[RISCV] Optimize INSERT_VECTOR_ELT sequences

[frasercrmck]

Summary

This patch optimizes the codegen for INSERT_VECTOR_ELT in various ways.
Primarily, it removes the use of vslidedown during lowering, and the
vector element is inserted entirely using vslideup with a custom VL and
slide index.

Additionally, lowering of i64-element vectors on RV32 has been optimized
in several ways. When the 64-bit value to insert is the same as the
sign-extension of the lower 32-bits, the codegen can follow the regular
path. When this is not possible, a new sequence of two i32 vslide1up
instructions is used to get the vector element into a vector. This
sequence was suggested by @craig.topper. From there, the value is slid
into the final position for more consistent lowering across RV32 and
RV64.

[craig.topper]

llvm/lib/Target/RISCV/RISCVISelLowering.cpp

2249

Would it be simpler to just check that Val is a ConstantSDNode that fits in sign extended 32 bits and if it is just call getConstant with the lower 32 bits?

2279

I can believe that doesn't work. Undef sources that aren't tied don't seem to really allocate and just pick the first register.

I'm now a little worried someone might try that in the intrinsic interface.

2287

Don't we need a bitcast here to get back to the ContainerVT to match the SDTypeProfile for the next operation?

[frasercrmck]

• rebase
• address review feedback

llvm/lib/Target/RISCV/RISCVISelLowering.cpp

2249

It would, thanks. I'll still extract ValLo and ValHi for the slow rv32/i64 case, but the lo/hi constant checking is unnecessarily complicated.

2279

Indeed. I went through some of the intrinsic tests to see if they were testing undef but didn't get any hits. I suspect we have problems there. I haven't worked with the builtin/intrinsic layers but presumably it'd be possible to do something like this in pseudo C:

nxvi32_t v;
v = rvv_vslide1up(v, a, m, vl);

and have that first operand lower to undef? I don't know how far we need to take this issue at this point, though I suspect it's not the last we've seen of it. Perhaps we'll just need to sanitize any problematic undef operands.

2287

Good catch, thanks. I've fixed that up.

On a related note, I noticed that it was perfectly content when using the original container mask, even though SDTCisSameNumEltsAs shouldn't have let (e.g.) nxv2i1 match against nxv4i1. The code for the TableGen constraint seems like it should be working, so maybe it's something in our patterns? I didn't have time to dig in more deeply.

[craig.topper]

llvm/lib/Target/RISCV/RISCVISelLowering.cpp

2287

The SDTypeProfiles cause tablegen to remove type checks from the isel table to minimize table size assuming the rules have been followed.. So if you tell it that something has the same number of elements as another result or operand and tell it the specific element type it won't emit a check.

[craig.topper]

llvm/lib/Target/RISCV/RISCVISelLowering.cpp

2250

I'm not sure how I feel about creating these extracts that might not get used. Having them here instead of just in the non-simple path only saves us from having to call getConstant where we assign Val = ValLo?

[frasercrmck]

• rebase
• address review feedback (pt. 2)

[frasercrmck]

llvm/lib/Target/RISCV/RISCVISelLowering.cpp

2250

Yeah fair enough, looking at it again in the cold light of day it doesn't seem like it brings any benefit and adds extra variables to scope in which they're never going to be used.

[craig.topper]

Thanks! LGTM

llvm/lib/Target/RISCV/RISCVISelLowering.h

enum NodeType : unsigned {
   VLEFF,
   VLEFF_MASK,
   // Matches the semantics of vslideup/vslidedown. The first operand is the
   // pass-thru operand, the second is the source vector, the third is the
   // XLenVT index (either constant or non-constant), the fourth is the mask
   // and the fifth the VL.
   VSLIDEUP_VL,
   VSLIDEDOWN_VL,
+  // Matches the semantics of vslide1up. The first operand is the source
+  // vector, the second is the XLenVT scalar value. The third and fourth
+  // operands are the mask and VL operands.
+  VSLIDE1UP_VL,
   // Matches the semantics of the vid.v instruction, with a mask and VL
   // operand.
   VID_VL,
   // Matches the semantics of the vfcnvt.rod function (Convert double-width
   // float to single-width float, rounding towards odd). Takes a double-width
   // float vector and produces a single-width float vector. Also has a mask and
   // VL operand.
   VFNCVT_ROD_VL,

llvm/lib/Target/RISCV/RISCVISelLowering.cpp

SDValue RISCVTargetLowering::lowerVectorMaskTrunc(SDValue Op,
   MVT MaskContainerVT = ContainerVT.changeVectorElementType(MVT::i1);
   SDValue Trunc =
       DAG.getNode(RISCVISD::AND_VL, DL, ContainerVT, Src, SplatOne, Mask, VL);
   Trunc = DAG.getNode(RISCVISD::SETCC_VL, DL, MaskContainerVT, Trunc, SplatZero,
                       DAG.getCondCode(ISD::SETNE), Mask, VL);
   return convertFromScalableVector(MaskVT, Trunc, DAG, Subtarget);
 }
 
+// Custom-legalize INSERT_VECTOR_ELT so that the value is inserted into the
+// first position of a vector, and that vector is slid up to the insert index.
+// By limiting the active vector length to index+1 and merging with the
+// original vector (with an undisturbed tail policy for elements >= VL), we
+// achieve the desired result of leaving all elements untouched except the one
+// at VL-1, which is replaced with the desired value.
 SDValue RISCVTargetLowering::lowerINSERT_VECTOR_ELT(SDValue Op,
                                                     SelectionDAG &DAG) const {
   SDLoc DL(Op);
   MVT VecVT = Op.getSimpleValueType();
   SDValue Vec = Op.getOperand(0);
   SDValue Val = Op.getOperand(1);
   SDValue Idx = Op.getOperand(2);
 
   MVT ContainerVT = VecVT;
   // If the operand is a fixed-length vector, convert to a scalable one.
   if (VecVT.isFixedLengthVector()) {
     ContainerVT = getContainerForFixedLengthVector(VecVT);
     Vec = convertToScalableVector(ContainerVT, Vec, DAG, Subtarget);
   }
 
+  MVT XLenVT = Subtarget.getXLenVT();
+
+  SDValue ValLo, ValHi;
+  SDValue Zero = DAG.getConstant(0, DL, XLenVT);
+  bool IsSimpleInsert = Subtarget.is64Bit() || Val.getValueType() != MVT::i64;
+  // Even i64-element vectors on RV32 can be lowered simply if the
+  // most-significant 32 bits of the value are not affected by the
+  // sign-extension of the lower 32 bits.
+  // TODO: We could also catch sign extensions of a 32-bit value.
+  if (!IsSimpleInsert) {
+    SDValue One = DAG.getConstant(1, DL, XLenVT);
+    ValLo = DAG.getNode(ISD::EXTRACT_ELEMENT, DL, MVT::i32, Val, Zero);
+    ValHi = DAG.getNode(ISD::EXTRACT_ELEMENT, DL, MVT::i32, Val, One);
+    if (isa<ConstantSDNode>(Val) &&
+        isInt<32>(cast<ConstantSDNode>(Val)->getSExtValue())) {
+      Val = ValLo;
+      IsSimpleInsert = true;
+    }
+  }
+
   SDValue Mask, VL;
   std::tie(Mask, VL) = getDefaultVLOps(VecVT, ContainerVT, DL, DAG, Subtarget);
 
-  // Custom-legalize INSERT_VECTOR_ELT where XLEN>=SEW, so that the vector is
-  // first slid down into position, the value is inserted into the first
-  // position, and the vector is slid back up. We do this to simplify patterns.
-  //   (slideup vec, (insertelt (slidedown impdef, vec, idx), val, 0), idx),
-  if (Subtarget.is64Bit() || Val.getValueType() != MVT::i64) {
+  SDValue ValInVec;
+
+  if (IsSimpleInsert) {
     if (isNullConstant(Idx))
       return DAG.getNode(RISCVISD::VMV_S_XF_VL, DL, ContainerVT, Vec, Val, VL);
-    SDValue Slidedown =
-        DAG.getNode(RISCVISD::VSLIDEDOWN_VL, DL, ContainerVT,
-                    DAG.getUNDEF(ContainerVT), Vec, Idx, Mask, VL);
-    SDValue InsertElt0 =
-        DAG.getNode(RISCVISD::VMV_S_XF_VL, DL, ContainerVT, Slidedown, Val, VL);
-    return DAG.getNode(RISCVISD::VSLIDEUP_VL, DL, ContainerVT, Vec, InsertElt0,
-                       Idx, Mask, VL);
-  }
-
-  // Custom-legalize INSERT_VECTOR_ELT where XLEN<SEW, as the SEW element type
-  // is illegal (currently only vXi64 RV32).
-  // Since there is no easy way of getting a single element into a vector when
-  // XLEN<SEW, we lower the operation to the following sequence:
-  //   splat      vVal, rVal
-  //   vid.v      vVid
-  //   vmseq.vx   mMask, vVid, rIdx
-  //   vmerge.vvm vDest, vSrc, vVal, mMask
-  // This essentially merges the original vector with the inserted element by
-  // using a mask whose only set bit is that corresponding to the insert
-  // index.
-  SDValue SplattedVal = DAG.getSplatVector(ContainerVT, DL, Val);
-  SDValue SplattedIdx =
-      DAG.getNode(RISCVISD::VMV_V_X_VL, DL, ContainerVT, Idx, VL);
-
-  SDValue VID = DAG.getNode(RISCVISD::VID_VL, DL, ContainerVT, Mask, VL);
-  auto SetCCVT =
-      getSetCCResultType(DAG.getDataLayout(), *DAG.getContext(), ContainerVT);
-  SDValue SelectCond =
-      DAG.getNode(RISCVISD::SETCC_VL, DL, SetCCVT, VID, SplattedIdx,
-                  DAG.getCondCode(ISD::SETEQ), Mask, VL);
-  SDValue Select = DAG.getNode(RISCVISD::VSELECT_VL, DL, ContainerVT,
-                               SelectCond, SplattedVal, Vec, VL);
+    ValInVec = DAG.getNode(RISCVISD::VMV_S_XF_VL, DL, ContainerVT,
+                           DAG.getUNDEF(ContainerVT), Val, VL);
+  } else {
+    // On RV32, i64-element vectors must be specially handled to place the
+    // value at element 0, by using two vslide1up instructions in sequence on
+    // the i32 split lo/hi value. Use an equivalently-sized i32 vector for
+    // this.
+    MVT I32ContainerVT =
+        MVT::getVectorVT(MVT::i32, ContainerVT.getVectorElementCount() * 2);
+    SDValue I32Mask =
+        getDefaultScalableVLOps(I32ContainerVT, DL, DAG, Subtarget).first;
+    // Limit the active VL to two.
+    SDValue InsertI64VL = DAG.getConstant(2, DL, XLenVT);
+    // Note: We can't pass a UNDEF to the first VSLIDE1UP_VL since an untied
+    // undef doesn't obey the earlyclobber constraint. Just splat a zero value.
+    ValInVec = DAG.getNode(RISCVISD::VMV_V_X_VL, DL, I32ContainerVT, Zero,
+                           InsertI64VL);
+    // First slide in the hi value, then the lo in underneath it.
+    ValInVec = DAG.getNode(RISCVISD::VSLIDE1UP_VL, DL, I32ContainerVT, ValInVec,
+                           ValHi, I32Mask, InsertI64VL);
+    ValInVec = DAG.getNode(RISCVISD::VSLIDE1UP_VL, DL, I32ContainerVT, ValInVec,
+                           ValLo, I32Mask, InsertI64VL);
+    // Bitcast back to the right container type.
+    ValInVec = DAG.getBitcast(ContainerVT, ValInVec);
+  }
+
+  // Now that the value is in a vector, slide it into position.
+  SDValue InsertVL =
+      DAG.getNode(ISD::ADD, DL, XLenVT, Idx, DAG.getConstant(1, DL, XLenVT));
+  SDValue Slideup = DAG.getNode(RISCVISD::VSLIDEUP_VL, DL, ContainerVT, Vec,
+                                ValInVec, Idx, Mask, InsertVL);
   if (!VecVT.isFixedLengthVector())
-    return Select;
-  return convertFromScalableVector(VecVT, Select, DAG, Subtarget);
+    return Slideup;
+  return convertFromScalableVector(VecVT, Slideup, DAG, Subtarget);
 }
 
 // Custom-lower EXTRACT_VECTOR_ELT operations to slide the vector down, then
 // extract the first element: (extractelt (slidedown vec, idx), 0). For integer
 // types this is done using VMV_X_S to allow us to glean information about the
 // sign bits of the result.
 SDValue RISCVTargetLowering::lowerEXTRACT_VECTOR_ELT(SDValue Op,
                                                      SelectionDAG &DAG) const {
#define NODE_NAME_CASE(NODE)                                                   \
   NODE_NAME_CASE(VMV_X_S)
   NODE_NAME_CASE(VMV_S_XF_VL)
   NODE_NAME_CASE(SPLAT_VECTOR_I64)
   NODE_NAME_CASE(READ_VLENB)
   NODE_NAME_CASE(TRUNCATE_VECTOR_VL)
   NODE_NAME_CASE(VLEFF)
   NODE_NAME_CASE(VLEFF_MASK)
   NODE_NAME_CASE(VSLIDEUP_VL)
+  NODE_NAME_CASE(VSLIDE1UP_VL)
   NODE_NAME_CASE(VSLIDEDOWN_VL)
   NODE_NAME_CASE(VID_VL)
   NODE_NAME_CASE(VFNCVT_ROD_VL)
   NODE_NAME_CASE(VECREDUCE_ADD_VL)
   NODE_NAME_CASE(VECREDUCE_UMAX_VL)
   NODE_NAME_CASE(VECREDUCE_SMAX_VL)
   NODE_NAME_CASE(VECREDUCE_UMIN_VL)
   NODE_NAME_CASE(VECREDUCE_SMIN_VL)

llvm/lib/Target/RISCV/RISCVInstrInfoVVLPatterns.td

 def riscv_vid_vl : SDNode<"RISCVISD::VID_VL", SDTypeProfile<1, 2,
                           [SDTCisVec<0>, SDTCVecEltisVT<1, i1>,
                            SDTCisSameNumEltsAs<0, 1>, SDTCisVT<2, XLenVT>]>, []>;
 
 def SDTRVVSlide : SDTypeProfile<1, 5, [
   SDTCisVec<0>, SDTCisSameAs<1, 0>, SDTCisSameAs<2, 0>, SDTCisVT<3, XLenVT>,
   SDTCVecEltisVT<4, i1>, SDTCisSameNumEltsAs<0, 4>, SDTCisVT<5, XLenVT>
 ]>;
+def SDTRVVSlide1 : SDTypeProfile<1, 4, [
+  SDTCisVec<0>, SDTCisSameAs<1, 0>, SDTCisVT<2, XLenVT>,
+  SDTCVecEltisVT<3, i1>, SDTCisSameNumEltsAs<0, 3>, SDTCisVT<4, XLenVT>
+]>;
 
 def riscv_slideup_vl   : SDNode<"RISCVISD::VSLIDEUP_VL", SDTRVVSlide, []>;
+def riscv_slide1up_vl  : SDNode<"RISCVISD::VSLIDE1UP_VL", SDTRVVSlide1, []>;
 def riscv_slidedown_vl : SDNode<"RISCVISD::VSLIDEDOWN_VL", SDTRVVSlide, []>;
 
 let Predicates = [HasStdExtV] in {
 
 foreach vti = AllIntegerVectors in
   def : Pat<(vti.Vector (riscv_vid_vl (vti.Mask true_mask),
                                       (XLenVT (VLOp GPR:$vl)))),
             (!cast<Instruction>("PseudoVID_V_"#vti.LMul.MX) GPR:$vl, vti.SEW)>;
foreach vti = !listconcat(AllIntegerVectors, AllFloatVectors) in {
   def : Pat<(vti.Vector (riscv_slideup_vl (vti.Vector vti.RegClass:$rs3),
                                           (vti.Vector vti.RegClass:$rs1),
                                           GPR:$rs2, (vti.Mask true_mask),
                                           (XLenVT (VLOp GPR:$vl)))),
             (!cast<Instruction>("PseudoVSLIDEUP_VX_"#vti.LMul.MX)
                 vti.RegClass:$rs3, vti.RegClass:$rs1, GPR:$rs2,
                 GPR:$vl, vti.SEW)>;
 
+  def : Pat<(vti.Vector (riscv_slide1up_vl (vti.Vector vti.RegClass:$rs1),
+                                           GPR:$rs2, (vti.Mask true_mask),
+                                           (XLenVT (VLOp GPR:$vl)))),
+            (!cast<Instruction>("PseudoVSLIDE1UP_VX_"#vti.LMul.MX)
+                vti.RegClass:$rs1, GPR:$rs2, GPR:$vl, vti.SEW)>;
+
   def : Pat<(vti.Vector (riscv_slidedown_vl (vti.Vector vti.RegClass:$rs3),
                                             (vti.Vector vti.RegClass:$rs1),
                                             uimm5:$rs2, (vti.Mask true_mask),
                                             (XLenVT (VLOp GPR:$vl)))),
             (!cast<Instruction>("PseudoVSLIDEDOWN_VI_"#vti.LMul.MX)
                 vti.RegClass:$rs3, vti.RegClass:$rs1, uimm5:$rs2,
                 GPR:$vl, vti.SEW)>;
 

llvm/test/CodeGen/RISCV/rvv/fixed-vectors-insert.ll

 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
 ; RUN: llc -mtriple=riscv32 -target-abi=ilp32d -mattr=+experimental-v,+experimental-zfh,+f,+d -verify-machineinstrs -riscv-v-vector-bits-min=128 -verify-machineinstrs < %s | FileCheck %s --check-prefixes=RV32
 ; RUN: llc -mtriple=riscv64 -target-abi=lp64d -mattr=+experimental-v,+experimental-zfh,+f,+d -verify-machineinstrs -riscv-v-vector-bits-min=128 -verify-machineinstrs < %s | FileCheck %s --check-prefixes=RV64
 
 ; FIXME: This codegen needs to be improved. These tests previously asserted
 ; type legalizing the i64 type on RV32.
 
 define void @insertelt_v4i64(<4 x i64>* %x, i64 %y) {
 ; RV32-LABEL: insertelt_v4i64:
 ; RV32:       # %bb.0:
 ; RV32-NEXT:    vsetivli a3, 4, e64,m2,ta,mu
 ; RV32-NEXT:    vle64.v v26, (a0)
-; RV32-NEXT:    vsetvli a3, zero, e64,m2,ta,mu
-; RV32-NEXT:    vmv.v.x v28, a2
-; RV32-NEXT:    addi a2, zero, 32
-; RV32-NEXT:    vsll.vx v28, v28, a2
-; RV32-NEXT:    vmv.v.x v30, a1
-; RV32-NEXT:    vsll.vx v30, v30, a2
-; RV32-NEXT:    vsrl.vx v30, v30, a2
-; RV32-NEXT:    vor.vv v28, v30, v28
+; RV32-NEXT:    vsetivli a3, 2, e32,m2,ta,mu
+; RV32-NEXT:    vmv.v.i v28, 0
+; RV32-NEXT:    vslide1up.vx v30, v28, a2
+; RV32-NEXT:    vslide1up.vx v28, v30, a1
+; RV32-NEXT:    vsetivli a1, 4, e64,m2,tu,mu
+; RV32-NEXT:    vslideup.vi v26, v28, 3
 ; RV32-NEXT:    vsetivli a1, 4, e64,m2,ta,mu
-; RV32-NEXT:    vid.v v30
-; RV32-NEXT:    vmseq.vi v0, v30, 3
-; RV32-NEXT:    vmerge.vvm v26, v26, v28, v0
 ; RV32-NEXT:    vse64.v v26, (a0)
 ; RV32-NEXT:    ret
 ;
 ; RV64-LABEL: insertelt_v4i64:
 ; RV64:       # %bb.0:
 ; RV64-NEXT:    vsetivli a2, 4, e64,m2,ta,mu
 ; RV64-NEXT:    vle64.v v26, (a0)
-; RV64-NEXT:    vslidedown.vi v28, v26, 3
 ; RV64-NEXT:    vmv.s.x v28, a1
 ; RV64-NEXT:    vsetivli a1, 4, e64,m2,tu,mu
 ; RV64-NEXT:    vslideup.vi v26, v28, 3
 ; RV64-NEXT:    vsetivli a1, 4, e64,m2,ta,mu
 ; RV64-NEXT:    vse64.v v26, (a0)
 ; RV64-NEXT:    ret
   %a = load <4 x i64>, <4 x i64>* %x
   %b = insertelement <4 x i64> %a, i64 %y, i32 3

