32-NEXT:    mul a1, a1, a2
+; RV32-NEXT:    slli a2, a1, 6
+; RV32-NEXT:    add a1, a2, a1
 ; RV32-NEXT:    add a1, sp, a1
 ; RV32-NEXT:    addi a1, a1, 16
 ; RV32-NEXT:    vs8r.v v8, (a1) # Unknown-size Folded Spill
 ; RV32-NEXT:    csrr a1, vlenb
-; RV32-NEXT:    li a2, 69
+; RV32-NEXT:    li a2, 73
 ; RV32-NEXT:    mul a1, a1, a2
 ; RV32-NEXT:    add a1, sp, a1
 ; RV32-NEXT:    addi a1, a1, 16
-; RV32-NEXT:    vl8r.v v16, (a1) # Unknown-size Folded Reload
-; RV32-NEXT:    vrgather.vv v8, v16, v24
+; RV32-NEXT:    vl8r.v v24, (a1) # Unknown-size Folded Reload
+; RV32-NEXT:    vrgather.vv v8, v24, v16
 ; RV32-NEXT:    vmv1r.v v0, v2
 ; RV32-NEXT:    csrr a1, vlenb
-; RV32-NEXT:    li a2, 45
+; RV32-NEXT:    li a2, 49
 ; RV32-NEXT:    mul a1, a1, a2
 ; RV32-NEXT:    add a1, sp, a1
 ; RV32-NEXT:    addi a1, a1, 16
 ; RV32-NEXT:    vl8r.v v16, (a1) # Unknown-size Folded Reload
 ; RV32-NEXT:    csrr a1, vlenb
-; RV32-NEXT:    li a2, 61
-; RV32-NEXT:    mul a1, a1, a2
+; RV32-NEXT:    slli a2, a1, 6
+; RV32-NEXT:    add a1, a2, a1
 ; RV32-NEXT:    add a1, sp, a1
 ; RV32-NEXT:    addi a1, a1, 16
 ; RV32-NEXT:    vl8r.v v24, (a1) # Unknown-size Folded Reload
 ; RV32-NEXT:    vrgather.vv v8, v16, v24, v0.t
 ; RV32-NEXT:    vsetivli zero, 10, e32, m4, tu, ma
 ; RV32-NEXT:    csrr a1, vlenb
-; RV32-NEXT:    li a2, 41
+; RV32-NEXT:    li a2, 45
 ; RV32-NEXT:    mul a1, a1, a2
 ; RV32-NEXT:    add a1, sp, a1
 ; RV32-NEXT:    addi a1, a1, 16
 ; RV32-NEXT:    vl4r.v v12, (a1) # Unknown-size Folded Reload
 ; RV32-NEXT:    vslideup.vi v12, v8, 0
 ; RV32-NEXT:    addi a1, a0, 320
 ; RV32-NEXT:    vsetivli zero, 16, e32, m4, ta, ma
 ; RV32-NEXT:    vse32.v v12, (a1)
 ; RV32-NEXT:    addi a1, a0, 256
 ; RV32-NEXT:    vse32.v v4, (a1)
 ; RV32-NEXT:    addi a1, a0, 192
 ; RV32-NEXT:    csrr a2, vlenb
-; RV32-NEXT:    li a3, 37
-; RV32-NEXT:    mul a2, a2, a3
+; RV32-NEXT:    slli a3, a2, 5
+; RV32-NEXT:    add a2, a3, a2
 ; RV32-NEXT:    add a2, sp, a2
 ; RV32-NEXT:    addi a2, a2, 16
 ; RV32-NEXT:    vl4r.v v8, (a2) # Unknown-size Folded Reload
 ; RV32-NEXT:    vse32.v v8, (a1)
 ; RV32-NEXT:    addi a1, a0, 128
 ; RV32-NEXT:    csrr a2, vlenb
-; RV32-NEXT:    slli a3, a2, 2
-; RV32-NEXT:    add a2, a3, a2
+; RV32-NEXT:    li a3, 37
+; RV32-NEXT:    mul a2, a2, a3
 ; RV32-NEXT:    add a2, sp, a2
 ; RV32-NEXT:    addi a2, a2, 16
 ; RV32-NEXT:    vl4r.v v8, (a2) # Unknown-size Folded Reload
 ; RV32-NEXT:    vse32.v v8, (a1)
 ; RV32-NEXT:    addi a1, a0, 64
 ; RV32-NEXT:    csrr a2, vlenb
-; RV32-NEXT:    slli a3, a2, 4
-; RV32-NEXT:    add a2, a3, a2
+; RV32-NEXT:    li a3, 25
+; RV32-NEXT:    mul a2, a2, a3
 ; RV32-NEXT:    add a2, sp, a2
 ; RV32-NEXT:    addi a2, a2, 16
 ; RV32-NEXT:    vl4r.v v8, (a2) # Unknown-size Folded Reload
 ; RV32-NEXT:    vse32.v v8, (a1)
 ; RV32-NEXT:    csrr a1, vlenb
-; RV32-NEXT:    li a2, 21
+; RV32-NEXT:    li a2, 41
 ; RV32-NEXT:    mul a1, a1, a2
 ; RV32-NEXT:    add a1, sp, a1
 ; RV32-NEXT:    addi a1, a1, 16
 ; RV32-NEXT:    vl4r.v v8, (a1) # Unknown-size Folded Reload
 ; RV32-NEXT:    vse32.v v8, (a0)
 ; RV32-NEXT:    csrr a0, vlenb
-; RV32-NEXT:    li a1, 78
+; RV32-NEXT:    li a1, 82
 ; RV32-NEXT:    mul a0, a0, a1
 ; RV32-NEXT:    add sp, sp, a0
 ; RV32-NEXT:    addi sp, sp, 16
 ; RV32-NEXT:    ret
 ;
 ; RV64-LABEL: load_factor6_too_big:
 ; RV64:       # %bb.0:
 ; RV64-NEXT:    addi sp, sp, -16

llvm/test/CodeGen/RISCV/rvv/fixed-vectors-masked-gather.ll

 ; RV64ZVE32F-NEXT:    andi a2, a1, 2
 ; RV64ZVE32F-NEXT:    beqz a2, .LBB98_4
 ; RV64ZVE32F-NEXT:  # %bb.3: # %cond.load1
 ; RV64ZVE32F-NEXT:    vsetivli zero, 1, e8, mf4, ta, ma
 ; RV64ZVE32F-NEXT:    vslidedown.vi v12, v8, 1
 ; RV64ZVE32F-NEXT:    vmv.x.s a2, v12
 ; RV64ZVE32F-NEXT:    add a2, a0, a2
 ; RV64ZVE32F-NEXT:    lbu a2, 0(a2)
 ; RV64ZVE32F-NEXT:    li a3, 32
-; RV64ZVE32F-NEXT:    vsetvli zero, a3, e8, m1, ta, ma
 ; RV64ZVE32F-NEXT:    vmv.s.x v12, a2
 ; RV64ZVE32F-NEXT:    vsetivli zero, 2, e8, m2, tu, ma
 ; RV64ZVE32F-NEXT:    vslideup.vi v10, v12, 1
 ; RV64ZVE32F-NEXT:  .LBB98_4: # %else2
 ; RV64ZVE32F-NEXT:    vsetivli zero, 2, e8, mf4, ta, ma
 ; RV64ZVE32F-NEXT:    andi a2, a1, 4
 ; RV64ZVE32F-NEXT:    vslidedown.vi v12, v8, 2
 ; RV64ZVE32F-NEXT:    beqz a2, .LBB98_6
 ; RV64ZVE32F-NEXT:  # %bb.5: # %cond.load4
 ; RV64ZVE32F-NEXT:    vmv.x.s a2, v12
 ; RV64ZVE32F-NEXT:    add a2, a0, a2
 ; RV64ZVE32F-NEXT:    lbu a2, 0(a2)
 ; RV64ZVE32F-NEXT:    li a3, 32
-; RV64ZVE32F-NEXT:    vsetvli zero, a3, e8, m1, ta, ma
 ; RV64ZVE32F-NEXT:    vmv.s.x v14, a2
 ; RV64ZVE32F-NEXT:    vsetivli zero, 3, e8, m2, tu, ma
 ; RV64ZVE32F-NEXT:    vslideup.vi v10, v14, 2
 ; RV64ZVE32F-NEXT:  .LBB98_6: # %else5
 ; RV64ZVE32F-NEXT:    vsetivli zero, 4, e8, mf2, ta, ma
 ; RV64ZVE32F-NEXT:    andi a2, a1, 8
 ; RV64ZVE32F-NEXT:    vslidedown.vi v13, v8, 4
 ; RV64ZVE32F-NEXT:    bnez a2, .LBB98_50
 ; RV64ZVE32F-NEXT:  # %bb.7: # %else8
 ; RV64ZVE32F-NEXT:    andi a2, a1, 16
 ; RV64ZVE32F-NEXT:    bnez a2, .LBB98_51
 ; RV64ZVE32F-NEXT:  .LBB98_8: # %else11
 ; RV64ZVE32F-NEXT:    andi a2, a1, 32
 ; RV64ZVE32F-NEXT:    beqz a2, .LBB98_10
 ; RV64ZVE32F-NEXT:  .LBB98_9: # %cond.load13
 ; RV64ZVE32F-NEXT:    vsetivli zero, 1, e8, mf4, ta, ma
 ; RV64ZVE32F-NEXT:    vslidedown.vi v12, v13, 1
 ; RV64ZVE32F-NEXT:    vmv.x.s a2, v12
 ; RV64ZVE32F-NEXT:    add a2, a0, a2
 ; RV64ZVE32F-NEXT:    lbu a2, 0(a2)
 ; RV64ZVE32F-NEXT:    li a3, 32
-; RV64ZVE32F-NEXT:    vsetvli zero, a3, e8, m1, ta, ma
 ; RV64ZVE32F-NEXT:    vmv.s.x v14, a2
 ; RV64ZVE32F-NEXT:    vsetivli zero, 6, e8, m2, tu, ma
 ; RV64ZVE32F-NEXT:    vslideup.vi v10, v14, 5
 ; RV64ZVE32F-NEXT:  .LBB98_10: # %else14
 ; RV64ZVE32F-NEXT:    vsetivli zero, 8, e8, m1, ta, ma
 ; RV64ZVE32F-NEXT:    vslidedown.vi v12, v8, 8
 ; RV64ZVE32F-NEXT:    vsetivli zero, 2, e8, mf4, ta, ma
 ; RV64ZVE32F-NEXT:    andi a2, a1, 64

 ; RV64ZVE32F-NEXT:    beqz a2, .LBB98_15
 ; RV64ZVE32F-NEXT:  .LBB98_14: # %cond.load25
 ; RV64ZVE32F-NEXT:    vsetivli zero, 1, e8, mf4, ta, ma
 ; RV64ZVE32F-NEXT:    vslidedown.vi v13, v12, 1
 ; RV64ZVE32F-NEXT:    vmv.x.s a2, v13
 ; RV64ZVE32F-NEXT:    add a2, a0, a2
 ; RV64ZVE32F-NEXT:    lbu a2, 0(a2)
 ; RV64ZVE32F-NEXT:    li a3, 32
-; RV64ZVE32F-NEXT:    vsetvli zero, a3, e8, m1, ta, ma
 ; RV64ZVE32F-NEXT:    vmv.s.x v14, a2
 ; RV64ZVE32F-NEXT:    vsetivli zero, 10, e8, m2, tu, ma
 ; RV64ZVE32F-NEXT:    vslideup.vi v10, v14, 9
 ; RV64ZVE32F-NEXT:  .LBB98_15: # %else26
 ; RV64ZVE32F-NEXT:    vsetivli zero, 2, e8, mf4, ta, ma
 ; RV64ZVE32F-NEXT:    andi a2, a1, 1024
 ; RV64ZVE32F-NEXT:    vslidedown.vi v13, v12, 2
 ; RV64ZVE32F-NEXT:    beqz a2, .LBB98_17