 ; RV32-NEXT:    vsetivli a3, 8, e32,m2,ta,mu
 ; RV32-NEXT:    vmv.v.i v28, 0
 ; RV32-NEXT:    vsetivli a3, 2, e64,m2,tu,mu
 ; RV32-NEXT:    vslideup.vi v28, v26, 0
 ; RV32-NEXT:    vsetivli a3, 4, e32,m1,ta,mu
 ; RV32-NEXT:    vle32.v v26, (sp)
 ; RV32-NEXT:    vsetivli a3, 4, e64,m2,tu,mu
 ; RV32-NEXT:    vslideup.vi v28, v26, 2
-; RV32-NEXT:    vsetvli a3, zero, e64,m2,ta,mu
-; RV32-NEXT:    vmv.v.x v26, a2
-; RV32-NEXT:    addi a3, zero, 32
-; RV32-NEXT:    vsll.vx v26, v26, a3
-; RV32-NEXT:    vmv.v.x v30, a1
-; RV32-NEXT:    vsll.vx v30, v30, a3
-; RV32-NEXT:    vsrl.vx v30, v30, a3
-; RV32-NEXT:    vor.vv v26, v30, v26
-; RV32-NEXT:    vsetivli a3, 4, e64,m2,ta,mu
-; RV32-NEXT:    vid.v v30
-; RV32-NEXT:    vmseq.vi v0, v30, 2
-; RV32-NEXT:    vmerge.vvm v26, v28, v26, v0
+; RV32-NEXT:    vsetivli a3, 2, e32,m2,ta,mu
+; RV32-NEXT:    vmv.v.i v26, 0
+; RV32-NEXT:    vslide1up.vx v30, v26, a2
+; RV32-NEXT:    vslide1up.vx v26, v30, a1
+; RV32-NEXT:    vsetivli a3, 3, e64,m2,tu,mu
+; RV32-NEXT:    vslideup.vi v28, v26, 2
 ; RV32-NEXT:    vsetivli a3, 2, e64,m1,ta,mu
-; RV32-NEXT:    vse64.v v26, (a0)
+; RV32-NEXT:    vse64.v v28, (a0)
 ; RV32-NEXT:    sw a1, 16(a0)
 ; RV32-NEXT:    sw a2, 20(a0)
 ; RV32-NEXT:    addi sp, sp, 16
 ; RV32-NEXT:    ret
 ;
 ; RV64-LABEL: insertelt_v3i64:
 ; RV64:       # %bb.0:
 ; RV64-NEXT:    sd a1, 16(a0)
 ; RV64-NEXT:    ret
   %a = load <3 x i64>, <3 x i64>* %x, align 8
   %b = insertelement <3 x i64> %a, i64 %y, i32 2
   store <3 x i64> %b, <3 x i64>* %x
   ret void
 }
 
 define void @insertelt_v16i8(<16 x i8>* %x, i8 %y) {
 ; RV32-LABEL: insertelt_v16i8:
 ; RV32:       # %bb.0:
 ; RV32-NEXT:    vsetivli a2, 16, e8,m1,ta,mu
 ; RV32-NEXT:    vle8.v v25, (a0)
-; RV32-NEXT:    vslidedown.vi v26, v25, 14
 ; RV32-NEXT:    vmv.s.x v26, a1
-; RV32-NEXT:    vsetivli a1, 16, e8,m1,tu,mu
+; RV32-NEXT:    vsetivli a1, 15, e8,m1,tu,mu
 ; RV32-NEXT:    vslideup.vi v25, v26, 14
 ; RV32-NEXT:    vsetivli a1, 16, e8,m1,ta,mu
 ; RV32-NEXT:    vse8.v v25, (a0)
 ; RV32-NEXT:    ret
 ;
 ; RV64-LABEL: insertelt_v16i8:
 ; RV64:       # %bb.0:
 ; RV64-NEXT:    vsetivli a2, 16, e8,m1,ta,mu
 ; RV64-NEXT:    vle8.v v25, (a0)
-; RV64-NEXT:    vslidedown.vi v26, v25, 14
 ; RV64-NEXT:    vmv.s.x v26, a1
-; RV64-NEXT:    vsetivli a1, 16, e8,m1,tu,mu
+; RV64-NEXT:    vsetivli a1, 15, e8,m1,tu,mu
 ; RV64-NEXT:    vslideup.vi v25, v26, 14
 ; RV64-NEXT:    vsetivli a1, 16, e8,m1,ta,mu
 ; RV64-NEXT:    vse8.v v25, (a0)
 ; RV64-NEXT:    ret
   %a = load <16 x i8>, <16 x i8>* %x
   %b = insertelement <16 x i8> %a, i8 %y, i32 14
   store <16 x i8> %b, <16 x i8>* %x
   ret void
 }
 
 define void @insertelt_v32i16(<32 x i16>* %x, i16 %y, i32 %idx) {
 ; RV32-LABEL: insertelt_v32i16:
 ; RV32:       # %bb.0:
 ; RV32-NEXT:    addi a3, zero, 32
 ; RV32-NEXT:    vsetvli a4, a3, e16,m4,ta,mu
 ; RV32-NEXT:    vle16.v v28, (a0)
-; RV32-NEXT:    vslidedown.vx v8, v28, a2
 ; RV32-NEXT:    vmv.s.x v8, a1
-; RV32-NEXT:    vsetvli a1, a3, e16,m4,tu,mu
+; RV32-NEXT:    addi a1, a2, 1
+; RV32-NEXT:    vsetvli a1, a1, e16,m4,tu,mu
 ; RV32-NEXT:    vslideup.vx v28, v8, a2
 ; RV32-NEXT:    vsetvli a1, a3, e16,m4,ta,mu
 ; RV32-NEXT:    vse16.v v28, (a0)
 ; RV32-NEXT:    ret
 ;
 ; RV64-LABEL: insertelt_v32i16:
 ; RV64:       # %bb.0:
 ; RV64-NEXT:    addi a3, zero, 32
 ; RV64-NEXT:    vsetvli a4, a3, e16,m4,ta,mu
 ; RV64-NEXT:    vle16.v v28, (a0)
-; RV64-NEXT:    sext.w a2, a2
-; RV64-NEXT:    vslidedown.vx v8, v28, a2
 ; RV64-NEXT:    vmv.s.x v8, a1
-; RV64-NEXT:    vsetvli a1, a3, e16,m4,tu,mu
-; RV64-NEXT:    vslideup.vx v28, v8, a2
+; RV64-NEXT:    sext.w a1, a2
+; RV64-NEXT:    addi a2, a1, 1
+; RV64-NEXT:    vsetvli a2, a2, e16,m4,tu,mu
+; RV64-NEXT:    vslideup.vx v28, v8, a1
 ; RV64-NEXT:    vsetvli a1, a3, e16,m4,ta,mu
 ; RV64-NEXT:    vse16.v v28, (a0)
 ; RV64-NEXT:    ret
   %a = load <32 x i16>, <32 x i16>* %x
   %b = insertelement <32 x i16> %a, i16 %y, i32 %idx
   store <32 x i16> %b, <32 x i16>* %x
   ret void
 }
 
 define void @insertelt_v8f32(<8 x float>* %x, float %y, i32 %idx) {
 ; RV32-LABEL: insertelt_v8f32:
 ; RV32:       # %bb.0:
 ; RV32-NEXT:    vsetivli a2, 8, e32,m2,ta,mu
 ; RV32-NEXT:    vle32.v v26, (a0)
-; RV32-NEXT:    vslidedown.vx v28, v26, a1
 ; RV32-NEXT:    vfmv.s.f v28, fa0
-; RV32-NEXT:    vsetivli a2, 8, e32,m2,tu,mu
+; RV32-NEXT:    addi a2, a1, 1
+; RV32-NEXT:    vsetvli a2, a2, e32,m2,tu,mu
 ; RV32-NEXT:    vslideup.vx v26, v28, a1
 ; RV32-NEXT:    vsetivli a1, 8, e32,m2,ta,mu
 ; RV32-NEXT:    vse32.v v26, (a0)
 ; RV32-NEXT:    ret
 ;
 ; RV64-LABEL: insertelt_v8f32:
 ; RV64:       # %bb.0:
 ; RV64-NEXT:    vsetivli a2, 8, e32,m2,ta,mu
 ; RV64-NEXT:    vle32.v v26, (a0)
-; RV64-NEXT:    sext.w a1, a1
-; RV64-NEXT:    vslidedown.vx v28, v26, a1
 ; RV64-NEXT:    vfmv.s.f v28, fa0
-; RV64-NEXT:    vsetivli a2, 8, e32,m2,tu,mu
+; RV64-NEXT:    sext.w a1, a1
+; RV64-NEXT:    addi a2, a1, 1
+; RV64-NEXT:    vsetvli a2, a2, e32,m2,tu,mu
 ; RV64-NEXT:    vslideup.vx v26, v28, a1
 ; RV64-NEXT:    vsetivli a1, 8, e32,m2,ta,mu
 ; RV64-NEXT:    vse32.v v26, (a0)
 ; RV64-NEXT:    ret
   %a = load <8 x float>, <8 x float>* %x
   %b = insertelement <8 x float> %a, float %y, i32 %idx
   store <8 x float> %b, <8 x float>* %x
   ret void
 }
+
+define void @insertelt_v8i64_0(<8 x i64>* %x) {
+; RV32-LABEL: insertelt_v8i64_0:
+; RV32:       # %bb.0:
+; RV32-NEXT:    vsetivli a1, 8, e64,m4,ta,mu
+; RV32-NEXT:    vle64.v v28, (a0)
+; RV32-NEXT:    addi a1, zero, -1
+; RV32-NEXT:    vmv.s.x v28, a1
+; RV32-NEXT:    vs4r.v v28, (a0)
+; RV32-NEXT:    ret
+;
+; RV64-LABEL: insertelt_v8i64_0:
+; RV64:       # %bb.0:
+; RV64-NEXT:    vsetivli a1, 8, e64,m4,ta,mu
+; RV64-NEXT:    vle64.v v28, (a0)
+; RV64-NEXT:    addi a1, zero, -1
+; RV64-NEXT:    vmv.s.x v28, a1
+; RV64-NEXT:    vse64.v v28, (a0)
+; RV64-NEXT:    ret
+  %a = load <8 x i64>, <8 x i64>* %x
+  %b = insertelement <8 x i64> %a, i64 -1, i32 0
+  store <8 x i64> %b, <8 x i64>* %x
+  ret void
+}
+
+define void @insertelt_v8i64(<8 x i64>* %x, i32 %idx) {
+; RV32-LABEL: insertelt_v8i64:
+; RV32:       # %bb.0:
+; RV32-NEXT:    vsetivli a2, 8, e64,m4,ta,mu
+; RV32-NEXT:    vle64.v v28, (a0)
+; RV32-NEXT:    addi a2, zero, -1
+; RV32-NEXT:    vmv.s.x v8, a2
+; RV32-NEXT:    addi a2, a1, 1
+; RV32-NEXT:    vsetvli a2, a2, e64,m4,tu,mu
+; RV32-NEXT:    vslideup.vx v28, v8, a1
+; RV32-NEXT:    vsetivli a1, 8, e64,m4,ta,mu
+; RV32-NEXT:    vse64.v v28, (a0)
+; RV32-NEXT:    ret
+;
+; RV64-LABEL: insertelt_v8i64:
+; RV64:       # %bb.0:
+; RV64-NEXT:    vsetivli a2, 8, e64,m4,ta,mu
+; RV64-NEXT:    vle64.v v28, (a0)
+; RV64-NEXT:    addi a2, zero, -1
+; RV64-NEXT:    vmv.s.x v8, a2
+; RV64-NEXT:    sext.w a1, a1
+; RV64-NEXT:    addi a2, a1, 1
+; RV64-NEXT:    vsetvli a2, a2, e64,m4,tu,mu
+; RV64-NEXT:    vslideup.vx v28, v8, a1
+; RV64-NEXT:    vsetivli a1, 8, e64,m4,ta,mu
+; RV64-NEXT:    vse64.v v28, (a0)
+; RV64-NEXT:    ret
+  %a = load <8 x i64>, <8 x i64>* %x
+  %b = insertelement <8 x i64> %a, i64 -1, i32 %idx
+  store <8 x i64> %b, <8 x i64>* %x
+  ret void
+}
+
+define void @insertelt_c6_v8i64_0(<8 x i64>* %x) {
+; RV32-LABEL: insertelt_c6_v8i64_0:
+; RV32:       # %bb.0:
+; RV32-NEXT:    vsetivli a1, 8, e64,m4,ta,mu
+; RV32-NEXT:    vle64.v v28, (a0)
+; RV32-NEXT:    addi a1, zero, 6
+; RV32-NEXT:    vmv.s.x v28, a1
+; RV32-NEXT:    vs4r.v v28, (a0)
+; RV32-NEXT:    ret
+;
+; RV64-LABEL: insertelt_c6_v8i64_0:
+; RV64:       # %bb.0:
+; RV64-NEXT:    vsetivli a1, 8, e64,m4,ta,mu
+; RV64-NEXT:    vle64.v v28, (a0)
+; RV64-NEXT:    addi a1, zero, 6
+; RV64-NEXT:    vmv.s.x v28, a1
+; RV64-NEXT:    vse64.v v28, (a0)
+; RV64-NEXT:    ret
+  %a = load <8 x i64>, <8 x i64>* %x
+  %b = insertelement <8 x i64> %a, i64 6, i32 0
+  store <8 x i64> %b, <8 x i64>* %x
+  ret void
+}
+
+define void @insertelt_c6_v8i64(<8 x i64>* %x, i32 %idx) {
+; RV32-LABEL: insertelt_c6_v8i64:
+; RV32:       # %bb.0:
+; RV32-NEXT:    vsetivli a2, 8, e64,m4,ta,mu
+; RV32-NEXT:    vle64.v v28, (a0)
+; RV32-NEXT:    addi a2, zero, 6
+; RV32-NEXT:    vmv.s.x v8, a2
+; RV32-NEXT:    addi a2, a1, 1
+; RV32-NEXT:    vsetvli a2, a2, e64,m4,tu,mu
+; RV32-NEXT:    vslideup.vx v28, v8, a1
+; RV32-NEXT:    vsetivli a1, 8, e64,m4,ta,mu
+; RV32-NEXT:    vse64.v v28, (a0)
+; RV32-NEXT:    ret
+;
+; RV64-LABEL: insertelt_c6_v8i64:
+; RV64:       # %bb.0:
+; RV64-NEXT:    vsetivli a2, 8, e64,m4,ta,mu
+; RV64-NEXT:    vle64.v v28, (a0)
+; RV64-NEXT:    addi a2, zero, 6
+; RV64-NEXT:    vmv.s.x v8, a2
+; RV64-NEXT:    sext.w a1, a1
+; RV64-NEXT:    addi a2, a1, 1
+; RV64-NEXT:    vsetvli a2, a2, e64,m4,tu,mu
+; RV64-NEXT:    vslideup.vx v28, v8, a1
+; RV64-NEXT:    vsetivli a1, 8, e64,m4,ta,mu
+; RV64-NEXT:    vse64.v v28, (a0)
+; RV64-NEXT:    ret
+  %a = load <8 x i64>, <8 x i64>* %x
+  %b = insertelement <8 x i64> %a, i64 6, i32 %idx
+  store <8 x i64> %b, <8 x i64>* %x
+  ret void
+}

llvm/test/CodeGen/RISCV/rvv/insertelt-fp-rv32.ll

; CHECK-NEXT:    ret
   %r = insertelement <vscale x 1 x half> %v, half %elt, i32 0
   ret <vscale x 1 x half> %r
 }
 
 define <vscale x 1 x half> @insertelt_nxv1f16_imm(<vscale x 1 x half> %v, half %elt) {
 ; CHECK-LABEL: insertelt_nxv1f16_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e16,mf4,ta,mu
-; CHECK-NEXT:    vslidedown.vi v25, v8, 3
 ; CHECK-NEXT:    vfmv.s.f v25, fa0
-; CHECK-NEXT:    vsetvli a0, zero, e16,mf4,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e16,mf4,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v25, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 1 x half> %v, half %elt, i32 3
   ret <vscale x 1 x half> %r
 }
 
 define <vscale x 1 x half> @insertelt_nxv1f16_idx(<vscale x 1 x half> %v, half %elt, i32 %idx) {
 ; CHECK-LABEL: insertelt_nxv1f16_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e16,mf4,ta,mu
-; CHECK-NEXT:    vslidedown.vx v25, v8, a0
 ; CHECK-NEXT:    vfmv.s.f v25, fa0
-; CHECK-NEXT:    vsetvli a1, zero, e16,mf4,tu,mu
+; CHECK-NEXT:    addi a1, a0, 1
+; CHECK-NEXT:    vsetvli a1, a1, e16,mf4,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v25, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 1 x half> %v, half %elt, i32 %idx
   ret <vscale x 1 x half> %r
 }
 
 define <vscale x 2 x half> @insertelt_nxv2f16_0(<vscale x 2 x half> %v, half %elt) {
 ; CHECK-LABEL: insertelt_nxv2f16_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e16,mf2,ta,mu
 ; CHECK-NEXT:    vfmv.s.f v8, fa0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 2 x half> %v, half %elt, i32 0
   ret <vscale x 2 x half> %r
 }
 
 define <vscale x 2 x half> @insertelt_nxv2f16_imm(<vscale x 2 x half> %v, half %elt) {
 ; CHECK-LABEL: insertelt_nxv2f16_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e16,mf2,ta,mu
-; CHECK-NEXT:    vslidedown.vi v25, v8, 3
 ; CHECK-NEXT:    vfmv.s.f v25, fa0
-; CHECK-NEXT:    vsetvli a0, zero, e16,mf2,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e16,mf2,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v25, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 2 x half> %v, half %elt, i32 3
   ret <vscale x 2 x half> %r
 }
 
 define <vscale x 2 x half> @insertelt_nxv2f16_idx(<vscale x 2 x half> %v, half %elt, i32 %idx) {
 ; CHECK-LABEL: insertelt_nxv2f16_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e16,mf2,ta,mu
-; CHECK-NEXT:    vslidedown.vx v25, v8, a0
 ; CHECK-NEXT:    vfmv.s.f v25, fa0
-; CHECK-NEXT:    vsetvli a1, zero, e16,mf2,tu,mu
+; CHECK-NEXT:    addi a1, a0, 1
+; CHECK-NEXT:    vsetvli a1, a1, e16,mf2,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v25, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 2 x half> %v, half %elt, i32 %idx
   ret <vscale x 2 x half> %r
 }
 
 define <vscale x 4 x half> @insertelt_nxv4f16_0(<vscale x 4 x half> %v, half %elt) {
 ; CHECK-LABEL: insertelt_nxv4f16_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e16,m1,ta,mu
 ; CHECK-NEXT:    vfmv.s.f v8, fa0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 4 x half> %v, half %elt, i32 0
   ret <vscale x 4 x half> %r
 }
 
 define <vscale x 4 x half> @insertelt_nxv4f16_imm(<vscale x 4 x half> %v, half %elt) {
 ; CHECK-LABEL: insertelt_nxv4f16_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e16,m1,ta,mu
-; CHECK-NEXT:    vslidedown.vi v25, v8, 3
 ; CHECK-NEXT:    vfmv.s.f v25, fa0
-; CHECK-NEXT:    vsetvli a0, zero, e16,m1,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e16,m1,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v25, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 4 x half> %v, half %elt, i32 3
   ret <vscale x 4 x half> %r
 }
 
 define <vscale x 4 x half> @insertelt_nxv4f16_idx(<vscale x 4 x half> %v, half %elt, i32 %idx) {
 ; CHECK-LABEL: insertelt_nxv4f16_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e16,m1,ta,mu
-; CHECK-NEXT:    vslidedown.vx v25, v8, a0
 ; CHECK-NEXT:    vfmv.s.f v25, fa0
-; CHECK-NEXT:    vsetvli a1, zero, e16,m1,tu,mu
+; CHECK-NEXT:    addi a1, a0, 1
+; CHECK-NEXT:    vsetvli a1, a1, e16,m1,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v25, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 4 x half> %v, half %elt, i32 %idx
   ret <vscale x 4 x half> %r
 }
 
 define <vscale x 8 x half> @insertelt_nxv8f16_0(<vscale x 8 x half> %v, half %elt) {
 ; CHECK-LABEL: insertelt_nxv8f16_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e16,m2,ta,mu
 ; CHECK-NEXT:    vfmv.s.f v8, fa0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 8 x half> %v, half %elt, i32 0
   ret <vscale x 8 x half> %r
 }
 
 define <vscale x 8 x half> @insertelt_nxv8f16_imm(<vscale x 8 x half> %v, half %elt) {
 ; CHECK-LABEL: insertelt_nxv8f16_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e16,m2,ta,mu
-; CHECK-NEXT:    vslidedown.vi v26, v8, 3
 ; CHECK-NEXT:    vfmv.s.f v26, fa0
-; CHECK-NEXT:    vsetvli a0, zero, e16,m2,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e16,m2,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v26, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 8 x half> %v, half %elt, i32 3
   ret <vscale x 8 x half> %r
 }
 
 define <vscale x 8 x half> @insertelt_nxv8f16_idx(<vscale x 8 x half> %v, half %elt, i32 %idx) {
 ; CHECK-LABEL: insertelt_nxv8f16_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e16,m2,ta,mu
-; CHECK-NEXT:    vslidedown.vx v26, v8, a0
 ; CHECK-NEXT:    vfmv.s.f v26, fa0
-; CHECK-NEXT:    vsetvli a1, zero, e16,m2,tu,mu
+; CHECK-NEXT:    addi a1, a0, 1
+; CHECK-NEXT:    vsetvli a1, a1, e16,m2,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v26, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 8 x half> %v, half %elt, i32 %idx
   ret <vscale x 8 x half> %r
 }
 
 define <vscale x 16 x half> @insertelt_nxv16f16_0(<vscale x 16 x half> %v, half %elt) {
 ; CHECK-LABEL: insertelt_nxv16f16_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e16,m4,ta,mu
 ; CHECK-NEXT:    vfmv.s.f v8, fa0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 16 x half> %v, half %elt, i32 0
   ret <vscale x 16 x half> %r
 }
 
 define <vscale x 16 x half> @insertelt_nxv16f16_imm(<vscale x 16 x half> %v, half %elt) {
 ; CHECK-LABEL: insertelt_nxv16f16_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e16,m4,ta,mu
-; CHECK-NEXT:    vslidedown.vi v28, v8, 3
 ; CHECK-NEXT:    vfmv.s.f v28, fa0
-; CHECK-NEXT:    vsetvli a0, zero, e16,m4,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e16,m4,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v28, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 16 x half> %v, half %elt, i32 3
   ret <vscale x 16 x half> %r
 }
 
 define <vscale x 16 x half> @insertelt_nxv16f16_idx(<vscale x 16 x half> %v, half %elt, i32 %idx) {
 ; CHECK-LABEL: insertelt_nxv16f16_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e16,m4,ta,mu
-; CHECK-NEXT:    vslidedown.vx v28, v8, a0
 ; CHECK-NEXT:    vfmv.s.f v28, fa0
-; CHECK-NEXT:    vsetvli a1, zero, e16,m4,tu,mu
+; CHECK-NEXT:    addi a1, a0, 1
+; CHECK-NEXT:    vsetvli a1, a1, e16,m4,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v28, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 16 x half> %v, half %elt, i32 %idx
   ret <vscale x 16 x half> %r
 }
 