 ; RV64ZVE32F-NEXT:  # %bb.16: # %cond.load28
 ; RV64ZVE32F-NEXT:    vmv.x.s a2, v13
 ; RV64ZVE32F-NEXT:    add a2, a0, a2
 ; RV64ZVE32F-NEXT:    lbu a2, 0(a2)
 ; RV64ZVE32F-NEXT:    li a3, 32
-; RV64ZVE32F-NEXT:    vsetvli zero, a3, e8, m1, ta, ma
 ; RV64ZVE32F-NEXT:    vmv.s.x v14, a2
 ; RV64ZVE32F-NEXT:    vsetivli zero, 11, e8, m2, tu, ma
 ; RV64ZVE32F-NEXT:    vslideup.vi v10, v14, 10
 ; RV64ZVE32F-NEXT:  .LBB98_17: # %else29
 ; RV64ZVE32F-NEXT:    vsetivli zero, 4, e8, mf2, ta, ma
 ; RV64ZVE32F-NEXT:    slli a2, a1, 52
 ; RV64ZVE32F-NEXT:    vslidedown.vi v12, v12, 4
 ; RV64ZVE32F-NEXT:    bgez a2, .LBB98_19
 ; RV64ZVE32F-NEXT:  # %bb.18: # %cond.load31
 ; RV64ZVE32F-NEXT:    vsetivli zero, 1, e8, mf4, ta, ma
 ; RV64ZVE32F-NEXT:    vslidedown.vi v13, v13, 1
 ; RV64ZVE32F-NEXT:    vmv.x.s a2, v13
 ; RV64ZVE32F-NEXT:    add a2, a0, a2
 ; RV64ZVE32F-NEXT:    lbu a2, 0(a2)
 ; RV64ZVE32F-NEXT:    li a3, 32
-; RV64ZVE32F-NEXT:    vsetvli zero, a3, e8, m1, ta, ma
 ; RV64ZVE32F-NEXT:    vmv.s.x v14, a2
 ; RV64ZVE32F-NEXT:    vsetivli zero, 12, e8, m2, tu, ma
 ; RV64ZVE32F-NEXT:    vslideup.vi v10, v14, 11
 ; RV64ZVE32F-NEXT:  .LBB98_19: # %else32
 ; RV64ZVE32F-NEXT:    vsetivli zero, 16, e8, m2, ta, ma
 ; RV64ZVE32F-NEXT:    slli a2, a1, 51
 ; RV64ZVE32F-NEXT:    vslidedown.vi v8, v8, 16
 ; RV64ZVE32F-NEXT:    bgez a2, .LBB98_21
 ; RV64ZVE32F-NEXT:  # %bb.20: # %cond.load34
 ; RV64ZVE32F-NEXT:    vmv.x.s a2, v12
 ; RV64ZVE32F-NEXT:    add a2, a0, a2
 ; RV64ZVE32F-NEXT:    lbu a2, 0(a2)
 ; RV64ZVE32F-NEXT:    li a3, 32
-; RV64ZVE32F-NEXT:    vsetvli zero, a3, e8, m1, ta, ma
 ; RV64ZVE32F-NEXT:    vmv.s.x v14, a2
 ; RV64ZVE32F-NEXT:    vsetivli zero, 13, e8, m2, tu, ma
 ; RV64ZVE32F-NEXT:    vslideup.vi v10, v14, 12
 ; RV64ZVE32F-NEXT:  .LBB98_21: # %else35
 ; RV64ZVE32F-NEXT:    slli a2, a1, 50
 ; RV64ZVE32F-NEXT:    bgez a2, .LBB98_23
 ; RV64ZVE32F-NEXT:  # %bb.22: # %cond.load37
 ; RV64ZVE32F-NEXT:    vsetivli zero, 1, e8, mf4, ta, ma
 ; RV64ZVE32F-NEXT:    vslidedown.vi v9, v12, 1
 ; RV64ZVE32F-NEXT:    vmv.x.s a2, v9
 ; RV64ZVE32F-NEXT:    add a2, a0, a2
 ; RV64ZVE32F-NEXT:    lbu a2, 0(a2)
 ; RV64ZVE32F-NEXT:    li a3, 32
-; RV64ZVE32F-NEXT:    vsetvli zero, a3, e8, m1, ta, ma
 ; RV64ZVE32F-NEXT:    vmv.s.x v14, a2
 ; RV64ZVE32F-NEXT:    vsetivli zero, 14, e8, m2, tu, ma
 ; RV64ZVE32F-NEXT:    vslideup.vi v10, v14, 13
 ; RV64ZVE32F-NEXT:  .LBB98_23: # %else38
 ; RV64ZVE32F-NEXT:    vsetivli zero, 2, e8, mf4, ta, ma
 ; RV64ZVE32F-NEXT:    slli a2, a1, 49
 ; RV64ZVE32F-NEXT:    vslidedown.vi v9, v12, 2
 ; RV64ZVE32F-NEXT:    bltz a2, .LBB98_55
 ; RV64ZVE32F-NEXT:  # %bb.24: # %else41
 ; RV64ZVE32F-NEXT:    slli a2, a1, 48
 ; RV64ZVE32F-NEXT:    bltz a2, .LBB98_56
 ; RV64ZVE32F-NEXT:  .LBB98_25: # %else44
 ; RV64ZVE32F-NEXT:    slli a2, a1, 47
 ; RV64ZVE32F-NEXT:    bltz a2, .LBB98_57
 ; RV64ZVE32F-NEXT:  .LBB98_26: # %else47
 ; RV64ZVE32F-NEXT:    slli a2, a1, 46
 ; RV64ZVE32F-NEXT:    bgez a2, .LBB98_28
 ; RV64ZVE32F-NEXT:  .LBB98_27: # %cond.load49
 ; RV64ZVE32F-NEXT:    vsetivli zero, 1, e8, mf4, ta, ma
 ; RV64ZVE32F-NEXT:    vslidedown.vi v9, v8, 1
 ; RV64ZVE32F-NEXT:    vmv.x.s a2, v9
 ; RV64ZVE32F-NEXT:    add a2, a0, a2
 ; RV64ZVE32F-NEXT:    lbu a2, 0(a2)
 ; RV64ZVE32F-NEXT:    li a3, 32
-; RV64ZVE32F-NEXT:    vsetvli zero, a3, e8, m1, ta, ma
 ; RV64ZVE32F-NEXT:    vmv.s.x v12, a2
 ; RV64ZVE32F-NEXT:    vsetivli zero, 18, e8, m2, tu, ma
 ; RV64ZVE32F-NEXT:    vslideup.vi v10, v12, 17
 ; RV64ZVE32F-NEXT:  .LBB98_28: # %else50
 ; RV64ZVE32F-NEXT:    vsetivli zero, 2, e8, mf4, ta, ma
 ; RV64ZVE32F-NEXT:    slli a2, a1, 45
 ; RV64ZVE32F-NEXT:    vslidedown.vi v12, v8, 2
 ; RV64ZVE32F-NEXT:    bgez a2, .LBB98_30
 ; RV64ZVE32F-NEXT:  # %bb.29: # %cond.load52
 ; RV64ZVE32F-NEXT:    vmv.x.s a2, v12
 ; RV64ZVE32F-NEXT:    add a2, a0, a2
 ; RV64ZVE32F-NEXT:    lbu a2, 0(a2)
 ; RV64ZVE32F-NEXT:    li a3, 32
-; RV64ZVE32F-NEXT:    vsetvli zero, a3, e8, m1, ta, ma
 ; RV64ZVE32F-NEXT:    vmv.s.x v14, a2
 ; RV64ZVE32F-NEXT:    vsetivli zero, 19, e8, m2, tu, ma
 ; RV64ZVE32F-NEXT:    vslideup.vi v10, v14, 18
 ; RV64ZVE32F-NEXT:  .LBB98_30: # %else53
 ; RV64ZVE32F-NEXT:    vsetivli zero, 4, e8, mf2, ta, ma
 ; RV64ZVE32F-NEXT:    slli a2, a1, 44
 ; RV64ZVE32F-NEXT:    vslidedown.vi v9, v8, 4
 ; RV64ZVE32F-NEXT:    bltz a2, .LBB98_58
 ; RV64ZVE32F-NEXT:  # %bb.31: # %else56
 ; RV64ZVE32F-NEXT:    slli a2, a1, 43
 ; RV64ZVE32F-NEXT:    bltz a2, .LBB98_59
 ; RV64ZVE32F-NEXT:  .LBB98_32: # %else59
 ; RV64ZVE32F-NEXT:    slli a2, a1, 42
 ; RV64ZVE32F-NEXT:    bgez a2, .LBB98_34
 ; RV64ZVE32F-NEXT:  .LBB98_33: # %cond.load61
 ; RV64ZVE32F-NEXT:    vsetivli zero, 1, e8, mf4, ta, ma
 ; RV64ZVE32F-NEXT:    vslidedown.vi v12, v9, 1
 ; RV64ZVE32F-NEXT:    vmv.x.s a2, v12
 ; RV64ZVE32F-NEXT:    add a2, a0, a2
 ; RV64ZVE32F-NEXT:    lbu a2, 0(a2)
 ; RV64ZVE32F-NEXT:    li a3, 32
-; RV64ZVE32F-NEXT:    vsetvli zero, a3, e8, m1, ta, ma
 ; RV64ZVE32F-NEXT:    vmv.s.x v12, a2
 ; RV64ZVE32F-NEXT:    vsetivli zero, 22, e8, m2, tu, ma
 ; RV64ZVE32F-NEXT:    vslideup.vi v10, v12, 21
 ; RV64ZVE32F-NEXT:  .LBB98_34: # %else62
 ; RV64ZVE32F-NEXT:    vsetivli zero, 8, e8, m1, ta, ma
 ; RV64ZVE32F-NEXT:    vslidedown.vi v8, v8, 8
 ; RV64ZVE32F-NEXT:    vsetivli zero, 2, e8, mf4, ta, ma
 ; RV64ZVE32F-NEXT:    slli a2, a1, 41

 ; RV64ZVE32F-NEXT:    bgez a2, .LBB98_39
 ; RV64ZVE32F-NEXT:  .LBB98_38: # %cond.load73
 ; RV64ZVE32F-NEXT:    vsetivli zero, 1, e8, mf4, ta, ma
 ; RV64ZVE32F-NEXT:    vslidedown.vi v9, v8, 1
 ; RV64ZVE32F-NEXT:    vmv.x.s a2, v9
 ; RV64ZVE32F-NEXT:    add a2, a0, a2
 ; RV64ZVE32F-NEXT:    lbu a2, 0(a2)
 ; RV64ZVE32F-NEXT:    li a3, 32
-; RV64ZVE32F-NEXT:    vsetvli zero, a3, e8, m1, ta, ma
 ; RV64ZVE32F-NEXT:    vmv.s.x v12, a2
 ; RV64ZVE32F-NEXT:    vsetivli zero, 26, e8, m2, tu, ma
 ; RV64ZVE32F-NEXT:    vslideup.vi v10, v12, 25
 ; RV64ZVE32F-NEXT:  .LBB98_39: # %else74
 ; RV64ZVE32F-NEXT:    vsetivli zero, 2, e8, mf4, ta, ma
 ; RV64ZVE32F-NEXT:    slli a2, a1, 37
 ; RV64ZVE32F-NEXT:    vslidedown.vi v9, v8, 2
 ; RV64ZVE32F-NEXT:    bgez a2, .LBB98_41
 ; RV64ZVE32F-NEXT:  # %bb.40: # %cond.load76
 ; RV64ZVE32F-NEXT:    vmv.x.s a2, v9
 ; RV64ZVE32F-NEXT:    add a2, a0, a2
 ; RV64ZVE32F-NEXT:    lbu a2, 0(a2)
 ; RV64ZVE32F-NEXT:    li a3, 32
-; RV64ZVE32F-NEXT:    vsetvli zero, a3, e8, m1, ta, ma
 ; RV64ZVE32F-NEXT:    vmv.s.x v12, a2
 ; RV64ZVE32F-NEXT:    vsetivli zero, 27, e8, m2, tu, ma
 ; RV64ZVE32F-NEXT:    vslideup.vi v10, v12, 26
 ; RV64ZVE32F-NEXT:  .LBB98_41: # %else77
 ; RV64ZVE32F-NEXT:    vsetivli zero, 4, e8, mf2, ta, ma
 ; RV64ZVE32F-NEXT:    slli a2, a1, 36
 ; RV64ZVE32F-NEXT:    vslidedown.vi v8, v8, 4