 define <vscale x 32 x half> @insertelt_nxv32f16_0(<vscale x 32 x half> %v, half %elt) {
 ; CHECK-LABEL: insertelt_nxv32f16_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e16,m8,ta,mu
 ; CHECK-NEXT:    vfmv.s.f v8, fa0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 32 x half> %v, half %elt, i32 0
   ret <vscale x 32 x half> %r
 }
 
 define <vscale x 32 x half> @insertelt_nxv32f16_imm(<vscale x 32 x half> %v, half %elt) {
 ; CHECK-LABEL: insertelt_nxv32f16_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e16,m8,ta,mu
-; CHECK-NEXT:    vslidedown.vi v16, v8, 3
 ; CHECK-NEXT:    vfmv.s.f v16, fa0
-; CHECK-NEXT:    vsetvli a0, zero, e16,m8,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e16,m8,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v16, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 32 x half> %v, half %elt, i32 3
   ret <vscale x 32 x half> %r
 }
 
 define <vscale x 32 x half> @insertelt_nxv32f16_idx(<vscale x 32 x half> %v, half %elt, i32 %idx) {
 ; CHECK-LABEL: insertelt_nxv32f16_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e16,m8,ta,mu
-; CHECK-NEXT:    vslidedown.vx v16, v8, a0
 ; CHECK-NEXT:    vfmv.s.f v16, fa0
-; CHECK-NEXT:    vsetvli a1, zero, e16,m8,tu,mu
+; CHECK-NEXT:    addi a1, a0, 1
+; CHECK-NEXT:    vsetvli a1, a1, e16,m8,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v16, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 32 x half> %v, half %elt, i32 %idx
   ret <vscale x 32 x half> %r
 }
 
 define <vscale x 1 x float> @insertelt_nxv1f32_0(<vscale x 1 x float> %v, float %elt) {
 ; CHECK-LABEL: insertelt_nxv1f32_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e32,mf2,ta,mu
 ; CHECK-NEXT:    vfmv.s.f v8, fa0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 1 x float> %v, float %elt, i32 0
   ret <vscale x 1 x float> %r
 }
 
 define <vscale x 1 x float> @insertelt_nxv1f32_imm(<vscale x 1 x float> %v, float %elt) {
 ; CHECK-LABEL: insertelt_nxv1f32_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e32,mf2,ta,mu
-; CHECK-NEXT:    vslidedown.vi v25, v8, 3
 ; CHECK-NEXT:    vfmv.s.f v25, fa0
-; CHECK-NEXT:    vsetvli a0, zero, e32,mf2,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e32,mf2,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v25, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 1 x float> %v, float %elt, i32 3
   ret <vscale x 1 x float> %r
 }
 
 define <vscale x 1 x float> @insertelt_nxv1f32_idx(<vscale x 1 x float> %v, float %elt, i32 %idx) {
 ; CHECK-LABEL: insertelt_nxv1f32_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e32,mf2,ta,mu
-; CHECK-NEXT:    vslidedown.vx v25, v8, a0
 ; CHECK-NEXT:    vfmv.s.f v25, fa0
-; CHECK-NEXT:    vsetvli a1, zero, e32,mf2,tu,mu
+; CHECK-NEXT:    addi a1, a0, 1
+; CHECK-NEXT:    vsetvli a1, a1, e32,mf2,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v25, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 1 x float> %v, float %elt, i32 %idx
   ret <vscale x 1 x float> %r
 }
 
 define <vscale x 2 x float> @insertelt_nxv2f32_0(<vscale x 2 x float> %v, float %elt) {
 ; CHECK-LABEL: insertelt_nxv2f32_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e32,m1,ta,mu
 ; CHECK-NEXT:    vfmv.s.f v8, fa0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 2 x float> %v, float %elt, i32 0
   ret <vscale x 2 x float> %r
 }
 
 define <vscale x 2 x float> @insertelt_nxv2f32_imm(<vscale x 2 x float> %v, float %elt) {
 ; CHECK-LABEL: insertelt_nxv2f32_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e32,m1,ta,mu
-; CHECK-NEXT:    vslidedown.vi v25, v8, 3
 ; CHECK-NEXT:    vfmv.s.f v25, fa0
-; CHECK-NEXT:    vsetvli a0, zero, e32,m1,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e32,m1,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v25, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 2 x float> %v, float %elt, i32 3
   ret <vscale x 2 x float> %r
 }
 
 define <vscale x 2 x float> @insertelt_nxv2f32_idx(<vscale x 2 x float> %v, float %elt, i32 %idx) {
 ; CHECK-LABEL: insertelt_nxv2f32_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e32,m1,ta,mu
-; CHECK-NEXT:    vslidedown.vx v25, v8, a0
 ; CHECK-NEXT:    vfmv.s.f v25, fa0
-; CHECK-NEXT:    vsetvli a1, zero, e32,m1,tu,mu
+; CHECK-NEXT:    addi a1, a0, 1
+; CHECK-NEXT:    vsetvli a1, a1, e32,m1,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v25, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 2 x float> %v, float %elt, i32 %idx
   ret <vscale x 2 x float> %r
 }
 
 define <vscale x 4 x float> @insertelt_nxv4f32_0(<vscale x 4 x float> %v, float %elt) {
 ; CHECK-LABEL: insertelt_nxv4f32_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e32,m2,ta,mu
 ; CHECK-NEXT:    vfmv.s.f v8, fa0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 4 x float> %v, float %elt, i32 0
   ret <vscale x 4 x float> %r
 }
 
 define <vscale x 4 x float> @insertelt_nxv4f32_imm(<vscale x 4 x float> %v, float %elt) {
 ; CHECK-LABEL: insertelt_nxv4f32_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e32,m2,ta,mu
-; CHECK-NEXT:    vslidedown.vi v26, v8, 3
 ; CHECK-NEXT:    vfmv.s.f v26, fa0
-; CHECK-NEXT:    vsetvli a0, zero, e32,m2,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e32,m2,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v26, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 4 x float> %v, float %elt, i32 3
   ret <vscale x 4 x float> %r
 }
 
 define <vscale x 4 x float> @insertelt_nxv4f32_idx(<vscale x 4 x float> %v, float %elt, i32 %idx) {
 ; CHECK-LABEL: insertelt_nxv4f32_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e32,m2,ta,mu
-; CHECK-NEXT:    vslidedown.vx v26, v8, a0
 ; CHECK-NEXT:    vfmv.s.f v26, fa0
-; CHECK-NEXT:    vsetvli a1, zero, e32,m2,tu,mu
+; CHECK-NEXT:    addi a1, a0, 1
+; CHECK-NEXT:    vsetvli a1, a1, e32,m2,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v26, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 4 x float> %v, float %elt, i32 %idx
   ret <vscale x 4 x float> %r
 }
 
 define <vscale x 8 x float> @insertelt_nxv8f32_0(<vscale x 8 x float> %v, float %elt) {
 ; CHECK-LABEL: insertelt_nxv8f32_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e32,m4,ta,mu
 ; CHECK-NEXT:    vfmv.s.f v8, fa0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 8 x float> %v, float %elt, i32 0
   ret <vscale x 8 x float> %r
 }
 
 define <vscale x 8 x float> @insertelt_nxv8f32_imm(<vscale x 8 x float> %v, float %elt) {
 ; CHECK-LABEL: insertelt_nxv8f32_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e32,m4,ta,mu
-; CHECK-NEXT:    vslidedown.vi v28, v8, 3
 ; CHECK-NEXT:    vfmv.s.f v28, fa0
-; CHECK-NEXT:    vsetvli a0, zero, e32,m4,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e32,m4,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v28, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 8 x float> %v, float %elt, i32 3
   ret <vscale x 8 x float> %r
 }
 
 define <vscale x 8 x float> @insertelt_nxv8f32_idx(<vscale x 8 x float> %v, float %elt, i32 %idx) {
 ; CHECK-LABEL: insertelt_nxv8f32_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e32,m4,ta,mu
-; CHECK-NEXT:    vslidedown.vx v28, v8, a0
 ; CHECK-NEXT:    vfmv.s.f v28, fa0
-; CHECK-NEXT:    vsetvli a1, zero, e32,m4,tu,mu
+; CHECK-NEXT:    addi a1, a0, 1
+; CHECK-NEXT:    vsetvli a1, a1, e32,m4,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v28, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 8 x float> %v, float %elt, i32 %idx
   ret <vscale x 8 x float> %r
 }
 
 define <vscale x 16 x float> @insertelt_nxv16f32_0(<vscale x 16 x float> %v, float %elt) {
 ; CHECK-LABEL: insertelt_nxv16f32_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e32,m8,ta,mu
 ; CHECK-NEXT:    vfmv.s.f v8, fa0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 16 x float> %v, float %elt, i32 0
   ret <vscale x 16 x float> %r
 }
 
 define <vscale x 16 x float> @insertelt_nxv16f32_imm(<vscale x 16 x float> %v, float %elt) {
 ; CHECK-LABEL: insertelt_nxv16f32_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e32,m8,ta,mu
-; CHECK-NEXT:    vslidedown.vi v16, v8, 3
 ; CHECK-NEXT:    vfmv.s.f v16, fa0
-; CHECK-NEXT:    vsetvli a0, zero, e32,m8,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e32,m8,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v16, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 16 x float> %v, float %elt, i32 3
   ret <vscale x 16 x float> %r
 }
 
 define <vscale x 16 x float> @insertelt_nxv16f32_idx(<vscale x 16 x float> %v, float %elt, i32 %idx) {
 ; CHECK-LABEL: insertelt_nxv16f32_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e32,m8,ta,mu
-; CHECK-NEXT:    vslidedown.vx v16, v8, a0
 ; CHECK-NEXT:    vfmv.s.f v16, fa0
-; CHECK-NEXT:    vsetvli a1, zero, e32,m8,tu,mu
+; CHECK-NEXT:    addi a1, a0, 1
+; CHECK-NEXT:    vsetvli a1, a1, e32,m8,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v16, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 16 x float> %v, float %elt, i32 %idx
   ret <vscale x 16 x float> %r
 }
 
 define <vscale x 1 x double> @insertelt_nxv1f64_0(<vscale x 1 x double> %v, double %elt) {
 ; CHECK-LABEL: insertelt_nxv1f64_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e64,m1,ta,mu
 ; CHECK-NEXT:    vfmv.s.f v8, fa0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 1 x double> %v, double %elt, i32 0
   ret <vscale x 1 x double> %r
 }
 
 define <vscale x 1 x double> @insertelt_nxv1f64_imm(<vscale x 1 x double> %v, double %elt) {
 ; CHECK-LABEL: insertelt_nxv1f64_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e64,m1,ta,mu
-; CHECK-NEXT:    vslidedown.vi v25, v8, 3
 ; CHECK-NEXT:    vfmv.s.f v25, fa0
-; CHECK-NEXT:    vsetvli a0, zero, e64,m1,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e64,m1,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v25, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 1 x double> %v, double %elt, i32 3
   ret <vscale x 1 x double> %r
 }
 
 define <vscale x 1 x double> @insertelt_nxv1f64_idx(<vscale x 1 x double> %v, double %elt, i32 %idx) {
 ; CHECK-LABEL: insertelt_nxv1f64_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e64,m1,ta,mu
-; CHECK-NEXT:    vslidedown.vx v25, v8, a0
 ; CHECK-NEXT:    vfmv.s.f v25, fa0
-; CHECK-NEXT:    vsetvli a1, zero, e64,m1,tu,mu
+; CHECK-NEXT:    addi a1, a0, 1
+; CHECK-NEXT:    vsetvli a1, a1, e64,m1,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v25, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 1 x double> %v, double %elt, i32 %idx
   ret <vscale x 1 x double> %r
 }
 
 define <vscale x 2 x double> @insertelt_nxv2f64_0(<vscale x 2 x double> %v, double %elt) {
 ; CHECK-LABEL: insertelt_nxv2f64_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e64,m2,ta,mu
 ; CHECK-NEXT:    vfmv.s.f v8, fa0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 2 x double> %v, double %elt, i32 0
   ret <vscale x 2 x double> %r
 }
 
 define <vscale x 2 x double> @insertelt_nxv2f64_imm(<vscale x 2 x double> %v, double %elt) {
 ; CHECK-LABEL: insertelt_nxv2f64_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e64,m2,ta,mu
-; CHECK-NEXT:    vslidedown.vi v26, v8, 3
 ; CHECK-NEXT:    vfmv.s.f v26, fa0
-; CHECK-NEXT:    vsetvli a0, zero, e64,m2,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e64,m2,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v26, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 2 x double> %v, double %elt, i32 3
   ret <vscale x 2 x double> %r
 }
 
 define <vscale x 2 x double> @insertelt_nxv2f64_idx(<vscale x 2 x double> %v, double %elt, i32 %idx) {
 ; CHECK-LABEL: insertelt_nxv2f64_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e64,m2,ta,mu
-; CHECK-NEXT:    vslidedown.vx v26, v8, a0
 ; CHECK-NEXT:    vfmv.s.f v26, fa0
-; CHECK-NEXT:    vsetvli a1, zero, e64,m2,tu,mu
+; CHECK-NEXT:    addi a1, a0, 1
+; CHECK-NEXT:    vsetvli a1, a1, e64,m2,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v26, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 2 x double> %v, double %elt, i32 %idx
   ret <vscale x 2 x double> %r
 }
 
 define <vscale x 4 x double> @insertelt_nxv4f64_0(<vscale x 4 x double> %v, double %elt) {
 ; CHECK-LABEL: insertelt_nxv4f64_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e64,m4,ta,mu
 ; CHECK-NEXT:    vfmv.s.f v8, fa0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 4 x double> %v, double %elt, i32 0
   ret <vscale x 4 x double> %r
 }
 
 define <vscale x 4 x double> @insertelt_nxv4f64_imm(<vscale x 4 x double> %v, double %elt) {
 ; CHECK-LABEL: insertelt_nxv4f64_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e64,m4,ta,mu
-; CHECK-NEXT:    vslidedown.vi v28, v8, 3
 ; CHECK-NEXT:    vfmv.s.f v28, fa0
-; CHECK-NEXT:    vsetvli a0, zero, e64,m4,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e64,m4,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v28, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 4 x double> %v, double %elt, i32 3
   ret <vscale x 4 x double> %r
 }
 
 define <vscale x 4 x double> @insertelt_nxv4f64_idx(<vscale x 4 x double> %v, double %elt, i32 %idx) {
 ; CHECK-LABEL: insertelt_nxv4f64_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e64,m4,ta,mu
-; CHECK-NEXT:    vslidedown.vx v28, v8, a0
 ; CHECK-NEXT:    vfmv.s.f v28, fa0
-; CHECK-NEXT:    vsetvli a1, zero, e64,m4,tu,mu
+; CHECK-NEXT:    addi a1, a0, 1
+; CHECK-NEXT:    vsetvli a1, a1, e64,m4,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v28, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 4 x double> %v, double %elt, i32 %idx
   ret <vscale x 4 x double> %r
 }
 
 define <vscale x 8 x double> @insertelt_nxv8f64_0(<vscale x 8 x double> %v, double %elt) {
 ; CHECK-LABEL: insertelt_nxv8f64_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e64,m8,ta,mu
 ; CHECK-NEXT:    vfmv.s.f v8, fa0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 8 x double> %v, double %elt, i32 0
   ret <vscale x 8 x double> %r
 }
 
 define <vscale x 8 x double> @insertelt_nxv8f64_imm(<vscale x 8 x double> %v, double %elt) {
 ; CHECK-LABEL: insertelt_nxv8f64_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e64,m8,ta,mu
-; CHECK-NEXT:    vslidedown.vi v16, v8, 3
 ; CHECK-NEXT:    vfmv.s.f v16, fa0
-; CHECK-NEXT:    vsetvli a0, zero, e64,m8,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e64,m8,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v16, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 8 x double> %v, double %elt, i32 3
   ret <vscale x 8 x double> %r
 }
 
 define <vscale x 8 x double> @insertelt_nxv8f64_idx(<vscale x 8 x double> %v, double %elt, i32 %idx) {
 ; CHECK-LABEL: insertelt_nxv8f64_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e64,m8,ta,mu
-; CHECK-NEXT:    vslidedown.vx v16, v8, a0
 ; CHECK-NEXT:    vfmv.s.f v16, fa0
-; CHECK-NEXT:    vsetvli a1, zero, e64,m8,tu,mu
+; CHECK-NEXT:    addi a1, a0, 1
+; CHECK-NEXT:    vsetvli a1, a1, e64,m8,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v16, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 8 x double> %v, double %elt, i32 %idx
   ret <vscale x 8 x double> %r
 }
 

llvm/test/CodeGen/RISCV/rvv/insertelt-fp-rv64.ll

; CHECK-NEXT:    ret
   %r = insertelement <vscale x 1 x half> %v, half %elt, i32 0
   ret <vscale x 1 x half> %r
 }
 
 define <vscale x 1 x half> @insertelt_nxv1f16_imm(<vscale x 1 x half> %v, half %elt) {
 ; CHECK-LABEL: insertelt_nxv1f16_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e16,mf4,ta,mu
-; CHECK-NEXT:    vslidedown.vi v25, v8, 3
 ; CHECK-NEXT:    vfmv.s.f v25, fa0
-; CHECK-NEXT:    vsetvli a0, zero, e16,mf4,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e16,mf4,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v25, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 1 x half> %v, half %elt, i32 3
   ret <vscale x 1 x half> %r
 }
 
 define <vscale x 1 x half> @insertelt_nxv1f16_idx(<vscale x 1 x half> %v, half %elt, i32 signext %idx) {
 ; CHECK-LABEL: insertelt_nxv1f16_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e16,mf4,ta,mu
-; CHECK-NEXT:    vslidedown.vx v25, v8, a0
 ; CHECK-NEXT:    vfmv.s.f v25, fa0
-; CHECK-NEXT:    vsetvli a1, zero, e16,mf4,tu,mu
+; CHECK-NEXT:    addi a1, a0, 1
+; CHECK-NEXT:    vsetvli a1, a1, e16,mf4,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v25, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 1 x half> %v, half %elt, i32 %idx
   ret <vscale x 1 x half> %r
 }
 
 define <vscale x 2 x half> @insertelt_nxv2f16_0(<vscale x 2 x half> %v, half %elt) {
 ; CHECK-LABEL: insertelt_nxv2f16_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e16,mf2,ta,mu
 ; CHECK-NEXT:    vfmv.s.f v8, fa0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 2 x half> %v, half %elt, i32 0
   ret <vscale x 2 x half> %r
 }
 
 define <vscale x 2 x half> @insertelt_nxv2f16_imm(<vscale x 2 x half> %v, half %elt) {
 ; CHECK-LABEL: insertelt_nxv2f16_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e16,mf2,ta,mu
-; CHECK-NEXT:    vslidedown.vi v25, v8, 3
 ; CHECK-NEXT:    vfmv.s.f v25, fa0
-; CHECK-NEXT:    vsetvli a0, zero, e16,mf2,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e16,mf2,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v25, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 2 x half> %v, half %elt, i32 3
   ret <vscale x 2 x half> %r
 }
 
 define <vscale x 2 x half> @insertelt_nxv2f16_idx(<vscale x 2 x half> %v, half %elt, i32 signext %idx) {
 ; CHECK-LABEL: insertelt_nxv2f16_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e16,mf2,ta,mu
-; CHECK-NEXT:    vslidedown.vx v25, v8, a0
 ; CHECK-NEXT:    vfmv.s.f v25, fa0
-; CHECK-NEXT:    vsetvli a1, zero, e16,mf2,tu,mu
+; CHECK-NEXT:    addi a1, a0, 1
+; CHECK-NEXT:    vsetvli a1, a1, e16,mf2,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v25, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 2 x half> %v, half %elt, i32 %idx
   ret <vscale x 2 x half> %r
 }
 
 define <vscale x 4 x half> @insertelt_nxv4f16_0(<vscale x 4 x half> %v, half %elt) {
 ; CHECK-LABEL: insertelt_nxv4f16_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e16,m1,ta,mu
 ; CHECK-NEXT:    vfmv.s.f v8, fa0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 4 x half> %v, half %elt, i32 0
   ret <vscale x 4 x half> %r
 }
 
 define <vscale x 4 x half> @insertelt_nxv4f16_imm(<vscale x 4 x half> %v, half %elt) {
 ; CHECK-LABEL: insertelt_nxv4f16_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e16,m1,ta,mu
-; CHECK-NEXT:    vslidedown.vi v25, v8, 3
 ; CHECK-NEXT:    vfmv.s.f v25, fa0
-; CHECK-NEXT:    vsetvli a0, zero, e16,m1,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e16,m1,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v25, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 4 x half> %v, half %elt, i32 3
   ret <vscale x 4 x half> %r
 }
 
 define <vscale x 4 x half> @insertelt_nxv4f16_idx(<vscale x 4 x half> %v, half %elt, i32 signext %idx) {
 ; CHECK-LABEL: insertelt_nxv4f16_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e16,m1,ta,mu
-; CHECK-NEXT:    vslidedown.vx v25, v8, a0
 ; CHECK-NEXT:    vfmv.s.f v25, fa0
-; CHECK-NEXT:    vsetvli a1, zero, e16,m1,tu,mu
+; CHECK-NEXT:    addi a1, a0, 1
+; CHECK-NEXT:    vsetvli a1, a1, e16,m1,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v25, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 4 x half> %v, half %elt, i32 %idx
   ret <vscale x 4 x half> %r
 }
 
 define <vscale x 8 x half> @insertelt_nxv8f16_0(<vscale x 8 x half> %v, half %elt) {
 ; CHECK-LABEL: insertelt_nxv8f16_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e16,m2,ta,mu
 ; CHECK-NEXT:    vfmv.s.f v8, fa0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 8 x half> %v, half %elt, i32 0
   ret <vscale x 8 x half> %r
 }
 