 ; RV64ZVE32F-NEXT:    bltz a2, .LBB98_63
 ; RV64ZVE32F-NEXT:  # %bb.42: # %else80
 ; RV64ZVE32F-NEXT:    slli a2, a1, 35
 ; RV64ZVE32F-NEXT:    bltz a2, .LBB98_64
 ; RV64ZVE32F-NEXT:  .LBB98_43: # %else83
 ; RV64ZVE32F-NEXT:    slli a2, a1, 34
 ; RV64ZVE32F-NEXT:    bgez a2, .LBB98_45
 ; RV64ZVE32F-NEXT:  .LBB98_44: # %cond.load85
 ; RV64ZVE32F-NEXT:    vsetivli zero, 1, e8, mf4, ta, ma
 ; RV64ZVE32F-NEXT:    vslidedown.vi v9, v8, 1
 ; RV64ZVE32F-NEXT:    vmv.x.s a2, v9
 ; RV64ZVE32F-NEXT:    add a2, a0, a2
 ; RV64ZVE32F-NEXT:    lbu a2, 0(a2)
 ; RV64ZVE32F-NEXT:    li a3, 32
-; RV64ZVE32F-NEXT:    vsetvli zero, a3, e8, m1, ta, ma
 ; RV64ZVE32F-NEXT:    vmv.s.x v12, a2
 ; RV64ZVE32F-NEXT:    vsetivli zero, 30, e8, m2, tu, ma
 ; RV64ZVE32F-NEXT:    vslideup.vi v10, v12, 29
 ; RV64ZVE32F-NEXT:  .LBB98_45: # %else86
 ; RV64ZVE32F-NEXT:    vsetivli zero, 2, e8, mf4, ta, ma
 ; RV64ZVE32F-NEXT:    slli a2, a1, 33
 ; RV64ZVE32F-NEXT:    vslidedown.vi v8, v8, 2
 ; RV64ZVE32F-NEXT:    bgez a2, .LBB98_47
 ; RV64ZVE32F-NEXT:  # %bb.46: # %cond.load88
 ; RV64ZVE32F-NEXT:    vmv.x.s a2, v8
 ; RV64ZVE32F-NEXT:    add a2, a0, a2
 ; RV64ZVE32F-NEXT:    lbu a2, 0(a2)
 ; RV64ZVE32F-NEXT:    li a3, 32
-; RV64ZVE32F-NEXT:    vsetvli zero, a3, e8, m1, ta, ma
 ; RV64ZVE32F-NEXT:    vmv.s.x v12, a2
 ; RV64ZVE32F-NEXT:    vsetivli zero, 31, e8, m2, tu, ma
 ; RV64ZVE32F-NEXT:    vslideup.vi v10, v12, 30
 ; RV64ZVE32F-NEXT:  .LBB98_47: # %else89
 ; RV64ZVE32F-NEXT:    lui a2, 524288
 ; RV64ZVE32F-NEXT:    and a1, a1, a2
 ; RV64ZVE32F-NEXT:    beqz a1, .LBB98_49
 ; RV64ZVE32F-NEXT:  # %bb.48: # %cond.load91
 ; RV64ZVE32F-NEXT:    vsetivli zero, 1, e8, mf4, ta, ma
 ; RV64ZVE32F-NEXT:    vslidedown.vi v8, v8, 1
 ; RV64ZVE32F-NEXT:    vmv.x.s a1, v8
 ; RV64ZVE32F-NEXT:    add a0, a0, a1
 ; RV64ZVE32F-NEXT:    lbu a0, 0(a0)
 ; RV64ZVE32F-NEXT:    li a1, 32
-; RV64ZVE32F-NEXT:    vsetvli zero, a1, e8, m1, ta, ma
 ; RV64ZVE32F-NEXT:    vmv.s.x v8, a0
 ; RV64ZVE32F-NEXT:    vsetvli zero, a1, e8, m2, ta, ma
 ; RV64ZVE32F-NEXT:    vslideup.vi v10, v8, 31
 ; RV64ZVE32F-NEXT:  .LBB98_49: # %else92
 ; RV64ZVE32F-NEXT:    vmv2r.v v8, v10
 ; RV64ZVE32F-NEXT:    ret
 ; RV64ZVE32F-NEXT:  .LBB98_50: # %cond.load7
 ; RV64ZVE32F-NEXT:    vsetivli zero, 1, e8, mf4, ta, ma
 ; RV64ZVE32F-NEXT:    vslidedown.vi v12, v12, 1
 ; RV64ZVE32F-NEXT:    vmv.x.s a2, v12
 ; RV64ZVE32F-NEXT:    add a2, a0, a2
 ; RV64ZVE32F-NEXT:    lbu a2, 0(a2)
 ; RV64ZVE32F-NEXT:    li a3, 32
-; RV64ZVE32F-NEXT:    vsetvli zero, a3, e8, m1, ta, ma
 ; RV64ZVE32F-NEXT:    vmv.s.x v14, a2
 ; RV64ZVE32F-NEXT:    vsetivli zero, 4, e8, m2, tu, ma
 ; RV64ZVE32F-NEXT:    vslideup.vi v10, v14, 3
 ; RV64ZVE32F-NEXT:    andi a2, a1, 16
 ; RV64ZVE32F-NEXT:    beqz a2, .LBB98_8
 ; RV64ZVE32F-NEXT:  .LBB98_51: # %cond.load10
 ; RV64ZVE32F-NEXT:    vsetivli zero, 0, e8, mf4, ta, ma
 ; RV64ZVE32F-NEXT:    vmv.x.s a2, v13
 ; RV64ZVE32F-NEXT:    add a2, a0, a2
 ; RV64ZVE32F-NEXT:    lbu a2, 0(a2)
 ; RV64ZVE32F-NEXT:    li a3, 32
 ; RV64ZVE32F-NEXT:    vsetvli zero, a3, e8, m1, ta, ma
 ; RV64ZVE32F-NEXT:    vmv.s.x v14, a2
 ; RV64ZVE32F-NEXT:    vsetivli zero, 5, e8, m2, tu, ma
 ; RV64ZVE32F-NEXT:    vslideup.vi v10, v14, 4
 ; RV64ZVE32F-NEXT:    andi a2, a1, 32
 ; RV64ZVE32F-NEXT:    bnez a2, .LBB98_9
 ; RV64ZVE32F-NEXT:    j .LBB98_10