 define <vscale x 8 x half> @insertelt_nxv8f16_imm(<vscale x 8 x half> %v, half %elt) {
 ; CHECK-LABEL: insertelt_nxv8f16_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e16,m2,ta,mu
-; CHECK-NEXT:    vslidedown.vi v26, v8, 3
 ; CHECK-NEXT:    vfmv.s.f v26, fa0
-; CHECK-NEXT:    vsetvli a0, zero, e16,m2,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e16,m2,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v26, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 8 x half> %v, half %elt, i32 3
   ret <vscale x 8 x half> %r
 }
 
 define <vscale x 8 x half> @insertelt_nxv8f16_idx(<vscale x 8 x half> %v, half %elt, i32 signext %idx) {
 ; CHECK-LABEL: insertelt_nxv8f16_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e16,m2,ta,mu
-; CHECK-NEXT:    vslidedown.vx v26, v8, a0
 ; CHECK-NEXT:    vfmv.s.f v26, fa0
-; CHECK-NEXT:    vsetvli a1, zero, e16,m2,tu,mu
+; CHECK-NEXT:    addi a1, a0, 1
+; CHECK-NEXT:    vsetvli a1, a1, e16,m2,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v26, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 8 x half> %v, half %elt, i32 %idx
   ret <vscale x 8 x half> %r
 }
 
 define <vscale x 16 x half> @insertelt_nxv16f16_0(<vscale x 16 x half> %v, half %elt) {
 ; CHECK-LABEL: insertelt_nxv16f16_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e16,m4,ta,mu
 ; CHECK-NEXT:    vfmv.s.f v8, fa0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 16 x half> %v, half %elt, i32 0
   ret <vscale x 16 x half> %r
 }
 
 define <vscale x 16 x half> @insertelt_nxv16f16_imm(<vscale x 16 x half> %v, half %elt) {
 ; CHECK-LABEL: insertelt_nxv16f16_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e16,m4,ta,mu
-; CHECK-NEXT:    vslidedown.vi v28, v8, 3
 ; CHECK-NEXT:    vfmv.s.f v28, fa0
-; CHECK-NEXT:    vsetvli a0, zero, e16,m4,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e16,m4,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v28, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 16 x half> %v, half %elt, i32 3
   ret <vscale x 16 x half> %r
 }
 
 define <vscale x 16 x half> @insertelt_nxv16f16_idx(<vscale x 16 x half> %v, half %elt, i32 signext %idx) {
 ; CHECK-LABEL: insertelt_nxv16f16_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e16,m4,ta,mu
-; CHECK-NEXT:    vslidedown.vx v28, v8, a0
 ; CHECK-NEXT:    vfmv.s.f v28, fa0
-; CHECK-NEXT:    vsetvli a1, zero, e16,m4,tu,mu
+; CHECK-NEXT:    addi a1, a0, 1
+; CHECK-NEXT:    vsetvli a1, a1, e16,m4,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v28, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 16 x half> %v, half %elt, i32 %idx
   ret <vscale x 16 x half> %r
 }
 
 define <vscale x 32 x half> @insertelt_nxv32f16_0(<vscale x 32 x half> %v, half %elt) {
 ; CHECK-LABEL: insertelt_nxv32f16_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e16,m8,ta,mu
 ; CHECK-NEXT:    vfmv.s.f v8, fa0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 32 x half> %v, half %elt, i32 0
   ret <vscale x 32 x half> %r
 }
 
 define <vscale x 32 x half> @insertelt_nxv32f16_imm(<vscale x 32 x half> %v, half %elt) {
 ; CHECK-LABEL: insertelt_nxv32f16_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e16,m8,ta,mu
-; CHECK-NEXT:    vslidedown.vi v16, v8, 3
 ; CHECK-NEXT:    vfmv.s.f v16, fa0
-; CHECK-NEXT:    vsetvli a0, zero, e16,m8,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e16,m8,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v16, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 32 x half> %v, half %elt, i32 3
   ret <vscale x 32 x half> %r
 }
 
 define <vscale x 32 x half> @insertelt_nxv32f16_idx(<vscale x 32 x half> %v, half %elt, i32 signext %idx) {
 ; CHECK-LABEL: insertelt_nxv32f16_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e16,m8,ta,mu
-; CHECK-NEXT:    vslidedown.vx v16, v8, a0
 ; CHECK-NEXT:    vfmv.s.f v16, fa0
-; CHECK-NEXT:    vsetvli a1, zero, e16,m8,tu,mu
+; CHECK-NEXT:    addi a1, a0, 1
+; CHECK-NEXT:    vsetvli a1, a1, e16,m8,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v16, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 32 x half> %v, half %elt, i32 %idx
   ret <vscale x 32 x half> %r
 }
 
 define <vscale x 1 x float> @insertelt_nxv1f32_0(<vscale x 1 x float> %v, float %elt) {
 ; CHECK-LABEL: insertelt_nxv1f32_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e32,mf2,ta,mu
 ; CHECK-NEXT:    vfmv.s.f v8, fa0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 1 x float> %v, float %elt, i32 0
   ret <vscale x 1 x float> %r
 }
 
 define <vscale x 1 x float> @insertelt_nxv1f32_imm(<vscale x 1 x float> %v, float %elt) {
 ; CHECK-LABEL: insertelt_nxv1f32_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e32,mf2,ta,mu
-; CHECK-NEXT:    vslidedown.vi v25, v8, 3
 ; CHECK-NEXT:    vfmv.s.f v25, fa0
-; CHECK-NEXT:    vsetvli a0, zero, e32,mf2,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e32,mf2,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v25, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 1 x float> %v, float %elt, i32 3
   ret <vscale x 1 x float> %r
 }
 
 define <vscale x 1 x float> @insertelt_nxv1f32_idx(<vscale x 1 x float> %v, float %elt, i32 signext %idx) {
 ; CHECK-LABEL: insertelt_nxv1f32_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e32,mf2,ta,mu
-; CHECK-NEXT:    vslidedown.vx v25, v8, a0
 ; CHECK-NEXT:    vfmv.s.f v25, fa0
-; CHECK-NEXT:    vsetvli a1, zero, e32,mf2,tu,mu
+; CHECK-NEXT:    addi a1, a0, 1
+; CHECK-NEXT:    vsetvli a1, a1, e32,mf2,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v25, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 1 x float> %v, float %elt, i32 %idx
   ret <vscale x 1 x float> %r
 }
 
 define <vscale x 2 x float> @insertelt_nxv2f32_0(<vscale x 2 x float> %v, float %elt) {
 ; CHECK-LABEL: insertelt_nxv2f32_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e32,m1,ta,mu
 ; CHECK-NEXT:    vfmv.s.f v8, fa0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 2 x float> %v, float %elt, i32 0
   ret <vscale x 2 x float> %r
 }
 
 define <vscale x 2 x float> @insertelt_nxv2f32_imm(<vscale x 2 x float> %v, float %elt) {
 ; CHECK-LABEL: insertelt_nxv2f32_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e32,m1,ta,mu
-; CHECK-NEXT:    vslidedown.vi v25, v8, 3
 ; CHECK-NEXT:    vfmv.s.f v25, fa0
-; CHECK-NEXT:    vsetvli a0, zero, e32,m1,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e32,m1,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v25, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 2 x float> %v, float %elt, i32 3
   ret <vscale x 2 x float> %r
 }
 
 define <vscale x 2 x float> @insertelt_nxv2f32_idx(<vscale x 2 x float> %v, float %elt, i32 signext %idx) {
 ; CHECK-LABEL: insertelt_nxv2f32_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e32,m1,ta,mu
-; CHECK-NEXT:    vslidedown.vx v25, v8, a0
 ; CHECK-NEXT:    vfmv.s.f v25, fa0
-; CHECK-NEXT:    vsetvli a1, zero, e32,m1,tu,mu
+; CHECK-NEXT:    addi a1, a0, 1
+; CHECK-NEXT:    vsetvli a1, a1, e32,m1,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v25, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 2 x float> %v, float %elt, i32 %idx
   ret <vscale x 2 x float> %r
 }
 
 define <vscale x 4 x float> @insertelt_nxv4f32_0(<vscale x 4 x float> %v, float %elt) {
 ; CHECK-LABEL: insertelt_nxv4f32_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e32,m2,ta,mu
 ; CHECK-NEXT:    vfmv.s.f v8, fa0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 4 x float> %v, float %elt, i32 0
   ret <vscale x 4 x float> %r
 }
 
 define <vscale x 4 x float> @insertelt_nxv4f32_imm(<vscale x 4 x float> %v, float %elt) {
 ; CHECK-LABEL: insertelt_nxv4f32_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e32,m2,ta,mu
-; CHECK-NEXT:    vslidedown.vi v26, v8, 3
 ; CHECK-NEXT:    vfmv.s.f v26, fa0
-; CHECK-NEXT:    vsetvli a0, zero, e32,m2,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e32,m2,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v26, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 4 x float> %v, float %elt, i32 3
   ret <vscale x 4 x float> %r
 }
 
 define <vscale x 4 x float> @insertelt_nxv4f32_idx(<vscale x 4 x float> %v, float %elt, i32 signext %idx) {
 ; CHECK-LABEL: insertelt_nxv4f32_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e32,m2,ta,mu
-; CHECK-NEXT:    vslidedown.vx v26, v8, a0
 ; CHECK-NEXT:    vfmv.s.f v26, fa0
-; CHECK-NEXT:    vsetvli a1, zero, e32,m2,tu,mu
+; CHECK-NEXT:    addi a1, a0, 1
+; CHECK-NEXT:    vsetvli a1, a1, e32,m2,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v26, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 4 x float> %v, float %elt, i32 %idx
   ret <vscale x 4 x float> %r
 }
 
 define <vscale x 8 x float> @insertelt_nxv8f32_0(<vscale x 8 x float> %v, float %elt) {
 ; CHECK-LABEL: insertelt_nxv8f32_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e32,m4,ta,mu
 ; CHECK-NEXT:    vfmv.s.f v8, fa0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 8 x float> %v, float %elt, i32 0
   ret <vscale x 8 x float> %r
 }
 
 define <vscale x 8 x float> @insertelt_nxv8f32_imm(<vscale x 8 x float> %v, float %elt) {
 ; CHECK-LABEL: insertelt_nxv8f32_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e32,m4,ta,mu
-; CHECK-NEXT:    vslidedown.vi v28, v8, 3
 ; CHECK-NEXT:    vfmv.s.f v28, fa0
-; CHECK-NEXT:    vsetvli a0, zero, e32,m4,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e32,m4,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v28, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 8 x float> %v, float %elt, i32 3
   ret <vscale x 8 x float> %r
 }
 
 define <vscale x 8 x float> @insertelt_nxv8f32_idx(<vscale x 8 x float> %v, float %elt, i32 signext %idx) {
 ; CHECK-LABEL: insertelt_nxv8f32_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e32,m4,ta,mu
-; CHECK-NEXT:    vslidedown.vx v28, v8, a0
 ; CHECK-NEXT:    vfmv.s.f v28, fa0
-; CHECK-NEXT:    vsetvli a1, zero, e32,m4,tu,mu
+; CHECK-NEXT:    addi a1, a0, 1
+; CHECK-NEXT:    vsetvli a1, a1, e32,m4,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v28, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 8 x float> %v, float %elt, i32 %idx
   ret <vscale x 8 x float> %r
 }
 
 define <vscale x 16 x float> @insertelt_nxv16f32_0(<vscale x 16 x float> %v, float %elt) {
 ; CHECK-LABEL: insertelt_nxv16f32_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e32,m8,ta,mu
 ; CHECK-NEXT:    vfmv.s.f v8, fa0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 16 x float> %v, float %elt, i32 0
   ret <vscale x 16 x float> %r
 }
 
 define <vscale x 16 x float> @insertelt_nxv16f32_imm(<vscale x 16 x float> %v, float %elt) {
 ; CHECK-LABEL: insertelt_nxv16f32_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e32,m8,ta,mu
-; CHECK-NEXT:    vslidedown.vi v16, v8, 3
 ; CHECK-NEXT:    vfmv.s.f v16, fa0
-; CHECK-NEXT:    vsetvli a0, zero, e32,m8,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e32,m8,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v16, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 16 x float> %v, float %elt, i32 3
   ret <vscale x 16 x float> %r
 }
 
 define <vscale x 16 x float> @insertelt_nxv16f32_idx(<vscale x 16 x float> %v, float %elt, i32 signext %idx) {
 ; CHECK-LABEL: insertelt_nxv16f32_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e32,m8,ta,mu
-; CHECK-NEXT:    vslidedown.vx v16, v8, a0
 ; CHECK-NEXT:    vfmv.s.f v16, fa0
-; CHECK-NEXT:    vsetvli a1, zero, e32,m8,tu,mu
+; CHECK-NEXT:    addi a1, a0, 1
+; CHECK-NEXT:    vsetvli a1, a1, e32,m8,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v16, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 16 x float> %v, float %elt, i32 %idx
   ret <vscale x 16 x float> %r
 }
 
 define <vscale x 1 x double> @insertelt_nxv1f64_0(<vscale x 1 x double> %v, double %elt) {
 ; CHECK-LABEL: insertelt_nxv1f64_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e64,m1,ta,mu
 ; CHECK-NEXT:    vfmv.s.f v8, fa0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 1 x double> %v, double %elt, i32 0
   ret <vscale x 1 x double> %r
 }
 
 define <vscale x 1 x double> @insertelt_nxv1f64_imm(<vscale x 1 x double> %v, double %elt) {
 ; CHECK-LABEL: insertelt_nxv1f64_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e64,m1,ta,mu
-; CHECK-NEXT:    vslidedown.vi v25, v8, 3
 ; CHECK-NEXT:    vfmv.s.f v25, fa0
-; CHECK-NEXT:    vsetvli a0, zero, e64,m1,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e64,m1,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v25, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 1 x double> %v, double %elt, i32 3
   ret <vscale x 1 x double> %r
 }
 
 define <vscale x 1 x double> @insertelt_nxv1f64_idx(<vscale x 1 x double> %v, double %elt, i32 signext %idx) {
 ; CHECK-LABEL: insertelt_nxv1f64_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e64,m1,ta,mu
-; CHECK-NEXT:    vslidedown.vx v25, v8, a0
 ; CHECK-NEXT:    vfmv.s.f v25, fa0
-; CHECK-NEXT:    vsetvli a1, zero, e64,m1,tu,mu
+; CHECK-NEXT:    addi a1, a0, 1
+; CHECK-NEXT:    vsetvli a1, a1, e64,m1,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v25, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 1 x double> %v, double %elt, i32 %idx
   ret <vscale x 1 x double> %r
 }
 
 define <vscale x 2 x double> @insertelt_nxv2f64_0(<vscale x 2 x double> %v, double %elt) {
 ; CHECK-LABEL: insertelt_nxv2f64_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e64,m2,ta,mu
 ; CHECK-NEXT:    vfmv.s.f v8, fa0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 2 x double> %v, double %elt, i32 0
   ret <vscale x 2 x double> %r
 }
 
 define <vscale x 2 x double> @insertelt_nxv2f64_imm(<vscale x 2 x double> %v, double %elt) {
 ; CHECK-LABEL: insertelt_nxv2f64_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e64,m2,ta,mu
-; CHECK-NEXT:    vslidedown.vi v26, v8, 3
 ; CHECK-NEXT:    vfmv.s.f v26, fa0
-; CHECK-NEXT:    vsetvli a0, zero, e64,m2,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e64,m2,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v26, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 2 x double> %v, double %elt, i32 3
   ret <vscale x 2 x double> %r
 }
 
 define <vscale x 2 x double> @insertelt_nxv2f64_idx(<vscale x 2 x double> %v, double %elt, i32 signext %idx) {
 ; CHECK-LABEL: insertelt_nxv2f64_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e64,m2,ta,mu
-; CHECK-NEXT:    vslidedown.vx v26, v8, a0
 ; CHECK-NEXT:    vfmv.s.f v26, fa0
-; CHECK-NEXT:    vsetvli a1, zero, e64,m2,tu,mu
+; CHECK-NEXT:    addi a1, a0, 1
+; CHECK-NEXT:    vsetvli a1, a1, e64,m2,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v26, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 2 x double> %v, double %elt, i32 %idx
   ret <vscale x 2 x double> %r
 }
 
 define <vscale x 4 x double> @insertelt_nxv4f64_0(<vscale x 4 x double> %v, double %elt) {
 ; CHECK-LABEL: insertelt_nxv4f64_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e64,m4,ta,mu
 ; CHECK-NEXT:    vfmv.s.f v8, fa0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 4 x double> %v, double %elt, i32 0
   ret <vscale x 4 x double> %r
 }
 
 define <vscale x 4 x double> @insertelt_nxv4f64_imm(<vscale x 4 x double> %v, double %elt) {
 ; CHECK-LABEL: insertelt_nxv4f64_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e64,m4,ta,mu
-; CHECK-NEXT:    vslidedown.vi v28, v8, 3
 ; CHECK-NEXT:    vfmv.s.f v28, fa0
-; CHECK-NEXT:    vsetvli a0, zero, e64,m4,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e64,m4,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v28, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 4 x double> %v, double %elt, i32 3
   ret <vscale x 4 x double> %r
 }
 
 define <vscale x 4 x double> @insertelt_nxv4f64_idx(<vscale x 4 x double> %v, double %elt, i32 signext %idx) {
 ; CHECK-LABEL: insertelt_nxv4f64_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e64,m4,ta,mu
-; CHECK-NEXT:    vslidedown.vx v28, v8, a0
 ; CHECK-NEXT:    vfmv.s.f v28, fa0
-; CHECK-NEXT:    vsetvli a1, zero, e64,m4,tu,mu
+; CHECK-NEXT:    addi a1, a0, 1
+; CHECK-NEXT:    vsetvli a1, a1, e64,m4,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v28, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 4 x double> %v, double %elt, i32 %idx
   ret <vscale x 4 x double> %r
 }
 
 define <vscale x 8 x double> @insertelt_nxv8f64_0(<vscale x 8 x double> %v, double %elt) {
 ; CHECK-LABEL: insertelt_nxv8f64_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e64,m8,ta,mu
 ; CHECK-NEXT:    vfmv.s.f v8, fa0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 8 x double> %v, double %elt, i32 0
   ret <vscale x 8 x double> %r
 }
 
 define <vscale x 8 x double> @insertelt_nxv8f64_imm(<vscale x 8 x double> %v, double %elt) {
 ; CHECK-LABEL: insertelt_nxv8f64_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a0, zero, e64,m8,ta,mu
-; CHECK-NEXT:    vslidedown.vi v16, v8, 3
 ; CHECK-NEXT:    vfmv.s.f v16, fa0
-; CHECK-NEXT:    vsetvli a0, zero, e64,m8,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e64,m8,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v16, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 8 x double> %v, double %elt, i32 3
   ret <vscale x 8 x double> %r
 }
 
 define <vscale x 8 x double> @insertelt_nxv8f64_idx(<vscale x 8 x double> %v, double %elt, i32 signext %idx) {
 ; CHECK-LABEL: insertelt_nxv8f64_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e64,m8,ta,mu
-; CHECK-NEXT:    vslidedown.vx v16, v8, a0
 ; CHECK-NEXT:    vfmv.s.f v16, fa0
-; CHECK-NEXT:    vsetvli a1, zero, e64,m8,tu,mu
+; CHECK-NEXT:    addi a1, a0, 1
+; CHECK-NEXT:    vsetvli a1, a1, e64,m8,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v16, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 8 x double> %v, double %elt, i32 %idx
   ret <vscale x 8 x double> %r
 }
 

llvm/test/CodeGen/RISCV/rvv/insertelt-int-rv32.ll

; CHECK-NEXT:    ret
   %r = insertelement <vscale x 1 x i8> %v, i8 %elt, i32 0
   ret <vscale x 1 x i8> %r
 }
 
 define <vscale x 1 x i8> @insertelt_nxv1i8_imm(<vscale x 1 x i8> %v, i8 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv1i8_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e8,mf8,ta,mu
-; CHECK-NEXT:    vslidedown.vi v25, v8, 3
 ; CHECK-NEXT:    vmv.s.x v25, a0
-; CHECK-NEXT:    vsetvli a0, zero, e8,mf8,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e8,mf8,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v25, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 1 x i8> %v, i8 %elt, i32 3
   ret <vscale x 1 x i8> %r
 }
 
 define <vscale x 1 x i8> @insertelt_nxv1i8_idx(<vscale x 1 x i8> %v, i8 signext %elt, i32 signext %idx) {
 ; CHECK-LABEL: insertelt_nxv1i8_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a2, zero, e8,mf8,ta,mu
-; CHECK-NEXT:    vslidedown.vx v25, v8, a1
 ; CHECK-NEXT:    vmv.s.x v25, a0
-; CHECK-NEXT:    vsetvli a0, zero, e8,mf8,tu,mu
+; CHECK-NEXT:    addi a0, a1, 1
+; CHECK-NEXT:    vsetvli a0, a0, e8,mf8,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v25, a1
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 1 x i8> %v, i8 %elt, i32 %idx
   ret <vscale x 1 x i8> %r
 }
 
 define <vscale x 2 x i8> @insertelt_nxv2i8_0(<vscale x 2 x i8> %v, i8 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv2i8_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e8,mf4,ta,mu
 ; CHECK-NEXT:    vmv.s.x v8, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 2 x i8> %v, i8 %elt, i32 0
   ret <vscale x 2 x i8> %r
 }
 
 define <vscale x 2 x i8> @insertelt_nxv2i8_imm(<vscale x 2 x i8> %v, i8 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv2i8_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e8,mf4,ta,mu
-; CHECK-NEXT:    vslidedown.vi v25, v8, 3
 ; CHECK-NEXT:    vmv.s.x v25, a0
-; CHECK-NEXT:    vsetvli a0, zero, e8,mf4,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e8,mf4,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v25, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 2 x i8> %v, i8 %elt, i32 3
   ret <vscale x 2 x i8> %r
 }
 
 define <vscale x 2 x i8> @insertelt_nxv2i8_idx(<vscale x 2 x i8> %v, i8 signext %elt, i32 signext %idx) {
 ; CHECK-LABEL: insertelt_nxv2i8_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a2, zero, e8,mf4,ta,mu
-; CHECK-NEXT:    vslidedown.vx v25, v8, a1
 ; CHECK-NEXT:    vmv.s.x v25, a0
-; CHECK-NEXT:    vsetvli a0, zero, e8,mf4,tu,mu
+; CHECK-NEXT:    addi a0, a1, 1
+; CHECK-NEXT:    vsetvli a0, a0, e8,mf4,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v25, a1
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 2 x i8> %v, i8 %elt, i32 %idx
   ret <vscale x 2 x i8> %r
 }
 
 define <vscale x 4 x i8> @insertelt_nxv4i8_0(<vscale x 4 x i8> %v, i8 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv4i8_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e8,mf2,ta,mu
 ; CHECK-NEXT:    vmv.s.x v8, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 4 x i8> %v, i8 %elt, i32 0
   ret <vscale x 4 x i8> %r
 }
 
 define <vscale x 4 x i8> @insertelt_nxv4i8_imm(<vscale x 4 x i8> %v, i8 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv4i8_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e8,mf2,ta,mu
-; CHECK-NEXT:    vslidedown.vi v25, v8, 3
 ; CHECK-NEXT:    vmv.s.x v25, a0
-; CHECK-NEXT:    vsetvli a0, zero, e8,mf2,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e8,mf2,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v25, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 4 x i8> %v, i8 %elt, i32 3
   ret <vscale x 4 x i8> %r
 }
 
 define <vscale x 4 x i8> @insertelt_nxv4i8_idx(<vscale x 4 x i8> %v, i8 signext %elt, i32 signext %idx) {
 ; CHECK-LABEL: insertelt_nxv4i8_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a2, zero, e8,mf2,ta,mu
-; CHECK-NEXT:    vslidedown.vx v25, v8, a1
 ; CHECK-NEXT:    vmv.s.x v25, a0
-; CHECK-NEXT:    vsetvli a0, zero, e8,mf2,tu,mu
+; CHECK-NEXT:    addi a0, a1, 1
+; CHECK-NEXT:    vsetvli a0, a0, e8,mf2,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v25, a1
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 4 x i8> %v, i8 %elt, i32 %idx
   ret <vscale x 4 x i8> %r
 }
 
 define <vscale x 8 x i8> @insertelt_nxv8i8_0(<vscale x 8 x i8> %v, i8 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv8i8_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e8,m1,ta,mu
 ; CHECK-NEXT:    vmv.s.x v8, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 8 x i8> %v, i8 %elt, i32 0
   ret <vscale x 8 x i8> %r
 }
 
 define <vscale x 8 x i8> @insertelt_nxv8i8_imm(<vscale x 8 x i8> %v, i8 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv8i8_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e8,m1,ta,mu
-; CHECK-NEXT:    vslidedown.vi v25, v8, 3
 ; CHECK-NEXT:    vmv.s.x v25, a0
-; CHECK-NEXT:    vsetvli a0, zero, e8,m1,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e8,m1,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v25, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 8 x i8> %v, i8 %elt, i32 3
   ret <vscale x 8 x i8> %r
 }
 
 define <vscale x 8 x i8> @insertelt_nxv8i8_idx(<vscale x 8 x i8> %v, i8 signext %elt, i32 signext %idx) {
 ; CHECK-LABEL: insertelt_nxv8i8_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a2, zero, e8,m1,ta,mu
-; CHECK-NEXT:    vslidedown.vx v25, v8, a1
 ; CHECK-NEXT:    vmv.s.x v25, a0
-; CHECK-NEXT:    vsetvli a0, zero, e8,m1,tu,mu
+; CHECK-NEXT:    addi a0, a1, 1
+; CHECK-NEXT:    vsetvli a0, a0, e8,m1,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v25, a1
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 8 x i8> %v, i8 %elt, i32 %idx
   ret <vscale x 8 x i8> %r
 }
 
 define <vscale x 16 x i8> @insertelt_nxv16i8_0(<vscale x 16 x i8> %v, i8 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv16i8_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e8,m2,ta,mu
 ; CHECK-NEXT:    vmv.s.x v8, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 16 x i8> %v, i8 %elt, i32 0
   ret <vscale x 16 x i8> %r
 }
 
 define <vscale x 16 x i8> @insertelt_nxv16i8_imm(<vscale x 16 x i8> %v, i8 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv16i8_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e8,m2,ta,mu
-; CHECK-NEXT:    vslidedown.vi v26, v8, 3
 ; CHECK-NEXT:    vmv.s.x v26, a0
-; CHECK-NEXT:    vsetvli a0, zero, e8,m2,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e8,m2,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v26, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 16 x i8> %v, i8 %elt, i32 3
   ret <vscale x 16 x i8> %r
 }
 
 define <vscale x 16 x i8> @insertelt_nxv16i8_idx(<vscale x 16 x i8> %v, i8 signext %elt, i32 signext %idx) {
 ; CHECK-LABEL: insertelt_nxv16i8_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a2, zero, e8,m2,ta,mu
-; CHECK-NEXT:    vslidedown.vx v26, v8, a1
 ; CHECK-NEXT:    vmv.s.x v26, a0
-; CHECK-NEXT:    vsetvli a0, zero, e8,m2,tu,mu
+; CHECK-NEXT:    addi a0, a1, 1
+; CHECK-NEXT:    vsetvli a0, a0, e8,m2,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v26, a1
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 16 x i8> %v, i8 %elt, i32 %idx
   ret <vscale x 16 x i8> %r
 }
 
 define <vscale x 32 x i8> @insertelt_nxv32i8_0(<vscale x 32 x i8> %v, i8 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv32i8_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e8,m4,ta,mu
 ; CHECK-NEXT:    vmv.s.x v8, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 32 x i8> %v, i8 %elt, i32 0
   ret <vscale x 32 x i8> %r
 }
 
 define <vscale x 32 x i8> @insertelt_nxv32i8_imm(<vscale x 32 x i8> %v, i8 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv32i8_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e8,m4,ta,mu
-; CHECK-NEXT:    vslidedown.vi v28, v8, 3
 ; CHECK-NEXT:    vmv.s.x v28, a0
-; CHECK-NEXT:    vsetvli a0, zero, e8,m4,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e8,m4,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v28, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 32 x i8> %v, i8 %elt, i32 3
   ret <vscale x 32 x i8> %r
 }
 
 define <vscale x 32 x i8> @insertelt_nxv32i8_idx(<vscale x 32 x i8> %v, i8 signext %elt, i32 signext %idx) {
 ; CHECK-LABEL: insertelt_nxv32i8_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a2, zero, e8,m4,ta,mu
-; CHECK-NEXT:    vslidedown.vx v28, v8, a1
 ; CHECK-NEXT:    vmv.s.x v28, a0
-; CHECK-NEXT:    vsetvli a0, zero, e8,m4,tu,mu
+; CHECK-NEXT:    addi a0, a1, 1
+; CHECK-NEXT:    vsetvli a0, a0, e8,m4,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v28, a1
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 32 x i8> %v, i8 %elt, i32 %idx
   ret <vscale x 32 x i8> %r
 }
 
 define <vscale x 64 x i8> @insertelt_nxv64i8_0(<vscale x 64 x i8> %v, i8 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv64i8_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e8,m8,ta,mu
 ; CHECK-NEXT:    vmv.s.x v8, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 64 x i8> %v, i8 %elt, i32 0
   ret <vscale x 64 x i8> %r
 }
 
 define <vscale x 64 x i8> @insertelt_nxv64i8_imm(<vscale x 64 x i8> %v, i8 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv64i8_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e8,m8,ta,mu
-; CHECK-NEXT:    vslidedown.vi v16, v8, 3
 ; CHECK-NEXT:    vmv.s.x v16, a0
-; CHECK-NEXT:    vsetvli a0, zero, e8,m8,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e8,m8,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v16, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 64 x i8> %v, i8 %elt, i32 3
   ret <vscale x 64 x i8> %r
 }
 
 define <vscale x 64 x i8> @insertelt_nxv64i8_idx(<vscale x 64 x i8> %v, i8 signext %elt, i32 signext %idx) {
 ; CHECK-LABEL: insertelt_nxv64i8_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a2, zero, e8,m8,ta,mu
-; CHECK-NEXT:    vslidedown.vx v16, v8, a1
 ; CHECK-NEXT:    vmv.s.x v16, a0
-; CHECK-NEXT:    vsetvli a0, zero, e8,m8,tu,mu
+; CHECK-NEXT:    addi a0, a1, 1
+; CHECK-NEXT:    vsetvli a0, a0, e8,m8,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v16, a1
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 64 x i8> %v, i8 %elt, i32 %idx
   ret <vscale x 64 x i8> %r
 }
 
 define <vscale x 1 x i16> @insertelt_nxv1i16_0(<vscale x 1 x i16> %v, i16 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv1i16_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e16,mf4,ta,mu
 ; CHECK-NEXT:    vmv.s.x v8, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 1 x i16> %v, i16 %elt, i32 0
   ret <vscale x 1 x i16> %r
 }
 
 define <vscale x 1 x i16> @insertelt_nxv1i16_imm(<vscale x 1 x i16> %v, i16 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv1i16_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e16,mf4,ta,mu
-; CHECK-NEXT:    vslidedown.vi v25, v8, 3
 ; CHECK-NEXT:    vmv.s.x v25, a0
-; CHECK-NEXT:    vsetvli a0, zero, e16,mf4,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e16,mf4,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v25, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 1 x i16> %v, i16 %elt, i32 3
   ret <vscale x 1 x i16> %r
 }
 
 define <vscale x 1 x i16> @insertelt_nxv1i16_idx(<vscale x 1 x i16> %v, i16 signext %elt, i32 signext %idx) {
 ; CHECK-LABEL: insertelt_nxv1i16_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a2, zero, e16,mf4,ta,mu
-; CHECK-NEXT:    vslidedown.vx v25, v8, a1
 ; CHECK-NEXT:    vmv.s.x v25, a0
-; CHECK-NEXT:    vsetvli a0, zero, e16,mf4,tu,mu
+; CHECK-NEXT:    addi a0, a1, 1
+; CHECK-NEXT:    vsetvli a0, a0, e16,mf4,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v25, a1
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 1 x i16> %v, i16 %elt, i32 %idx
   ret <vscale x 1 x i16> %r
 }
 
 define <vscale x 2 x i16> @insertelt_nxv2i16_0(<vscale x 2 x i16> %v, i16 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv2i16_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e16,mf2,ta,mu
 ; CHECK-NEXT:    vmv.s.x v8, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 2 x i16> %v, i16 %elt, i32 0
   ret <vscale x 2 x i16> %r
 }
 
 define <vscale x 2 x i16> @insertelt_nxv2i16_imm(<vscale x 2 x i16> %v, i16 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv2i16_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e16,mf2,ta,mu
-; CHECK-NEXT:    vslidedown.vi v25, v8, 3
 ; CHECK-NEXT:    vmv.s.x v25, a0
-; CHECK-NEXT:    vsetvli a0, zero, e16,mf2,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e16,mf2,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v25, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 2 x i16> %v, i16 %elt, i32 3
   ret <vscale x 2 x i16> %r
 }
 
 define <vscale x 2 x i16> @insertelt_nxv2i16_idx(<vscale x 2 x i16> %v, i16 signext %elt, i32 signext %idx) {
 ; CHECK-LABEL: insertelt_nxv2i16_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a2, zero, e16,mf2,ta,mu
-; CHECK-NEXT:    vslidedown.vx v25, v8, a1
 ; CHECK-NEXT:    vmv.s.x v25, a0
-; CHECK-NEXT:    vsetvli a0, zero, e16,mf2,tu,mu
+; CHECK-NEXT:    addi a0, a1, 1
+; CHECK-NEXT:    vsetvli a0, a0, e16,mf2,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v25, a1
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 2 x i16> %v, i16 %elt, i32 %idx
   ret <vscale x 2 x i16> %r
 }
 
 define <vscale x 4 x i16> @insertelt_nxv4i16_0(<vscale x 4 x i16> %v, i16 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv4i16_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e16,m1,ta,mu
 ; CHECK-NEXT:    vmv.s.x v8, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 4 x i16> %v, i16 %elt, i32 0
   ret <vscale x 4 x i16> %r
 }
 
 define <vscale x 4 x i16> @insertelt_nxv4i16_imm(<vscale x 4 x i16> %v, i16 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv4i16_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e16,m1,ta,mu
-; CHECK-NEXT:    vslidedown.vi v25, v8, 3
 ; CHECK-NEXT:    vmv.s.x v25, a0
-; CHECK-NEXT:    vsetvli a0, zero, e16,m1,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e16,m1,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v25, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 4 x i16> %v, i16 %elt, i32 3
   ret <vscale x 4 x i16> %r
 }
 
 define <vscale x 4 x i16> @insertelt_nxv4i16_idx(<vscale x 4 x i16> %v, i16 signext %elt, i32 signext %idx) {
 ; CHECK-LABEL: insertelt_nxv4i16_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a2, zero, e16,m1,ta,mu
-; CHECK-NEXT:    vslidedown.vx v25, v8, a1
 ; CHECK-NEXT:    vmv.s.x v25, a0
-; CHECK-NEXT:    vsetvli a0, zero, e16,m1,tu,mu
+; CHECK-NEXT:    addi a0, a1, 1
+; CHECK-NEXT:    vsetvli a0, a0, e16,m1,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v25, a1
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 4 x i16> %v, i16 %elt, i32 %idx
   ret <vscale x 4 x i16> %r
 }
 
 define <vscale x 8 x i16> @insertelt_nxv8i16_0(<vscale x 8 x i16> %v, i16 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv8i16_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e16,m2,ta,mu
 ; CHECK-NEXT:    vmv.s.x v8, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 8 x i16> %v, i16 %elt, i32 0
   ret <vscale x 8 x i16> %r
 }
 
 define <vscale x 8 x i16> @insertelt_nxv8i16_imm(<vscale x 8 x i16> %v, i16 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv8i16_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e16,m2,ta,mu
-; CHECK-NEXT:    vslidedown.vi v26, v8, 3
 ; CHECK-NEXT:    vmv.s.x v26, a0
-; CHECK-NEXT:    vsetvli a0, zero, e16,m2,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e16,m2,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v26, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 8 x i16> %v, i16 %elt, i32 3
   ret <vscale x 8 x i16> %r
 }
 
 define <vscale x 8 x i16> @insertelt_nxv8i16_idx(<vscale x 8 x i16> %v, i16 signext %elt, i32 signext %idx) {
 ; CHECK-LABEL: insertelt_nxv8i16_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a2, zero, e16,m2,ta,mu
-; CHECK-NEXT:    vslidedown.vx v26, v8, a1
 ; CHECK-NEXT:    vmv.s.x v26, a0
-; CHECK-NEXT:    vsetvli a0, zero, e16,m2,tu,mu
+; CHECK-NEXT:    addi a0, a1, 1
+; CHECK-NEXT:    vsetvli a0, a0, e16,m2,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v26, a1
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 8 x i16> %v, i16 %elt, i32 %idx
   ret <vscale x 8 x i16> %r
 }
 
 define <vscale x 16 x i16> @insertelt_nxv16i16_0(<vscale x 16 x i16> %v, i16 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv16i16_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e16,m4,ta,mu
 ; CHECK-NEXT:    vmv.s.x v8, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 16 x i16> %v, i16 %elt, i32 0
   ret <vscale x 16 x i16> %r
 }
 
 define <vscale x 16 x i16> @insertelt_nxv16i16_imm(<vscale x 16 x i16> %v, i16 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv16i16_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e16,m4,ta,mu
-; CHECK-NEXT:    vslidedown.vi v28, v8, 3
 ; CHECK-NEXT:    vmv.s.x v28, a0
-; CHECK-NEXT:    vsetvli a0, zero, e16,m4,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e16,m4,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v28, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 16 x i16> %v, i16 %elt, i32 3
   ret <vscale x 16 x i16> %r
 }
 
 define <vscale x 16 x i16> @insertelt_nxv16i16_idx(<vscale x 16 x i16> %v, i16 signext %elt, i32 signext %idx) {
 ; CHECK-LABEL: insertelt_nxv16i16_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a2, zero, e16,m4,ta,mu
-; CHECK-NEXT:    vslidedown.vx v28, v8, a1
 ; CHECK-NEXT:    vmv.s.x v28, a0
-; CHECK-NEXT:    vsetvli a0, zero, e16,m4,tu,mu
+; CHECK-NEXT:    addi a0, a1, 1
+; CHECK-NEXT:    vsetvli a0, a0, e16,m4,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v28, a1
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 16 x i16> %v, i16 %elt, i32 %idx
   ret <vscale x 16 x i16> %r
 }
 
 define <vscale x 32 x i16> @insertelt_nxv32i16_0(<vscale x 32 x i16> %v, i16 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv32i16_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e16,m8,ta,mu
 ; CHECK-NEXT:    vmv.s.x v8, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 32 x i16> %v, i16 %elt, i32 0
   ret <vscale x 32 x i16> %r
 }
 
 define <vscale x 32 x i16> @insertelt_nxv32i16_imm(<vscale x 32 x i16> %v, i16 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv32i16_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e16,m8,ta,mu
-; CHECK-NEXT:    vslidedown.vi v16, v8, 3
 ; CHECK-NEXT:    vmv.s.x v16, a0
-; CHECK-NEXT:    vsetvli a0, zero, e16,m8,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e16,m8,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v16, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 32 x i16> %v, i16 %elt, i32 3
   ret <vscale x 32 x i16> %r
 }
 
 define <vscale x 32 x i16> @insertelt_nxv32i16_idx(<vscale x 32 x i16> %v, i16 signext %elt, i32 signext %idx) {
 ; CHECK-LABEL: insertelt_nxv32i16_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a2, zero, e16,m8,ta,mu
-; CHECK-NEXT:    vslidedown.vx v16, v8, a1
 ; CHECK-NEXT:    vmv.s.x v16, a0
-; CHECK-NEXT:    vsetvli a0, zero, e16,m8,tu,mu
+; CHECK-NEXT:    addi a0, a1, 1
+; CHECK-NEXT:    vsetvli a0, a0, e16,m8,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v16, a1
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 32 x i16> %v, i16 %elt, i32 %idx
   ret <vscale x 32 x i16> %r
 }
 
 define <vscale x 1 x i32> @insertelt_nxv1i32_0(<vscale x 1 x i32> %v, i32 %elt) {
 ; CHECK-LABEL: insertelt_nxv1i32_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e32,mf2,ta,mu
 ; CHECK-NEXT:    vmv.s.x v8, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 1 x i32> %v, i32 %elt, i32 0
   ret <vscale x 1 x i32> %r
 }
 
 define <vscale x 1 x i32> @insertelt_nxv1i32_imm(<vscale x 1 x i32> %v, i32 %elt) {
 ; CHECK-LABEL: insertelt_nxv1i32_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e32,mf2,ta,mu
-; CHECK-NEXT:    vslidedown.vi v25, v8, 3
 ; CHECK-NEXT:    vmv.s.x v25, a0
-; CHECK-NEXT:    vsetvli a0, zero, e32,mf2,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e32,mf2,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v25, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 1 x i32> %v, i32 %elt, i32 3
   ret <vscale x 1 x i32> %r
 }
 
 define <vscale x 1 x i32> @insertelt_nxv1i32_idx(<vscale x 1 x i32> %v, i32 %elt, i32 %idx) {
 ; CHECK-LABEL: insertelt_nxv1i32_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a2, zero, e32,mf2,ta,mu
-; CHECK-NEXT:    vslidedown.vx v25, v8, a1
 ; CHECK-NEXT:    vmv.s.x v25, a0
-; CHECK-NEXT:    vsetvli a0, zero, e32,mf2,tu,mu
+; CHECK-NEXT:    addi a0, a1, 1
+; CHECK-NEXT:    vsetvli a0, a0, e32,mf2,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v25, a1
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 1 x i32> %v, i32 %elt, i32 %idx
   ret <vscale x 1 x i32> %r
 }
 
 define <vscale x 2 x i32> @insertelt_nxv2i32_0(<vscale x 2 x i32> %v, i32 %elt) {
 ; CHECK-LABEL: insertelt_nxv2i32_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e32,m1,ta,mu
 ; CHECK-NEXT:    vmv.s.x v8, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 2 x i32> %v, i32 %elt, i32 0
   ret <vscale x 2 x i32> %r
 }
 
 define <vscale x 2 x i32> @insertelt_nxv2i32_imm(<vscale x 2 x i32> %v, i32 %elt) {
 ; CHECK-LABEL: insertelt_nxv2i32_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e32,m1,ta,mu
-; CHECK-NEXT:    vslidedown.vi v25, v8, 3
 ; CHECK-NEXT:    vmv.s.x v25, a0
-; CHECK-NEXT:    vsetvli a0, zero, e32,m1,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e32,m1,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v25, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 2 x i32> %v, i32 %elt, i32 3
   ret <vscale x 2 x i32> %r
 }
 
 define <vscale x 2 x i32> @insertelt_nxv2i32_idx(<vscale x 2 x i32> %v, i32 %elt, i32 %idx) {
 ; CHECK-LABEL: insertelt_nxv2i32_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a2, zero, e32,m1,ta,mu
-; CHECK-NEXT:    vslidedown.vx v25, v8, a1
 ; CHECK-NEXT:    vmv.s.x v25, a0
-; CHECK-NEXT:    vsetvli a0, zero, e32,m1,tu,mu
+; CHECK-NEXT:    addi a0, a1, 1
+; CHECK-NEXT:    vsetvli a0, a0, e32,m1,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v25, a1
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 2 x i32> %v, i32 %elt, i32 %idx
   ret <vscale x 2 x i32> %r
 }
 