 ; RV64ZVE32F-NEXT:  .LBB98_52: # %cond.load16
 ; RV64ZVE32F-NEXT:    vmv.x.s a2, v13
 ; RV64ZVE32F-NEXT:    add a2, a0, a2
 ; RV64ZVE32F-NEXT:    lbu a2, 0(a2)
 ; RV64ZVE32F-NEXT:    li a3, 32
-; RV64ZVE32F-NEXT:    vsetvli zero, a3, e8, m1, ta, ma
 ; RV64ZVE32F-NEXT:    vmv.s.x v14, a2
 ; RV64ZVE32F-NEXT:    vsetivli zero, 7, e8, m2, tu, ma
 ; RV64ZVE32F-NEXT:    vslideup.vi v10, v14, 6
 ; RV64ZVE32F-NEXT:    andi a2, a1, 128
 ; RV64ZVE32F-NEXT:    beqz a2, .LBB98_12
 ; RV64ZVE32F-NEXT:  .LBB98_53: # %cond.load19
 ; RV64ZVE32F-NEXT:    vsetivli zero, 1, e8, mf4, ta, ma
 ; RV64ZVE32F-NEXT:    vslidedown.vi v13, v13, 1
 ; RV64ZVE32F-NEXT:    vmv.x.s a2, v13
 ; RV64ZVE32F-NEXT:    add a2, a0, a2
 ; RV64ZVE32F-NEXT:    lbu a2, 0(a2)
 ; RV64ZVE32F-NEXT:    li a3, 32
-; RV64ZVE32F-NEXT:    vsetvli zero, a3, e8, m1, ta, ma
 ; RV64ZVE32F-NEXT:    vmv.s.x v14, a2
 ; RV64ZVE32F-NEXT:    vsetivli zero, 8, e8, m2, tu, ma
 ; RV64ZVE32F-NEXT:    vslideup.vi v10, v14, 7
 ; RV64ZVE32F-NEXT:    andi a2, a1, 256
 ; RV64ZVE32F-NEXT:    beqz a2, .LBB98_13
 ; RV64ZVE32F-NEXT:  .LBB98_54: # %cond.load22
 ; RV64ZVE32F-NEXT:    vsetivli zero, 0, e8, mf4, ta, ma
 ; RV64ZVE32F-NEXT:    vmv.x.s a2, v12
 ; RV64ZVE32F-NEXT:    add a2, a0, a2
 ; RV64ZVE32F-NEXT:    lbu a2, 0(a2)
 ; RV64ZVE32F-NEXT:    li a3, 32
 ; RV64ZVE32F-NEXT:    vsetvli zero, a3, e8, m1, ta, ma
 ; RV64ZVE32F-NEXT:    vmv.s.x v14, a2
 ; RV64ZVE32F-NEXT:    vsetivli zero, 9, e8, m2, tu, ma
 ; RV64ZVE32F-NEXT:    vslideup.vi v10, v14, 8
 ; RV64ZVE32F-NEXT:    andi a2, a1, 512
 ; RV64ZVE32F-NEXT:    bnez a2, .LBB98_14
 ; RV64ZVE32F-NEXT:    j .LBB98_15
 ; RV64ZVE32F-NEXT:  .LBB98_55: # %cond.load40
 ; RV64ZVE32F-NEXT:    vmv.x.s a2, v9
 ; RV64ZVE32F-NEXT:    add a2, a0, a2
 ; RV64ZVE32F-NEXT:    lbu a2, 0(a2)
 ; RV64ZVE32F-NEXT:    li a3, 32
-; RV64ZVE32F-NEXT:    vsetvli zero, a3, e8, m1, ta, ma
 ; RV64ZVE32F-NEXT:    vmv.s.x v12, a2
 ; RV64ZVE32F-NEXT:    vsetivli zero, 15, e8, m2, tu, ma
 ; RV64ZVE32F-NEXT:    vslideup.vi v10, v12, 14
 ; RV64ZVE32F-NEXT:    slli a2, a1, 48
 ; RV64ZVE32F-NEXT:    bgez a2, .LBB98_25
 ; RV64ZVE32F-NEXT:  .LBB98_56: # %cond.load43
 ; RV64ZVE32F-NEXT:    vsetivli zero, 1, e8, mf4, ta, ma
 ; RV64ZVE32F-NEXT:    vslidedown.vi v9, v9, 1
 ; RV64ZVE32F-NEXT:    vmv.x.s a2, v9
 ; RV64ZVE32F-NEXT:    add a2, a0, a2
 ; RV64ZVE32F-NEXT:    lbu a2, 0(a2)
 ; RV64ZVE32F-NEXT:    li a3, 32
-; RV64ZVE32F-NEXT:    vsetvli zero, a3, e8, m1, ta, ma
 ; RV64ZVE32F-NEXT:    vmv.s.x v12, a2
 ; RV64ZVE32F-NEXT:    vsetivli zero, 16, e8, m2, tu, ma
 ; RV64ZVE32F-NEXT:    vslideup.vi v10, v12, 15
 ; RV64ZVE32F-NEXT:    slli a2, a1, 47
 ; RV64ZVE32F-NEXT:    bgez a2, .LBB98_26
 ; RV64ZVE32F-NEXT:  .LBB98_57: # %cond.load46
 ; RV64ZVE32F-NEXT:    vsetivli zero, 0, e8, mf4, ta, ma
 ; RV64ZVE32F-NEXT:    vmv.x.s a2, v8

 ; RV64ZVE32F-NEXT:    j .LBB98_28
 ; RV64ZVE32F-NEXT:  .LBB98_58: # %cond.load55
 ; RV64ZVE32F-NEXT:    vsetivli zero, 1, e8, mf4, ta, ma
 ; RV64ZVE32F-NEXT:    vslidedown.vi v12, v12, 1
 ; RV64ZVE32F-NEXT:    vmv.x.s a2, v12
 ; RV64ZVE32F-NEXT:    add a2, a0, a2
 ; RV64ZVE32F-NEXT:    lbu a2, 0(a2)
 ; RV64ZVE32F-NEXT:    li a3, 32
-; RV64ZVE32F-NEXT:    vsetvli zero, a3, e8, m1, ta, ma
 ; RV64ZVE32F-NEXT:    vmv.s.x v12, a2
 ; RV64ZVE32F-NEXT:    vsetivli zero, 20, e8, m2, tu, ma
 ; RV64ZVE32F-NEXT:    vslideup.vi v10, v12, 19
 ; RV64ZVE32F-NEXT:    slli a2, a1, 43
 ; RV64ZVE32F-NEXT:    bgez a2, .LBB98_32
 ; RV64ZVE32F-NEXT:  .LBB98_59: # %cond.load58
 ; RV64ZVE32F-NEXT:    vsetivli zero, 0, e8, mf4, ta, ma
 ; RV64ZVE32F-NEXT:    vmv.x.s a2, v9
 ; RV64ZVE32F-NEXT:    add a2, a0, a2
 ; RV64ZVE32F-NEXT:    lbu a2, 0(a2)
 ; RV64ZVE32F-NEXT:    li a3, 32
 ; RV64ZVE32F-NEXT:    vsetvli zero, a3, e8, m1, ta, ma
 ; RV64ZVE32F-NEXT:    vmv.s.x v12, a2
 ; RV64ZVE32F-NEXT:    vsetivli zero, 21, e8, m2, tu, ma
 ; RV64ZVE32F-NEXT:    vslideup.vi v10, v12, 20
 ; RV64ZVE32F-NEXT:    slli a2, a1, 42
 ; RV64ZVE32F-NEXT:    bltz a2, .LBB98_33
 ; RV64ZVE32F-NEXT:    j .LBB98_34
 ; RV64ZVE32F-NEXT:  .LBB98_60: # %cond.load64
 ; RV64ZVE32F-NEXT:    vmv.x.s a2, v9
 ; RV64ZVE32F-NEXT:    add a2, a0, a2
 ; RV64ZVE32F-NEXT:    lbu a2, 0(a2)
 ; RV64ZVE32F-NEXT:    li a3, 32
-; RV64ZVE32F-NEXT:    vsetvli zero, a3, e8, m1, ta, ma
 ; RV64ZVE32F-NEXT:    vmv.s.x v12, a2
 ; RV64ZVE32F-NEXT:    vsetivli zero, 23, e8, m2, tu, ma
 ; RV64ZVE32F-NEXT:    vslideup.vi v10, v12, 22
 ; RV64ZVE32F-NEXT:    slli a2, a1, 40
 ; RV64ZVE32F-NEXT:    bgez a2, .LBB98_36
 ; RV64ZVE32F-NEXT:  .LBB98_61: # %cond.load67
 ; RV64ZVE32F-NEXT:    vsetivli zero, 1, e8, mf4, ta, ma
 ; RV64ZVE32F-NEXT:    vslidedown.vi v9, v9, 1
 ; RV64ZVE32F-NEXT:    vmv.x.s a2, v9
 ; RV64ZVE32F-NEXT:    add a2, a0, a2
 ; RV64ZVE32F-NEXT:    lbu a2, 0(a2)
 ; RV64ZVE32F-NEXT:    li a3, 32
-; RV64ZVE32F-NEXT:    vsetvli zero, a3, e8, m1, ta, ma
 ; RV64ZVE32F-NEXT:    vmv.s.x v12, a2
 ; RV64ZVE32F-NEXT:    vsetivli zero, 24, e8, m2, tu, ma
 ; RV64ZVE32F-NEXT:    vslideup.vi v10, v12, 23
 ; RV64ZVE32F-NEXT:    slli a2, a1, 39
 ; RV64ZVE32F-NEXT:    bgez a2, .LBB98_37
 ; RV64ZVE32F-NEXT:  .LBB98_62: # %cond.load70
 ; RV64ZVE32F-NEXT:    vsetivli zero, 0, e8, mf4, ta, ma
 ; RV64ZVE32F-NEXT:    vmv.x.s a2, v8

 ; RV64ZVE32F-NEXT:    j .LBB98_39
 ; RV64ZVE32F-NEXT:  .LBB98_63: # %cond.load79
 ; RV64ZVE32F-NEXT:    vsetivli zero, 1, e8, mf4, ta, ma
 ; RV64ZVE32F-NEXT:    vslidedown.vi v9, v9, 1
 ; RV64ZVE32F-NEXT:    vmv.x.s a2, v9
 ; RV64ZVE32F-NEXT:    add a2, a0, a2
 ; RV64ZVE32F-NEXT:    lbu a2, 0(a2)
 ; RV64ZVE32F-NEXT:    li a3, 32
-; RV64ZVE32F-NEXT:    vsetvli zero, a3, e8, m1, ta, ma
 ; RV64ZVE32F-NEXT:    vmv.s.x v12, a2
 ; RV64ZVE32F-NEXT:    vsetivli zero, 28, e8, m2, tu, ma
 ; RV64ZVE32F-NEXT:    vslideup.vi v10, v12, 27
 ; RV64ZVE32F-NEXT:    slli a2, a1, 35
 ; RV64ZVE32F-NEXT:    bgez a2, .LBB98_43
 ; RV64ZVE32F-NEXT:  .LBB98_64: # %cond.load82
 ; RV64ZVE32F-NEXT:    vsetivli zero, 0, e8, mf4, ta, ma
 ; RV64ZVE32F-NEXT:    vmv.x.s a2, v8

llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vand-vp.ll

 }
 
 define <11 x i64> @vand_vx_v11i64(<11 x i64> %va, i64 %b, <11 x i1> %m, i32 zeroext %evl) {
 ; RV32-LABEL: vand_vx_v11i64:
 ; RV32:       # %bb.0:
 ; RV32-NEXT:    vmv1r.v v16, v0
 ; RV32-NEXT:    li a3, 32
 ; RV32-NEXT:    vsetvli zero, a3, e32, m8, ta, ma
+; RV32-NEXT:    lui a3, 341
+; RV32-NEXT:    addi a3, a3, 1365
+; RV32-NEXT:    vmv.s.x v0, a3
 ; RV32-NEXT:    vmv.v.x v24, a1
-; RV32-NEXT:    lui a1, 341
-; RV32-NEXT:    addi a1, a1, 1365
-; RV32-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
-; RV32-NEXT:    vmv.s.x v0, a1
-; RV32-NEXT:    vsetvli zero, a3, e32, m8, ta, ma