 define <vscale x 4 x i32> @insertelt_nxv4i32_0(<vscale x 4 x i32> %v, i32 %elt) {
 ; CHECK-LABEL: insertelt_nxv4i32_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e32,m2,ta,mu
 ; CHECK-NEXT:    vmv.s.x v8, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 4 x i32> %v, i32 %elt, i32 0
   ret <vscale x 4 x i32> %r
 }
 
 define <vscale x 4 x i32> @insertelt_nxv4i32_imm(<vscale x 4 x i32> %v, i32 %elt) {
 ; CHECK-LABEL: insertelt_nxv4i32_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e32,m2,ta,mu
-; CHECK-NEXT:    vslidedown.vi v26, v8, 3
 ; CHECK-NEXT:    vmv.s.x v26, a0
-; CHECK-NEXT:    vsetvli a0, zero, e32,m2,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e32,m2,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v26, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 4 x i32> %v, i32 %elt, i32 3
   ret <vscale x 4 x i32> %r
 }
 
 define <vscale x 4 x i32> @insertelt_nxv4i32_idx(<vscale x 4 x i32> %v, i32 %elt, i32 %idx) {
 ; CHECK-LABEL: insertelt_nxv4i32_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a2, zero, e32,m2,ta,mu
-; CHECK-NEXT:    vslidedown.vx v26, v8, a1
 ; CHECK-NEXT:    vmv.s.x v26, a0
-; CHECK-NEXT:    vsetvli a0, zero, e32,m2,tu,mu
+; CHECK-NEXT:    addi a0, a1, 1
+; CHECK-NEXT:    vsetvli a0, a0, e32,m2,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v26, a1
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 4 x i32> %v, i32 %elt, i32 %idx
   ret <vscale x 4 x i32> %r
 }
 
 define <vscale x 8 x i32> @insertelt_nxv8i32_0(<vscale x 8 x i32> %v, i32 %elt) {
 ; CHECK-LABEL: insertelt_nxv8i32_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e32,m4,ta,mu
 ; CHECK-NEXT:    vmv.s.x v8, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 8 x i32> %v, i32 %elt, i32 0
   ret <vscale x 8 x i32> %r
 }
 
 define <vscale x 8 x i32> @insertelt_nxv8i32_imm(<vscale x 8 x i32> %v, i32 %elt) {
 ; CHECK-LABEL: insertelt_nxv8i32_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e32,m4,ta,mu
-; CHECK-NEXT:    vslidedown.vi v28, v8, 3
 ; CHECK-NEXT:    vmv.s.x v28, a0
-; CHECK-NEXT:    vsetvli a0, zero, e32,m4,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e32,m4,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v28, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 8 x i32> %v, i32 %elt, i32 3
   ret <vscale x 8 x i32> %r
 }
 
 define <vscale x 8 x i32> @insertelt_nxv8i32_idx(<vscale x 8 x i32> %v, i32 %elt, i32 %idx) {
 ; CHECK-LABEL: insertelt_nxv8i32_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a2, zero, e32,m4,ta,mu
-; CHECK-NEXT:    vslidedown.vx v28, v8, a1
 ; CHECK-NEXT:    vmv.s.x v28, a0
-; CHECK-NEXT:    vsetvli a0, zero, e32,m4,tu,mu
+; CHECK-NEXT:    addi a0, a1, 1
+; CHECK-NEXT:    vsetvli a0, a0, e32,m4,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v28, a1
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 8 x i32> %v, i32 %elt, i32 %idx
   ret <vscale x 8 x i32> %r
 }
 
 define <vscale x 16 x i32> @insertelt_nxv16i32_0(<vscale x 16 x i32> %v, i32 %elt) {
 ; CHECK-LABEL: insertelt_nxv16i32_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e32,m8,ta,mu
 ; CHECK-NEXT:    vmv.s.x v8, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 16 x i32> %v, i32 %elt, i32 0
   ret <vscale x 16 x i32> %r
 }
 
 define <vscale x 16 x i32> @insertelt_nxv16i32_imm(<vscale x 16 x i32> %v, i32 %elt) {
 ; CHECK-LABEL: insertelt_nxv16i32_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e32,m8,ta,mu
-; CHECK-NEXT:    vslidedown.vi v16, v8, 3
 ; CHECK-NEXT:    vmv.s.x v16, a0
-; CHECK-NEXT:    vsetvli a0, zero, e32,m8,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e32,m8,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v16, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 16 x i32> %v, i32 %elt, i32 3
   ret <vscale x 16 x i32> %r
 }
 
 define <vscale x 16 x i32> @insertelt_nxv16i32_idx(<vscale x 16 x i32> %v, i32 %elt, i32 %idx) {
 ; CHECK-LABEL: insertelt_nxv16i32_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a2, zero, e32,m8,ta,mu
-; CHECK-NEXT:    vslidedown.vx v16, v8, a1
 ; CHECK-NEXT:    vmv.s.x v16, a0
-; CHECK-NEXT:    vsetvli a0, zero, e32,m8,tu,mu
+; CHECK-NEXT:    addi a0, a1, 1
+; CHECK-NEXT:    vsetvli a0, a0, e32,m8,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v16, a1
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 16 x i32> %v, i32 %elt, i32 %idx
   ret <vscale x 16 x i32> %r
 }
 
 define <vscale x 1 x i64> @insertelt_nxv1i64_0(<vscale x 1 x i64> %v, i64 %elt) {
 ; CHECK-LABEL: insertelt_nxv1i64_0:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a2, zero, e64,m1,ta,mu
-; CHECK-NEXT:    vmv.v.x v25, a1
-; CHECK-NEXT:    addi a1, zero, 32
-; CHECK-NEXT:    vsll.vx v25, v25, a1
-; CHECK-NEXT:    vmv.v.x v26, a0
-; CHECK-NEXT:    vsll.vx v26, v26, a1
-; CHECK-NEXT:    vsrl.vx v26, v26, a1
-; CHECK-NEXT:    vor.vv v25, v26, v25
-; CHECK-NEXT:    vid.v v26
-; CHECK-NEXT:    vmseq.vi v0, v26, 0
-; CHECK-NEXT:    vmerge.vvm v8, v8, v25, v0
+; CHECK-NEXT:    vsetivli a2, 2, e32,m1,ta,mu
+; CHECK-NEXT:    vmv.v.i v25, 0
+; CHECK-NEXT:    vslide1up.vx v26, v25, a1
+; CHECK-NEXT:    vslide1up.vx v25, v26, a0
+; CHECK-NEXT:    vsetivli a0, 1, e64,m1,tu,mu
+; CHECK-NEXT:    vslideup.vi v8, v25, 0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 1 x i64> %v, i64 %elt, i32 0
   ret <vscale x 1 x i64> %r
 }
 
 define <vscale x 1 x i64> @insertelt_nxv1i64_imm(<vscale x 1 x i64> %v, i64 %elt) {
 ; CHECK-LABEL: insertelt_nxv1i64_imm:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a2, zero, e64,m1,ta,mu
-; CHECK-NEXT:    vmv.v.x v25, a1
-; CHECK-NEXT:    addi a1, zero, 32
-; CHECK-NEXT:    vsll.vx v25, v25, a1
-; CHECK-NEXT:    vmv.v.x v26, a0
-; CHECK-NEXT:    vsll.vx v26, v26, a1
-; CHECK-NEXT:    vsrl.vx v26, v26, a1
-; CHECK-NEXT:    vor.vv v25, v26, v25
-; CHECK-NEXT:    vid.v v26
-; CHECK-NEXT:    vmseq.vi v0, v26, 3
-; CHECK-NEXT:    vmerge.vvm v8, v8, v25, v0
+; CHECK-NEXT:    vsetivli a2, 2, e32,m1,ta,mu
+; CHECK-NEXT:    vmv.v.i v25, 0
+; CHECK-NEXT:    vslide1up.vx v26, v25, a1
+; CHECK-NEXT:    vslide1up.vx v25, v26, a0
+; CHECK-NEXT:    vsetivli a0, 4, e64,m1,tu,mu
+; CHECK-NEXT:    vslideup.vi v8, v25, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 1 x i64> %v, i64 %elt, i32 3
   ret <vscale x 1 x i64> %r
 }
 
 define <vscale x 1 x i64> @insertelt_nxv1i64_idx(<vscale x 1 x i64> %v, i64 %elt, i32 %idx) {
 ; CHECK-LABEL: insertelt_nxv1i64_idx:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a3, zero, e64,m1,ta,mu
-; CHECK-NEXT:    vmv.v.x v25, a1
-; CHECK-NEXT:    addi a1, zero, 32
-; CHECK-NEXT:    vsll.vx v25, v25, a1
-; CHECK-NEXT:    vmv.v.x v26, a0
-; CHECK-NEXT:    vsll.vx v26, v26, a1
-; CHECK-NEXT:    vsrl.vx v26, v26, a1
-; CHECK-NEXT:    vor.vv v25, v26, v25
-; CHECK-NEXT:    vid.v v26
-; CHECK-NEXT:    vmseq.vx v0, v26, a2
-; CHECK-NEXT:    vmerge.vvm v8, v8, v25, v0
+; CHECK-NEXT:    vsetivli a3, 2, e32,m1,ta,mu
+; CHECK-NEXT:    vmv.v.i v25, 0
+; CHECK-NEXT:    vslide1up.vx v26, v25, a1
+; CHECK-NEXT:    vslide1up.vx v25, v26, a0
+; CHECK-NEXT:    addi a0, a2, 1
+; CHECK-NEXT:    vsetvli a0, a0, e64,m1,tu,mu
+; CHECK-NEXT:    vslideup.vx v8, v25, a2
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 1 x i64> %v, i64 %elt, i32 %idx
   ret <vscale x 1 x i64> %r
 }
 
 define <vscale x 2 x i64> @insertelt_nxv2i64_0(<vscale x 2 x i64> %v, i64 %elt) {
 ; CHECK-LABEL: insertelt_nxv2i64_0:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a2, zero, e64,m2,ta,mu
-; CHECK-NEXT:    vmv.v.x v26, a1
-; CHECK-NEXT:    addi a1, zero, 32
-; CHECK-NEXT:    vsll.vx v26, v26, a1
-; CHECK-NEXT:    vmv.v.x v28, a0
-; CHECK-NEXT:    vsll.vx v28, v28, a1
-; CHECK-NEXT:    vsrl.vx v28, v28, a1
-; CHECK-NEXT:    vor.vv v26, v28, v26
-; CHECK-NEXT:    vid.v v28
-; CHECK-NEXT:    vmseq.vi v0, v28, 0
-; CHECK-NEXT:    vmerge.vvm v8, v8, v26, v0
+; CHECK-NEXT:    vsetivli a2, 2, e32,m2,ta,mu
+; CHECK-NEXT:    vmv.v.i v26, 0
+; CHECK-NEXT:    vslide1up.vx v28, v26, a1
+; CHECK-NEXT:    vslide1up.vx v26, v28, a0
+; CHECK-NEXT:    vsetivli a0, 1, e64,m2,tu,mu
+; CHECK-NEXT:    vslideup.vi v8, v26, 0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 2 x i64> %v, i64 %elt, i32 0
   ret <vscale x 2 x i64> %r
 }
 
 define <vscale x 2 x i64> @insertelt_nxv2i64_imm(<vscale x 2 x i64> %v, i64 %elt) {
 ; CHECK-LABEL: insertelt_nxv2i64_imm:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a2, zero, e64,m2,ta,mu
-; CHECK-NEXT:    vmv.v.x v26, a1
-; CHECK-NEXT:    addi a1, zero, 32
-; CHECK-NEXT:    vsll.vx v26, v26, a1
-; CHECK-NEXT:    vmv.v.x v28, a0
-; CHECK-NEXT:    vsll.vx v28, v28, a1
-; CHECK-NEXT:    vsrl.vx v28, v28, a1
-; CHECK-NEXT:    vor.vv v26, v28, v26
-; CHECK-NEXT:    vid.v v28
-; CHECK-NEXT:    vmseq.vi v0, v28, 3
-; CHECK-NEXT:    vmerge.vvm v8, v8, v26, v0
+; CHECK-NEXT:    vsetivli a2, 2, e32,m2,ta,mu
+; CHECK-NEXT:    vmv.v.i v26, 0
+; CHECK-NEXT:    vslide1up.vx v28, v26, a1
+; CHECK-NEXT:    vslide1up.vx v26, v28, a0
+; CHECK-NEXT:    vsetivli a0, 4, e64,m2,tu,mu
+; CHECK-NEXT:    vslideup.vi v8, v26, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 2 x i64> %v, i64 %elt, i32 3
   ret <vscale x 2 x i64> %r
 }
 
 define <vscale x 2 x i64> @insertelt_nxv2i64_idx(<vscale x 2 x i64> %v, i64 %elt, i32 %idx) {
 ; CHECK-LABEL: insertelt_nxv2i64_idx:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a3, zero, e64,m2,ta,mu
-; CHECK-NEXT:    vmv.v.x v26, a1
-; CHECK-NEXT:    addi a1, zero, 32
-; CHECK-NEXT:    vsll.vx v26, v26, a1
-; CHECK-NEXT:    vmv.v.x v28, a0
-; CHECK-NEXT:    vsll.vx v28, v28, a1
-; CHECK-NEXT:    vsrl.vx v28, v28, a1
-; CHECK-NEXT:    vor.vv v26, v28, v26
-; CHECK-NEXT:    vid.v v28
-; CHECK-NEXT:    vmseq.vx v0, v28, a2
-; CHECK-NEXT:    vmerge.vvm v8, v8, v26, v0
+; CHECK-NEXT:    vsetivli a3, 2, e32,m2,ta,mu
+; CHECK-NEXT:    vmv.v.i v26, 0
+; CHECK-NEXT:    vslide1up.vx v28, v26, a1
+; CHECK-NEXT:    vslide1up.vx v26, v28, a0
+; CHECK-NEXT:    addi a0, a2, 1
+; CHECK-NEXT:    vsetvli a0, a0, e64,m2,tu,mu
+; CHECK-NEXT:    vslideup.vx v8, v26, a2
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 2 x i64> %v, i64 %elt, i32 %idx
   ret <vscale x 2 x i64> %r
 }
 
 define <vscale x 4 x i64> @insertelt_nxv4i64_0(<vscale x 4 x i64> %v, i64 %elt) {
 ; CHECK-LABEL: insertelt_nxv4i64_0:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a2, zero, e64,m4,ta,mu
-; CHECK-NEXT:    vmv.v.x v28, a1
-; CHECK-NEXT:    addi a1, zero, 32
-; CHECK-NEXT:    vsll.vx v28, v28, a1
-; CHECK-NEXT:    vmv.v.x v12, a0
-; CHECK-NEXT:    vsll.vx v12, v12, a1
-; CHECK-NEXT:    vsrl.vx v12, v12, a1
-; CHECK-NEXT:    vor.vv v28, v12, v28
-; CHECK-NEXT:    vid.v v12
-; CHECK-NEXT:    vmseq.vi v0, v12, 0
-; CHECK-NEXT:    vmerge.vvm v8, v8, v28, v0
+; CHECK-NEXT:    vsetivli a2, 2, e32,m4,ta,mu
+; CHECK-NEXT:    vmv.v.i v28, 0
+; CHECK-NEXT:    vslide1up.vx v12, v28, a1
+; CHECK-NEXT:    vslide1up.vx v28, v12, a0
+; CHECK-NEXT:    vsetivli a0, 1, e64,m4,tu,mu
+; CHECK-NEXT:    vslideup.vi v8, v28, 0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 4 x i64> %v, i64 %elt, i32 0
   ret <vscale x 4 x i64> %r
 }
 
 define <vscale x 4 x i64> @insertelt_nxv4i64_imm(<vscale x 4 x i64> %v, i64 %elt) {
 ; CHECK-LABEL: insertelt_nxv4i64_imm:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a2, zero, e64,m4,ta,mu
-; CHECK-NEXT:    vmv.v.x v28, a1
-; CHECK-NEXT:    addi a1, zero, 32
-; CHECK-NEXT:    vsll.vx v28, v28, a1
-; CHECK-NEXT:    vmv.v.x v12, a0
-; CHECK-NEXT:    vsll.vx v12, v12, a1
-; CHECK-NEXT:    vsrl.vx v12, v12, a1
-; CHECK-NEXT:    vor.vv v28, v12, v28
-; CHECK-NEXT:    vid.v v12
-; CHECK-NEXT:    vmseq.vi v0, v12, 3
-; CHECK-NEXT:    vmerge.vvm v8, v8, v28, v0
+; CHECK-NEXT:    vsetivli a2, 2, e32,m4,ta,mu
+; CHECK-NEXT:    vmv.v.i v28, 0
+; CHECK-NEXT:    vslide1up.vx v12, v28, a1
+; CHECK-NEXT:    vslide1up.vx v28, v12, a0
+; CHECK-NEXT:    vsetivli a0, 4, e64,m4,tu,mu
+; CHECK-NEXT:    vslideup.vi v8, v28, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 4 x i64> %v, i64 %elt, i32 3
   ret <vscale x 4 x i64> %r
 }
 
 define <vscale x 4 x i64> @insertelt_nxv4i64_idx(<vscale x 4 x i64> %v, i64 %elt, i32 %idx) {
 ; CHECK-LABEL: insertelt_nxv4i64_idx:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a3, zero, e64,m4,ta,mu
-; CHECK-NEXT:    vmv.v.x v28, a1
-; CHECK-NEXT:    addi a1, zero, 32
-; CHECK-NEXT:    vsll.vx v28, v28, a1
-; CHECK-NEXT:    vmv.v.x v12, a0
-; CHECK-NEXT:    vsll.vx v12, v12, a1
-; CHECK-NEXT:    vsrl.vx v12, v12, a1
-; CHECK-NEXT:    vor.vv v28, v12, v28
-; CHECK-NEXT:    vid.v v12
-; CHECK-NEXT:    vmseq.vx v0, v12, a2
-; CHECK-NEXT:    vmerge.vvm v8, v8, v28, v0
+; CHECK-NEXT:    vsetivli a3, 2, e32,m4,ta,mu
+; CHECK-NEXT:    vmv.v.i v28, 0
+; CHECK-NEXT:    vslide1up.vx v12, v28, a1
+; CHECK-NEXT:    vslide1up.vx v28, v12, a0
+; CHECK-NEXT:    addi a0, a2, 1
+; CHECK-NEXT:    vsetvli a0, a0, e64,m4,tu,mu
+; CHECK-NEXT:    vslideup.vx v8, v28, a2
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 4 x i64> %v, i64 %elt, i32 %idx
   ret <vscale x 4 x i64> %r
 }
 
 define <vscale x 8 x i64> @insertelt_nxv8i64_0(<vscale x 8 x i64> %v, i64 %elt) {
 ; CHECK-LABEL: insertelt_nxv8i64_0:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a2, zero, e64,m8,ta,mu
-; CHECK-NEXT:    vmv.v.x v16, a1
-; CHECK-NEXT:    addi a1, zero, 32
-; CHECK-NEXT:    vsll.vx v16, v16, a1
-; CHECK-NEXT:    vmv.v.x v24, a0
-; CHECK-NEXT:    vsll.vx v24, v24, a1
-; CHECK-NEXT:    vsrl.vx v24, v24, a1
-; CHECK-NEXT:    vor.vv v16, v24, v16
-; CHECK-NEXT:    vid.v v24
-; CHECK-NEXT:    vmseq.vi v0, v24, 0
-; CHECK-NEXT:    vmerge.vvm v8, v8, v16, v0
+; CHECK-NEXT:    vsetivli a2, 2, e32,m8,ta,mu
+; CHECK-NEXT:    vmv.v.i v16, 0
+; CHECK-NEXT:    vslide1up.vx v24, v16, a1
+; CHECK-NEXT:    vslide1up.vx v16, v24, a0
+; CHECK-NEXT:    vsetivli a0, 1, e64,m8,tu,mu
+; CHECK-NEXT:    vslideup.vi v8, v16, 0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 8 x i64> %v, i64 %elt, i32 0
   ret <vscale x 8 x i64> %r
 }
 
 define <vscale x 8 x i64> @insertelt_nxv8i64_imm(<vscale x 8 x i64> %v, i64 %elt) {
 ; CHECK-LABEL: insertelt_nxv8i64_imm:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a2, zero, e64,m8,ta,mu
-; CHECK-NEXT:    vmv.v.x v16, a1
-; CHECK-NEXT:    addi a1, zero, 32
-; CHECK-NEXT:    vsll.vx v16, v16, a1
-; CHECK-NEXT:    vmv.v.x v24, a0
-; CHECK-NEXT:    vsll.vx v24, v24, a1
-; CHECK-NEXT:    vsrl.vx v24, v24, a1
-; CHECK-NEXT:    vor.vv v16, v24, v16
-; CHECK-NEXT:    vid.v v24
-; CHECK-NEXT:    vmseq.vi v0, v24, 3
-; CHECK-NEXT:    vmerge.vvm v8, v8, v16, v0
+; CHECK-NEXT:    vsetivli a2, 2, e32,m8,ta,mu
+; CHECK-NEXT:    vmv.v.i v16, 0
+; CHECK-NEXT:    vslide1up.vx v24, v16, a1
+; CHECK-NEXT:    vslide1up.vx v16, v24, a0
+; CHECK-NEXT:    vsetivli a0, 4, e64,m8,tu,mu
+; CHECK-NEXT:    vslideup.vi v8, v16, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 8 x i64> %v, i64 %elt, i32 3
   ret <vscale x 8 x i64> %r
 }
 
 define <vscale x 8 x i64> @insertelt_nxv8i64_idx(<vscale x 8 x i64> %v, i64 %elt, i32 %idx) {
 ; CHECK-LABEL: insertelt_nxv8i64_idx:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a3, zero, e64,m8,ta,mu
-; CHECK-NEXT:    vmv.v.x v16, a1
-; CHECK-NEXT:    addi a1, zero, 32
-; CHECK-NEXT:    vsll.vx v16, v16, a1
-; CHECK-NEXT:    vmv.v.x v24, a0
-; CHECK-NEXT:    vsll.vx v24, v24, a1
-; CHECK-NEXT:    vsrl.vx v24, v24, a1
-; CHECK-NEXT:    vor.vv v16, v24, v16
-; CHECK-NEXT:    vid.v v24
-; CHECK-NEXT:    vmseq.vx v0, v24, a2
-; CHECK-NEXT:    vmerge.vvm v8, v8, v16, v0
+; CHECK-NEXT:    vsetivli a3, 2, e32,m8,ta,mu
+; CHECK-NEXT:    vmv.v.i v16, 0
+; CHECK-NEXT:    vslide1up.vx v24, v16, a1
+; CHECK-NEXT:    vslide1up.vx v16, v24, a0
+; CHECK-NEXT:    addi a0, a2, 1
+; CHECK-NEXT:    vsetvli a0, a0, e64,m8,tu,mu
+; CHECK-NEXT:    vslideup.vx v8, v16, a2
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 8 x i64> %v, i64 %elt, i32 %idx
   ret <vscale x 8 x i64> %r
 }
 
 ; Extra tests to check lowering of constant values
 define <vscale x 2 x i64> @insertelt_nxv2i64_0_c10(<vscale x 2 x i64> %v) {
 ; CHECK-LABEL: insertelt_nxv2i64_0_c10:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a0, zero, e64,m2,ta,mu
-; CHECK-NEXT:    vid.v v26
-; CHECK-NEXT:    vmseq.vi v0, v26, 0
-; CHECK-NEXT:    vmerge.vim v8, v8, 10, v0
+; CHECK-NEXT:    addi a0, zero, 10
+; CHECK-NEXT:    vsetvli a1, zero, e64,m2,ta,mu
+; CHECK-NEXT:    vmv.s.x v8, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 2 x i64> %v, i64 10, i32 0
   ret <vscale x 2 x i64> %r
 }
 
 define <vscale x 2 x i64> @insertelt_nxv2i64_imm_c10(<vscale x 2 x i64> %v) {
 ; CHECK-LABEL: insertelt_nxv2i64_imm_c10:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a0, zero, e64,m2,ta,mu
-; CHECK-NEXT:    vid.v v26
-; CHECK-NEXT:    vmseq.vi v0, v26, 3
-; CHECK-NEXT:    vmerge.vim v8, v8, 10, v0
+; CHECK-NEXT:    addi a0, zero, 10
+; CHECK-NEXT:    vsetvli a1, zero, e64,m2,ta,mu
+; CHECK-NEXT:    vmv.s.x v26, a0
+; CHECK-NEXT:    vsetivli a0, 4, e64,m2,tu,mu
+; CHECK-NEXT:    vslideup.vi v8, v26, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 2 x i64> %v, i64 10, i32 3
   ret <vscale x 2 x i64> %r
 }
 
 define <vscale x 2 x i64> @insertelt_nxv2i64_idx_c10(<vscale x 2 x i64> %v, i32 %idx) {
 ; CHECK-LABEL: insertelt_nxv2i64_idx_c10:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e64,m2,ta,mu
-; CHECK-NEXT:    vid.v v26
-; CHECK-NEXT:    vmseq.vx v0, v26, a0
-; CHECK-NEXT:    vmerge.vim v8, v8, 10, v0
+; CHECK-NEXT:    addi a1, zero, 10
+; CHECK-NEXT:    vsetvli a2, zero, e64,m2,ta,mu
+; CHECK-NEXT:    vmv.s.x v26, a1
+; CHECK-NEXT:    addi a1, a0, 1
+; CHECK-NEXT:    vsetvli a1, a1, e64,m2,tu,mu
+; CHECK-NEXT:    vslideup.vx v8, v26, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 2 x i64> %v, i64 10, i32 %idx
   ret <vscale x 2 x i64> %r
 }
 
 define <vscale x 2 x i64> @insertelt_nxv2i64_0_cn1(<vscale x 2 x i64> %v) {
 ; CHECK-LABEL: insertelt_nxv2i64_0_cn1:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a0, zero, e64,m2,ta,mu
-; CHECK-NEXT:    vid.v v26
-; CHECK-NEXT:    vmseq.vi v0, v26, 0
-; CHECK-NEXT:    vmerge.vim v8, v8, -1, v0
+; CHECK-NEXT:    addi a0, zero, -1
+; CHECK-NEXT:    vsetvli a1, zero, e64,m2,ta,mu
+; CHECK-NEXT:    vmv.s.x v8, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 2 x i64> %v, i64 -1, i32 0
   ret <vscale x 2 x i64> %r
 }
 
 define <vscale x 2 x i64> @insertelt_nxv2i64_imm_cn1(<vscale x 2 x i64> %v) {
 ; CHECK-LABEL: insertelt_nxv2i64_imm_cn1:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a0, zero, e64,m2,ta,mu
-; CHECK-NEXT:    vid.v v26
-; CHECK-NEXT:    vmseq.vi v0, v26, 3
-; CHECK-NEXT:    vmerge.vim v8, v8, -1, v0
+; CHECK-NEXT:    addi a0, zero, -1
+; CHECK-NEXT:    vsetvli a1, zero, e64,m2,ta,mu
+; CHECK-NEXT:    vmv.s.x v26, a0
+; CHECK-NEXT:    vsetivli a0, 4, e64,m2,tu,mu
+; CHECK-NEXT:    vslideup.vi v8, v26, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 2 x i64> %v, i64 -1, i32 3
   ret <vscale x 2 x i64> %r
 }
 
 define <vscale x 2 x i64> @insertelt_nxv2i64_idx_cn1(<vscale x 2 x i64> %v, i32 %idx) {
 ; CHECK-LABEL: insertelt_nxv2i64_idx_cn1:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e64,m2,ta,mu
-; CHECK-NEXT:    vid.v v26
-; CHECK-NEXT:    vmseq.vx v0, v26, a0
-; CHECK-NEXT:    vmerge.vim v8, v8, -1, v0
+; CHECK-NEXT:    addi a1, zero, -1
+; CHECK-NEXT:    vsetvli a2, zero, e64,m2,ta,mu
+; CHECK-NEXT:    vmv.s.x v26, a1
+; CHECK-NEXT:    addi a1, a0, 1
+; CHECK-NEXT:    vsetvli a1, a1, e64,m2,tu,mu
+; CHECK-NEXT:    vslideup.vx v8, v26, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 2 x i64> %v, i64 -1, i32 %idx
   ret <vscale x 2 x i64> %r
 }

llvm/test/CodeGen/RISCV/rvv/insertelt-int-rv64.ll

; CHECK-NEXT:    ret
   %r = insertelement <vscale x 1 x i8> %v, i8 %elt, i32 0
   ret <vscale x 1 x i8> %r
 }
 
 define <vscale x 1 x i8> @insertelt_nxv1i8_imm(<vscale x 1 x i8> %v, i8 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv1i8_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e8,mf8,ta,mu
-; CHECK-NEXT:    vslidedown.vi v25, v8, 3
 ; CHECK-NEXT:    vmv.s.x v25, a0
-; CHECK-NEXT:    vsetvli a0, zero, e8,mf8,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e8,mf8,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v25, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 1 x i8> %v, i8 %elt, i32 3
   ret <vscale x 1 x i8> %r
 }
 
 define <vscale x 1 x i8> @insertelt_nxv1i8_idx(<vscale x 1 x i8> %v, i8 signext %elt, i32 signext %idx) {
 ; CHECK-LABEL: insertelt_nxv1i8_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a2, zero, e8,mf8,ta,mu
-; CHECK-NEXT:    vslidedown.vx v25, v8, a1
 ; CHECK-NEXT:    vmv.s.x v25, a0
-; CHECK-NEXT:    vsetvli a0, zero, e8,mf8,tu,mu
+; CHECK-NEXT:    addi a0, a1, 1
+; CHECK-NEXT:    vsetvli a0, a0, e8,mf8,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v25, a1
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 1 x i8> %v, i8 %elt, i32 %idx
   ret <vscale x 1 x i8> %r
 }
 
 define <vscale x 2 x i8> @insertelt_nxv2i8_0(<vscale x 2 x i8> %v, i8 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv2i8_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e8,mf4,ta,mu
 ; CHECK-NEXT:    vmv.s.x v8, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 2 x i8> %v, i8 %elt, i32 0
   ret <vscale x 2 x i8> %r
 }
 
 define <vscale x 2 x i8> @insertelt_nxv2i8_imm(<vscale x 2 x i8> %v, i8 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv2i8_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e8,mf4,ta,mu
-; CHECK-NEXT:    vslidedown.vi v25, v8, 3
 ; CHECK-NEXT:    vmv.s.x v25, a0
-; CHECK-NEXT:    vsetvli a0, zero, e8,mf4,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e8,mf4,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v25, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 2 x i8> %v, i8 %elt, i32 3
   ret <vscale x 2 x i8> %r
 }
 
 define <vscale x 2 x i8> @insertelt_nxv2i8_idx(<vscale x 2 x i8> %v, i8 signext %elt, i32 signext %idx) {
 ; CHECK-LABEL: insertelt_nxv2i8_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a2, zero, e8,mf4,ta,mu
-; CHECK-NEXT:    vslidedown.vx v25, v8, a1
 ; CHECK-NEXT:    vmv.s.x v25, a0
-; CHECK-NEXT:    vsetvli a0, zero, e8,mf4,tu,mu
+; CHECK-NEXT:    addi a0, a1, 1
+; CHECK-NEXT:    vsetvli a0, a0, e8,mf4,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v25, a1
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 2 x i8> %v, i8 %elt, i32 %idx
   ret <vscale x 2 x i8> %r
 }
 
 define <vscale x 4 x i8> @insertelt_nxv4i8_0(<vscale x 4 x i8> %v, i8 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv4i8_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e8,mf2,ta,mu
 ; CHECK-NEXT:    vmv.s.x v8, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 4 x i8> %v, i8 %elt, i32 0
   ret <vscale x 4 x i8> %r
 }
 
 define <vscale x 4 x i8> @insertelt_nxv4i8_imm(<vscale x 4 x i8> %v, i8 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv4i8_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e8,mf2,ta,mu
-; CHECK-NEXT:    vslidedown.vi v25, v8, 3
 ; CHECK-NEXT:    vmv.s.x v25, a0
-; CHECK-NEXT:    vsetvli a0, zero, e8,mf2,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e8,mf2,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v25, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 4 x i8> %v, i8 %elt, i32 3
   ret <vscale x 4 x i8> %r
 }
 
 define <vscale x 4 x i8> @insertelt_nxv4i8_idx(<vscale x 4 x i8> %v, i8 signext %elt, i32 signext %idx) {
 ; CHECK-LABEL: insertelt_nxv4i8_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a2, zero, e8,mf2,ta,mu
-; CHECK-NEXT:    vslidedown.vx v25, v8, a1
 ; CHECK-NEXT:    vmv.s.x v25, a0
-; CHECK-NEXT:    vsetvli a0, zero, e8,mf2,tu,mu
+; CHECK-NEXT:    addi a0, a1, 1
+; CHECK-NEXT:    vsetvli a0, a0, e8,mf2,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v25, a1
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 4 x i8> %v, i8 %elt, i32 %idx
   ret <vscale x 4 x i8> %r
 }
 
 define <vscale x 8 x i8> @insertelt_nxv8i8_0(<vscale x 8 x i8> %v, i8 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv8i8_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e8,m1,ta,mu
 ; CHECK-NEXT:    vmv.s.x v8, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 8 x i8> %v, i8 %elt, i32 0
   ret <vscale x 8 x i8> %r
 }
 
 define <vscale x 8 x i8> @insertelt_nxv8i8_imm(<vscale x 8 x i8> %v, i8 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv8i8_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e8,m1,ta,mu
-; CHECK-NEXT:    vslidedown.vi v25, v8, 3
 ; CHECK-NEXT:    vmv.s.x v25, a0
-; CHECK-NEXT:    vsetvli a0, zero, e8,m1,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e8,m1,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v25, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 8 x i8> %v, i8 %elt, i32 3
   ret <vscale x 8 x i8> %r
 }
 
 define <vscale x 8 x i8> @insertelt_nxv8i8_idx(<vscale x 8 x i8> %v, i8 signext %elt, i32 signext %idx) {
 ; CHECK-LABEL: insertelt_nxv8i8_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a2, zero, e8,m1,ta,mu
-; CHECK-NEXT:    vslidedown.vx v25, v8, a1
 ; CHECK-NEXT:    vmv.s.x v25, a0
-; CHECK-NEXT:    vsetvli a0, zero, e8,m1,tu,mu
+; CHECK-NEXT:    addi a0, a1, 1
+; CHECK-NEXT:    vsetvli a0, a0, e8,m1,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v25, a1
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 8 x i8> %v, i8 %elt, i32 %idx
   ret <vscale x 8 x i8> %r
 }
 
 define <vscale x 16 x i8> @insertelt_nxv16i8_0(<vscale x 16 x i8> %v, i8 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv16i8_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e8,m2,ta,mu
 ; CHECK-NEXT:    vmv.s.x v8, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 16 x i8> %v, i8 %elt, i32 0
   ret <vscale x 16 x i8> %r
 }
 
 define <vscale x 16 x i8> @insertelt_nxv16i8_imm(<vscale x 16 x i8> %v, i8 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv16i8_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e8,m2,ta,mu
-; CHECK-NEXT:    vslidedown.vi v26, v8, 3
 ; CHECK-NEXT:    vmv.s.x v26, a0
-; CHECK-NEXT:    vsetvli a0, zero, e8,m2,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e8,m2,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v26, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 16 x i8> %v, i8 %elt, i32 3
   ret <vscale x 16 x i8> %r
 }
 
 define <vscale x 16 x i8> @insertelt_nxv16i8_idx(<vscale x 16 x i8> %v, i8 signext %elt, i32 signext %idx) {
 ; CHECK-LABEL: insertelt_nxv16i8_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a2, zero, e8,m2,ta,mu
-; CHECK-NEXT:    vslidedown.vx v26, v8, a1
 ; CHECK-NEXT:    vmv.s.x v26, a0
-; CHECK-NEXT:    vsetvli a0, zero, e8,m2,tu,mu
+; CHECK-NEXT:    addi a0, a1, 1
+; CHECK-NEXT:    vsetvli a0, a0, e8,m2,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v26, a1
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 16 x i8> %v, i8 %elt, i32 %idx
   ret <vscale x 16 x i8> %r
 }
 
 define <vscale x 32 x i8> @insertelt_nxv32i8_0(<vscale x 32 x i8> %v, i8 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv32i8_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e8,m4,ta,mu
 ; CHECK-NEXT:    vmv.s.x v8, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 32 x i8> %v, i8 %elt, i32 0
   ret <vscale x 32 x i8> %r
 }
 
 define <vscale x 32 x i8> @insertelt_nxv32i8_imm(<vscale x 32 x i8> %v, i8 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv32i8_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e8,m4,ta,mu
-; CHECK-NEXT:    vslidedown.vi v28, v8, 3
 ; CHECK-NEXT:    vmv.s.x v28, a0
-; CHECK-NEXT:    vsetvli a0, zero, e8,m4,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e8,m4,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v28, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 32 x i8> %v, i8 %elt, i32 3
   ret <vscale x 32 x i8> %r
 }
 
 define <vscale x 32 x i8> @insertelt_nxv32i8_idx(<vscale x 32 x i8> %v, i8 signext %elt, i32 signext %idx) {
 ; CHECK-LABEL: insertelt_nxv32i8_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a2, zero, e8,m4,ta,mu
-; CHECK-NEXT:    vslidedown.vx v28, v8, a1
 ; CHECK-NEXT:    vmv.s.x v28, a0
-; CHECK-NEXT:    vsetvli a0, zero, e8,m4,tu,mu
+; CHECK-NEXT:    addi a0, a1, 1
+; CHECK-NEXT:    vsetvli a0, a0, e8,m4,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v28, a1
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 32 x i8> %v, i8 %elt, i32 %idx
   ret <vscale x 32 x i8> %r
 }
 
 define <vscale x 64 x i8> @insertelt_nxv64i8_0(<vscale x 64 x i8> %v, i8 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv64i8_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e8,m8,ta,mu
 ; CHECK-NEXT:    vmv.s.x v8, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 64 x i8> %v, i8 %elt, i32 0
   ret <vscale x 64 x i8> %r
 }
 
 define <vscale x 64 x i8> @insertelt_nxv64i8_imm(<vscale x 64 x i8> %v, i8 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv64i8_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e8,m8,ta,mu
-; CHECK-NEXT:    vslidedown.vi v16, v8, 3
 ; CHECK-NEXT:    vmv.s.x v16, a0
-; CHECK-NEXT:    vsetvli a0, zero, e8,m8,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e8,m8,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v16, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 64 x i8> %v, i8 %elt, i32 3
   ret <vscale x 64 x i8> %r
 }
 
 define <vscale x 64 x i8> @insertelt_nxv64i8_idx(<vscale x 64 x i8> %v, i8 signext %elt, i32 signext %idx) {
 ; CHECK-LABEL: insertelt_nxv64i8_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a2, zero, e8,m8,ta,mu
-; CHECK-NEXT:    vslidedown.vx v16, v8, a1
 ; CHECK-NEXT:    vmv.s.x v16, a0
-; CHECK-NEXT:    vsetvli a0, zero, e8,m8,tu,mu
+; CHECK-NEXT:    addi a0, a1, 1
+; CHECK-NEXT:    vsetvli a0, a0, e8,m8,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v16, a1
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 64 x i8> %v, i8 %elt, i32 %idx
   ret <vscale x 64 x i8> %r
 }
 
 define <vscale x 1 x i16> @insertelt_nxv1i16_0(<vscale x 1 x i16> %v, i16 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv1i16_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e16,mf4,ta,mu
 ; CHECK-NEXT:    vmv.s.x v8, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 1 x i16> %v, i16 %elt, i32 0
   ret <vscale x 1 x i16> %r
 }
 
 define <vscale x 1 x i16> @insertelt_nxv1i16_imm(<vscale x 1 x i16> %v, i16 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv1i16_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e16,mf4,ta,mu
-; CHECK-NEXT:    vslidedown.vi v25, v8, 3
 ; CHECK-NEXT:    vmv.s.x v25, a0
-; CHECK-NEXT:    vsetvli a0, zero, e16,mf4,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e16,mf4,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v25, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 1 x i16> %v, i16 %elt, i32 3
   ret <vscale x 1 x i16> %r
 }
 
 define <vscale x 1 x i16> @insertelt_nxv1i16_idx(<vscale x 1 x i16> %v, i16 signext %elt, i32 signext %idx) {
 ; CHECK-LABEL: insertelt_nxv1i16_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a2, zero, e16,mf4,ta,mu
-; CHECK-NEXT:    vslidedown.vx v25, v8, a1
 ; CHECK-NEXT:    vmv.s.x v25, a0
-; CHECK-NEXT:    vsetvli a0, zero, e16,mf4,tu,mu
+; CHECK-NEXT:    addi a0, a1, 1
+; CHECK-NEXT:    vsetvli a0, a0, e16,mf4,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v25, a1
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 1 x i16> %v, i16 %elt, i32 %idx
   ret <vscale x 1 x i16> %r
 }
 
 define <vscale x 2 x i16> @insertelt_nxv2i16_0(<vscale x 2 x i16> %v, i16 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv2i16_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e16,mf2,ta,mu
 ; CHECK-NEXT:    vmv.s.x v8, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 2 x i16> %v, i16 %elt, i32 0
   ret <vscale x 2 x i16> %r
 }
 
 define <vscale x 2 x i16> @insertelt_nxv2i16_imm(<vscale x 2 x i16> %v, i16 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv2i16_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e16,mf2,ta,mu
-; CHECK-NEXT:    vslidedown.vi v25, v8, 3
 ; CHECK-NEXT:    vmv.s.x v25, a0
-; CHECK-NEXT:    vsetvli a0, zero, e16,mf2,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e16,mf2,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v25, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 2 x i16> %v, i16 %elt, i32 3
   ret <vscale x 2 x i16> %r
 }
 
 define <vscale x 2 x i16> @insertelt_nxv2i16_idx(<vscale x 2 x i16> %v, i16 signext %elt, i32 signext %idx) {
 ; CHECK-LABEL: insertelt_nxv2i16_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a2, zero, e16,mf2,ta,mu
-; CHECK-NEXT:    vslidedown.vx v25, v8, a1
 ; CHECK-NEXT:    vmv.s.x v25, a0
-; CHECK-NEXT:    vsetvli a0, zero, e16,mf2,tu,mu
+; CHECK-NEXT:    addi a0, a1, 1
+; CHECK-NEXT:    vsetvli a0, a0, e16,mf2,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v25, a1
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 2 x i16> %v, i16 %elt, i32 %idx
   ret <vscale x 2 x i16> %r
 }
 
 define <vscale x 4 x i16> @insertelt_nxv4i16_0(<vscale x 4 x i16> %v, i16 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv4i16_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e16,m1,ta,mu
 ; CHECK-NEXT:    vmv.s.x v8, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 4 x i16> %v, i16 %elt, i32 0
   ret <vscale x 4 x i16> %r
 }
 
 define <vscale x 4 x i16> @insertelt_nxv4i16_imm(<vscale x 4 x i16> %v, i16 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv4i16_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e16,m1,ta,mu
-; CHECK-NEXT:    vslidedown.vi v25, v8, 3
 ; CHECK-NEXT:    vmv.s.x v25, a0
-; CHECK-NEXT:    vsetvli a0, zero, e16,m1,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e16,m1,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v25, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 4 x i16> %v, i16 %elt, i32 3
   ret <vscale x 4 x i16> %r
 }
 
 define <vscale x 4 x i16> @insertelt_nxv4i16_idx(<vscale x 4 x i16> %v, i16 signext %elt, i32 signext %idx) {
 ; CHECK-LABEL: insertelt_nxv4i16_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a2, zero, e16,m1,ta,mu
-; CHECK-NEXT:    vslidedown.vx v25, v8, a1
 ; CHECK-NEXT:    vmv.s.x v25, a0
-; CHECK-NEXT:    vsetvli a0, zero, e16,m1,tu,mu
+; CHECK-NEXT:    addi a0, a1, 1
+; CHECK-NEXT:    vsetvli a0, a0, e16,m1,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v25, a1
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 4 x i16> %v, i16 %elt, i32 %idx
   ret <vscale x 4 x i16> %r
 }
 
 define <vscale x 8 x i16> @insertelt_nxv8i16_0(<vscale x 8 x i16> %v, i16 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv8i16_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e16,m2,ta,mu
 ; CHECK-NEXT:    vmv.s.x v8, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 8 x i16> %v, i16 %elt, i32 0
   ret <vscale x 8 x i16> %r
 }
 
 define <vscale x 8 x i16> @insertelt_nxv8i16_imm(<vscale x 8 x i16> %v, i16 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv8i16_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e16,m2,ta,mu
-; CHECK-NEXT:    vslidedown.vi v26, v8, 3
 ; CHECK-NEXT:    vmv.s.x v26, a0
-; CHECK-NEXT:    vsetvli a0, zero, e16,m2,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e16,m2,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v26, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 8 x i16> %v, i16 %elt, i32 3
   ret <vscale x 8 x i16> %r
 }
 
 define <vscale x 8 x i16> @insertelt_nxv8i16_idx(<vscale x 8 x i16> %v, i16 signext %elt, i32 signext %idx) {
 ; CHECK-LABEL: insertelt_nxv8i16_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a2, zero, e16,m2,ta,mu
-; CHECK-NEXT:    vslidedown.vx v26, v8, a1
 ; CHECK-NEXT:    vmv.s.x v26, a0
-; CHECK-NEXT:    vsetvli a0, zero, e16,m2,tu,mu
+; CHECK-NEXT:    addi a0, a1, 1
+; CHECK-NEXT:    vsetvli a0, a0, e16,m2,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v26, a1
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 8 x i16> %v, i16 %elt, i32 %idx
   ret <vscale x 8 x i16> %r
 }
 
 define <vscale x 16 x i16> @insertelt_nxv16i16_0(<vscale x 16 x i16> %v, i16 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv16i16_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e16,m4,ta,mu
 ; CHECK-NEXT:    vmv.s.x v8, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 16 x i16> %v, i16 %elt, i32 0
   ret <vscale x 16 x i16> %r
 }
 
 define <vscale x 16 x i16> @insertelt_nxv16i16_imm(<vscale x 16 x i16> %v, i16 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv16i16_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e16,m4,ta,mu
-; CHECK-NEXT:    vslidedown.vi v28, v8, 3
 ; CHECK-NEXT:    vmv.s.x v28, a0
-; CHECK-NEXT:    vsetvli a0, zero, e16,m4,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e16,m4,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v28, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 16 x i16> %v, i16 %elt, i32 3
   ret <vscale x 16 x i16> %r
 }
 
 define <vscale x 16 x i16> @insertelt_nxv16i16_idx(<vscale x 16 x i16> %v, i16 signext %elt, i32 signext %idx) {
 ; CHECK-LABEL: insertelt_nxv16i16_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a2, zero, e16,m4,ta,mu
-; CHECK-NEXT:    vslidedown.vx v28, v8, a1
 ; CHECK-NEXT:    vmv.s.x v28, a0
-; CHECK-NEXT:    vsetvli a0, zero, e16,m4,tu,mu
+; CHECK-NEXT:    addi a0, a1, 1
+; CHECK-NEXT:    vsetvli a0, a0, e16,m4,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v28, a1
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 16 x i16> %v, i16 %elt, i32 %idx
   ret <vscale x 16 x i16> %r
 }
 
 define <vscale x 32 x i16> @insertelt_nxv32i16_0(<vscale x 32 x i16> %v, i16 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv32i16_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e16,m8,ta,mu
 ; CHECK-NEXT:    vmv.s.x v8, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 32 x i16> %v, i16 %elt, i32 0
   ret <vscale x 32 x i16> %r
 }
 
 define <vscale x 32 x i16> @insertelt_nxv32i16_imm(<vscale x 32 x i16> %v, i16 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv32i16_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e16,m8,ta,mu
-; CHECK-NEXT:    vslidedown.vi v16, v8, 3
 ; CHECK-NEXT:    vmv.s.x v16, a0
-; CHECK-NEXT:    vsetvli a0, zero, e16,m8,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e16,m8,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v16, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 32 x i16> %v, i16 %elt, i32 3
   ret <vscale x 32 x i16> %r
 }
 
 define <vscale x 32 x i16> @insertelt_nxv32i16_idx(<vscale x 32 x i16> %v, i16 signext %elt, i32 signext %idx) {
 ; CHECK-LABEL: insertelt_nxv32i16_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a2, zero, e16,m8,ta,mu
-; CHECK-NEXT:    vslidedown.vx v16, v8, a1
 ; CHECK-NEXT:    vmv.s.x v16, a0
-; CHECK-NEXT:    vsetvli a0, zero, e16,m8,tu,mu
+; CHECK-NEXT:    addi a0, a1, 1
+; CHECK-NEXT:    vsetvli a0, a0, e16,m8,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v16, a1
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 32 x i16> %v, i16 %elt, i32 %idx
   ret <vscale x 32 x i16> %r
 }
 
 define <vscale x 1 x i32> @insertelt_nxv1i32_0(<vscale x 1 x i32> %v, i32 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv1i32_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e32,mf2,ta,mu
 ; CHECK-NEXT:    vmv.s.x v8, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 1 x i32> %v, i32 %elt, i32 0
   ret <vscale x 1 x i32> %r
 }
 
 define <vscale x 1 x i32> @insertelt_nxv1i32_imm(<vscale x 1 x i32> %v, i32 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv1i32_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e32,mf2,ta,mu
-; CHECK-NEXT:    vslidedown.vi v25, v8, 3
 ; CHECK-NEXT:    vmv.s.x v25, a0
-; CHECK-NEXT:    vsetvli a0, zero, e32,mf2,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e32,mf2,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v25, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 1 x i32> %v, i32 %elt, i32 3
   ret <vscale x 1 x i32> %r
 }
 
 define <vscale x 1 x i32> @insertelt_nxv1i32_idx(<vscale x 1 x i32> %v, i32 signext %elt, i32 signext %idx) {
 ; CHECK-LABEL: insertelt_nxv1i32_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a2, zero, e32,mf2,ta,mu
-; CHECK-NEXT:    vslidedown.vx v25, v8, a1
 ; CHECK-NEXT:    vmv.s.x v25, a0
-; CHECK-NEXT:    vsetvli a0, zero, e32,mf2,tu,mu
+; CHECK-NEXT:    addi a0, a1, 1
+; CHECK-NEXT:    vsetvli a0, a0, e32,mf2,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v25, a1
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 1 x i32> %v, i32 %elt, i32 %idx
   ret <vscale x 1 x i32> %r
 }
 
 define <vscale x 2 x i32> @insertelt_nxv2i32_0(<vscale x 2 x i32> %v, i32 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv2i32_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e32,m1,ta,mu
 ; CHECK-NEXT:    vmv.s.x v8, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 2 x i32> %v, i32 %elt, i32 0
   ret <vscale x 2 x i32> %r
 }
 
 define <vscale x 2 x i32> @insertelt_nxv2i32_imm(<vscale x 2 x i32> %v, i32 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv2i32_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e32,m1,ta,mu
-; CHECK-NEXT:    vslidedown.vi v25, v8, 3
 ; CHECK-NEXT:    vmv.s.x v25, a0
-; CHECK-NEXT:    vsetvli a0, zero, e32,m1,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e32,m1,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v25, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 2 x i32> %v, i32 %elt, i32 3
   ret <vscale x 2 x i32> %r
 }
 
 define <vscale x 2 x i32> @insertelt_nxv2i32_idx(<vscale x 2 x i32> %v, i32 signext %elt, i32 signext %idx) {
 ; CHECK-LABEL: insertelt_nxv2i32_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a2, zero, e32,m1,ta,mu
-; CHECK-NEXT:    vslidedown.vx v25, v8, a1
 ; CHECK-NEXT:    vmv.s.x v25, a0
-; CHECK-NEXT:    vsetvli a0, zero, e32,m1,tu,mu
+; CHECK-NEXT:    addi a0, a1, 1
+; CHECK-NEXT:    vsetvli a0, a0, e32,m1,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v25, a1
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 2 x i32> %v, i32 %elt, i32 %idx
   ret <vscale x 2 x i32> %r
 }
 
 define <vscale x 4 x i32> @insertelt_nxv4i32_0(<vscale x 4 x i32> %v, i32 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv4i32_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e32,m2,ta,mu
 ; CHECK-NEXT:    vmv.s.x v8, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 4 x i32> %v, i32 %elt, i32 0
   ret <vscale x 4 x i32> %r
 }
 
 define <vscale x 4 x i32> @insertelt_nxv4i32_imm(<vscale x 4 x i32> %v, i32 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv4i32_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e32,m2,ta,mu
-; CHECK-NEXT:    vslidedown.vi v26, v8, 3
 ; CHECK-NEXT:    vmv.s.x v26, a0
-; CHECK-NEXT:    vsetvli a0, zero, e32,m2,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e32,m2,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v26, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 4 x i32> %v, i32 %elt, i32 3
   ret <vscale x 4 x i32> %r
 }
 
 define <vscale x 4 x i32> @insertelt_nxv4i32_idx(<vscale x 4 x i32> %v, i32 signext %elt, i32 signext %idx) {
 ; CHECK-LABEL: insertelt_nxv4i32_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a2, zero, e32,m2,ta,mu
-; CHECK-NEXT:    vslidedown.vx v26, v8, a1
 ; CHECK-NEXT:    vmv.s.x v26, a0
-; CHECK-NEXT:    vsetvli a0, zero, e32,m2,tu,mu
+; CHECK-NEXT:    addi a0, a1, 1
+; CHECK-NEXT:    vsetvli a0, a0, e32,m2,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v26, a1
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 4 x i32> %v, i32 %elt, i32 %idx
   ret <vscale x 4 x i32> %r
 }
 
 define <vscale x 8 x i32> @insertelt_nxv8i32_0(<vscale x 8 x i32> %v, i32 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv8i32_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e32,m4,ta,mu
 ; CHECK-NEXT:    vmv.s.x v8, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 8 x i32> %v, i32 %elt, i32 0
   ret <vscale x 8 x i32> %r
 }
 
 define <vscale x 8 x i32> @insertelt_nxv8i32_imm(<vscale x 8 x i32> %v, i32 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv8i32_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e32,m4,ta,mu
-; CHECK-NEXT:    vslidedown.vi v28, v8, 3
 ; CHECK-NEXT:    vmv.s.x v28, a0
-; CHECK-NEXT:    vsetvli a0, zero, e32,m4,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e32,m4,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v28, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 8 x i32> %v, i32 %elt, i32 3
   ret <vscale x 8 x i32> %r
 }
 
 define <vscale x 8 x i32> @insertelt_nxv8i32_idx(<vscale x 8 x i32> %v, i32 signext %elt, i32 signext %idx) {
 ; CHECK-LABEL: insertelt_nxv8i32_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a2, zero, e32,m4,ta,mu
-; CHECK-NEXT:    vslidedown.vx v28, v8, a1
 ; CHECK-NEXT:    vmv.s.x v28, a0
-; CHECK-NEXT:    vsetvli a0, zero, e32,m4,tu,mu
+; CHECK-NEXT:    addi a0, a1, 1
+; CHECK-NEXT:    vsetvli a0, a0, e32,m4,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v28, a1
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 8 x i32> %v, i32 %elt, i32 %idx
   ret <vscale x 8 x i32> %r
 }
 
 define <vscale x 16 x i32> @insertelt_nxv16i32_0(<vscale x 16 x i32> %v, i32 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv16i32_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e32,m8,ta,mu
 ; CHECK-NEXT:    vmv.s.x v8, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 16 x i32> %v, i32 %elt, i32 0
   ret <vscale x 16 x i32> %r
 }
 
 define <vscale x 16 x i32> @insertelt_nxv16i32_imm(<vscale x 16 x i32> %v, i32 signext %elt) {
 ; CHECK-LABEL: insertelt_nxv16i32_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e32,m8,ta,mu
-; CHECK-NEXT:    vslidedown.vi v16, v8, 3
 ; CHECK-NEXT:    vmv.s.x v16, a0
-; CHECK-NEXT:    vsetvli a0, zero, e32,m8,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e32,m8,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v16, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 16 x i32> %v, i32 %elt, i32 3
   ret <vscale x 16 x i32> %r
 }
 
 define <vscale x 16 x i32> @insertelt_nxv16i32_idx(<vscale x 16 x i32> %v, i32 signext %elt, i32 signext %idx) {
 ; CHECK-LABEL: insertelt_nxv16i32_idx:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a2, zero, e32,m8,ta,mu
-; CHECK-NEXT:    vslidedown.vx v16, v8, a1
 ; CHECK-NEXT:    vmv.s.x v16, a0
-; CHECK-NEXT:    vsetvli a0, zero, e32,m8,tu,mu
+; CHECK-NEXT:    addi a0, a1, 1
+; CHECK-NEXT:    vsetvli a0, a0, e32,m8,tu,mu
 ; CHECK-NEXT:    vslideup.vx v8, v16, a1
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 16 x i32> %v, i32 %elt, i32 %idx
   ret <vscale x 16 x i32> %r
 }
 
 define <vscale x 1 x i64> @insertelt_nxv1i64_0(<vscale x 1 x i64> %v, i64 %elt) {
 ; CHECK-LABEL: insertelt_nxv1i64_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e64,m1,ta,mu
 ; CHECK-NEXT:    vmv.s.x v8, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 1 x i64> %v, i64 %elt, i32 0
   ret <vscale x 1 x i64> %r
 }
 
 define <vscale x 1 x i64> @insertelt_nxv1i64_imm(<vscale x 1 x i64> %v, i64 %elt) {
 ; CHECK-LABEL: insertelt_nxv1i64_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e64,m1,ta,mu
-; CHECK-NEXT:    vslidedown.vi v25, v8, 3
 ; CHECK-NEXT:    vmv.s.x v25, a0
-; CHECK-NEXT:    vsetvli a0, zero, e64,m1,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e64,m1,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v25, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 1 x i64> %v, i64 %elt, i32 3
   ret <vscale x 1 x i64> %r
 }
 
 define <vscale x 1 x i64> @insertelt_nxv1i64_idx(<vscale x 1 x i64> %v, i64 %elt, i32 %idx) {
 ; CHECK-LABEL: insertelt_nxv1i64_idx:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    sext.w a1, a1
 ; CHECK-NEXT:    vsetvli a2, zero, e64,m1,ta,mu
-; CHECK-NEXT:    vslidedown.vx v25, v8, a1
 ; CHECK-NEXT:    vmv.s.x v25, a0
-; CHECK-NEXT:    vsetvli a0, zero, e64,m1,tu,mu
-; CHECK-NEXT:    vslideup.vx v8, v25, a1
+; CHECK-NEXT:    sext.w a0, a1
+; CHECK-NEXT:    addi a1, a0, 1
+; CHECK-NEXT:    vsetvli a1, a1, e64,m1,tu,mu
+; CHECK-NEXT:    vslideup.vx v8, v25, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 1 x i64> %v, i64 %elt, i32 %idx
   ret <vscale x 1 x i64> %r
 }
 
 define <vscale x 2 x i64> @insertelt_nxv2i64_0(<vscale x 2 x i64> %v, i64 %elt) {
 ; CHECK-LABEL: insertelt_nxv2i64_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e64,m2,ta,mu
 ; CHECK-NEXT:    vmv.s.x v8, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 2 x i64> %v, i64 %elt, i32 0
   ret <vscale x 2 x i64> %r
 }
 
 define <vscale x 2 x i64> @insertelt_nxv2i64_imm(<vscale x 2 x i64> %v, i64 %elt) {
 ; CHECK-LABEL: insertelt_nxv2i64_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e64,m2,ta,mu
-; CHECK-NEXT:    vslidedown.vi v26, v8, 3
 ; CHECK-NEXT:    vmv.s.x v26, a0
-; CHECK-NEXT:    vsetvli a0, zero, e64,m2,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e64,m2,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v26, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 2 x i64> %v, i64 %elt, i32 3
   ret <vscale x 2 x i64> %r
 }
 
 define <vscale x 2 x i64> @insertelt_nxv2i64_idx(<vscale x 2 x i64> %v, i64 %elt, i32 %idx) {
 ; CHECK-LABEL: insertelt_nxv2i64_idx:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    sext.w a1, a1
 ; CHECK-NEXT:    vsetvli a2, zero, e64,m2,ta,mu
-; CHECK-NEXT:    vslidedown.vx v26, v8, a1
 ; CHECK-NEXT:    vmv.s.x v26, a0
-; CHECK-NEXT:    vsetvli a0, zero, e64,m2,tu,mu
-; CHECK-NEXT:    vslideup.vx v8, v26, a1
+; CHECK-NEXT:    sext.w a0, a1
+; CHECK-NEXT:    addi a1, a0, 1
+; CHECK-NEXT:    vsetvli a1, a1, e64,m2,tu,mu
+; CHECK-NEXT:    vslideup.vx v8, v26, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 2 x i64> %v, i64 %elt, i32 %idx
   ret <vscale x 2 x i64> %r
 }
 
 define <vscale x 4 x i64> @insertelt_nxv4i64_0(<vscale x 4 x i64> %v, i64 %elt) {
 ; CHECK-LABEL: insertelt_nxv4i64_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e64,m4,ta,mu
 ; CHECK-NEXT:    vmv.s.x v8, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 4 x i64> %v, i64 %elt, i32 0
   ret <vscale x 4 x i64> %r
 }
 
 define <vscale x 4 x i64> @insertelt_nxv4i64_imm(<vscale x 4 x i64> %v, i64 %elt) {
 ; CHECK-LABEL: insertelt_nxv4i64_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e64,m4,ta,mu
-; CHECK-NEXT:    vslidedown.vi v28, v8, 3
 ; CHECK-NEXT:    vmv.s.x v28, a0
-; CHECK-NEXT:    vsetvli a0, zero, e64,m4,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e64,m4,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v28, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 4 x i64> %v, i64 %elt, i32 3
   ret <vscale x 4 x i64> %r
 }
 
 define <vscale x 4 x i64> @insertelt_nxv4i64_idx(<vscale x 4 x i64> %v, i64 %elt, i32 %idx) {
 ; CHECK-LABEL: insertelt_nxv4i64_idx:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    sext.w a1, a1
 ; CHECK-NEXT:    vsetvli a2, zero, e64,m4,ta,mu
-; CHECK-NEXT:    vslidedown.vx v28, v8, a1
 ; CHECK-NEXT:    vmv.s.x v28, a0
-; CHECK-NEXT:    vsetvli a0, zero, e64,m4,tu,mu
-; CHECK-NEXT:    vslideup.vx v8, v28, a1
+; CHECK-NEXT:    sext.w a0, a1
+; CHECK-NEXT:    addi a1, a0, 1
+; CHECK-NEXT:    vsetvli a1, a1, e64,m4,tu,mu
+; CHECK-NEXT:    vslideup.vx v8, v28, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 4 x i64> %v, i64 %elt, i32 %idx
   ret <vscale x 4 x i64> %r
 }
 
 define <vscale x 8 x i64> @insertelt_nxv8i64_0(<vscale x 8 x i64> %v, i64 %elt) {
 ; CHECK-LABEL: insertelt_nxv8i64_0:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e64,m8,ta,mu
 ; CHECK-NEXT:    vmv.s.x v8, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 8 x i64> %v, i64 %elt, i32 0
   ret <vscale x 8 x i64> %r
 }
 
 define <vscale x 8 x i64> @insertelt_nxv8i64_imm(<vscale x 8 x i64> %v, i64 %elt) {
 ; CHECK-LABEL: insertelt_nxv8i64_imm:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetvli a1, zero, e64,m8,ta,mu
-; CHECK-NEXT:    vslidedown.vi v16, v8, 3
 ; CHECK-NEXT:    vmv.s.x v16, a0
-; CHECK-NEXT:    vsetvli a0, zero, e64,m8,tu,mu
+; CHECK-NEXT:    vsetivli a0, 4, e64,m8,tu,mu
 ; CHECK-NEXT:    vslideup.vi v8, v16, 3
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 8 x i64> %v, i64 %elt, i32 3
   ret <vscale x 8 x i64> %r
 }
 
 define <vscale x 8 x i64> @insertelt_nxv8i64_idx(<vscale x 8 x i64> %v, i64 %elt, i32 %idx) {
 ; CHECK-LABEL: insertelt_nxv8i64_idx:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    sext.w a1, a1
 ; CHECK-NEXT:    vsetvli a2, zero, e64,m8,ta,mu
-; CHECK-NEXT:    vslidedown.vx v16, v8, a1
 ; CHECK-NEXT:    vmv.s.x v16, a0
-; CHECK-NEXT:    vsetvli a0, zero, e64,m8,tu,mu
-; CHECK-NEXT:    vslideup.vx v8, v16, a1
+; CHECK-NEXT:    sext.w a0, a1
+; CHECK-NEXT:    addi a1, a0, 1
+; CHECK-NEXT:    vsetvli a1, a1, e64,m8,tu,mu
+; CHECK-NEXT:    vslideup.vx v8, v16, a0
 ; CHECK-NEXT:    ret
   %r = insertelement <vscale x 8 x i64> %v, i64 %elt, i32 %idx
   ret <vscale x 8 x i64> %r
 }