 ; RV32-NEXT:    vmerge.vxm v24, v24, a0, v0
 ; RV32-NEXT:    vsetvli zero, a2, e64, m8, ta, ma
 ; RV32-NEXT:    vmv1r.v v0, v16
 ; RV32-NEXT:    vand.vv v8, v8, v24, v0.t
 ; RV32-NEXT:    ret
 ;
 ; RV64-LABEL: vand_vx_v11i64:
 ; RV64:       # %bb.0:
 ; RV64-NEXT:    vsetvli zero, a1, e64, m8, ta, ma
 ; RV64-NEXT:    vand.vx v8, v8, a0, v0.t
 ; RV64-NEXT:    ret
   %elt.head = insertelement <11 x i64> poison, i64 %b, i32 0
   %vb = shufflevector <11 x i64> %elt.head, <11 x i64> poison, <11 x i32> zeroinitializer
   %v = call <11 x i64> @llvm.vp.and.v11i64(<11 x i64> %va, <11 x i64> %vb, <11 x i1> %m, i32 %evl)
   ret <11 x i64> %v
 }
 
 define <11 x i64> @vand_vx_v11i64_unmasked(<11 x i64> %va, i64 %b, i32 zeroext %evl) {
 ; RV32-LABEL: vand_vx_v11i64_unmasked:
 ; RV32:       # %bb.0:
 ; RV32-NEXT:    li a3, 32
 ; RV32-NEXT:    vsetvli zero, a3, e32, m8, ta, ma
+; RV32-NEXT:    lui a3, 341
+; RV32-NEXT:    addi a3, a3, 1365
+; RV32-NEXT:    vmv.s.x v0, a3
 ; RV32-NEXT:    vmv.v.x v16, a1
-; RV32-NEXT:    lui a1, 341
-; RV32-NEXT:    addi a1, a1, 1365
-; RV32-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
-; RV32-NEXT:    vmv.s.x v0, a1
-; RV32-NEXT:    vsetvli zero, a3, e32, m8, ta, ma
 ; RV32-NEXT:    vmerge.vxm v16, v16, a0, v0
 ; RV32-NEXT:    vsetvli zero, a2, e64, m8, ta, ma
 ; RV32-NEXT:    vand.vv v8, v8, v16
 ; RV32-NEXT:    ret
 ;
 ; RV64-LABEL: vand_vx_v11i64_unmasked:
 ; RV64:       # %bb.0:
 ; RV64-NEXT:    vsetvli zero, a1, e64, m8, ta, ma

llvm/test/CodeGen/RISCV/rvv/shuffle-reverse.ll

 }
 
 define <32 x i32> @v16i32_2(<16 x i32> %a, <16 x i32> %b) {
 ; RV32-LABEL: v16i32_2:
 ; RV32:       # %bb.0:
 ; RV32-NEXT:    lui a0, %hi(.LCPI23_0)
 ; RV32-NEXT:    addi a0, a0, %lo(.LCPI23_0)
 ; RV32-NEXT:    li a1, 32
-; RV32-NEXT:    vsetvli zero, a1, e32, m8, ta, ma
+; RV32-NEXT:    vsetvli zero, a1, e32, m8, ta, mu
 ; RV32-NEXT:    vle32.v v0, (a0)
 ; RV32-NEXT:    vmv4r.v v24, v12
 ; RV32-NEXT:    vmv4r.v v16, v8
 ; RV32-NEXT:    vrgather.vv v8, v16, v0
 ; RV32-NEXT:    vid.v v16
 ; RV32-NEXT:    vrsub.vi v16, v16, 15
 ; RV32-NEXT:    lui a0, 16
 ; RV32-NEXT:    addi a0, a0, -1
-; RV32-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
 ; RV32-NEXT:    vmv.s.x v0, a0
-; RV32-NEXT:    vsetvli zero, a1, e32, m8, ta, mu
 ; RV32-NEXT:    vrgather.vv v8, v24, v16, v0.t
 ; RV32-NEXT:    ret
 ;
 ; RV64-LABEL: v16i32_2:
 ; RV64:       # %bb.0:
 ; RV64-NEXT:    lui a0, %hi(.LCPI23_0)
 ; RV64-NEXT:    addi a0, a0, %lo(.LCPI23_0)
 ; RV64-NEXT:    li a1, 32
-; RV64-NEXT:    vsetvli zero, a1, e32, m8, ta, ma
+; RV64-NEXT:    vsetvli zero, a1, e32, m8, ta, mu
 ; RV64-NEXT:    vle32.v v0, (a0)
 ; RV64-NEXT:    vmv4r.v v24, v12
 ; RV64-NEXT:    vmv4r.v v16, v8
 ; RV64-NEXT:    vrgather.vv v8, v16, v0
 ; RV64-NEXT:    vid.v v16
 ; RV64-NEXT:    vrsub.vi v16, v16, 15
 ; RV64-NEXT:    lui a0, 16
 ; RV64-NEXT:    addiw a0, a0, -1
-; RV64-NEXT:    vsetivli zero, 1, e32, mf2, ta, ma
 ; RV64-NEXT:    vmv.s.x v0, a0
-; RV64-NEXT:    vsetvli zero, a1, e32, m8, ta, mu
 ; RV64-NEXT:    vrgather.vv v8, v24, v16, v0.t
 ; RV64-NEXT:    ret
   %v32i32 = shufflevector <16 x i32> %a, <16 x i32> %b,  <32 x i32> <i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
   ret <32 x i32> %v32i32
 }
 
 define <2 x i64> @v2i64(<2 x i64> %a) {
 ; CHECK-LABEL: v2i64: