diff --git a/llvm/lib/Target/RISCV/RISCVVLOptimizer.cpp b/llvm/lib/Target/RISCV/RISCVVLOptimizer.cpp
index 69ee210071286..8156eaff8a04c 100644
--- a/llvm/lib/Target/RISCV/RISCVVLOptimizer.cpp
+++ b/llvm/lib/Target/RISCV/RISCVVLOptimizer.cpp
@@ -551,6 +551,7 @@ getOperandLog2EEW(const MachineOperand &MO, const MachineRegisterInfo *MRI) {
   case RISCV::VFWCVT_F_XU_V:
   case RISCV::VFWCVT_F_X_V:
   case RISCV::VFWCVT_F_F_V:
+  case RISCV::VFWCVTBF16_F_F_V:
     return IsMODef ? MILog2SEW + 1 : MILog2SEW;
 
   // Def and Op1 uses EEW=2*SEW. Op2 uses EEW=SEW.
@@ -607,7 +608,8 @@ getOperandLog2EEW(const MachineOperand &MO, const MachineRegisterInfo *MRI) {
   case RISCV::VFNCVT_F_XU_W:
   case RISCV::VFNCVT_F_X_W:
   case RISCV::VFNCVT_F_F_W:
-  case RISCV::VFNCVT_ROD_F_F_W: {
+  case RISCV::VFNCVT_ROD_F_F_W:
+  case RISCV::VFNCVTBF16_F_F_W: {
     bool IsOp1 = HasPassthru ? MO.getOperandNo() == 2 : MO.getOperandNo() == 1;
     bool TwoTimes = IsOp1;
     return TwoTimes ? MILog2SEW + 1 : MILog2SEW;
@@ -1045,6 +1047,7 @@ static bool isSupportedInstr(const MachineInstr &MI) {
   case RISCV::VFWCVT_F_XU_V:
   case RISCV::VFWCVT_F_X_V:
   case RISCV::VFWCVT_F_F_V:
+  case RISCV::VFWCVTBF16_F_F_V:
   // Narrowing Floating-Point/Integer Type-Convert Instructions
   case RISCV::VFNCVT_XU_F_W:
   case RISCV::VFNCVT_X_F_W:
@@ -1054,6 +1057,7 @@ static bool isSupportedInstr(const MachineInstr &MI) {
   case RISCV::VFNCVT_F_X_W:
   case RISCV::VFNCVT_F_F_W:
   case RISCV::VFNCVT_ROD_F_F_W:
+  case RISCV::VFNCVTBF16_F_F_W:
     return true;
   }
 
diff --git a/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-fp.ll b/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-fp.ll
index b8710a518287a..49db94e1a02df 100644
--- a/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-fp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-fp.ll
@@ -32,12 +32,11 @@ define void @fadd_v6bf16(ptr %x, ptr %y) {
 ; CHECK-NEXT:    vsetivli zero, 6, e16, m1, ta, ma
 ; CHECK-NEXT:    vle16.v v8, (a1)
 ; CHECK-NEXT:    vle16.v v9, (a0)
-; CHECK-NEXT:    vsetivli zero, 8, e16, m1, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v9
 ; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
 ; CHECK-NEXT:    vfadd.vv v8, v12, v10
-; CHECK-NEXT:    vsetivli zero, 6, e16, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e16, m1, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v10, v8
 ; CHECK-NEXT:    vse16.v v10, (a0)
 ; CHECK-NEXT:    ret
@@ -167,12 +166,11 @@ define void @fsub_v6bf16(ptr %x, ptr %y) {
 ; CHECK-NEXT:    vsetivli zero, 6, e16, m1, ta, ma
 ; CHECK-NEXT:    vle16.v v8, (a1)
 ; CHECK-NEXT:    vle16.v v9, (a0)
-; CHECK-NEXT:    vsetivli zero, 8, e16, m1, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v9
 ; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
 ; CHECK-NEXT:    vfsub.vv v8, v12, v10
-; CHECK-NEXT:    vsetivli zero, 6, e16, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e16, m1, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v10, v8
 ; CHECK-NEXT:    vse16.v v10, (a0)
 ; CHECK-NEXT:    ret
@@ -302,12 +300,11 @@ define void @fmul_v6bf16(ptr %x, ptr %y) {
 ; CHECK-NEXT:    vsetivli zero, 6, e16, m1, ta, ma
 ; CHECK-NEXT:    vle16.v v8, (a1)
 ; CHECK-NEXT:    vle16.v v9, (a0)
-; CHECK-NEXT:    vsetivli zero, 8, e16, m1, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v9
 ; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
 ; CHECK-NEXT:    vfmul.vv v8, v12, v10
-; CHECK-NEXT:    vsetivli zero, 6, e16, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e16, m1, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v10, v8
 ; CHECK-NEXT:    vse16.v v10, (a0)
 ; CHECK-NEXT:    ret
@@ -437,12 +434,11 @@ define void @fdiv_v6bf16(ptr %x, ptr %y) {
 ; CHECK-NEXT:    vsetivli zero, 6, e16, m1, ta, ma
 ; CHECK-NEXT:    vle16.v v8, (a1)
 ; CHECK-NEXT:    vle16.v v9, (a0)
-; CHECK-NEXT:    vsetivli zero, 8, e16, m1, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v9
 ; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
 ; CHECK-NEXT:    vfdiv.vv v8, v12, v10
-; CHECK-NEXT:    vsetivli zero, 6, e16, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e16, m1, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v10, v8
 ; CHECK-NEXT:    vse16.v v10, (a0)
 ; CHECK-NEXT:    ret
@@ -1196,9 +1192,7 @@ define void @copysign_neg_trunc_v3bf16_v3f32(ptr %x, ptr %y) {
 ; CHECK-NEXT:    lui a1, 8
 ; CHECK-NEXT:    addi a2, a1, -1
 ; CHECK-NEXT:    vand.vx v8, v8, a2
-; CHECK-NEXT:    vsetivli zero, 4, e16, mf2, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v10, v9
-; CHECK-NEXT:    vsetivli zero, 3, e16, mf2, ta, ma
 ; CHECK-NEXT:    vxor.vx v9, v10, a1
 ; CHECK-NEXT:    vand.vx v9, v9, a1
 ; CHECK-NEXT:    vor.vv v8, v8, v9
@@ -2263,13 +2257,12 @@ define void @fadd_vf_v6bf16(ptr %x, bfloat %y) {
 ; CHECK-NEXT:    vsetivli zero, 6, e16, m1, ta, ma
 ; CHECK-NEXT:    vle16.v v8, (a0)
 ; CHECK-NEXT:    fmv.x.w a1, fa0
-; CHECK-NEXT:    vsetivli zero, 8, e16, m1, ta, ma
 ; CHECK-NEXT:    vmv.v.x v9, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v9
 ; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
 ; CHECK-NEXT:    vfadd.vv v8, v10, v12
-; CHECK-NEXT:    vsetivli zero, 6, e16, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e16, m1, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v10, v8
 ; CHECK-NEXT:    vse16.v v10, (a0)
 ; CHECK-NEXT:    ret
@@ -2404,13 +2397,12 @@ define void @fadd_fv_v6bf16(ptr %x, bfloat %y) {
 ; CHECK-NEXT:    vsetivli zero, 6, e16, m1, ta, ma
 ; CHECK-NEXT:    vle16.v v8, (a0)
 ; CHECK-NEXT:    fmv.x.w a1, fa0
-; CHECK-NEXT:    vsetivli zero, 8, e16, m1, ta, ma
 ; CHECK-NEXT:    vmv.v.x v9, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v9
 ; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
 ; CHECK-NEXT:    vfadd.vv v8, v12, v10
-; CHECK-NEXT:    vsetivli zero, 6, e16, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e16, m1, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v10, v8
 ; CHECK-NEXT:    vse16.v v10, (a0)
 ; CHECK-NEXT:    ret
@@ -2545,13 +2537,12 @@ define void @fsub_vf_v6bf16(ptr %x, bfloat %y) {
 ; CHECK-NEXT:    vsetivli zero, 6, e16, m1, ta, ma
 ; CHECK-NEXT:    vle16.v v8, (a0)
 ; CHECK-NEXT:    fmv.x.w a1, fa0
-; CHECK-NEXT:    vsetivli zero, 8, e16, m1, ta, ma
 ; CHECK-NEXT:    vmv.v.x v9, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v9
 ; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
 ; CHECK-NEXT:    vfsub.vv v8, v10, v12
-; CHECK-NEXT:    vsetivli zero, 6, e16, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e16, m1, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v10, v8
 ; CHECK-NEXT:    vse16.v v10, (a0)
 ; CHECK-NEXT:    ret
@@ -2686,13 +2677,12 @@ define void @fsub_fv_v6bf16(ptr %x, bfloat %y) {
 ; CHECK-NEXT:    vsetivli zero, 6, e16, m1, ta, ma
 ; CHECK-NEXT:    vle16.v v8, (a0)
 ; CHECK-NEXT:    fmv.x.w a1, fa0
-; CHECK-NEXT:    vsetivli zero, 8, e16, m1, ta, ma
 ; CHECK-NEXT:    vmv.v.x v9, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v9
 ; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
 ; CHECK-NEXT:    vfsub.vv v8, v12, v10
-; CHECK-NEXT:    vsetivli zero, 6, e16, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e16, m1, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v10, v8
 ; CHECK-NEXT:    vse16.v v10, (a0)
 ; CHECK-NEXT:    ret
@@ -2827,13 +2817,12 @@ define void @fmul_vf_v6bf16(ptr %x, bfloat %y) {
 ; CHECK-NEXT:    vsetivli zero, 6, e16, m1, ta, ma
 ; CHECK-NEXT:    vle16.v v8, (a0)
 ; CHECK-NEXT:    fmv.x.w a1, fa0
-; CHECK-NEXT:    vsetivli zero, 8, e16, m1, ta, ma
 ; CHECK-NEXT:    vmv.v.x v9, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v9
 ; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
 ; CHECK-NEXT:    vfmul.vv v8, v10, v12
-; CHECK-NEXT:    vsetivli zero, 6, e16, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e16, m1, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v10, v8
 ; CHECK-NEXT:    vse16.v v10, (a0)
 ; CHECK-NEXT:    ret
@@ -2968,13 +2957,12 @@ define void @fmul_fv_v6bf16(ptr %x, bfloat %y) {
 ; CHECK-NEXT:    vsetivli zero, 6, e16, m1, ta, ma
 ; CHECK-NEXT:    vle16.v v8, (a0)
 ; CHECK-NEXT:    fmv.x.w a1, fa0
-; CHECK-NEXT:    vsetivli zero, 8, e16, m1, ta, ma
 ; CHECK-NEXT:    vmv.v.x v9, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v9
 ; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
 ; CHECK-NEXT:    vfmul.vv v8, v12, v10
-; CHECK-NEXT:    vsetivli zero, 6, e16, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e16, m1, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v10, v8
 ; CHECK-NEXT:    vse16.v v10, (a0)
 ; CHECK-NEXT:    ret
@@ -3109,13 +3097,12 @@ define void @fdiv_vf_v6bf16(ptr %x, bfloat %y) {
 ; CHECK-NEXT:    vsetivli zero, 6, e16, m1, ta, ma
 ; CHECK-NEXT:    vle16.v v8, (a0)
 ; CHECK-NEXT:    fmv.x.w a1, fa0
-; CHECK-NEXT:    vsetivli zero, 8, e16, m1, ta, ma
 ; CHECK-NEXT:    vmv.v.x v9, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v9
 ; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
 ; CHECK-NEXT:    vfdiv.vv v8, v10, v12
-; CHECK-NEXT:    vsetivli zero, 6, e16, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e16, m1, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v10, v8
 ; CHECK-NEXT:    vse16.v v10, (a0)
 ; CHECK-NEXT:    ret
@@ -3250,13 +3237,12 @@ define void @fdiv_fv_v6bf16(ptr %x, bfloat %y) {
 ; CHECK-NEXT:    vsetivli zero, 6, e16, m1, ta, ma
 ; CHECK-NEXT:    vle16.v v8, (a0)
 ; CHECK-NEXT:    fmv.x.w a1, fa0
-; CHECK-NEXT:    vsetivli zero, 8, e16, m1, ta, ma
 ; CHECK-NEXT:    vmv.v.x v9, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v9
 ; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
 ; CHECK-NEXT:    vfdiv.vv v8, v12, v10
-; CHECK-NEXT:    vsetivli zero, 6, e16, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e16, m1, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v10, v8
 ; CHECK-NEXT:    vse16.v v10, (a0)
 ; CHECK-NEXT:    ret
@@ -4908,7 +4894,6 @@ define void @fmuladd_v6bf16(ptr %x, ptr %y, ptr %z) {
 ; CHECK-NEXT:    vle16.v v8, (a1)
 ; CHECK-NEXT:    vle16.v v9, (a0)
 ; CHECK-NEXT:    vle16.v v10, (a2)
-; CHECK-NEXT:    vsetivli zero, 8, e16, m1, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v14, v9
 ; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
@@ -4919,7 +4904,7 @@ define void @fmuladd_v6bf16(ptr %x, ptr %y, ptr %z) {
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v10
 ; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
 ; CHECK-NEXT:    vfadd.vv v8, v8, v12
-; CHECK-NEXT:    vsetivli zero, 6, e16, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e16, m1, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v10, v8
 ; CHECK-NEXT:    vse16.v v10, (a0)
 ; CHECK-NEXT:    ret
@@ -5082,7 +5067,6 @@ define void @fmsub_fmuladd_v6bf16(ptr %x, ptr %y, ptr %z) {
 ; CHECK-NEXT:    vle16.v v8, (a1)
 ; CHECK-NEXT:    vle16.v v9, (a0)
 ; CHECK-NEXT:    vle16.v v10, (a2)
-; CHECK-NEXT:    vsetivli zero, 8, e16, m1, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v14, v9
 ; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
@@ -5093,7 +5077,7 @@ define void @fmsub_fmuladd_v6bf16(ptr %x, ptr %y, ptr %z) {
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v10
 ; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
 ; CHECK-NEXT:    vfsub.vv v8, v8, v12
-; CHECK-NEXT:    vsetivli zero, 6, e16, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e16, m1, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v10, v8
 ; CHECK-NEXT:    vse16.v v10, (a0)
 ; CHECK-NEXT:    ret
diff --git a/llvm/test/CodeGen/RISCV/rvv/fmaximum-vp.ll b/llvm/test/CodeGen/RISCV/rvv/fmaximum-vp.ll
index 33fe73a097e32..7e0c3f45de463 100644
--- a/llvm/test/CodeGen/RISCV/rvv/fmaximum-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/fmaximum-vp.ll
@@ -19,14 +19,14 @@ declare <vscale x 1 x bfloat> @llvm.vp.maximum.nxv1bf16(<vscale x 1 x bfloat>, <
 define <vscale x 1 x bfloat> @vfmax_vv_nxv1bf16(<vscale x 1 x bfloat> %va, <vscale x 1 x bfloat> %vb, <vscale x 1 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: vfmax_vv_nxv1bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vmv1r.v v10, v0
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v11, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmfeq.vv v0, v11, v11, v0.t
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmerge.vvm v9, v11, v8, v0
 ; CHECK-NEXT:    vmv1r.v v0, v10
 ; CHECK-NEXT:    vmfeq.vv v0, v8, v8, v0.t
@@ -43,13 +43,13 @@ define <vscale x 1 x bfloat> @vfmax_vv_nxv1bf16(<vscale x 1 x bfloat> %va, <vsca
 define <vscale x 1 x bfloat> @vfmax_vv_nxv1bf16_unmasked(<vscale x 1 x bfloat> %va, <vscale x 1 x bfloat> %vb, i32 zeroext %evl) {
 ; CHECK-LABEL: vfmax_vv_nxv1bf16_unmasked:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmfeq.vv v0, v10, v10
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmerge.vvm v9, v10, v8, v0
 ; CHECK-NEXT:    vmfeq.vv v0, v8, v8
 ; CHECK-NEXT:    vmerge.vvm v8, v8, v10, v0
@@ -66,14 +66,14 @@ declare <vscale x 2 x bfloat> @llvm.vp.maximum.nxv2bf16(<vscale x 2 x bfloat>, <
 define <vscale x 2 x bfloat> @vfmax_vv_nxv2bf16(<vscale x 2 x bfloat> %va, <vscale x 2 x bfloat> %vb, <vscale x 2 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: vfmax_vv_nxv2bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, ma
 ; CHECK-NEXT:    vmv1r.v v10, v0
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v11, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m1, ta, ma
 ; CHECK-NEXT:    vmfeq.vv v0, v11, v11, v0.t
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e16, mf2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m1, ta, ma
 ; CHECK-NEXT:    vmerge.vvm v9, v11, v8, v0
 ; CHECK-NEXT:    vmv1r.v v0, v10
 ; CHECK-NEXT:    vmfeq.vv v0, v8, v8, v0.t
@@ -90,13 +90,13 @@ define <vscale x 2 x bfloat> @vfmax_vv_nxv2bf16(<vscale x 2 x bfloat> %va, <vsca
 define <vscale x 2 x bfloat> @vfmax_vv_nxv2bf16_unmasked(<vscale x 2 x bfloat> %va, <vscale x 2 x bfloat> %vb, i32 zeroext %evl) {
 ; CHECK-LABEL: vfmax_vv_nxv2bf16_unmasked:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m1, ta, ma
 ; CHECK-NEXT:    vmfeq.vv v0, v10, v10
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e16, mf2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m1, ta, ma
 ; CHECK-NEXT:    vmerge.vvm v9, v10, v8, v0
 ; CHECK-NEXT:    vmfeq.vv v0, v8, v8
 ; CHECK-NEXT:    vmerge.vvm v8, v8, v10, v0
@@ -113,15 +113,15 @@ declare <vscale x 4 x bfloat> @llvm.vp.maximum.nxv4bf16(<vscale x 4 x bfloat>, <
 define <vscale x 4 x bfloat> @vfmax_vv_nxv4bf16(<vscale x 4 x bfloat> %va, <vscale x 4 x bfloat> %vb, <vscale x 4 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: vfmax_vv_nxv4bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m1, ta, ma
 ; CHECK-NEXT:    vmv1r.v v10, v0
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
 ; CHECK-NEXT:    vmfeq.vv v8, v12, v12, v0.t
-; CHECK-NEXT:    vsetvli a1, zero, e16, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e16, m1, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v14, v9
 ; CHECK-NEXT:    vmv1r.v v0, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
 ; CHECK-NEXT:    vmerge.vvm v16, v12, v14, v0
 ; CHECK-NEXT:    vmv1r.v v0, v10
 ; CHECK-NEXT:    vmfeq.vv v8, v14, v14, v0.t
@@ -139,13 +139,13 @@ define <vscale x 4 x bfloat> @vfmax_vv_nxv4bf16(<vscale x 4 x bfloat> %va, <vsca
 define <vscale x 4 x bfloat> @vfmax_vv_nxv4bf16_unmasked(<vscale x 4 x bfloat> %va, <vscale x 4 x bfloat> %vb, i32 zeroext %evl) {
 ; CHECK-LABEL: vfmax_vv_nxv4bf16_unmasked:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m1, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
 ; CHECK-NEXT:    vmfeq.vv v0, v10, v10
-; CHECK-NEXT:    vsetvli a1, zero, e16, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e16, m1, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
 ; CHECK-NEXT:    vmerge.vvm v8, v10, v12, v0
 ; CHECK-NEXT:    vmfeq.vv v0, v12, v12
 ; CHECK-NEXT:    vmerge.vvm v10, v12, v10, v0
@@ -162,15 +162,15 @@ declare <vscale x 8 x bfloat> @llvm.vp.maximum.nxv8bf16(<vscale x 8 x bfloat>, <
 define <vscale x 8 x bfloat> @vfmax_vv_nxv8bf16(<vscale x 8 x bfloat> %va, <vscale x 8 x bfloat> %vb, <vscale x 8 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: vfmax_vv_nxv8bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vmv1r.v v12, v0
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmfeq.vv v8, v16, v16, v0.t
-; CHECK-NEXT:    vsetvli a1, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e16, m2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v20, v10
 ; CHECK-NEXT:    vmv1r.v v0, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmerge.vvm v24, v16, v20, v0
 ; CHECK-NEXT:    vmv1r.v v0, v12
 ; CHECK-NEXT:    vmfeq.vv v8, v20, v20, v0.t
@@ -188,13 +188,13 @@ define <vscale x 8 x bfloat> @vfmax_vv_nxv8bf16(<vscale x 8 x bfloat> %va, <vsca
 define <vscale x 8 x bfloat> @vfmax_vv_nxv8bf16_unmasked(<vscale x 8 x bfloat> %va, <vscale x 8 x bfloat> %vb, i32 zeroext %evl) {
 ; CHECK-LABEL: vfmax_vv_nxv8bf16_unmasked:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmfeq.vv v0, v12, v12
-; CHECK-NEXT:    vsetvli a1, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e16, m2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v10
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmerge.vvm v8, v12, v16, v0
 ; CHECK-NEXT:    vmfeq.vv v0, v16, v16
 ; CHECK-NEXT:    vmerge.vvm v12, v16, v12, v0
@@ -217,15 +217,15 @@ define <vscale x 16 x bfloat> @vfmax_vv_nxv16bf16(<vscale x 16 x bfloat> %va, <v
 ; CHECK-NEXT:    slli a1, a1, 3
 ; CHECK-NEXT:    sub sp, sp, a1
 ; CHECK-NEXT:    .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x08, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 8 * vlenb
-; CHECK-NEXT:    vsetvli a1, zero, e16, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m4, ta, ma
 ; CHECK-NEXT:    vmv1r.v v7, v0
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vmfeq.vv v8, v24, v24, v0.t
-; CHECK-NEXT:    vsetvli a1, zero, e16, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v12
 ; CHECK-NEXT:    vmv1r.v v0, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vmerge.vvm v8, v24, v16, v0
 ; CHECK-NEXT:    addi a0, sp, 16
 ; CHECK-NEXT:    vs8r.v v8, (a0) # Unknown-size Folded Spill
@@ -258,10 +258,10 @@ define <vscale x 16 x bfloat> @vfmax_vv_nxv16bf16_unmasked(<vscale x 16 x bfloat
 ; CHECK-NEXT:    slli a1, a1, 3
 ; CHECK-NEXT:    sub sp, sp, a1
 ; CHECK-NEXT:    .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x08, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 8 * vlenb
-; CHECK-NEXT:    vsetvli a1, zero, e16, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v12
-; CHECK-NEXT:    vsetvli zero, a0, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vmfeq.vv v0, v16, v16
 ; CHECK-NEXT:    vmfeq.vv v7, v24, v24
 ; CHECK-NEXT:    vmerge.vvm v8, v16, v24, v0
@@ -296,64 +296,62 @@ define <vscale x 32 x bfloat> @vfmax_vv_nxv32bf16(<vscale x 32 x bfloat> %va, <v
 ; CHECK-NEXT:    add a1, a2, a1
 ; CHECK-NEXT:    sub sp, sp, a1
 ; CHECK-NEXT:    .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x21, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 33 * vlenb
-; CHECK-NEXT:    csrr a2, vlenb
-; CHECK-NEXT:    vsetvli a1, zero, e16, m4, ta, ma
-; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v12
 ; CHECK-NEXT:    csrr a1, vlenb
-; CHECK-NEXT:    li a3, 24
-; CHECK-NEXT:    mul a1, a1, a3
+; CHECK-NEXT:    li a2, 25
+; CHECK-NEXT:    mul a1, a1, a2
 ; CHECK-NEXT:    add a1, sp, a1
 ; CHECK-NEXT:    addi a1, a1, 16
-; CHECK-NEXT:    vs8r.v v24, (a1) # Unknown-size Folded Spill
+; CHECK-NEXT:    vs8r.v v16, (a1) # Unknown-size Folded Spill
+; CHECK-NEXT:    vsetvli a1, zero, e8, mf2, ta, ma
+; CHECK-NEXT:    vmv8r.v v16, v8
+; CHECK-NEXT:    csrr a2, vlenb
 ; CHECK-NEXT:    slli a1, a2, 1
 ; CHECK-NEXT:    srli a2, a2, 2
 ; CHECK-NEXT:    sub a3, a0, a1
 ; CHECK-NEXT:    csrr a4, vlenb
-; CHECK-NEXT:    slli a4, a4, 5
+; CHECK-NEXT:    li a5, 24
+; CHECK-NEXT:    mul a4, a4, a5
 ; CHECK-NEXT:    add a4, sp, a4
 ; CHECK-NEXT:    addi a4, a4, 16
 ; CHECK-NEXT:    vs1r.v v0, (a4) # Unknown-size Folded Spill
-; CHECK-NEXT:    vsetvli a4, zero, e8, mf2, ta, ma
-; CHECK-NEXT:    vslidedown.vx v12, v0, a2
+; CHECK-NEXT:    vslidedown.vx v0, v0, a2
+; CHECK-NEXT:    addi a2, sp, 16
+; CHECK-NEXT:    vs1r.v v0, (a2) # Unknown-size Folded Spill
 ; CHECK-NEXT:    sltu a2, a0, a3
 ; CHECK-NEXT:    addi a2, a2, -1
 ; CHECK-NEXT:    and a2, a2, a3
-; CHECK-NEXT:    vmv1r.v v0, v12
-; CHECK-NEXT:    vsetvli zero, a2, e32, m8, ta, ma
-; CHECK-NEXT:    vmfeq.vv v13, v24, v24, v0.t
-; CHECK-NEXT:    vmv8r.v v0, v16
 ; CHECK-NEXT:    csrr a3, vlenb
 ; CHECK-NEXT:    slli a3, a3, 4
 ; CHECK-NEXT:    add a3, sp, a3
 ; CHECK-NEXT:    addi a3, a3, 16
-; CHECK-NEXT:    vs8r.v v16, (a3) # Unknown-size Folded Spill
-; CHECK-NEXT:    vsetvli a3, zero, e16, m4, ta, ma
-; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v4
-; CHECK-NEXT:    vmv1r.v v0, v13
-; CHECK-NEXT:    csrr a3, vlenb
-; CHECK-NEXT:    li a4, 24
-; CHECK-NEXT:    mul a3, a3, a4
-; CHECK-NEXT:    add a3, sp, a3
-; CHECK-NEXT:    addi a3, a3, 16
-; CHECK-NEXT:    vl8r.v v24, (a3) # Unknown-size Folded Reload
-; CHECK-NEXT:    vsetvli zero, a2, e32, m8, ta, ma
-; CHECK-NEXT:    vmerge.vvm v24, v24, v16, v0
+; CHECK-NEXT:    vs8r.v v8, (a3) # Unknown-size Folded Spill
+; CHECK-NEXT:    vsetvli zero, a2, e16, m4, ta, ma
+; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v20
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
+; CHECK-NEXT:    vmfeq.vv v12, v24, v24, v0.t
 ; CHECK-NEXT:    csrr a2, vlenb
-; CHECK-NEXT:    slli a2, a2, 3
+; CHECK-NEXT:    li a3, 25
+; CHECK-NEXT:    mul a2, a2, a3
 ; CHECK-NEXT:    add a2, sp, a2
 ; CHECK-NEXT:    addi a2, a2, 16
-; CHECK-NEXT:    vs8r.v v24, (a2) # Unknown-size Folded Spill
+; CHECK-NEXT:    vl8r.v v0, (a2) # Unknown-size Folded Reload
+; CHECK-NEXT:    vsetvli zero, zero, e16, m4, ta, ma
+; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v4
 ; CHECK-NEXT:    vmv1r.v v0, v12
-; CHECK-NEXT:    vmfeq.vv v13, v16, v16, v0.t
-; CHECK-NEXT:    vmv1r.v v0, v13
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
+; CHECK-NEXT:    vmerge.vvm v8, v24, v16, v0
 ; CHECK-NEXT:    csrr a2, vlenb
-; CHECK-NEXT:    li a3, 24
-; CHECK-NEXT:    mul a2, a2, a3
+; CHECK-NEXT:    slli a2, a2, 3
 ; CHECK-NEXT:    add a2, sp, a2
 ; CHECK-NEXT:    addi a2, a2, 16
-; CHECK-NEXT:    vl8r.v v24, (a2) # Unknown-size Folded Reload
-; CHECK-NEXT:    vmerge.vvm v16, v16, v24, v0
+; CHECK-NEXT:    vs8r.v v8, (a2) # Unknown-size Folded Spill
+; CHECK-NEXT:    addi a2, sp, 16
+; CHECK-NEXT:    vl1r.v v8, (a2) # Unknown-size Folded Reload
+; CHECK-NEXT:    vmv1r.v v0, v8
+; CHECK-NEXT:    vmfeq.vv v12, v16, v16, v0.t
 ; CHECK-NEXT:    vmv1r.v v0, v12
+; CHECK-NEXT:    vmerge.vvm v16, v16, v24, v0
+; CHECK-NEXT:    vmv1r.v v0, v8
 ; CHECK-NEXT:    csrr a2, vlenb
 ; CHECK-NEXT:    slli a2, a2, 3
 ; CHECK-NEXT:    add a2, sp, a2
@@ -371,35 +369,43 @@ define <vscale x 32 x bfloat> @vfmax_vv_nxv32bf16(<vscale x 32 x bfloat> %va, <v
 ; CHECK-NEXT:  # %bb.1:
 ; CHECK-NEXT:    mv a0, a1
 ; CHECK-NEXT:  .LBB10_2:
-; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v8
-; CHECK-NEXT:    csrr a1, vlenb
-; CHECK-NEXT:    slli a1, a1, 5
-; CHECK-NEXT:    add a1, sp, a1
-; CHECK-NEXT:    addi a1, a1, 16
-; CHECK-NEXT:    vl1r.v v0, (a1) # Unknown-size Folded Reload
-; CHECK-NEXT:    vsetvli zero, a0, e32, m8, ta, ma
-; CHECK-NEXT:    vmfeq.vv v24, v16, v16, v0.t
-; CHECK-NEXT:    vmv8r.v v8, v16
-; CHECK-NEXT:    addi a1, sp, 16
-; CHECK-NEXT:    vs8r.v v16, (a1) # Unknown-size Folded Spill
 ; CHECK-NEXT:    csrr a1, vlenb
 ; CHECK-NEXT:    slli a1, a1, 4
 ; CHECK-NEXT:    add a1, sp, a1
 ; CHECK-NEXT:    addi a1, a1, 16
-; CHECK-NEXT:    vl8r.v v0, (a1) # Unknown-size Folded Reload
-; CHECK-NEXT:    vsetvli a1, zero, e16, m4, ta, ma
-; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v0
-; CHECK-NEXT:    vmv1r.v v0, v24
-; CHECK-NEXT:    vsetvli zero, a0, e32, m8, ta, ma
-; CHECK-NEXT:    vmerge.vvm v24, v8, v16, v0
+; CHECK-NEXT:    vl8r.v v16, (a1) # Unknown-size Folded Reload
+; CHECK-NEXT:    vsetvli zero, a0, e16, m4, ta, ma
+; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v16
 ; CHECK-NEXT:    csrr a0, vlenb
 ; CHECK-NEXT:    li a1, 24
 ; CHECK-NEXT:    mul a0, a0, a1
 ; CHECK-NEXT:    add a0, sp, a0
 ; CHECK-NEXT:    addi a0, a0, 16
+; CHECK-NEXT:    vl1r.v v0, (a0) # Unknown-size Folded Reload
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
+; CHECK-NEXT:    vmfeq.vv v8, v24, v24, v0.t
+; CHECK-NEXT:    addi a0, sp, 16
 ; CHECK-NEXT:    vs8r.v v24, (a0) # Unknown-size Folded Spill
 ; CHECK-NEXT:    csrr a0, vlenb
-; CHECK-NEXT:    slli a0, a0, 5
+; CHECK-NEXT:    li a1, 25
+; CHECK-NEXT:    mul a0, a0, a1
+; CHECK-NEXT:    add a0, sp, a0
+; CHECK-NEXT:    addi a0, a0, 16
+; CHECK-NEXT:    vl8r.v v0, (a0) # Unknown-size Folded Reload
+; CHECK-NEXT:    vsetvli zero, zero, e16, m4, ta, ma
+; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v0
+; CHECK-NEXT:    vmv1r.v v0, v8
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
+; CHECK-NEXT:    vmerge.vvm v24, v24, v16, v0
+; CHECK-NEXT:    csrr a0, vlenb
+; CHECK-NEXT:    li a1, 25
+; CHECK-NEXT:    mul a0, a0, a1
+; CHECK-NEXT:    add a0, sp, a0
+; CHECK-NEXT:    addi a0, a0, 16
+; CHECK-NEXT:    vs8r.v v24, (a0) # Unknown-size Folded Spill
+; CHECK-NEXT:    csrr a0, vlenb
+; CHECK-NEXT:    li a1, 24
+; CHECK-NEXT:    mul a0, a0, a1
 ; CHECK-NEXT:    add a0, sp, a0
 ; CHECK-NEXT:    addi a0, a0, 16
 ; CHECK-NEXT:    vl1r.v v9, (a0) # Unknown-size Folded Reload
@@ -411,7 +417,7 @@ define <vscale x 32 x bfloat> @vfmax_vv_nxv32bf16(<vscale x 32 x bfloat> %va, <v
 ; CHECK-NEXT:    vmerge.vvm v16, v16, v24, v0
 ; CHECK-NEXT:    vmv1r.v v0, v9
 ; CHECK-NEXT:    csrr a0, vlenb
-; CHECK-NEXT:    li a1, 24
+; CHECK-NEXT:    li a1, 25
 ; CHECK-NEXT:    mul a0, a0, a1
 ; CHECK-NEXT:    add a0, sp, a0
 ; CHECK-NEXT:    addi a0, a0, 16
@@ -442,68 +448,61 @@ define <vscale x 32 x bfloat> @vfmax_vv_nxv32bf16_unmasked(<vscale x 32 x bfloat
 ; CHECK-NEXT:    addi sp, sp, -16
 ; CHECK-NEXT:    .cfi_def_cfa_offset 16
 ; CHECK-NEXT:    csrr a1, vlenb
-; CHECK-NEXT:    slli a1, a1, 5
+; CHECK-NEXT:    li a2, 25
+; CHECK-NEXT:    mul a1, a1, a2
 ; CHECK-NEXT:    sub sp, sp, a1
-; CHECK-NEXT:    .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x20, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 32 * vlenb
-; CHECK-NEXT:    csrr a2, vlenb
+; CHECK-NEXT:    .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x19, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 25 * vlenb
 ; CHECK-NEXT:    vsetvli a1, zero, e8, m4, ta, ma
-; CHECK-NEXT:    vmset.m v7
-; CHECK-NEXT:    addi a1, sp, 16
-; CHECK-NEXT:    vs8r.v v8, (a1) # Unknown-size Folded Spill
-; CHECK-NEXT:    vsetvli a1, zero, e16, m4, ta, ma
-; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v12
-; CHECK-NEXT:    csrr a1, vlenb
-; CHECK-NEXT:    li a3, 24
-; CHECK-NEXT:    mul a1, a1, a3
-; CHECK-NEXT:    add a1, sp, a1
-; CHECK-NEXT:    addi a1, a1, 16
-; CHECK-NEXT:    vs8r.v v24, (a1) # Unknown-size Folded Spill
+; CHECK-NEXT:    vmv8r.v v0, v8
+; CHECK-NEXT:    csrr a2, vlenb
+; CHECK-NEXT:    vmset.m v24
 ; CHECK-NEXT:    slli a1, a2, 1
 ; CHECK-NEXT:    srli a2, a2, 2
 ; CHECK-NEXT:    sub a3, a0, a1
 ; CHECK-NEXT:    vsetvli a4, zero, e8, mf2, ta, ma
-; CHECK-NEXT:    vslidedown.vx v12, v7, a2
+; CHECK-NEXT:    vslidedown.vx v8, v24, a2
+; CHECK-NEXT:    addi a2, sp, 16
+; CHECK-NEXT:    vs1r.v v8, (a2) # Unknown-size Folded Spill
 ; CHECK-NEXT:    sltu a2, a0, a3
 ; CHECK-NEXT:    addi a2, a2, -1
 ; CHECK-NEXT:    and a2, a2, a3
-; CHECK-NEXT:    vmv1r.v v0, v12
-; CHECK-NEXT:    vsetvli zero, a2, e32, m8, ta, ma
-; CHECK-NEXT:    vmfeq.vv v13, v24, v24, v0.t
-; CHECK-NEXT:    vmv8r.v v0, v16
 ; CHECK-NEXT:    csrr a3, vlenb
-; CHECK-NEXT:    slli a3, a3, 4
 ; CHECK-NEXT:    add a3, sp, a3
 ; CHECK-NEXT:    addi a3, a3, 16
-; CHECK-NEXT:    vs8r.v v16, (a3) # Unknown-size Folded Spill
-; CHECK-NEXT:    vsetvli a3, zero, e16, m4, ta, ma
-; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v4
-; CHECK-NEXT:    vmv1r.v v0, v13
-; CHECK-NEXT:    csrr a3, vlenb
-; CHECK-NEXT:    li a4, 24
-; CHECK-NEXT:    mul a3, a3, a4
-; CHECK-NEXT:    add a3, sp, a3
-; CHECK-NEXT:    addi a3, a3, 16
-; CHECK-NEXT:    vl8r.v v24, (a3) # Unknown-size Folded Reload
-; CHECK-NEXT:    vsetvli zero, a2, e32, m8, ta, ma
-; CHECK-NEXT:    vmerge.vvm v24, v24, v16, v0
+; CHECK-NEXT:    vs8r.v v0, (a3) # Unknown-size Folded Spill
+; CHECK-NEXT:    vsetvli zero, a2, e16, m4, ta, ma
+; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v4
+; CHECK-NEXT:    vmv1r.v v0, v8
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
+; CHECK-NEXT:    vmfeq.vv v12, v24, v24, v0.t
+; CHECK-NEXT:    vmv8r.v v0, v16
 ; CHECK-NEXT:    csrr a2, vlenb
-; CHECK-NEXT:    slli a2, a2, 3
+; CHECK-NEXT:    slli a3, a2, 4
+; CHECK-NEXT:    add a2, a3, a2
 ; CHECK-NEXT:    add a2, sp, a2
 ; CHECK-NEXT:    addi a2, a2, 16
-; CHECK-NEXT:    vs8r.v v24, (a2) # Unknown-size Folded Spill
+; CHECK-NEXT:    vs8r.v v16, (a2) # Unknown-size Folded Spill
+; CHECK-NEXT:    vsetvli zero, zero, e16, m4, ta, ma
+; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v4
 ; CHECK-NEXT:    vmv1r.v v0, v12
-; CHECK-NEXT:    vmfeq.vv v13, v16, v16, v0.t
-; CHECK-NEXT:    vmv1r.v v0, v13
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
+; CHECK-NEXT:    vmerge.vvm v8, v24, v16, v0
 ; CHECK-NEXT:    csrr a2, vlenb
-; CHECK-NEXT:    li a3, 24
-; CHECK-NEXT:    mul a2, a2, a3
+; CHECK-NEXT:    slli a3, a2, 3
+; CHECK-NEXT:    add a2, a3, a2
 ; CHECK-NEXT:    add a2, sp, a2
 ; CHECK-NEXT:    addi a2, a2, 16
-; CHECK-NEXT:    vl8r.v v24, (a2) # Unknown-size Folded Reload
-; CHECK-NEXT:    vmerge.vvm v16, v16, v24, v0
+; CHECK-NEXT:    vs8r.v v8, (a2) # Unknown-size Folded Spill
+; CHECK-NEXT:    addi a2, sp, 16
+; CHECK-NEXT:    vl1r.v v13, (a2) # Unknown-size Folded Reload
+; CHECK-NEXT:    vmv1r.v v0, v13
+; CHECK-NEXT:    vmfeq.vv v12, v16, v16, v0.t
 ; CHECK-NEXT:    vmv1r.v v0, v12
+; CHECK-NEXT:    vmerge.vvm v16, v16, v24, v0
+; CHECK-NEXT:    vmv1r.v v0, v13
 ; CHECK-NEXT:    csrr a2, vlenb
-; CHECK-NEXT:    slli a2, a2, 3
+; CHECK-NEXT:    slli a3, a2, 3
+; CHECK-NEXT:    add a2, a3, a2
 ; CHECK-NEXT:    add a2, sp, a2
 ; CHECK-NEXT:    addi a2, a2, 16
 ; CHECK-NEXT:    vl8r.v v24, (a2) # Unknown-size Folded Reload
@@ -511,7 +510,8 @@ define <vscale x 32 x bfloat> @vfmax_vv_nxv32bf16_unmasked(<vscale x 32 x bfloat
 ; CHECK-NEXT:    vsetvli a2, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v12, v16
 ; CHECK-NEXT:    csrr a2, vlenb
-; CHECK-NEXT:    slli a2, a2, 3
+; CHECK-NEXT:    slli a3, a2, 3
+; CHECK-NEXT:    add a2, a3, a2
 ; CHECK-NEXT:    add a2, sp, a2
 ; CHECK-NEXT:    addi a2, a2, 16
 ; CHECK-NEXT:    vs8r.v v8, (a2) # Unknown-size Folded Spill
@@ -519,43 +519,49 @@ define <vscale x 32 x bfloat> @vfmax_vv_nxv32bf16_unmasked(<vscale x 32 x bfloat
 ; CHECK-NEXT:  # %bb.1:
 ; CHECK-NEXT:    mv a0, a1
 ; CHECK-NEXT:  .LBB11_2:
-; CHECK-NEXT:    addi a1, sp, 16
-; CHECK-NEXT:    vl8r.v v16, (a1) # Unknown-size Folded Reload
-; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v16
 ; CHECK-NEXT:    csrr a1, vlenb
-; CHECK-NEXT:    slli a1, a1, 4
 ; CHECK-NEXT:    add a1, sp, a1
 ; CHECK-NEXT:    addi a1, a1, 16
-; CHECK-NEXT:    vl8r.v v0, (a1) # Unknown-size Folded Reload
+; CHECK-NEXT:    vl8r.v v16, (a1) # Unknown-size Folded Reload
+; CHECK-NEXT:    vsetvli zero, a0, e16, m4, ta, ma
+; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v16
+; CHECK-NEXT:    csrr a0, vlenb
+; CHECK-NEXT:    slli a1, a0, 4
+; CHECK-NEXT:    add a0, a1, a0
+; CHECK-NEXT:    add a0, sp, a0
+; CHECK-NEXT:    addi a0, a0, 16
+; CHECK-NEXT:    vl8r.v v0, (a0) # Unknown-size Folded Reload
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v0
-; CHECK-NEXT:    vsetvli zero, a0, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vmfeq.vv v0, v8, v8
 ; CHECK-NEXT:    vmfeq.vv v7, v16, v16
 ; CHECK-NEXT:    vmerge.vvm v24, v8, v16, v0
 ; CHECK-NEXT:    csrr a0, vlenb
-; CHECK-NEXT:    li a1, 24
-; CHECK-NEXT:    mul a0, a0, a1
+; CHECK-NEXT:    slli a1, a0, 4
+; CHECK-NEXT:    add a0, a1, a0
 ; CHECK-NEXT:    add a0, sp, a0
 ; CHECK-NEXT:    addi a0, a0, 16
 ; CHECK-NEXT:    vs8r.v v24, (a0) # Unknown-size Folded Spill
 ; CHECK-NEXT:    vmv1r.v v0, v7
 ; CHECK-NEXT:    vmerge.vvm v16, v16, v8, v0
 ; CHECK-NEXT:    csrr a0, vlenb
-; CHECK-NEXT:    li a1, 24
-; CHECK-NEXT:    mul a0, a0, a1
+; CHECK-NEXT:    slli a1, a0, 4
+; CHECK-NEXT:    add a0, a1, a0
 ; CHECK-NEXT:    add a0, sp, a0
 ; CHECK-NEXT:    addi a0, a0, 16
 ; CHECK-NEXT:    vl8r.v v24, (a0) # Unknown-size Folded Reload
 ; CHECK-NEXT:    vfmax.vv v16, v16, v24
 ; CHECK-NEXT:    csrr a0, vlenb
-; CHECK-NEXT:    slli a0, a0, 3
+; CHECK-NEXT:    slli a1, a0, 3
+; CHECK-NEXT:    add a0, a1, a0
 ; CHECK-NEXT:    add a0, sp, a0
 ; CHECK-NEXT:    addi a0, a0, 16
 ; CHECK-NEXT:    vl8r.v v8, (a0) # Unknown-size Folded Reload
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v16
 ; CHECK-NEXT:    csrr a0, vlenb
-; CHECK-NEXT:    slli a0, a0, 5
+; CHECK-NEXT:    li a1, 25
+; CHECK-NEXT:    mul a0, a0, a1
 ; CHECK-NEXT:    add sp, sp, a0
 ; CHECK-NEXT:    .cfi_def_cfa sp, 16
 ; CHECK-NEXT:    addi sp, sp, 16
diff --git a/llvm/test/CodeGen/RISCV/rvv/fminimum-vp.ll b/llvm/test/CodeGen/RISCV/rvv/fminimum-vp.ll
index c65712e9965aa..d877073c3487d 100644
--- a/llvm/test/CodeGen/RISCV/rvv/fminimum-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/fminimum-vp.ll
@@ -19,14 +19,14 @@ declare <vscale x 1 x bfloat> @llvm.vp.minimum.nxv1bf16(<vscale x 1 x bfloat>, <
 define <vscale x 1 x bfloat> @vfmin_vv_nxv1bf16(<vscale x 1 x bfloat> %va, <vscale x 1 x bfloat> %vb, <vscale x 1 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: vfmin_vv_nxv1bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vmv1r.v v10, v0
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v11, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmfeq.vv v0, v11, v11, v0.t
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmerge.vvm v9, v11, v8, v0
 ; CHECK-NEXT:    vmv1r.v v0, v10
 ; CHECK-NEXT:    vmfeq.vv v0, v8, v8, v0.t
@@ -43,13 +43,13 @@ define <vscale x 1 x bfloat> @vfmin_vv_nxv1bf16(<vscale x 1 x bfloat> %va, <vsca
 define <vscale x 1 x bfloat> @vfmin_vv_nxv1bf16_unmasked(<vscale x 1 x bfloat> %va, <vscale x 1 x bfloat> %vb, i32 zeroext %evl) {
 ; CHECK-LABEL: vfmin_vv_nxv1bf16_unmasked:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmfeq.vv v0, v10, v10
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmerge.vvm v9, v10, v8, v0
 ; CHECK-NEXT:    vmfeq.vv v0, v8, v8
 ; CHECK-NEXT:    vmerge.vvm v8, v8, v10, v0
@@ -66,14 +66,14 @@ declare <vscale x 2 x bfloat> @llvm.vp.minimum.nxv2bf16(<vscale x 2 x bfloat>, <
 define <vscale x 2 x bfloat> @vfmin_vv_nxv2bf16(<vscale x 2 x bfloat> %va, <vscale x 2 x bfloat> %vb, <vscale x 2 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: vfmin_vv_nxv2bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, ma
 ; CHECK-NEXT:    vmv1r.v v10, v0
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v11, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m1, ta, ma
 ; CHECK-NEXT:    vmfeq.vv v0, v11, v11, v0.t
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e16, mf2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m1, ta, ma
 ; CHECK-NEXT:    vmerge.vvm v9, v11, v8, v0
 ; CHECK-NEXT:    vmv1r.v v0, v10
 ; CHECK-NEXT:    vmfeq.vv v0, v8, v8, v0.t
@@ -90,13 +90,13 @@ define <vscale x 2 x bfloat> @vfmin_vv_nxv2bf16(<vscale x 2 x bfloat> %va, <vsca
 define <vscale x 2 x bfloat> @vfmin_vv_nxv2bf16_unmasked(<vscale x 2 x bfloat> %va, <vscale x 2 x bfloat> %vb, i32 zeroext %evl) {
 ; CHECK-LABEL: vfmin_vv_nxv2bf16_unmasked:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m1, ta, ma
 ; CHECK-NEXT:    vmfeq.vv v0, v10, v10
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e16, mf2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m1, ta, ma
 ; CHECK-NEXT:    vmerge.vvm v9, v10, v8, v0
 ; CHECK-NEXT:    vmfeq.vv v0, v8, v8
 ; CHECK-NEXT:    vmerge.vvm v8, v8, v10, v0
@@ -113,15 +113,15 @@ declare <vscale x 4 x bfloat> @llvm.vp.minimum.nxv4bf16(<vscale x 4 x bfloat>, <
 define <vscale x 4 x bfloat> @vfmin_vv_nxv4bf16(<vscale x 4 x bfloat> %va, <vscale x 4 x bfloat> %vb, <vscale x 4 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: vfmin_vv_nxv4bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m1, ta, ma
 ; CHECK-NEXT:    vmv1r.v v10, v0
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
 ; CHECK-NEXT:    vmfeq.vv v8, v12, v12, v0.t
-; CHECK-NEXT:    vsetvli a1, zero, e16, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e16, m1, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v14, v9
 ; CHECK-NEXT:    vmv1r.v v0, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
 ; CHECK-NEXT:    vmerge.vvm v16, v12, v14, v0
 ; CHECK-NEXT:    vmv1r.v v0, v10
 ; CHECK-NEXT:    vmfeq.vv v8, v14, v14, v0.t
@@ -139,13 +139,13 @@ define <vscale x 4 x bfloat> @vfmin_vv_nxv4bf16(<vscale x 4 x bfloat> %va, <vsca
 define <vscale x 4 x bfloat> @vfmin_vv_nxv4bf16_unmasked(<vscale x 4 x bfloat> %va, <vscale x 4 x bfloat> %vb, i32 zeroext %evl) {
 ; CHECK-LABEL: vfmin_vv_nxv4bf16_unmasked:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m1, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
 ; CHECK-NEXT:    vmfeq.vv v0, v10, v10
-; CHECK-NEXT:    vsetvli a1, zero, e16, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e16, m1, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
 ; CHECK-NEXT:    vmerge.vvm v8, v10, v12, v0
 ; CHECK-NEXT:    vmfeq.vv v0, v12, v12
 ; CHECK-NEXT:    vmerge.vvm v10, v12, v10, v0
@@ -162,15 +162,15 @@ declare <vscale x 8 x bfloat> @llvm.vp.minimum.nxv8bf16(<vscale x 8 x bfloat>, <
 define <vscale x 8 x bfloat> @vfmin_vv_nxv8bf16(<vscale x 8 x bfloat> %va, <vscale x 8 x bfloat> %vb, <vscale x 8 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: vfmin_vv_nxv8bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vmv1r.v v12, v0
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmfeq.vv v8, v16, v16, v0.t
-; CHECK-NEXT:    vsetvli a1, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e16, m2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v20, v10
 ; CHECK-NEXT:    vmv1r.v v0, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmerge.vvm v24, v16, v20, v0
 ; CHECK-NEXT:    vmv1r.v v0, v12
 ; CHECK-NEXT:    vmfeq.vv v8, v20, v20, v0.t
@@ -188,13 +188,13 @@ define <vscale x 8 x bfloat> @vfmin_vv_nxv8bf16(<vscale x 8 x bfloat> %va, <vsca
 define <vscale x 8 x bfloat> @vfmin_vv_nxv8bf16_unmasked(<vscale x 8 x bfloat> %va, <vscale x 8 x bfloat> %vb, i32 zeroext %evl) {
 ; CHECK-LABEL: vfmin_vv_nxv8bf16_unmasked:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmfeq.vv v0, v12, v12
-; CHECK-NEXT:    vsetvli a1, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e16, m2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v10
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmerge.vvm v8, v12, v16, v0
 ; CHECK-NEXT:    vmfeq.vv v0, v16, v16
 ; CHECK-NEXT:    vmerge.vvm v12, v16, v12, v0
@@ -217,15 +217,15 @@ define <vscale x 16 x bfloat> @vfmin_vv_nxv16bf16(<vscale x 16 x bfloat> %va, <v
 ; CHECK-NEXT:    slli a1, a1, 3
 ; CHECK-NEXT:    sub sp, sp, a1
 ; CHECK-NEXT:    .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x08, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 8 * vlenb
-; CHECK-NEXT:    vsetvli a1, zero, e16, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m4, ta, ma
 ; CHECK-NEXT:    vmv1r.v v7, v0
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vmfeq.vv v8, v24, v24, v0.t
-; CHECK-NEXT:    vsetvli a1, zero, e16, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v12
 ; CHECK-NEXT:    vmv1r.v v0, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vmerge.vvm v8, v24, v16, v0
 ; CHECK-NEXT:    addi a0, sp, 16
 ; CHECK-NEXT:    vs8r.v v8, (a0) # Unknown-size Folded Spill
@@ -258,10 +258,10 @@ define <vscale x 16 x bfloat> @vfmin_vv_nxv16bf16_unmasked(<vscale x 16 x bfloat
 ; CHECK-NEXT:    slli a1, a1, 3
 ; CHECK-NEXT:    sub sp, sp, a1
 ; CHECK-NEXT:    .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x08, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 8 * vlenb
-; CHECK-NEXT:    vsetvli a1, zero, e16, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v12
-; CHECK-NEXT:    vsetvli zero, a0, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vmfeq.vv v0, v16, v16
 ; CHECK-NEXT:    vmfeq.vv v7, v24, v24
 ; CHECK-NEXT:    vmerge.vvm v8, v16, v24, v0
@@ -296,64 +296,62 @@ define <vscale x 32 x bfloat> @vfmin_vv_nxv32bf16(<vscale x 32 x bfloat> %va, <v
 ; CHECK-NEXT:    add a1, a2, a1
 ; CHECK-NEXT:    sub sp, sp, a1
 ; CHECK-NEXT:    .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x21, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 33 * vlenb
-; CHECK-NEXT:    csrr a2, vlenb
-; CHECK-NEXT:    vsetvli a1, zero, e16, m4, ta, ma
-; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v12
 ; CHECK-NEXT:    csrr a1, vlenb
-; CHECK-NEXT:    li a3, 24
-; CHECK-NEXT:    mul a1, a1, a3
+; CHECK-NEXT:    li a2, 25
+; CHECK-NEXT:    mul a1, a1, a2
 ; CHECK-NEXT:    add a1, sp, a1
 ; CHECK-NEXT:    addi a1, a1, 16
-; CHECK-NEXT:    vs8r.v v24, (a1) # Unknown-size Folded Spill
+; CHECK-NEXT:    vs8r.v v16, (a1) # Unknown-size Folded Spill
+; CHECK-NEXT:    vsetvli a1, zero, e8, mf2, ta, ma
+; CHECK-NEXT:    vmv8r.v v16, v8
+; CHECK-NEXT:    csrr a2, vlenb
 ; CHECK-NEXT:    slli a1, a2, 1
 ; CHECK-NEXT:    srli a2, a2, 2
 ; CHECK-NEXT:    sub a3, a0, a1
 ; CHECK-NEXT:    csrr a4, vlenb
-; CHECK-NEXT:    slli a4, a4, 5
+; CHECK-NEXT:    li a5, 24
+; CHECK-NEXT:    mul a4, a4, a5
 ; CHECK-NEXT:    add a4, sp, a4
 ; CHECK-NEXT:    addi a4, a4, 16
 ; CHECK-NEXT:    vs1r.v v0, (a4) # Unknown-size Folded Spill
-; CHECK-NEXT:    vsetvli a4, zero, e8, mf2, ta, ma
-; CHECK-NEXT:    vslidedown.vx v12, v0, a2
+; CHECK-NEXT:    vslidedown.vx v0, v0, a2
+; CHECK-NEXT:    addi a2, sp, 16
+; CHECK-NEXT:    vs1r.v v0, (a2) # Unknown-size Folded Spill
 ; CHECK-NEXT:    sltu a2, a0, a3
 ; CHECK-NEXT:    addi a2, a2, -1
 ; CHECK-NEXT:    and a2, a2, a3
-; CHECK-NEXT:    vmv1r.v v0, v12
-; CHECK-NEXT:    vsetvli zero, a2, e32, m8, ta, ma
-; CHECK-NEXT:    vmfeq.vv v13, v24, v24, v0.t
-; CHECK-NEXT:    vmv8r.v v0, v16
 ; CHECK-NEXT:    csrr a3, vlenb
 ; CHECK-NEXT:    slli a3, a3, 4
 ; CHECK-NEXT:    add a3, sp, a3
 ; CHECK-NEXT:    addi a3, a3, 16
-; CHECK-NEXT:    vs8r.v v16, (a3) # Unknown-size Folded Spill
-; CHECK-NEXT:    vsetvli a3, zero, e16, m4, ta, ma
-; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v4
-; CHECK-NEXT:    vmv1r.v v0, v13
-; CHECK-NEXT:    csrr a3, vlenb
-; CHECK-NEXT:    li a4, 24
-; CHECK-NEXT:    mul a3, a3, a4
-; CHECK-NEXT:    add a3, sp, a3
-; CHECK-NEXT:    addi a3, a3, 16
-; CHECK-NEXT:    vl8r.v v24, (a3) # Unknown-size Folded Reload
-; CHECK-NEXT:    vsetvli zero, a2, e32, m8, ta, ma
-; CHECK-NEXT:    vmerge.vvm v24, v24, v16, v0
+; CHECK-NEXT:    vs8r.v v8, (a3) # Unknown-size Folded Spill
+; CHECK-NEXT:    vsetvli zero, a2, e16, m4, ta, ma
+; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v20
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
+; CHECK-NEXT:    vmfeq.vv v12, v24, v24, v0.t
 ; CHECK-NEXT:    csrr a2, vlenb
-; CHECK-NEXT:    slli a2, a2, 3
+; CHECK-NEXT:    li a3, 25
+; CHECK-NEXT:    mul a2, a2, a3
 ; CHECK-NEXT:    add a2, sp, a2
 ; CHECK-NEXT:    addi a2, a2, 16
-; CHECK-NEXT:    vs8r.v v24, (a2) # Unknown-size Folded Spill
+; CHECK-NEXT:    vl8r.v v0, (a2) # Unknown-size Folded Reload
+; CHECK-NEXT:    vsetvli zero, zero, e16, m4, ta, ma
+; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v4
 ; CHECK-NEXT:    vmv1r.v v0, v12
-; CHECK-NEXT:    vmfeq.vv v13, v16, v16, v0.t
-; CHECK-NEXT:    vmv1r.v v0, v13
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
+; CHECK-NEXT:    vmerge.vvm v8, v24, v16, v0
 ; CHECK-NEXT:    csrr a2, vlenb
-; CHECK-NEXT:    li a3, 24
-; CHECK-NEXT:    mul a2, a2, a3
+; CHECK-NEXT:    slli a2, a2, 3
 ; CHECK-NEXT:    add a2, sp, a2
 ; CHECK-NEXT:    addi a2, a2, 16
-; CHECK-NEXT:    vl8r.v v24, (a2) # Unknown-size Folded Reload
-; CHECK-NEXT:    vmerge.vvm v16, v16, v24, v0
+; CHECK-NEXT:    vs8r.v v8, (a2) # Unknown-size Folded Spill
+; CHECK-NEXT:    addi a2, sp, 16
+; CHECK-NEXT:    vl1r.v v8, (a2) # Unknown-size Folded Reload
+; CHECK-NEXT:    vmv1r.v v0, v8
+; CHECK-NEXT:    vmfeq.vv v12, v16, v16, v0.t
 ; CHECK-NEXT:    vmv1r.v v0, v12
+; CHECK-NEXT:    vmerge.vvm v16, v16, v24, v0
+; CHECK-NEXT:    vmv1r.v v0, v8
 ; CHECK-NEXT:    csrr a2, vlenb
 ; CHECK-NEXT:    slli a2, a2, 3
 ; CHECK-NEXT:    add a2, sp, a2
@@ -371,35 +369,43 @@ define <vscale x 32 x bfloat> @vfmin_vv_nxv32bf16(<vscale x 32 x bfloat> %va, <v
 ; CHECK-NEXT:  # %bb.1:
 ; CHECK-NEXT:    mv a0, a1
 ; CHECK-NEXT:  .LBB10_2:
-; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v8
-; CHECK-NEXT:    csrr a1, vlenb
-; CHECK-NEXT:    slli a1, a1, 5
-; CHECK-NEXT:    add a1, sp, a1
-; CHECK-NEXT:    addi a1, a1, 16
-; CHECK-NEXT:    vl1r.v v0, (a1) # Unknown-size Folded Reload
-; CHECK-NEXT:    vsetvli zero, a0, e32, m8, ta, ma
-; CHECK-NEXT:    vmfeq.vv v24, v16, v16, v0.t
-; CHECK-NEXT:    vmv8r.v v8, v16
-; CHECK-NEXT:    addi a1, sp, 16
-; CHECK-NEXT:    vs8r.v v16, (a1) # Unknown-size Folded Spill
 ; CHECK-NEXT:    csrr a1, vlenb
 ; CHECK-NEXT:    slli a1, a1, 4
 ; CHECK-NEXT:    add a1, sp, a1
 ; CHECK-NEXT:    addi a1, a1, 16
-; CHECK-NEXT:    vl8r.v v0, (a1) # Unknown-size Folded Reload
-; CHECK-NEXT:    vsetvli a1, zero, e16, m4, ta, ma
-; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v0
-; CHECK-NEXT:    vmv1r.v v0, v24
-; CHECK-NEXT:    vsetvli zero, a0, e32, m8, ta, ma
-; CHECK-NEXT:    vmerge.vvm v24, v8, v16, v0
+; CHECK-NEXT:    vl8r.v v16, (a1) # Unknown-size Folded Reload
+; CHECK-NEXT:    vsetvli zero, a0, e16, m4, ta, ma
+; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v16
 ; CHECK-NEXT:    csrr a0, vlenb
 ; CHECK-NEXT:    li a1, 24
 ; CHECK-NEXT:    mul a0, a0, a1
 ; CHECK-NEXT:    add a0, sp, a0
 ; CHECK-NEXT:    addi a0, a0, 16
+; CHECK-NEXT:    vl1r.v v0, (a0) # Unknown-size Folded Reload
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
+; CHECK-NEXT:    vmfeq.vv v8, v24, v24, v0.t
+; CHECK-NEXT:    addi a0, sp, 16
 ; CHECK-NEXT:    vs8r.v v24, (a0) # Unknown-size Folded Spill
 ; CHECK-NEXT:    csrr a0, vlenb
-; CHECK-NEXT:    slli a0, a0, 5
+; CHECK-NEXT:    li a1, 25
+; CHECK-NEXT:    mul a0, a0, a1
+; CHECK-NEXT:    add a0, sp, a0
+; CHECK-NEXT:    addi a0, a0, 16
+; CHECK-NEXT:    vl8r.v v0, (a0) # Unknown-size Folded Reload
+; CHECK-NEXT:    vsetvli zero, zero, e16, m4, ta, ma
+; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v0
+; CHECK-NEXT:    vmv1r.v v0, v8
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
+; CHECK-NEXT:    vmerge.vvm v24, v24, v16, v0
+; CHECK-NEXT:    csrr a0, vlenb
+; CHECK-NEXT:    li a1, 25
+; CHECK-NEXT:    mul a0, a0, a1
+; CHECK-NEXT:    add a0, sp, a0
+; CHECK-NEXT:    addi a0, a0, 16
+; CHECK-NEXT:    vs8r.v v24, (a0) # Unknown-size Folded Spill
+; CHECK-NEXT:    csrr a0, vlenb
+; CHECK-NEXT:    li a1, 24
+; CHECK-NEXT:    mul a0, a0, a1
 ; CHECK-NEXT:    add a0, sp, a0
 ; CHECK-NEXT:    addi a0, a0, 16
 ; CHECK-NEXT:    vl1r.v v9, (a0) # Unknown-size Folded Reload
@@ -411,7 +417,7 @@ define <vscale x 32 x bfloat> @vfmin_vv_nxv32bf16(<vscale x 32 x bfloat> %va, <v
 ; CHECK-NEXT:    vmerge.vvm v16, v16, v24, v0
 ; CHECK-NEXT:    vmv1r.v v0, v9
 ; CHECK-NEXT:    csrr a0, vlenb
-; CHECK-NEXT:    li a1, 24
+; CHECK-NEXT:    li a1, 25
 ; CHECK-NEXT:    mul a0, a0, a1
 ; CHECK-NEXT:    add a0, sp, a0
 ; CHECK-NEXT:    addi a0, a0, 16
@@ -442,68 +448,61 @@ define <vscale x 32 x bfloat> @vfmin_vv_nxv32bf16_unmasked(<vscale x 32 x bfloat
 ; CHECK-NEXT:    addi sp, sp, -16
 ; CHECK-NEXT:    .cfi_def_cfa_offset 16
 ; CHECK-NEXT:    csrr a1, vlenb
-; CHECK-NEXT:    slli a1, a1, 5
+; CHECK-NEXT:    li a2, 25
+; CHECK-NEXT:    mul a1, a1, a2
 ; CHECK-NEXT:    sub sp, sp, a1
-; CHECK-NEXT:    .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x20, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 32 * vlenb
-; CHECK-NEXT:    csrr a2, vlenb
+; CHECK-NEXT:    .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x19, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 25 * vlenb
 ; CHECK-NEXT:    vsetvli a1, zero, e8, m4, ta, ma
-; CHECK-NEXT:    vmset.m v7
-; CHECK-NEXT:    addi a1, sp, 16
-; CHECK-NEXT:    vs8r.v v8, (a1) # Unknown-size Folded Spill
-; CHECK-NEXT:    vsetvli a1, zero, e16, m4, ta, ma
-; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v12
-; CHECK-NEXT:    csrr a1, vlenb
-; CHECK-NEXT:    li a3, 24
-; CHECK-NEXT:    mul a1, a1, a3
-; CHECK-NEXT:    add a1, sp, a1
-; CHECK-NEXT:    addi a1, a1, 16
-; CHECK-NEXT:    vs8r.v v24, (a1) # Unknown-size Folded Spill
+; CHECK-NEXT:    vmv8r.v v0, v8
+; CHECK-NEXT:    csrr a2, vlenb
+; CHECK-NEXT:    vmset.m v24
 ; CHECK-NEXT:    slli a1, a2, 1
 ; CHECK-NEXT:    srli a2, a2, 2
 ; CHECK-NEXT:    sub a3, a0, a1
 ; CHECK-NEXT:    vsetvli a4, zero, e8, mf2, ta, ma
-; CHECK-NEXT:    vslidedown.vx v12, v7, a2
+; CHECK-NEXT:    vslidedown.vx v8, v24, a2
+; CHECK-NEXT:    addi a2, sp, 16
+; CHECK-NEXT:    vs1r.v v8, (a2) # Unknown-size Folded Spill
 ; CHECK-NEXT:    sltu a2, a0, a3
 ; CHECK-NEXT:    addi a2, a2, -1
 ; CHECK-NEXT:    and a2, a2, a3
-; CHECK-NEXT:    vmv1r.v v0, v12
-; CHECK-NEXT:    vsetvli zero, a2, e32, m8, ta, ma
-; CHECK-NEXT:    vmfeq.vv v13, v24, v24, v0.t
-; CHECK-NEXT:    vmv8r.v v0, v16
 ; CHECK-NEXT:    csrr a3, vlenb
-; CHECK-NEXT:    slli a3, a3, 4
 ; CHECK-NEXT:    add a3, sp, a3
 ; CHECK-NEXT:    addi a3, a3, 16
-; CHECK-NEXT:    vs8r.v v16, (a3) # Unknown-size Folded Spill
-; CHECK-NEXT:    vsetvli a3, zero, e16, m4, ta, ma
-; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v4
-; CHECK-NEXT:    vmv1r.v v0, v13
-; CHECK-NEXT:    csrr a3, vlenb
-; CHECK-NEXT:    li a4, 24
-; CHECK-NEXT:    mul a3, a3, a4
-; CHECK-NEXT:    add a3, sp, a3
-; CHECK-NEXT:    addi a3, a3, 16
-; CHECK-NEXT:    vl8r.v v24, (a3) # Unknown-size Folded Reload
-; CHECK-NEXT:    vsetvli zero, a2, e32, m8, ta, ma
-; CHECK-NEXT:    vmerge.vvm v24, v24, v16, v0
+; CHECK-NEXT:    vs8r.v v0, (a3) # Unknown-size Folded Spill
+; CHECK-NEXT:    vsetvli zero, a2, e16, m4, ta, ma
+; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v4
+; CHECK-NEXT:    vmv1r.v v0, v8
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
+; CHECK-NEXT:    vmfeq.vv v12, v24, v24, v0.t
+; CHECK-NEXT:    vmv8r.v v0, v16
 ; CHECK-NEXT:    csrr a2, vlenb
-; CHECK-NEXT:    slli a2, a2, 3
+; CHECK-NEXT:    slli a3, a2, 4
+; CHECK-NEXT:    add a2, a3, a2
 ; CHECK-NEXT:    add a2, sp, a2
 ; CHECK-NEXT:    addi a2, a2, 16
-; CHECK-NEXT:    vs8r.v v24, (a2) # Unknown-size Folded Spill
+; CHECK-NEXT:    vs8r.v v16, (a2) # Unknown-size Folded Spill
+; CHECK-NEXT:    vsetvli zero, zero, e16, m4, ta, ma
+; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v4
 ; CHECK-NEXT:    vmv1r.v v0, v12
-; CHECK-NEXT:    vmfeq.vv v13, v16, v16, v0.t
-; CHECK-NEXT:    vmv1r.v v0, v13
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
+; CHECK-NEXT:    vmerge.vvm v8, v24, v16, v0
 ; CHECK-NEXT:    csrr a2, vlenb
-; CHECK-NEXT:    li a3, 24
-; CHECK-NEXT:    mul a2, a2, a3
+; CHECK-NEXT:    slli a3, a2, 3
+; CHECK-NEXT:    add a2, a3, a2
 ; CHECK-NEXT:    add a2, sp, a2
 ; CHECK-NEXT:    addi a2, a2, 16
-; CHECK-NEXT:    vl8r.v v24, (a2) # Unknown-size Folded Reload
-; CHECK-NEXT:    vmerge.vvm v16, v16, v24, v0
+; CHECK-NEXT:    vs8r.v v8, (a2) # Unknown-size Folded Spill
+; CHECK-NEXT:    addi a2, sp, 16
+; CHECK-NEXT:    vl1r.v v13, (a2) # Unknown-size Folded Reload
+; CHECK-NEXT:    vmv1r.v v0, v13
+; CHECK-NEXT:    vmfeq.vv v12, v16, v16, v0.t
 ; CHECK-NEXT:    vmv1r.v v0, v12
+; CHECK-NEXT:    vmerge.vvm v16, v16, v24, v0
+; CHECK-NEXT:    vmv1r.v v0, v13
 ; CHECK-NEXT:    csrr a2, vlenb
-; CHECK-NEXT:    slli a2, a2, 3
+; CHECK-NEXT:    slli a3, a2, 3
+; CHECK-NEXT:    add a2, a3, a2
 ; CHECK-NEXT:    add a2, sp, a2
 ; CHECK-NEXT:    addi a2, a2, 16
 ; CHECK-NEXT:    vl8r.v v24, (a2) # Unknown-size Folded Reload
@@ -511,7 +510,8 @@ define <vscale x 32 x bfloat> @vfmin_vv_nxv32bf16_unmasked(<vscale x 32 x bfloat
 ; CHECK-NEXT:    vsetvli a2, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v12, v16
 ; CHECK-NEXT:    csrr a2, vlenb
-; CHECK-NEXT:    slli a2, a2, 3
+; CHECK-NEXT:    slli a3, a2, 3
+; CHECK-NEXT:    add a2, a3, a2
 ; CHECK-NEXT:    add a2, sp, a2
 ; CHECK-NEXT:    addi a2, a2, 16
 ; CHECK-NEXT:    vs8r.v v8, (a2) # Unknown-size Folded Spill
@@ -519,43 +519,49 @@ define <vscale x 32 x bfloat> @vfmin_vv_nxv32bf16_unmasked(<vscale x 32 x bfloat
 ; CHECK-NEXT:  # %bb.1:
 ; CHECK-NEXT:    mv a0, a1
 ; CHECK-NEXT:  .LBB11_2:
-; CHECK-NEXT:    addi a1, sp, 16
-; CHECK-NEXT:    vl8r.v v16, (a1) # Unknown-size Folded Reload
-; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v16
 ; CHECK-NEXT:    csrr a1, vlenb
-; CHECK-NEXT:    slli a1, a1, 4
 ; CHECK-NEXT:    add a1, sp, a1
 ; CHECK-NEXT:    addi a1, a1, 16
-; CHECK-NEXT:    vl8r.v v0, (a1) # Unknown-size Folded Reload
+; CHECK-NEXT:    vl8r.v v16, (a1) # Unknown-size Folded Reload
+; CHECK-NEXT:    vsetvli zero, a0, e16, m4, ta, ma
+; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v16
+; CHECK-NEXT:    csrr a0, vlenb
+; CHECK-NEXT:    slli a1, a0, 4
+; CHECK-NEXT:    add a0, a1, a0
+; CHECK-NEXT:    add a0, sp, a0
+; CHECK-NEXT:    addi a0, a0, 16
+; CHECK-NEXT:    vl8r.v v0, (a0) # Unknown-size Folded Reload
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v0
-; CHECK-NEXT:    vsetvli zero, a0, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vmfeq.vv v0, v8, v8
 ; CHECK-NEXT:    vmfeq.vv v7, v16, v16
 ; CHECK-NEXT:    vmerge.vvm v24, v8, v16, v0
 ; CHECK-NEXT:    csrr a0, vlenb
-; CHECK-NEXT:    li a1, 24
-; CHECK-NEXT:    mul a0, a0, a1
+; CHECK-NEXT:    slli a1, a0, 4
+; CHECK-NEXT:    add a0, a1, a0
 ; CHECK-NEXT:    add a0, sp, a0
 ; CHECK-NEXT:    addi a0, a0, 16
 ; CHECK-NEXT:    vs8r.v v24, (a0) # Unknown-size Folded Spill
 ; CHECK-NEXT:    vmv1r.v v0, v7
 ; CHECK-NEXT:    vmerge.vvm v16, v16, v8, v0
 ; CHECK-NEXT:    csrr a0, vlenb
-; CHECK-NEXT:    li a1, 24
-; CHECK-NEXT:    mul a0, a0, a1
+; CHECK-NEXT:    slli a1, a0, 4
+; CHECK-NEXT:    add a0, a1, a0
 ; CHECK-NEXT:    add a0, sp, a0
 ; CHECK-NEXT:    addi a0, a0, 16
 ; CHECK-NEXT:    vl8r.v v24, (a0) # Unknown-size Folded Reload
 ; CHECK-NEXT:    vfmin.vv v16, v16, v24
 ; CHECK-NEXT:    csrr a0, vlenb
-; CHECK-NEXT:    slli a0, a0, 3
+; CHECK-NEXT:    slli a1, a0, 3
+; CHECK-NEXT:    add a0, a1, a0
 ; CHECK-NEXT:    add a0, sp, a0
 ; CHECK-NEXT:    addi a0, a0, 16
 ; CHECK-NEXT:    vl8r.v v8, (a0) # Unknown-size Folded Reload
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v16
 ; CHECK-NEXT:    csrr a0, vlenb
-; CHECK-NEXT:    slli a0, a0, 5
+; CHECK-NEXT:    li a1, 25
+; CHECK-NEXT:    mul a0, a0, a1
 ; CHECK-NEXT:    add sp, sp, a0
 ; CHECK-NEXT:    .cfi_def_cfa sp, 16
 ; CHECK-NEXT:    addi sp, sp, 16
diff --git a/llvm/test/CodeGen/RISCV/rvv/setcc-fp-vp.ll b/llvm/test/CodeGen/RISCV/rvv/setcc-fp-vp.ll
index 06f48762e24c2..30ef3dccd426b 100644
--- a/llvm/test/CodeGen/RISCV/rvv/setcc-fp-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/setcc-fp-vp.ll
@@ -17,10 +17,10 @@ declare <vscale x 1 x i1> @llvm.vp.fcmp.nxv1bf16(<vscale x 1 x bfloat>, <vscale
 define <vscale x 1 x i1> @fcmp_oeq_vv_nxv1bf16(<vscale x 1 x bfloat> %va, <vscale x 1 x bfloat> %vb, <vscale x 1 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: fcmp_oeq_vv_nxv1bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v9
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmfeq.vv v0, v9, v10, v0.t
 ; CHECK-NEXT:    ret
   %v = call <vscale x 1 x i1> @llvm.vp.fcmp.nxv1bf16(<vscale x 1 x bfloat> %va, <vscale x 1 x bfloat> %vb, metadata !"oeq", <vscale x 1 x i1> %m, i32 %evl)
@@ -31,11 +31,11 @@ define <vscale x 1 x i1> @fcmp_oeq_vf_nxv1bf16(<vscale x 1 x bfloat> %va, bfloat
 ; CHECK-LABEL: fcmp_oeq_vf_nxv1bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vmv.v.x v9, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmfeq.vv v0, v10, v8, v0.t
 ; CHECK-NEXT:    ret
   %elt.head = insertelement <vscale x 1 x bfloat> poison, bfloat %b, i32 0
@@ -48,11 +48,11 @@ define <vscale x 1 x i1> @fcmp_oeq_vf_swap_nxv1bf16(<vscale x 1 x bfloat> %va, b
 ; CHECK-LABEL: fcmp_oeq_vf_swap_nxv1bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vmv.v.x v9, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmfeq.vv v0, v8, v10, v0.t
 ; CHECK-NEXT:    ret
   %elt.head = insertelement <vscale x 1 x bfloat> poison, bfloat %b, i32 0
@@ -64,10 +64,10 @@ define <vscale x 1 x i1> @fcmp_oeq_vf_swap_nxv1bf16(<vscale x 1 x bfloat> %va, b
 define <vscale x 1 x i1> @fcmp_ogt_vv_nxv1bf16(<vscale x 1 x bfloat> %va, <vscale x 1 x bfloat> %vb, <vscale x 1 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: fcmp_ogt_vv_nxv1bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmflt.vv v0, v8, v10, v0.t
 ; CHECK-NEXT:    ret
   %v = call <vscale x 1 x i1> @llvm.vp.fcmp.nxv1bf16(<vscale x 1 x bfloat> %va, <vscale x 1 x bfloat> %vb, metadata !"ogt", <vscale x 1 x i1> %m, i32 %evl)
@@ -78,11 +78,11 @@ define <vscale x 1 x i1> @fcmp_ogt_vf_nxv1bf16(<vscale x 1 x bfloat> %va, bfloat
 ; CHECK-LABEL: fcmp_ogt_vf_nxv1bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vmv.v.x v9, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmflt.vv v0, v8, v10, v0.t
 ; CHECK-NEXT:    ret
   %elt.head = insertelement <vscale x 1 x bfloat> poison, bfloat %b, i32 0
@@ -95,11 +95,11 @@ define <vscale x 1 x i1> @fcmp_ogt_vf_swap_nxv1bf16(<vscale x 1 x bfloat> %va, b
 ; CHECK-LABEL: fcmp_ogt_vf_swap_nxv1bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vmv.v.x v9, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmflt.vv v0, v10, v8, v0.t
 ; CHECK-NEXT:    ret
   %elt.head = insertelement <vscale x 1 x bfloat> poison, bfloat %b, i32 0
@@ -111,10 +111,10 @@ define <vscale x 1 x i1> @fcmp_ogt_vf_swap_nxv1bf16(<vscale x 1 x bfloat> %va, b
 define <vscale x 1 x i1> @fcmp_oge_vv_nxv1bf16(<vscale x 1 x bfloat> %va, <vscale x 1 x bfloat> %vb, <vscale x 1 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: fcmp_oge_vv_nxv1bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmfle.vv v0, v8, v10, v0.t
 ; CHECK-NEXT:    ret
   %v = call <vscale x 1 x i1> @llvm.vp.fcmp.nxv1bf16(<vscale x 1 x bfloat> %va, <vscale x 1 x bfloat> %vb, metadata !"oge", <vscale x 1 x i1> %m, i32 %evl)
@@ -125,11 +125,11 @@ define <vscale x 1 x i1> @fcmp_oge_vf_nxv1bf16(<vscale x 1 x bfloat> %va, bfloat
 ; CHECK-LABEL: fcmp_oge_vf_nxv1bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vmv.v.x v9, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmfle.vv v0, v8, v10, v0.t
 ; CHECK-NEXT:    ret
   %elt.head = insertelement <vscale x 1 x bfloat> poison, bfloat %b, i32 0
@@ -142,11 +142,11 @@ define <vscale x 1 x i1> @fcmp_oge_vf_swap_nxv1bf16(<vscale x 1 x bfloat> %va, b
 ; CHECK-LABEL: fcmp_oge_vf_swap_nxv1bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vmv.v.x v9, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmfle.vv v0, v10, v8, v0.t
 ; CHECK-NEXT:    ret
   %elt.head = insertelement <vscale x 1 x bfloat> poison, bfloat %b, i32 0
@@ -158,10 +158,10 @@ define <vscale x 1 x i1> @fcmp_oge_vf_swap_nxv1bf16(<vscale x 1 x bfloat> %va, b
 define <vscale x 1 x i1> @fcmp_olt_vv_nxv1bf16(<vscale x 1 x bfloat> %va, <vscale x 1 x bfloat> %vb, <vscale x 1 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: fcmp_olt_vv_nxv1bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v9
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmflt.vv v0, v9, v10, v0.t
 ; CHECK-NEXT:    ret
   %v = call <vscale x 1 x i1> @llvm.vp.fcmp.nxv1bf16(<vscale x 1 x bfloat> %va, <vscale x 1 x bfloat> %vb, metadata !"olt", <vscale x 1 x i1> %m, i32 %evl)
@@ -172,11 +172,11 @@ define <vscale x 1 x i1> @fcmp_olt_vf_nxv1bf16(<vscale x 1 x bfloat> %va, bfloat
 ; CHECK-LABEL: fcmp_olt_vf_nxv1bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vmv.v.x v9, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmflt.vv v0, v10, v8, v0.t
 ; CHECK-NEXT:    ret
   %elt.head = insertelement <vscale x 1 x bfloat> poison, bfloat %b, i32 0
@@ -189,11 +189,11 @@ define <vscale x 1 x i1> @fcmp_olt_vf_swap_nxv1bf16(<vscale x 1 x bfloat> %va, b
 ; CHECK-LABEL: fcmp_olt_vf_swap_nxv1bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vmv.v.x v9, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmflt.vv v0, v8, v10, v0.t
 ; CHECK-NEXT:    ret
   %elt.head = insertelement <vscale x 1 x bfloat> poison, bfloat %b, i32 0
@@ -205,10 +205,10 @@ define <vscale x 1 x i1> @fcmp_olt_vf_swap_nxv1bf16(<vscale x 1 x bfloat> %va, b
 define <vscale x 1 x i1> @fcmp_ole_vv_nxv1bf16(<vscale x 1 x bfloat> %va, <vscale x 1 x bfloat> %vb, <vscale x 1 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: fcmp_ole_vv_nxv1bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v9
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmfle.vv v0, v9, v10, v0.t
 ; CHECK-NEXT:    ret
   %v = call <vscale x 1 x i1> @llvm.vp.fcmp.nxv1bf16(<vscale x 1 x bfloat> %va, <vscale x 1 x bfloat> %vb, metadata !"ole", <vscale x 1 x i1> %m, i32 %evl)
@@ -219,11 +219,11 @@ define <vscale x 1 x i1> @fcmp_ole_vf_nxv1bf16(<vscale x 1 x bfloat> %va, bfloat
 ; CHECK-LABEL: fcmp_ole_vf_nxv1bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vmv.v.x v9, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmfle.vv v0, v10, v8, v0.t
 ; CHECK-NEXT:    ret
   %elt.head = insertelement <vscale x 1 x bfloat> poison, bfloat %b, i32 0
@@ -236,11 +236,11 @@ define <vscale x 1 x i1> @fcmp_ole_vf_swap_nxv1bf16(<vscale x 1 x bfloat> %va, b
 ; CHECK-LABEL: fcmp_ole_vf_swap_nxv1bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vmv.v.x v9, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmfle.vv v0, v8, v10, v0.t
 ; CHECK-NEXT:    ret
   %elt.head = insertelement <vscale x 1 x bfloat> poison, bfloat %b, i32 0
@@ -252,10 +252,10 @@ define <vscale x 1 x i1> @fcmp_ole_vf_swap_nxv1bf16(<vscale x 1 x bfloat> %va, b
 define <vscale x 1 x i1> @fcmp_one_vv_nxv1bf16(<vscale x 1 x bfloat> %va, <vscale x 1 x bfloat> %vb, <vscale x 1 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: fcmp_one_vv_nxv1bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v9
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmflt.vv v8, v9, v10, v0.t
 ; CHECK-NEXT:    vmflt.vv v9, v10, v9, v0.t
 ; CHECK-NEXT:    vmor.mm v0, v9, v8
@@ -268,11 +268,11 @@ define <vscale x 1 x i1> @fcmp_one_vf_nxv1bf16(<vscale x 1 x bfloat> %va, bfloat
 ; CHECK-LABEL: fcmp_one_vf_nxv1bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
 ; CHECK-NEXT:    vmv.v.x v8, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmflt.vv v8, v9, v10, v0.t
 ; CHECK-NEXT:    vmflt.vv v9, v10, v9, v0.t
 ; CHECK-NEXT:    vmor.mm v0, v9, v8
@@ -287,11 +287,11 @@ define <vscale x 1 x i1> @fcmp_one_vf_swap_nxv1bf16(<vscale x 1 x bfloat> %va, b
 ; CHECK-LABEL: fcmp_one_vf_swap_nxv1bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
 ; CHECK-NEXT:    vmv.v.x v8, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmflt.vv v8, v10, v9, v0.t
 ; CHECK-NEXT:    vmflt.vv v9, v9, v10, v0.t
 ; CHECK-NEXT:    vmor.mm v0, v9, v8
@@ -305,10 +305,10 @@ define <vscale x 1 x i1> @fcmp_one_vf_swap_nxv1bf16(<vscale x 1 x bfloat> %va, b
 define <vscale x 1 x i1> @fcmp_ord_vv_nxv1bf16(<vscale x 1 x bfloat> %va, <vscale x 1 x bfloat> %vb, <vscale x 1 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: fcmp_ord_vv_nxv1bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v9
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmfeq.vv v8, v10, v10, v0.t
 ; CHECK-NEXT:    vmfeq.vv v9, v9, v9, v0.t
 ; CHECK-NEXT:    vmand.mm v0, v9, v8
@@ -321,14 +321,14 @@ define <vscale x 1 x i1> @fcmp_ord_vf_nxv1bf16(<vscale x 1 x bfloat> %va, bfloat
 ; CHECK-LABEL: fcmp_ord_vf_nxv1bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
 ; CHECK-NEXT:    vmv.v.x v8, a1
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmfeq.vv v9, v9, v9, v0.t
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmfeq.vv v8, v10, v10, v0.t
 ; CHECK-NEXT:    vmand.mm v0, v9, v8
 ; CHECK-NEXT:    ret
@@ -342,14 +342,14 @@ define <vscale x 1 x i1> @fcmp_ord_vf_swap_nxv1bf16(<vscale x 1 x bfloat> %va, b
 ; CHECK-LABEL: fcmp_ord_vf_swap_nxv1bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
 ; CHECK-NEXT:    vmv.v.x v8, a1
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmfeq.vv v9, v9, v9, v0.t
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmfeq.vv v8, v10, v10, v0.t
 ; CHECK-NEXT:    vmand.mm v0, v8, v9
 ; CHECK-NEXT:    ret
@@ -362,10 +362,10 @@ define <vscale x 1 x i1> @fcmp_ord_vf_swap_nxv1bf16(<vscale x 1 x bfloat> %va, b
 define <vscale x 1 x i1> @fcmp_ueq_vv_nxv1bf16(<vscale x 1 x bfloat> %va, <vscale x 1 x bfloat> %vb, <vscale x 1 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: fcmp_ueq_vv_nxv1bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v9
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmflt.vv v8, v9, v10, v0.t
 ; CHECK-NEXT:    vmflt.vv v9, v10, v9, v0.t
 ; CHECK-NEXT:    vmnor.mm v0, v9, v8
@@ -378,11 +378,11 @@ define <vscale x 1 x i1> @fcmp_ueq_vf_nxv1bf16(<vscale x 1 x bfloat> %va, bfloat
 ; CHECK-LABEL: fcmp_ueq_vf_nxv1bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
 ; CHECK-NEXT:    vmv.v.x v8, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmflt.vv v8, v9, v10, v0.t
 ; CHECK-NEXT:    vmflt.vv v9, v10, v9, v0.t
 ; CHECK-NEXT:    vmnor.mm v0, v9, v8
@@ -397,11 +397,11 @@ define <vscale x 1 x i1> @fcmp_ueq_vf_swap_nxv1bf16(<vscale x 1 x bfloat> %va, b
 ; CHECK-LABEL: fcmp_ueq_vf_swap_nxv1bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
 ; CHECK-NEXT:    vmv.v.x v8, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmflt.vv v8, v10, v9, v0.t
 ; CHECK-NEXT:    vmflt.vv v9, v9, v10, v0.t
 ; CHECK-NEXT:    vmnor.mm v0, v9, v8
@@ -415,10 +415,10 @@ define <vscale x 1 x i1> @fcmp_ueq_vf_swap_nxv1bf16(<vscale x 1 x bfloat> %va, b
 define <vscale x 1 x i1> @fcmp_ugt_vv_nxv1bf16(<vscale x 1 x bfloat> %va, <vscale x 1 x bfloat> %vb, <vscale x 1 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: fcmp_ugt_vv_nxv1bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v9
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmfle.vv v8, v9, v10, v0.t
 ; CHECK-NEXT:    vmnot.m v0, v8
 ; CHECK-NEXT:    ret
@@ -430,11 +430,11 @@ define <vscale x 1 x i1> @fcmp_ugt_vf_nxv1bf16(<vscale x 1 x bfloat> %va, bfloat
 ; CHECK-LABEL: fcmp_ugt_vf_nxv1bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vmv.v.x v9, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmfle.vv v8, v10, v8, v0.t
 ; CHECK-NEXT:    vmnot.m v0, v8
 ; CHECK-NEXT:    ret
@@ -448,11 +448,11 @@ define <vscale x 1 x i1> @fcmp_ugt_vf_swap_nxv1bf16(<vscale x 1 x bfloat> %va, b
 ; CHECK-LABEL: fcmp_ugt_vf_swap_nxv1bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vmv.v.x v9, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmfle.vv v8, v8, v10, v0.t
 ; CHECK-NEXT:    vmnot.m v0, v8
 ; CHECK-NEXT:    ret
@@ -465,10 +465,10 @@ define <vscale x 1 x i1> @fcmp_ugt_vf_swap_nxv1bf16(<vscale x 1 x bfloat> %va, b
 define <vscale x 1 x i1> @fcmp_uge_vv_nxv1bf16(<vscale x 1 x bfloat> %va, <vscale x 1 x bfloat> %vb, <vscale x 1 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: fcmp_uge_vv_nxv1bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v9
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmflt.vv v8, v9, v10, v0.t
 ; CHECK-NEXT:    vmnot.m v0, v8
 ; CHECK-NEXT:    ret
@@ -480,11 +480,11 @@ define <vscale x 1 x i1> @fcmp_uge_vf_nxv1bf16(<vscale x 1 x bfloat> %va, bfloat
 ; CHECK-LABEL: fcmp_uge_vf_nxv1bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vmv.v.x v9, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmflt.vv v8, v10, v8, v0.t
 ; CHECK-NEXT:    vmnot.m v0, v8
 ; CHECK-NEXT:    ret
@@ -498,11 +498,11 @@ define <vscale x 1 x i1> @fcmp_uge_vf_swap_nxv1bf16(<vscale x 1 x bfloat> %va, b
 ; CHECK-LABEL: fcmp_uge_vf_swap_nxv1bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vmv.v.x v9, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmflt.vv v8, v8, v10, v0.t
 ; CHECK-NEXT:    vmnot.m v0, v8
 ; CHECK-NEXT:    ret
@@ -515,10 +515,10 @@ define <vscale x 1 x i1> @fcmp_uge_vf_swap_nxv1bf16(<vscale x 1 x bfloat> %va, b
 define <vscale x 1 x i1> @fcmp_ult_vv_nxv1bf16(<vscale x 1 x bfloat> %va, <vscale x 1 x bfloat> %vb, <vscale x 1 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: fcmp_ult_vv_nxv1bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmfle.vv v8, v8, v10, v0.t
 ; CHECK-NEXT:    vmnot.m v0, v8
 ; CHECK-NEXT:    ret
@@ -530,11 +530,11 @@ define <vscale x 1 x i1> @fcmp_ult_vf_nxv1bf16(<vscale x 1 x bfloat> %va, bfloat
 ; CHECK-LABEL: fcmp_ult_vf_nxv1bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vmv.v.x v9, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmfle.vv v8, v8, v10, v0.t
 ; CHECK-NEXT:    vmnot.m v0, v8
 ; CHECK-NEXT:    ret
@@ -548,11 +548,11 @@ define <vscale x 1 x i1> @fcmp_ult_vf_swap_nxv1bf16(<vscale x 1 x bfloat> %va, b
 ; CHECK-LABEL: fcmp_ult_vf_swap_nxv1bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vmv.v.x v9, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmfle.vv v8, v10, v8, v0.t
 ; CHECK-NEXT:    vmnot.m v0, v8
 ; CHECK-NEXT:    ret
@@ -565,10 +565,10 @@ define <vscale x 1 x i1> @fcmp_ult_vf_swap_nxv1bf16(<vscale x 1 x bfloat> %va, b
 define <vscale x 1 x i1> @fcmp_ule_vv_nxv1bf16(<vscale x 1 x bfloat> %va, <vscale x 1 x bfloat> %vb, <vscale x 1 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: fcmp_ule_vv_nxv1bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmflt.vv v8, v8, v10, v0.t
 ; CHECK-NEXT:    vmnot.m v0, v8
 ; CHECK-NEXT:    ret
@@ -580,11 +580,11 @@ define <vscale x 1 x i1> @fcmp_ule_vf_nxv1bf16(<vscale x 1 x bfloat> %va, bfloat
 ; CHECK-LABEL: fcmp_ule_vf_nxv1bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vmv.v.x v9, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmflt.vv v8, v8, v10, v0.t
 ; CHECK-NEXT:    vmnot.m v0, v8
 ; CHECK-NEXT:    ret
@@ -598,11 +598,11 @@ define <vscale x 1 x i1> @fcmp_ule_vf_swap_nxv1bf16(<vscale x 1 x bfloat> %va, b
 ; CHECK-LABEL: fcmp_ule_vf_swap_nxv1bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vmv.v.x v9, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmflt.vv v8, v10, v8, v0.t
 ; CHECK-NEXT:    vmnot.m v0, v8
 ; CHECK-NEXT:    ret
@@ -615,10 +615,10 @@ define <vscale x 1 x i1> @fcmp_ule_vf_swap_nxv1bf16(<vscale x 1 x bfloat> %va, b
 define <vscale x 1 x i1> @fcmp_une_vv_nxv1bf16(<vscale x 1 x bfloat> %va, <vscale x 1 x bfloat> %vb, <vscale x 1 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: fcmp_une_vv_nxv1bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v9
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmfne.vv v0, v9, v10, v0.t
 ; CHECK-NEXT:    ret
   %v = call <vscale x 1 x i1> @llvm.vp.fcmp.nxv1bf16(<vscale x 1 x bfloat> %va, <vscale x 1 x bfloat> %vb, metadata !"une", <vscale x 1 x i1> %m, i32 %evl)
@@ -629,11 +629,11 @@ define <vscale x 1 x i1> @fcmp_une_vf_nxv1bf16(<vscale x 1 x bfloat> %va, bfloat
 ; CHECK-LABEL: fcmp_une_vf_nxv1bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vmv.v.x v9, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmfne.vv v0, v10, v8, v0.t
 ; CHECK-NEXT:    ret
   %elt.head = insertelement <vscale x 1 x bfloat> poison, bfloat %b, i32 0
@@ -646,11 +646,11 @@ define <vscale x 1 x i1> @fcmp_une_vf_swap_nxv1bf16(<vscale x 1 x bfloat> %va, b
 ; CHECK-LABEL: fcmp_une_vf_swap_nxv1bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vmv.v.x v9, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmfne.vv v0, v8, v10, v0.t
 ; CHECK-NEXT:    ret
   %elt.head = insertelement <vscale x 1 x bfloat> poison, bfloat %b, i32 0
@@ -662,10 +662,10 @@ define <vscale x 1 x i1> @fcmp_une_vf_swap_nxv1bf16(<vscale x 1 x bfloat> %va, b
 define <vscale x 1 x i1> @fcmp_uno_vv_nxv1bf16(<vscale x 1 x bfloat> %va, <vscale x 1 x bfloat> %vb, <vscale x 1 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: fcmp_uno_vv_nxv1bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v9
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmfne.vv v8, v10, v10, v0.t
 ; CHECK-NEXT:    vmfne.vv v9, v9, v9, v0.t
 ; CHECK-NEXT:    vmor.mm v0, v9, v8
@@ -678,14 +678,14 @@ define <vscale x 1 x i1> @fcmp_uno_vf_nxv1bf16(<vscale x 1 x bfloat> %va, bfloat
 ; CHECK-LABEL: fcmp_uno_vf_nxv1bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
 ; CHECK-NEXT:    vmv.v.x v8, a1
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmfne.vv v9, v9, v9, v0.t
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmfne.vv v8, v10, v10, v0.t
 ; CHECK-NEXT:    vmor.mm v0, v9, v8
 ; CHECK-NEXT:    ret
@@ -699,14 +699,14 @@ define <vscale x 1 x i1> @fcmp_uno_vf_swap_nxv1bf16(<vscale x 1 x bfloat> %va, b
 ; CHECK-LABEL: fcmp_uno_vf_swap_nxv1bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
 ; CHECK-NEXT:    vmv.v.x v8, a1
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmfne.vv v9, v9, v9, v0.t
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vmfne.vv v8, v10, v10, v0.t
 ; CHECK-NEXT:    vmor.mm v0, v8, v9
 ; CHECK-NEXT:    ret
@@ -721,10 +721,10 @@ declare <vscale x 3 x i1> @llvm.vp.fcmp.nxv3bf16(<vscale x 3 x bfloat>, <vscale
 define <vscale x 3 x i1> @fcmp_oeq_vv_nxv3bf16(<vscale x 3 x bfloat> %va, <vscale x 3 x bfloat> %vb, <vscale x 3 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: fcmp_oeq_vv_nxv3bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m1, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v9
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
 ; CHECK-NEXT:    vmfeq.vv v8, v12, v10, v0.t
 ; CHECK-NEXT:    vmv1r.v v0, v8
 ; CHECK-NEXT:    ret
@@ -737,10 +737,10 @@ declare <vscale x 8 x i1> @llvm.vp.fcmp.nxv8bf16(<vscale x 8 x bfloat>, <vscale
 define <vscale x 8 x i1> @fcmp_oeq_vv_nxv8bf16(<vscale x 8 x bfloat> %va, <vscale x 8 x bfloat> %vb, <vscale x 8 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: fcmp_oeq_vv_nxv8bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v10
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmfeq.vv v8, v16, v12, v0.t
 ; CHECK-NEXT:    vmv1r.v v0, v8
 ; CHECK-NEXT:    ret
@@ -752,11 +752,11 @@ define <vscale x 8 x i1> @fcmp_oeq_vf_nxv8bf16(<vscale x 8 x bfloat> %va, bfloat
 ; CHECK-LABEL: fcmp_oeq_vf_nxv8bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vmv.v.x v10, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v10
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmfeq.vv v8, v12, v16, v0.t
 ; CHECK-NEXT:    vmv1r.v v0, v8
 ; CHECK-NEXT:    ret
@@ -770,11 +770,11 @@ define <vscale x 8 x i1> @fcmp_oeq_vf_swap_nxv8bf16(<vscale x 8 x bfloat> %va, b
 ; CHECK-LABEL: fcmp_oeq_vf_swap_nxv8bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vmv.v.x v10, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v10
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmfeq.vv v8, v16, v12, v0.t
 ; CHECK-NEXT:    vmv1r.v v0, v8
 ; CHECK-NEXT:    ret
@@ -787,10 +787,10 @@ define <vscale x 8 x i1> @fcmp_oeq_vf_swap_nxv8bf16(<vscale x 8 x bfloat> %va, b
 define <vscale x 8 x i1> @fcmp_ogt_vv_nxv8bf16(<vscale x 8 x bfloat> %va, <vscale x 8 x bfloat> %vb, <vscale x 8 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: fcmp_ogt_vv_nxv8bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v10
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmflt.vv v8, v16, v12, v0.t
 ; CHECK-NEXT:    vmv1r.v v0, v8
 ; CHECK-NEXT:    ret
@@ -802,11 +802,11 @@ define <vscale x 8 x i1> @fcmp_ogt_vf_nxv8bf16(<vscale x 8 x bfloat> %va, bfloat
 ; CHECK-LABEL: fcmp_ogt_vf_nxv8bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vmv.v.x v10, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v10
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmflt.vv v8, v16, v12, v0.t
 ; CHECK-NEXT:    vmv1r.v v0, v8
 ; CHECK-NEXT:    ret
@@ -820,11 +820,11 @@ define <vscale x 8 x i1> @fcmp_ogt_vf_swap_nxv8bf16(<vscale x 8 x bfloat> %va, b
 ; CHECK-LABEL: fcmp_ogt_vf_swap_nxv8bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vmv.v.x v10, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v10
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmflt.vv v8, v12, v16, v0.t
 ; CHECK-NEXT:    vmv1r.v v0, v8
 ; CHECK-NEXT:    ret
@@ -837,10 +837,10 @@ define <vscale x 8 x i1> @fcmp_ogt_vf_swap_nxv8bf16(<vscale x 8 x bfloat> %va, b
 define <vscale x 8 x i1> @fcmp_oge_vv_nxv8bf16(<vscale x 8 x bfloat> %va, <vscale x 8 x bfloat> %vb, <vscale x 8 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: fcmp_oge_vv_nxv8bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v10
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmfle.vv v8, v16, v12, v0.t
 ; CHECK-NEXT:    vmv1r.v v0, v8
 ; CHECK-NEXT:    ret
@@ -852,11 +852,11 @@ define <vscale x 8 x i1> @fcmp_oge_vf_nxv8bf16(<vscale x 8 x bfloat> %va, bfloat
 ; CHECK-LABEL: fcmp_oge_vf_nxv8bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vmv.v.x v10, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v10
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmfle.vv v8, v16, v12, v0.t
 ; CHECK-NEXT:    vmv1r.v v0, v8
 ; CHECK-NEXT:    ret
@@ -870,11 +870,11 @@ define <vscale x 8 x i1> @fcmp_oge_vf_swap_nxv8bf16(<vscale x 8 x bfloat> %va, b
 ; CHECK-LABEL: fcmp_oge_vf_swap_nxv8bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vmv.v.x v10, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v10
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmfle.vv v8, v12, v16, v0.t
 ; CHECK-NEXT:    vmv1r.v v0, v8
 ; CHECK-NEXT:    ret
@@ -887,10 +887,10 @@ define <vscale x 8 x i1> @fcmp_oge_vf_swap_nxv8bf16(<vscale x 8 x bfloat> %va, b
 define <vscale x 8 x i1> @fcmp_olt_vv_nxv8bf16(<vscale x 8 x bfloat> %va, <vscale x 8 x bfloat> %vb, <vscale x 8 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: fcmp_olt_vv_nxv8bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v10
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmflt.vv v8, v16, v12, v0.t
 ; CHECK-NEXT:    vmv1r.v v0, v8
 ; CHECK-NEXT:    ret
@@ -902,11 +902,11 @@ define <vscale x 8 x i1> @fcmp_olt_vf_nxv8bf16(<vscale x 8 x bfloat> %va, bfloat
 ; CHECK-LABEL: fcmp_olt_vf_nxv8bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vmv.v.x v10, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v10
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmflt.vv v8, v12, v16, v0.t
 ; CHECK-NEXT:    vmv1r.v v0, v8
 ; CHECK-NEXT:    ret
@@ -920,11 +920,11 @@ define <vscale x 8 x i1> @fcmp_olt_vf_swap_nxv8bf16(<vscale x 8 x bfloat> %va, b
 ; CHECK-LABEL: fcmp_olt_vf_swap_nxv8bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vmv.v.x v10, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v10
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmflt.vv v8, v16, v12, v0.t
 ; CHECK-NEXT:    vmv1r.v v0, v8
 ; CHECK-NEXT:    ret
@@ -937,10 +937,10 @@ define <vscale x 8 x i1> @fcmp_olt_vf_swap_nxv8bf16(<vscale x 8 x bfloat> %va, b
 define <vscale x 8 x i1> @fcmp_ole_vv_nxv8bf16(<vscale x 8 x bfloat> %va, <vscale x 8 x bfloat> %vb, <vscale x 8 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: fcmp_ole_vv_nxv8bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v10
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmfle.vv v8, v16, v12, v0.t
 ; CHECK-NEXT:    vmv1r.v v0, v8
 ; CHECK-NEXT:    ret
@@ -952,11 +952,11 @@ define <vscale x 8 x i1> @fcmp_ole_vf_nxv8bf16(<vscale x 8 x bfloat> %va, bfloat
 ; CHECK-LABEL: fcmp_ole_vf_nxv8bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vmv.v.x v10, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v10
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmfle.vv v8, v12, v16, v0.t
 ; CHECK-NEXT:    vmv1r.v v0, v8
 ; CHECK-NEXT:    ret
@@ -970,11 +970,11 @@ define <vscale x 8 x i1> @fcmp_ole_vf_swap_nxv8bf16(<vscale x 8 x bfloat> %va, b
 ; CHECK-LABEL: fcmp_ole_vf_swap_nxv8bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vmv.v.x v10, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v10
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmfle.vv v8, v16, v12, v0.t
 ; CHECK-NEXT:    vmv1r.v v0, v8
 ; CHECK-NEXT:    ret
@@ -987,10 +987,10 @@ define <vscale x 8 x i1> @fcmp_ole_vf_swap_nxv8bf16(<vscale x 8 x bfloat> %va, b
 define <vscale x 8 x i1> @fcmp_one_vv_nxv8bf16(<vscale x 8 x bfloat> %va, <vscale x 8 x bfloat> %vb, <vscale x 8 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: fcmp_one_vv_nxv8bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v10
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmflt.vv v8, v16, v12, v0.t
 ; CHECK-NEXT:    vmflt.vv v9, v12, v16, v0.t
 ; CHECK-NEXT:    vmor.mm v0, v9, v8
@@ -1003,11 +1003,11 @@ define <vscale x 8 x i1> @fcmp_one_vf_nxv8bf16(<vscale x 8 x bfloat> %va, bfloat
 ; CHECK-LABEL: fcmp_one_vf_nxv8bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
 ; CHECK-NEXT:    vmv.v.x v8, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmflt.vv v8, v12, v16, v0.t
 ; CHECK-NEXT:    vmflt.vv v9, v16, v12, v0.t
 ; CHECK-NEXT:    vmor.mm v0, v9, v8
@@ -1022,11 +1022,11 @@ define <vscale x 8 x i1> @fcmp_one_vf_swap_nxv8bf16(<vscale x 8 x bfloat> %va, b
 ; CHECK-LABEL: fcmp_one_vf_swap_nxv8bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
 ; CHECK-NEXT:    vmv.v.x v8, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmflt.vv v8, v16, v12, v0.t
 ; CHECK-NEXT:    vmflt.vv v9, v12, v16, v0.t
 ; CHECK-NEXT:    vmor.mm v0, v9, v8
@@ -1040,13 +1040,13 @@ define <vscale x 8 x i1> @fcmp_one_vf_swap_nxv8bf16(<vscale x 8 x bfloat> %va, b
 define <vscale x 8 x i1> @fcmp_ord_vv_nxv8bf16(<vscale x 8 x bfloat> %va, <vscale x 8 x bfloat> %vb, <vscale x 8 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: fcmp_ord_vv_nxv8bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v10
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmfeq.vv v10, v12, v12, v0.t
-; CHECK-NEXT:    vsetvli a1, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e16, m2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmfeq.vv v8, v12, v12, v0.t
 ; CHECK-NEXT:    vmand.mm v0, v8, v10
 ; CHECK-NEXT:    ret
@@ -1058,14 +1058,14 @@ define <vscale x 8 x i1> @fcmp_ord_vf_nxv8bf16(<vscale x 8 x bfloat> %va, bfloat
 ; CHECK-LABEL: fcmp_ord_vf_nxv8bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
 ; CHECK-NEXT:    vmv.v.x v8, a1
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmfeq.vv v10, v12, v12, v0.t
-; CHECK-NEXT:    vsetvli a1, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e16, m2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmfeq.vv v8, v12, v12, v0.t
 ; CHECK-NEXT:    vmand.mm v0, v10, v8
 ; CHECK-NEXT:    ret
@@ -1079,14 +1079,14 @@ define <vscale x 8 x i1> @fcmp_ord_vf_swap_nxv8bf16(<vscale x 8 x bfloat> %va, b
 ; CHECK-LABEL: fcmp_ord_vf_swap_nxv8bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
 ; CHECK-NEXT:    vmv.v.x v8, a1
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmfeq.vv v10, v12, v12, v0.t
-; CHECK-NEXT:    vsetvli a1, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e16, m2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmfeq.vv v8, v12, v12, v0.t
 ; CHECK-NEXT:    vmand.mm v0, v8, v10
 ; CHECK-NEXT:    ret
@@ -1099,10 +1099,10 @@ define <vscale x 8 x i1> @fcmp_ord_vf_swap_nxv8bf16(<vscale x 8 x bfloat> %va, b
 define <vscale x 8 x i1> @fcmp_ueq_vv_nxv8bf16(<vscale x 8 x bfloat> %va, <vscale x 8 x bfloat> %vb, <vscale x 8 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: fcmp_ueq_vv_nxv8bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v10
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmflt.vv v8, v16, v12, v0.t
 ; CHECK-NEXT:    vmflt.vv v9, v12, v16, v0.t
 ; CHECK-NEXT:    vmnor.mm v0, v9, v8
@@ -1115,11 +1115,11 @@ define <vscale x 8 x i1> @fcmp_ueq_vf_nxv8bf16(<vscale x 8 x bfloat> %va, bfloat
 ; CHECK-LABEL: fcmp_ueq_vf_nxv8bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
 ; CHECK-NEXT:    vmv.v.x v8, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmflt.vv v8, v12, v16, v0.t
 ; CHECK-NEXT:    vmflt.vv v9, v16, v12, v0.t
 ; CHECK-NEXT:    vmnor.mm v0, v9, v8
@@ -1134,11 +1134,11 @@ define <vscale x 8 x i1> @fcmp_ueq_vf_swap_nxv8bf16(<vscale x 8 x bfloat> %va, b
 ; CHECK-LABEL: fcmp_ueq_vf_swap_nxv8bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
 ; CHECK-NEXT:    vmv.v.x v8, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmflt.vv v8, v16, v12, v0.t
 ; CHECK-NEXT:    vmflt.vv v9, v12, v16, v0.t
 ; CHECK-NEXT:    vmnor.mm v0, v9, v8
@@ -1152,10 +1152,10 @@ define <vscale x 8 x i1> @fcmp_ueq_vf_swap_nxv8bf16(<vscale x 8 x bfloat> %va, b
 define <vscale x 8 x i1> @fcmp_ugt_vv_nxv8bf16(<vscale x 8 x bfloat> %va, <vscale x 8 x bfloat> %vb, <vscale x 8 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: fcmp_ugt_vv_nxv8bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v10
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmfle.vv v8, v16, v12, v0.t
 ; CHECK-NEXT:    vmnot.m v0, v8
 ; CHECK-NEXT:    ret
@@ -1167,11 +1167,11 @@ define <vscale x 8 x i1> @fcmp_ugt_vf_nxv8bf16(<vscale x 8 x bfloat> %va, bfloat
 ; CHECK-LABEL: fcmp_ugt_vf_nxv8bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vmv.v.x v10, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v10
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmfle.vv v8, v12, v16, v0.t
 ; CHECK-NEXT:    vmnot.m v0, v8
 ; CHECK-NEXT:    ret
@@ -1185,11 +1185,11 @@ define <vscale x 8 x i1> @fcmp_ugt_vf_swap_nxv8bf16(<vscale x 8 x bfloat> %va, b
 ; CHECK-LABEL: fcmp_ugt_vf_swap_nxv8bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vmv.v.x v10, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v10
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmfle.vv v8, v16, v12, v0.t
 ; CHECK-NEXT:    vmnot.m v0, v8
 ; CHECK-NEXT:    ret
@@ -1202,10 +1202,10 @@ define <vscale x 8 x i1> @fcmp_ugt_vf_swap_nxv8bf16(<vscale x 8 x bfloat> %va, b
 define <vscale x 8 x i1> @fcmp_uge_vv_nxv8bf16(<vscale x 8 x bfloat> %va, <vscale x 8 x bfloat> %vb, <vscale x 8 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: fcmp_uge_vv_nxv8bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v10
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmflt.vv v8, v16, v12, v0.t
 ; CHECK-NEXT:    vmnot.m v0, v8
 ; CHECK-NEXT:    ret
@@ -1217,11 +1217,11 @@ define <vscale x 8 x i1> @fcmp_uge_vf_nxv8bf16(<vscale x 8 x bfloat> %va, bfloat
 ; CHECK-LABEL: fcmp_uge_vf_nxv8bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vmv.v.x v10, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v10
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmflt.vv v8, v12, v16, v0.t
 ; CHECK-NEXT:    vmnot.m v0, v8
 ; CHECK-NEXT:    ret
@@ -1235,11 +1235,11 @@ define <vscale x 8 x i1> @fcmp_uge_vf_swap_nxv8bf16(<vscale x 8 x bfloat> %va, b
 ; CHECK-LABEL: fcmp_uge_vf_swap_nxv8bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vmv.v.x v10, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v10
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmflt.vv v8, v16, v12, v0.t
 ; CHECK-NEXT:    vmnot.m v0, v8
 ; CHECK-NEXT:    ret
@@ -1252,10 +1252,10 @@ define <vscale x 8 x i1> @fcmp_uge_vf_swap_nxv8bf16(<vscale x 8 x bfloat> %va, b
 define <vscale x 8 x i1> @fcmp_ult_vv_nxv8bf16(<vscale x 8 x bfloat> %va, <vscale x 8 x bfloat> %vb, <vscale x 8 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: fcmp_ult_vv_nxv8bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v10
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmfle.vv v8, v16, v12, v0.t
 ; CHECK-NEXT:    vmnot.m v0, v8
 ; CHECK-NEXT:    ret
@@ -1267,11 +1267,11 @@ define <vscale x 8 x i1> @fcmp_ult_vf_nxv8bf16(<vscale x 8 x bfloat> %va, bfloat
 ; CHECK-LABEL: fcmp_ult_vf_nxv8bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vmv.v.x v10, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v10
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmfle.vv v8, v16, v12, v0.t
 ; CHECK-NEXT:    vmnot.m v0, v8
 ; CHECK-NEXT:    ret
@@ -1285,11 +1285,11 @@ define <vscale x 8 x i1> @fcmp_ult_vf_swap_nxv8bf16(<vscale x 8 x bfloat> %va, b
 ; CHECK-LABEL: fcmp_ult_vf_swap_nxv8bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vmv.v.x v10, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v10
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmfle.vv v8, v12, v16, v0.t
 ; CHECK-NEXT:    vmnot.m v0, v8
 ; CHECK-NEXT:    ret
@@ -1302,10 +1302,10 @@ define <vscale x 8 x i1> @fcmp_ult_vf_swap_nxv8bf16(<vscale x 8 x bfloat> %va, b
 define <vscale x 8 x i1> @fcmp_ule_vv_nxv8bf16(<vscale x 8 x bfloat> %va, <vscale x 8 x bfloat> %vb, <vscale x 8 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: fcmp_ule_vv_nxv8bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v10
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmflt.vv v8, v16, v12, v0.t
 ; CHECK-NEXT:    vmnot.m v0, v8
 ; CHECK-NEXT:    ret
@@ -1317,11 +1317,11 @@ define <vscale x 8 x i1> @fcmp_ule_vf_nxv8bf16(<vscale x 8 x bfloat> %va, bfloat
 ; CHECK-LABEL: fcmp_ule_vf_nxv8bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vmv.v.x v10, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v10
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmflt.vv v8, v16, v12, v0.t
 ; CHECK-NEXT:    vmnot.m v0, v8
 ; CHECK-NEXT:    ret
@@ -1335,11 +1335,11 @@ define <vscale x 8 x i1> @fcmp_ule_vf_swap_nxv8bf16(<vscale x 8 x bfloat> %va, b
 ; CHECK-LABEL: fcmp_ule_vf_swap_nxv8bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vmv.v.x v10, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v10
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmflt.vv v8, v12, v16, v0.t
 ; CHECK-NEXT:    vmnot.m v0, v8
 ; CHECK-NEXT:    ret
@@ -1352,10 +1352,10 @@ define <vscale x 8 x i1> @fcmp_ule_vf_swap_nxv8bf16(<vscale x 8 x bfloat> %va, b
 define <vscale x 8 x i1> @fcmp_une_vv_nxv8bf16(<vscale x 8 x bfloat> %va, <vscale x 8 x bfloat> %vb, <vscale x 8 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: fcmp_une_vv_nxv8bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v10
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmfne.vv v8, v16, v12, v0.t
 ; CHECK-NEXT:    vmv1r.v v0, v8
 ; CHECK-NEXT:    ret
@@ -1367,11 +1367,11 @@ define <vscale x 8 x i1> @fcmp_une_vf_nxv8bf16(<vscale x 8 x bfloat> %va, bfloat
 ; CHECK-LABEL: fcmp_une_vf_nxv8bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vmv.v.x v10, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v10
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmfne.vv v8, v12, v16, v0.t
 ; CHECK-NEXT:    vmv1r.v v0, v8
 ; CHECK-NEXT:    ret
@@ -1385,11 +1385,11 @@ define <vscale x 8 x i1> @fcmp_une_vf_swap_nxv8bf16(<vscale x 8 x bfloat> %va, b
 ; CHECK-LABEL: fcmp_une_vf_swap_nxv8bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vmv.v.x v10, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v10
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmfne.vv v8, v16, v12, v0.t
 ; CHECK-NEXT:    vmv1r.v v0, v8
 ; CHECK-NEXT:    ret
@@ -1402,13 +1402,13 @@ define <vscale x 8 x i1> @fcmp_une_vf_swap_nxv8bf16(<vscale x 8 x bfloat> %va, b
 define <vscale x 8 x i1> @fcmp_uno_vv_nxv8bf16(<vscale x 8 x bfloat> %va, <vscale x 8 x bfloat> %vb, <vscale x 8 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: fcmp_uno_vv_nxv8bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v10
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmfne.vv v10, v12, v12, v0.t
-; CHECK-NEXT:    vsetvli a1, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e16, m2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmfne.vv v8, v12, v12, v0.t
 ; CHECK-NEXT:    vmor.mm v0, v8, v10
 ; CHECK-NEXT:    ret
@@ -1420,14 +1420,14 @@ define <vscale x 8 x i1> @fcmp_uno_vf_nxv8bf16(<vscale x 8 x bfloat> %va, bfloat
 ; CHECK-LABEL: fcmp_uno_vf_nxv8bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
 ; CHECK-NEXT:    vmv.v.x v8, a1
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmfne.vv v10, v12, v12, v0.t
-; CHECK-NEXT:    vsetvli a1, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e16, m2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmfne.vv v8, v12, v12, v0.t
 ; CHECK-NEXT:    vmor.mm v0, v10, v8
 ; CHECK-NEXT:    ret
@@ -1441,14 +1441,14 @@ define <vscale x 8 x i1> @fcmp_uno_vf_swap_nxv8bf16(<vscale x 8 x bfloat> %va, b
 ; CHECK-LABEL: fcmp_uno_vf_swap_nxv8bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
 ; CHECK-NEXT:    vmv.v.x v8, a1
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmfne.vv v10, v12, v12, v0.t
-; CHECK-NEXT:    vsetvli a1, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e16, m2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vmfne.vv v8, v12, v12, v0.t
 ; CHECK-NEXT:    vmor.mm v0, v8, v10
 ; CHECK-NEXT:    ret
@@ -1521,14 +1521,14 @@ define <vscale x 64 x i1> @fcmp_oeq_vv_nxv64bf16(<vscale x 64 x bfloat> %va, <vs
 ; CHECK-NEXT:    add a0, sp, a0
 ; CHECK-NEXT:    addi a0, a0, 16
 ; CHECK-NEXT:    vs8r.v v0, (a0) # Unknown-size Folded Spill
-; CHECK-NEXT:    vsetvli a0, zero, e16, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a7, e16, m4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v4
 ; CHECK-NEXT:    addi a0, sp, 16
 ; CHECK-NEXT:    vs8r.v v8, (a0) # Unknown-size Folded Spill
 ; CHECK-NEXT:    vmv8r.v v8, v16
 ; CHECK-NEXT:    csrr a0, vlenb
-; CHECK-NEXT:    slli t0, a0, 4
-; CHECK-NEXT:    add a0, t0, a0
+; CHECK-NEXT:    slli a7, a0, 4
+; CHECK-NEXT:    add a0, a7, a0
 ; CHECK-NEXT:    add a0, sp, a0
 ; CHECK-NEXT:    addi a0, a0, 16
 ; CHECK-NEXT:    vs8r.v v16, (a0) # Unknown-size Folded Spill
@@ -1536,7 +1536,7 @@ define <vscale x 64 x i1> @fcmp_oeq_vv_nxv64bf16(<vscale x 64 x bfloat> %va, <vs
 ; CHECK-NEXT:    vmv1r.v v0, v24
 ; CHECK-NEXT:    addi a0, sp, 16
 ; CHECK-NEXT:    vl8r.v v8, (a0) # Unknown-size Folded Reload
-; CHECK-NEXT:    vsetvli zero, a7, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vmfeq.vv v6, v16, v8, v0.t
 ; CHECK-NEXT:    bltu a6, a4, .LBB85_2
 ; CHECK-NEXT:  # %bb.1:
@@ -1549,16 +1549,16 @@ define <vscale x 64 x i1> @fcmp_oeq_vv_nxv64bf16(<vscale x 64 x bfloat> %va, <vs
 ; CHECK-NEXT:    add a0, sp, a0
 ; CHECK-NEXT:    addi a0, a0, 16
 ; CHECK-NEXT:    vl8r.v v16, (a0) # Unknown-size Folded Reload
-; CHECK-NEXT:    vsetvli a0, zero, e16, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a6, e16, m4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v16
 ; CHECK-NEXT:    csrr a0, vlenb
-; CHECK-NEXT:    slli a7, a0, 4
-; CHECK-NEXT:    add a0, a7, a0
+; CHECK-NEXT:    slli a6, a0, 4
+; CHECK-NEXT:    add a0, a6, a0
 ; CHECK-NEXT:    add a0, sp, a0
 ; CHECK-NEXT:    addi a0, a0, 16
 ; CHECK-NEXT:    vl8r.v v16, (a0) # Unknown-size Folded Reload
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v16
-; CHECK-NEXT:    vsetvli zero, a6, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vmfeq.vv v5, v24, v8, v0.t
 ; CHECK-NEXT:    add a0, a3, a3
 ; CHECK-NEXT:    bltu a2, a5, .LBB85_4
@@ -1573,6 +1573,9 @@ define <vscale x 64 x i1> @fcmp_oeq_vv_nxv64bf16(<vscale x 64 x bfloat> %va, <vs
 ; CHECK-NEXT:    vl1r.v v7, (a6) # Unknown-size Folded Reload
 ; CHECK-NEXT:    vsetvli a6, zero, e8, mf2, ta, ma
 ; CHECK-NEXT:    vslidedown.vx v0, v7, a3
+; CHECK-NEXT:    sltu a6, a2, a5
+; CHECK-NEXT:    addi a6, a6, -1
+; CHECK-NEXT:    and a5, a6, a5
 ; CHECK-NEXT:    csrr a6, vlenb
 ; CHECK-NEXT:    mv a7, a6
 ; CHECK-NEXT:    slli a6, a6, 3
@@ -1582,31 +1585,28 @@ define <vscale x 64 x i1> @fcmp_oeq_vv_nxv64bf16(<vscale x 64 x bfloat> %va, <vs
 ; CHECK-NEXT:    add a6, sp, a6
 ; CHECK-NEXT:    addi a6, a6, 16
 ; CHECK-NEXT:    vl8r.v v16, (a6) # Unknown-size Folded Reload
-; CHECK-NEXT:    vsetvli a6, zero, e16, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a5, e16, m4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v20
-; CHECK-NEXT:    csrr a6, vlenb
-; CHECK-NEXT:    slli a7, a6, 4
-; CHECK-NEXT:    add a6, a7, a6
-; CHECK-NEXT:    add a6, sp, a6
-; CHECK-NEXT:    addi a6, a6, 16
-; CHECK-NEXT:    vs8r.v v8, (a6) # Unknown-size Folded Spill
-; CHECK-NEXT:    csrr a6, vlenb
-; CHECK-NEXT:    slli a7, a6, 5
-; CHECK-NEXT:    add a6, a7, a6
-; CHECK-NEXT:    add a6, sp, a6
-; CHECK-NEXT:    addi a6, a6, 16
-; CHECK-NEXT:    vl8r.v v24, (a6) # Unknown-size Folded Reload
+; CHECK-NEXT:    csrr a5, vlenb
+; CHECK-NEXT:    slli a6, a5, 4
+; CHECK-NEXT:    add a5, a6, a5
+; CHECK-NEXT:    add a5, sp, a5
+; CHECK-NEXT:    addi a5, a5, 16
+; CHECK-NEXT:    vs8r.v v8, (a5) # Unknown-size Folded Spill
+; CHECK-NEXT:    csrr a5, vlenb
+; CHECK-NEXT:    slli a6, a5, 5
+; CHECK-NEXT:    add a5, a6, a5
+; CHECK-NEXT:    add a5, sp, a5
+; CHECK-NEXT:    addi a5, a5, 16
+; CHECK-NEXT:    vl8r.v v24, (a5) # Unknown-size Folded Reload
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v28
-; CHECK-NEXT:    sltu a6, a2, a5
-; CHECK-NEXT:    addi a6, a6, -1
-; CHECK-NEXT:    and a5, a6, a5
-; CHECK-NEXT:    csrr a6, vlenb
-; CHECK-NEXT:    slli a7, a6, 4
-; CHECK-NEXT:    add a6, a7, a6
-; CHECK-NEXT:    add a6, sp, a6
-; CHECK-NEXT:    addi a6, a6, 16
-; CHECK-NEXT:    vl8r.v v24, (a6) # Unknown-size Folded Reload
-; CHECK-NEXT:    vsetvli zero, a5, e32, m8, ta, ma
+; CHECK-NEXT:    csrr a5, vlenb
+; CHECK-NEXT:    slli a6, a5, 4
+; CHECK-NEXT:    add a5, a6, a5
+; CHECK-NEXT:    add a5, sp, a5
+; CHECK-NEXT:    addi a5, a5, 16
+; CHECK-NEXT:    vl8r.v v24, (a5) # Unknown-size Folded Reload
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vmfeq.vv v4, v24, v8, v0.t
 ; CHECK-NEXT:    vsetvli zero, a0, e8, mf2, ta, ma
 ; CHECK-NEXT:    vslideup.vx v5, v6, a3
@@ -1615,16 +1615,16 @@ define <vscale x 64 x i1> @fcmp_oeq_vv_nxv64bf16(<vscale x 64 x bfloat> %va, <vs
 ; CHECK-NEXT:    mv a2, a4
 ; CHECK-NEXT:  .LBB85_6:
 ; CHECK-NEXT:    vmv1r.v v0, v7
-; CHECK-NEXT:    vsetvli a4, zero, e16, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a2, e16, m4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v16
-; CHECK-NEXT:    csrr a4, vlenb
-; CHECK-NEXT:    slli a5, a4, 5
-; CHECK-NEXT:    add a4, a5, a4
-; CHECK-NEXT:    add a4, sp, a4
-; CHECK-NEXT:    addi a4, a4, 16
-; CHECK-NEXT:    vl8r.v v8, (a4) # Unknown-size Folded Reload
+; CHECK-NEXT:    csrr a2, vlenb
+; CHECK-NEXT:    slli a4, a2, 5
+; CHECK-NEXT:    add a2, a4, a2
+; CHECK-NEXT:    add a2, sp, a2
+; CHECK-NEXT:    addi a2, a2, 16
+; CHECK-NEXT:    vl8r.v v8, (a2) # Unknown-size Folded Reload
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v8
-; CHECK-NEXT:    vsetvli zero, a2, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vmfeq.vv v8, v24, v16, v0.t
 ; CHECK-NEXT:    vsetvli zero, a0, e8, mf2, ta, ma
 ; CHECK-NEXT:    vslideup.vx v8, v4, a3
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfadd-vp.ll b/llvm/test/CodeGen/RISCV/rvv/vfadd-vp.ll
index 31359c3f68ec7..4a8ee54d87eba 100644
--- a/llvm/test/CodeGen/RISCV/rvv/vfadd-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/vfadd-vp.ll
@@ -17,10 +17,10 @@ declare <vscale x 1 x bfloat> @llvm.vp.fadd.nxv1bf16(<vscale x 1 x bfloat>, <vsc
 define <vscale x 1 x bfloat> @vfadd_vv_nxv1bf16(<vscale x 1 x bfloat> %va, <vscale x 1 x bfloat> %b, <vscale x 1 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: vfadd_vv_nxv1bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v9
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vfadd.vv v9, v9, v10, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v9
@@ -32,10 +32,10 @@ define <vscale x 1 x bfloat> @vfadd_vv_nxv1bf16(<vscale x 1 x bfloat> %va, <vsca
 define <vscale x 1 x bfloat> @vfadd_vv_nxv1bf16_unmasked(<vscale x 1 x bfloat> %va, <vscale x 1 x bfloat> %b, i32 zeroext %evl) {
 ; CHECK-LABEL: vfadd_vv_nxv1bf16_unmasked:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v9
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vfadd.vv v9, v9, v10
 ; CHECK-NEXT:    vsetvli a0, zero, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v9
@@ -48,11 +48,11 @@ define <vscale x 1 x bfloat> @vfadd_vf_nxv1bf16(<vscale x 1 x bfloat> %va, bfloa
 ; CHECK-LABEL: vfadd_vf_nxv1bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vmv.v.x v9, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vfadd.vv v9, v10, v8, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v9
@@ -67,11 +67,11 @@ define <vscale x 1 x bfloat> @vfadd_vf_nxv1bf16_commute(<vscale x 1 x bfloat> %v
 ; CHECK-LABEL: vfadd_vf_nxv1bf16_commute:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vmv.v.x v9, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vfadd.vv v9, v8, v10, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v9
@@ -86,11 +86,11 @@ define <vscale x 1 x bfloat> @vfadd_vf_nxv1bf16_unmasked(<vscale x 1 x bfloat> %
 ; CHECK-LABEL: vfadd_vf_nxv1bf16_unmasked:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vmv.v.x v9, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vfadd.vv v9, v10, v8
 ; CHECK-NEXT:    vsetvli a0, zero, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v9
@@ -105,11 +105,11 @@ define <vscale x 1 x bfloat> @vfadd_vf_nxv1bf16_unmasked_commute(<vscale x 1 x b
 ; CHECK-LABEL: vfadd_vf_nxv1bf16_unmasked_commute:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vmv.v.x v9, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vfadd.vv v9, v8, v10
 ; CHECK-NEXT:    vsetvli a0, zero, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v9
@@ -125,10 +125,10 @@ declare <vscale x 2 x bfloat> @llvm.vp.fadd.nxv2bf16(<vscale x 2 x bfloat>, <vsc
 define <vscale x 2 x bfloat> @vfadd_vv_nxv2bf16(<vscale x 2 x bfloat> %va, <vscale x 2 x bfloat> %b, <vscale x 2 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: vfadd_vv_nxv2bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v9
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m1, ta, ma
 ; CHECK-NEXT:    vfadd.vv v9, v9, v10, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, mf2, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v9
@@ -140,10 +140,10 @@ define <vscale x 2 x bfloat> @vfadd_vv_nxv2bf16(<vscale x 2 x bfloat> %va, <vsca
 define <vscale x 2 x bfloat> @vfadd_vv_nxv2bf16_unmasked(<vscale x 2 x bfloat> %va, <vscale x 2 x bfloat> %b, i32 zeroext %evl) {
 ; CHECK-LABEL: vfadd_vv_nxv2bf16_unmasked:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v9
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m1, ta, ma
 ; CHECK-NEXT:    vfadd.vv v9, v9, v10
 ; CHECK-NEXT:    vsetvli a0, zero, e16, mf2, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v9
@@ -156,11 +156,11 @@ define <vscale x 2 x bfloat> @vfadd_vf_nxv2bf16(<vscale x 2 x bfloat> %va, bfloa
 ; CHECK-LABEL: vfadd_vf_nxv2bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, ma
 ; CHECK-NEXT:    vmv.v.x v9, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m1, ta, ma
 ; CHECK-NEXT:    vfadd.vv v9, v10, v8, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, mf2, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v9
@@ -175,11 +175,11 @@ define <vscale x 2 x bfloat> @vfadd_vf_nxv2bf16_unmasked(<vscale x 2 x bfloat> %
 ; CHECK-LABEL: vfadd_vf_nxv2bf16_unmasked:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, ma
 ; CHECK-NEXT:    vmv.v.x v9, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m1, ta, ma
 ; CHECK-NEXT:    vfadd.vv v9, v10, v8
 ; CHECK-NEXT:    vsetvli a0, zero, e16, mf2, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v9
@@ -195,10 +195,10 @@ declare <vscale x 4 x bfloat> @llvm.vp.fadd.nxv4bf16(<vscale x 4 x bfloat>, <vsc
 define <vscale x 4 x bfloat> @vfadd_vv_nxv4bf16(<vscale x 4 x bfloat> %va, <vscale x 4 x bfloat> %b, <vscale x 4 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: vfadd_vv_nxv4bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m1, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v9
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
 ; CHECK-NEXT:    vfadd.vv v10, v12, v10, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m1, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v10
@@ -210,10 +210,10 @@ define <vscale x 4 x bfloat> @vfadd_vv_nxv4bf16(<vscale x 4 x bfloat> %va, <vsca
 define <vscale x 4 x bfloat> @vfadd_vv_nxv4bf16_unmasked(<vscale x 4 x bfloat> %va, <vscale x 4 x bfloat> %b, i32 zeroext %evl) {
 ; CHECK-LABEL: vfadd_vv_nxv4bf16_unmasked:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m1, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v9
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
 ; CHECK-NEXT:    vfadd.vv v10, v12, v10
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m1, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v10
@@ -226,11 +226,11 @@ define <vscale x 4 x bfloat> @vfadd_vf_nxv4bf16(<vscale x 4 x bfloat> %va, bfloa
 ; CHECK-LABEL: vfadd_vf_nxv4bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m1, ta, ma
 ; CHECK-NEXT:    vmv.v.x v9, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
 ; CHECK-NEXT:    vfadd.vv v10, v10, v12, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m1, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v10
@@ -245,11 +245,11 @@ define <vscale x 4 x bfloat> @vfadd_vf_nxv4bf16_unmasked(<vscale x 4 x bfloat> %
 ; CHECK-LABEL: vfadd_vf_nxv4bf16_unmasked:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m1, ta, ma
 ; CHECK-NEXT:    vmv.v.x v9, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
 ; CHECK-NEXT:    vfadd.vv v10, v10, v12
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m1, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v10
@@ -265,10 +265,10 @@ declare <vscale x 8 x bfloat> @llvm.vp.fadd.nxv8bf16(<vscale x 8 x bfloat>, <vsc
 define <vscale x 8 x bfloat> @vfadd_vv_nxv8bf16(<vscale x 8 x bfloat> %va, <vscale x 8 x bfloat> %b, <vscale x 8 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: vfadd_vv_nxv8bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v10
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vfadd.vv v12, v16, v12, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m2, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v12
@@ -280,10 +280,10 @@ define <vscale x 8 x bfloat> @vfadd_vv_nxv8bf16(<vscale x 8 x bfloat> %va, <vsca
 define <vscale x 8 x bfloat> @vfadd_vv_nxv8bf16_unmasked(<vscale x 8 x bfloat> %va, <vscale x 8 x bfloat> %b, i32 zeroext %evl) {
 ; CHECK-LABEL: vfadd_vv_nxv8bf16_unmasked:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v10
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vfadd.vv v12, v16, v12
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m2, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v12
@@ -296,11 +296,11 @@ define <vscale x 8 x bfloat> @vfadd_vf_nxv8bf16(<vscale x 8 x bfloat> %va, bfloa
 ; CHECK-LABEL: vfadd_vf_nxv8bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vmv.v.x v10, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v10
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vfadd.vv v12, v12, v16, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m2, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v12
@@ -315,11 +315,11 @@ define <vscale x 8 x bfloat> @vfadd_vf_nxv8bf16_unmasked(<vscale x 8 x bfloat> %
 ; CHECK-LABEL: vfadd_vf_nxv8bf16_unmasked:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vmv.v.x v10, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v10
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vfadd.vv v12, v12, v16
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m2, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v12
@@ -335,10 +335,10 @@ declare <vscale x 16 x bfloat> @llvm.vp.fadd.nxv16bf16(<vscale x 16 x bfloat>, <
 define <vscale x 16 x bfloat> @vfadd_vv_nxv16bf16(<vscale x 16 x bfloat> %va, <vscale x 16 x bfloat> %b, <vscale x 16 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: vfadd_vv_nxv16bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v12
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfadd.vv v16, v24, v16, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v16
@@ -350,10 +350,10 @@ define <vscale x 16 x bfloat> @vfadd_vv_nxv16bf16(<vscale x 16 x bfloat> %va, <v
 define <vscale x 16 x bfloat> @vfadd_vv_nxv16bf16_unmasked(<vscale x 16 x bfloat> %va, <vscale x 16 x bfloat> %b, i32 zeroext %evl) {
 ; CHECK-LABEL: vfadd_vv_nxv16bf16_unmasked:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v12
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfadd.vv v16, v24, v16
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v16
@@ -366,11 +366,11 @@ define <vscale x 16 x bfloat> @vfadd_vf_nxv16bf16(<vscale x 16 x bfloat> %va, bf
 ; CHECK-LABEL: vfadd_vf_nxv16bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m4, ta, ma
 ; CHECK-NEXT:    vmv.v.x v12, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v12
-; CHECK-NEXT:    vsetvli zero, a0, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfadd.vv v16, v16, v24, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v16
@@ -385,11 +385,11 @@ define <vscale x 16 x bfloat> @vfadd_vf_nxv16bf16_unmasked(<vscale x 16 x bfloat
 ; CHECK-LABEL: vfadd_vf_nxv16bf16_unmasked:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m4, ta, ma
 ; CHECK-NEXT:    vmv.v.x v12, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v12
-; CHECK-NEXT:    vsetvli zero, a0, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfadd.vv v16, v16, v24
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v16
@@ -411,23 +411,22 @@ define <vscale x 32 x bfloat> @vfadd_vv_nxv32bf16(<vscale x 32 x bfloat> %va, <v
 ; CHECK-NEXT:    slli a1, a1, 3
 ; CHECK-NEXT:    sub sp, sp, a1
 ; CHECK-NEXT:    .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x08, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 8 * vlenb
-; CHECK-NEXT:    vsetvli a1, zero, e16, m4, ta, ma
+; CHECK-NEXT:    vsetvli a1, zero, e8, mf2, ta, ma
 ; CHECK-NEXT:    vmv1r.v v7, v0
 ; CHECK-NEXT:    csrr a2, vlenb
-; CHECK-NEXT:    addi a1, sp, 16
-; CHECK-NEXT:    vs8r.v v16, (a1) # Unknown-size Folded Spill
-; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v20
 ; CHECK-NEXT:    slli a1, a2, 1
 ; CHECK-NEXT:    srli a2, a2, 2
 ; CHECK-NEXT:    sub a3, a0, a1
-; CHECK-NEXT:    vsetvli a4, zero, e8, mf2, ta, ma
 ; CHECK-NEXT:    vslidedown.vx v0, v0, a2
 ; CHECK-NEXT:    sltu a2, a0, a3
 ; CHECK-NEXT:    addi a2, a2, -1
 ; CHECK-NEXT:    and a2, a2, a3
-; CHECK-NEXT:    vsetvli a3, zero, e16, m4, ta, ma
+; CHECK-NEXT:    addi a3, sp, 16
+; CHECK-NEXT:    vs8r.v v16, (a3) # Unknown-size Folded Spill
+; CHECK-NEXT:    vsetvli zero, a2, e16, m4, ta, ma
+; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v20
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v12
-; CHECK-NEXT:    vsetvli zero, a2, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfadd.vv v16, v16, v24, v0.t
 ; CHECK-NEXT:    vsetvli a2, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v12, v16
@@ -437,10 +436,11 @@ define <vscale x 32 x bfloat> @vfadd_vv_nxv32bf16(<vscale x 32 x bfloat> %va, <v
 ; CHECK-NEXT:  .LBB22_2:
 ; CHECK-NEXT:    addi a1, sp, 16
 ; CHECK-NEXT:    vl8r.v v24, (a1) # Unknown-size Folded Reload
+; CHECK-NEXT:    vsetvli zero, a0, e16, m4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v24
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v8
 ; CHECK-NEXT:    vmv1r.v v0, v7
-; CHECK-NEXT:    vsetvli zero, a0, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfadd.vv v16, v24, v16, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v16
@@ -466,22 +466,21 @@ define <vscale x 32 x bfloat> @vfadd_vv_nxv32bf16_unmasked(<vscale x 32 x bfloat
 ; CHECK-NEXT:    .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x08, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 8 * vlenb
 ; CHECK-NEXT:    csrr a2, vlenb
 ; CHECK-NEXT:    vsetvli a1, zero, e8, m4, ta, ma
-; CHECK-NEXT:    vmset.m v7
-; CHECK-NEXT:    addi a1, sp, 16
-; CHECK-NEXT:    vs8r.v v16, (a1) # Unknown-size Folded Spill
-; CHECK-NEXT:    vsetvli a1, zero, e16, m4, ta, ma
-; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v20
+; CHECK-NEXT:    vmset.m v24
 ; CHECK-NEXT:    slli a1, a2, 1
 ; CHECK-NEXT:    srli a2, a2, 2
 ; CHECK-NEXT:    sub a3, a0, a1
 ; CHECK-NEXT:    vsetvli a4, zero, e8, mf2, ta, ma
-; CHECK-NEXT:    vslidedown.vx v0, v7, a2
+; CHECK-NEXT:    vslidedown.vx v0, v24, a2
 ; CHECK-NEXT:    sltu a2, a0, a3
 ; CHECK-NEXT:    addi a2, a2, -1
 ; CHECK-NEXT:    and a2, a2, a3
-; CHECK-NEXT:    vsetvli a3, zero, e16, m4, ta, ma
+; CHECK-NEXT:    addi a3, sp, 16
+; CHECK-NEXT:    vs8r.v v16, (a3) # Unknown-size Folded Spill
+; CHECK-NEXT:    vsetvli zero, a2, e16, m4, ta, ma
+; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v20
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v12
-; CHECK-NEXT:    vsetvli zero, a2, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfadd.vv v16, v16, v24, v0.t
 ; CHECK-NEXT:    vsetvli a2, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v12, v16
@@ -491,9 +490,10 @@ define <vscale x 32 x bfloat> @vfadd_vv_nxv32bf16_unmasked(<vscale x 32 x bfloat
 ; CHECK-NEXT:  .LBB23_2:
 ; CHECK-NEXT:    addi a1, sp, 16
 ; CHECK-NEXT:    vl8r.v v24, (a1) # Unknown-size Folded Reload
+; CHECK-NEXT:    vsetvli zero, a0, e16, m4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v24
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfadd.vv v16, v24, v16
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v16
@@ -518,14 +518,10 @@ define <vscale x 32 x bfloat> @vfadd_vf_nxv32bf16(<vscale x 32 x bfloat> %va, bf
 ; CHECK-NEXT:    add a1, a2, a1
 ; CHECK-NEXT:    sub sp, sp, a1
 ; CHECK-NEXT:    .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x11, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 17 * vlenb
-; CHECK-NEXT:    vsetvli a1, zero, e16, m4, ta, ma
-; CHECK-NEXT:    vmv8r.v v16, v8
+; CHECK-NEXT:    vsetvli a1, zero, e16, m8, ta, ma
+; CHECK-NEXT:    vmv8r.v v24, v8
 ; CHECK-NEXT:    fmv.x.h a1, fa0
 ; CHECK-NEXT:    csrr a2, vlenb
-; CHECK-NEXT:    addi a3, sp, 16
-; CHECK-NEXT:    vs8r.v v8, (a3) # Unknown-size Folded Spill
-; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v20
-; CHECK-NEXT:    vsetvli a3, zero, e16, m8, ta, ma
 ; CHECK-NEXT:    vmv.v.x v16, a1
 ; CHECK-NEXT:    csrr a1, vlenb
 ; CHECK-NEXT:    slli a3, a1, 3
@@ -546,15 +542,18 @@ define <vscale x 32 x bfloat> @vfadd_vf_nxv32bf16(<vscale x 32 x bfloat> %va, bf
 ; CHECK-NEXT:    sltu a2, a0, a3
 ; CHECK-NEXT:    addi a2, a2, -1
 ; CHECK-NEXT:    and a2, a2, a3
-; CHECK-NEXT:    csrr a3, vlenb
-; CHECK-NEXT:    slli a4, a3, 3
-; CHECK-NEXT:    add a3, a4, a3
-; CHECK-NEXT:    add a3, sp, a3
-; CHECK-NEXT:    addi a3, a3, 16
-; CHECK-NEXT:    vl8r.v v24, (a3) # Unknown-size Folded Reload
-; CHECK-NEXT:    vsetvli a3, zero, e16, m4, ta, ma
+; CHECK-NEXT:    addi a3, sp, 16
+; CHECK-NEXT:    vs8r.v v8, (a3) # Unknown-size Folded Spill
+; CHECK-NEXT:    vsetvli zero, a2, e16, m4, ta, ma
+; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v28
+; CHECK-NEXT:    csrr a2, vlenb
+; CHECK-NEXT:    slli a3, a2, 3
+; CHECK-NEXT:    add a2, a3, a2
+; CHECK-NEXT:    add a2, sp, a2
+; CHECK-NEXT:    addi a2, a2, 16
+; CHECK-NEXT:    vl8r.v v24, (a2) # Unknown-size Folded Reload
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v28
-; CHECK-NEXT:    vsetvli zero, a2, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfadd.vv v16, v8, v16, v0.t
 ; CHECK-NEXT:    vsetvli a2, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v12, v16
@@ -564,20 +563,21 @@ define <vscale x 32 x bfloat> @vfadd_vf_nxv32bf16(<vscale x 32 x bfloat> %va, bf
 ; CHECK-NEXT:  .LBB24_2:
 ; CHECK-NEXT:    addi a1, sp, 16
 ; CHECK-NEXT:    vl8r.v v24, (a1) # Unknown-size Folded Reload
+; CHECK-NEXT:    vsetvli zero, a0, e16, m4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v24
-; CHECK-NEXT:    csrr a1, vlenb
-; CHECK-NEXT:    slli a2, a1, 3
-; CHECK-NEXT:    add a1, a2, a1
-; CHECK-NEXT:    add a1, sp, a1
-; CHECK-NEXT:    addi a1, a1, 16
-; CHECK-NEXT:    vl8r.v v0, (a1) # Unknown-size Folded Reload
+; CHECK-NEXT:    csrr a0, vlenb
+; CHECK-NEXT:    slli a1, a0, 3
+; CHECK-NEXT:    add a0, a1, a0
+; CHECK-NEXT:    add a0, sp, a0
+; CHECK-NEXT:    addi a0, a0, 16
+; CHECK-NEXT:    vl8r.v v0, (a0) # Unknown-size Folded Reload
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v0
-; CHECK-NEXT:    csrr a1, vlenb
-; CHECK-NEXT:    slli a1, a1, 3
-; CHECK-NEXT:    add a1, sp, a1
-; CHECK-NEXT:    addi a1, a1, 16
-; CHECK-NEXT:    vl1r.v v0, (a1) # Unknown-size Folded Reload
-; CHECK-NEXT:    vsetvli zero, a0, e32, m8, ta, ma
+; CHECK-NEXT:    csrr a0, vlenb
+; CHECK-NEXT:    slli a0, a0, 3
+; CHECK-NEXT:    add a0, sp, a0
+; CHECK-NEXT:    addi a0, a0, 16
+; CHECK-NEXT:    vl1r.v v0, (a0) # Unknown-size Folded Reload
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfadd.vv v16, v16, v24, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v16
@@ -604,16 +604,10 @@ define <vscale x 32 x bfloat> @vfadd_vf_nxv32bf16_unmasked(<vscale x 32 x bfloat
 ; CHECK-NEXT:    slli a1, a1, 4
 ; CHECK-NEXT:    sub sp, sp, a1
 ; CHECK-NEXT:    .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x10, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 16 * vlenb
-; CHECK-NEXT:    vsetvli a1, zero, e8, m4, ta, ma
-; CHECK-NEXT:    vmv8r.v v16, v8
 ; CHECK-NEXT:    fmv.x.h a1, fa0
 ; CHECK-NEXT:    csrr a2, vlenb
-; CHECK-NEXT:    vmset.m v7
-; CHECK-NEXT:    addi a3, sp, 16
-; CHECK-NEXT:    vs8r.v v8, (a3) # Unknown-size Folded Spill
-; CHECK-NEXT:    vsetvli a3, zero, e16, m4, ta, ma
-; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v20
 ; CHECK-NEXT:    vsetvli a3, zero, e16, m8, ta, ma
+; CHECK-NEXT:    vmset.m v24
 ; CHECK-NEXT:    vmv.v.x v16, a1
 ; CHECK-NEXT:    csrr a1, vlenb
 ; CHECK-NEXT:    slli a1, a1, 3
@@ -624,18 +618,22 @@ define <vscale x 32 x bfloat> @vfadd_vf_nxv32bf16_unmasked(<vscale x 32 x bfloat
 ; CHECK-NEXT:    srli a2, a2, 2
 ; CHECK-NEXT:    sub a3, a0, a1
 ; CHECK-NEXT:    vsetvli a4, zero, e8, mf2, ta, ma
-; CHECK-NEXT:    vslidedown.vx v0, v7, a2
+; CHECK-NEXT:    vslidedown.vx v0, v24, a2
 ; CHECK-NEXT:    sltu a2, a0, a3
 ; CHECK-NEXT:    addi a2, a2, -1
 ; CHECK-NEXT:    and a2, a2, a3
-; CHECK-NEXT:    csrr a3, vlenb
-; CHECK-NEXT:    slli a3, a3, 3
-; CHECK-NEXT:    add a3, sp, a3
-; CHECK-NEXT:    addi a3, a3, 16
-; CHECK-NEXT:    vl8r.v v24, (a3) # Unknown-size Folded Reload
-; CHECK-NEXT:    vsetvli a3, zero, e16, m4, ta, ma
+; CHECK-NEXT:    vmv8r.v v16, v8
+; CHECK-NEXT:    addi a3, sp, 16
+; CHECK-NEXT:    vs8r.v v8, (a3) # Unknown-size Folded Spill
+; CHECK-NEXT:    vsetvli zero, a2, e16, m4, ta, ma
+; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v20
+; CHECK-NEXT:    csrr a2, vlenb
+; CHECK-NEXT:    slli a2, a2, 3
+; CHECK-NEXT:    add a2, sp, a2
+; CHECK-NEXT:    addi a2, a2, 16
+; CHECK-NEXT:    vl8r.v v24, (a2) # Unknown-size Folded Reload
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v28
-; CHECK-NEXT:    vsetvli zero, a2, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfadd.vv v16, v8, v16, v0.t
 ; CHECK-NEXT:    vsetvli a2, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v12, v16
@@ -645,14 +643,15 @@ define <vscale x 32 x bfloat> @vfadd_vf_nxv32bf16_unmasked(<vscale x 32 x bfloat
 ; CHECK-NEXT:  .LBB25_2:
 ; CHECK-NEXT:    addi a1, sp, 16
 ; CHECK-NEXT:    vl8r.v v24, (a1) # Unknown-size Folded Reload
+; CHECK-NEXT:    vsetvli zero, a0, e16, m4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v24
-; CHECK-NEXT:    csrr a1, vlenb
-; CHECK-NEXT:    slli a1, a1, 3
-; CHECK-NEXT:    add a1, sp, a1
-; CHECK-NEXT:    addi a1, a1, 16
-; CHECK-NEXT:    vl8r.v v0, (a1) # Unknown-size Folded Reload
+; CHECK-NEXT:    csrr a0, vlenb
+; CHECK-NEXT:    slli a0, a0, 3
+; CHECK-NEXT:    add a0, sp, a0
+; CHECK-NEXT:    addi a0, a0, 16
+; CHECK-NEXT:    vl8r.v v0, (a0) # Unknown-size Folded Reload
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v0
-; CHECK-NEXT:    vsetvli zero, a0, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfadd.vv v16, v16, v24
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v16
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfdiv-vp.ll b/llvm/test/CodeGen/RISCV/rvv/vfdiv-vp.ll
index 2205769d3494a..2997fad370e46 100644
--- a/llvm/test/CodeGen/RISCV/rvv/vfdiv-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/vfdiv-vp.ll
@@ -17,10 +17,10 @@ declare <vscale x 1 x bfloat> @llvm.vp.fdiv.nxv1bf16(<vscale x 1 x bfloat>, <vsc
 define <vscale x 1 x bfloat> @vfdiv_vv_nxv1bf16(<vscale x 1 x bfloat> %va, <vscale x 1 x bfloat> %b, <vscale x 1 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: vfdiv_vv_nxv1bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v9
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vfdiv.vv v9, v9, v10, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v9
@@ -32,10 +32,10 @@ define <vscale x 1 x bfloat> @vfdiv_vv_nxv1bf16(<vscale x 1 x bfloat> %va, <vsca
 define <vscale x 1 x bfloat> @vfdiv_vv_nxv1bf16_unmasked(<vscale x 1 x bfloat> %va, <vscale x 1 x bfloat> %b, i32 zeroext %evl) {
 ; CHECK-LABEL: vfdiv_vv_nxv1bf16_unmasked:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v9
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vfdiv.vv v9, v9, v10
 ; CHECK-NEXT:    vsetvli a0, zero, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v9
@@ -48,11 +48,11 @@ define <vscale x 1 x bfloat> @vfdiv_vf_nxv1bf16(<vscale x 1 x bfloat> %va, bfloa
 ; CHECK-LABEL: vfdiv_vf_nxv1bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vmv.v.x v9, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vfdiv.vv v9, v10, v8, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v9
@@ -67,11 +67,11 @@ define <vscale x 1 x bfloat> @vfdiv_vf_nxv1bf16_unmasked(<vscale x 1 x bfloat> %
 ; CHECK-LABEL: vfdiv_vf_nxv1bf16_unmasked:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vmv.v.x v9, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vfdiv.vv v9, v10, v8
 ; CHECK-NEXT:    vsetvli a0, zero, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v9
@@ -87,10 +87,10 @@ declare <vscale x 2 x bfloat> @llvm.vp.fdiv.nxv2bf16(<vscale x 2 x bfloat>, <vsc
 define <vscale x 2 x bfloat> @vfdiv_vv_nxv2bf16(<vscale x 2 x bfloat> %va, <vscale x 2 x bfloat> %b, <vscale x 2 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: vfdiv_vv_nxv2bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v9
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m1, ta, ma
 ; CHECK-NEXT:    vfdiv.vv v9, v9, v10, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, mf2, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v9
@@ -102,10 +102,10 @@ define <vscale x 2 x bfloat> @vfdiv_vv_nxv2bf16(<vscale x 2 x bfloat> %va, <vsca
 define <vscale x 2 x bfloat> @vfdiv_vv_nxv2bf16_unmasked(<vscale x 2 x bfloat> %va, <vscale x 2 x bfloat> %b, i32 zeroext %evl) {
 ; CHECK-LABEL: vfdiv_vv_nxv2bf16_unmasked:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v9
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m1, ta, ma
 ; CHECK-NEXT:    vfdiv.vv v9, v9, v10
 ; CHECK-NEXT:    vsetvli a0, zero, e16, mf2, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v9
@@ -118,11 +118,11 @@ define <vscale x 2 x bfloat> @vfdiv_vf_nxv2bf16(<vscale x 2 x bfloat> %va, bfloa
 ; CHECK-LABEL: vfdiv_vf_nxv2bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, ma
 ; CHECK-NEXT:    vmv.v.x v9, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m1, ta, ma
 ; CHECK-NEXT:    vfdiv.vv v9, v10, v8, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, mf2, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v9
@@ -137,11 +137,11 @@ define <vscale x 2 x bfloat> @vfdiv_vf_nxv2bf16_unmasked(<vscale x 2 x bfloat> %
 ; CHECK-LABEL: vfdiv_vf_nxv2bf16_unmasked:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, ma
 ; CHECK-NEXT:    vmv.v.x v9, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m1, ta, ma
 ; CHECK-NEXT:    vfdiv.vv v9, v10, v8
 ; CHECK-NEXT:    vsetvli a0, zero, e16, mf2, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v9
@@ -157,10 +157,10 @@ declare <vscale x 4 x bfloat> @llvm.vp.fdiv.nxv4bf16(<vscale x 4 x bfloat>, <vsc
 define <vscale x 4 x bfloat> @vfdiv_vv_nxv4bf16(<vscale x 4 x bfloat> %va, <vscale x 4 x bfloat> %b, <vscale x 4 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: vfdiv_vv_nxv4bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m1, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v9
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
 ; CHECK-NEXT:    vfdiv.vv v10, v12, v10, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m1, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v10
@@ -172,10 +172,10 @@ define <vscale x 4 x bfloat> @vfdiv_vv_nxv4bf16(<vscale x 4 x bfloat> %va, <vsca
 define <vscale x 4 x bfloat> @vfdiv_vv_nxv4bf16_unmasked(<vscale x 4 x bfloat> %va, <vscale x 4 x bfloat> %b, i32 zeroext %evl) {
 ; CHECK-LABEL: vfdiv_vv_nxv4bf16_unmasked:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m1, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v9
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
 ; CHECK-NEXT:    vfdiv.vv v10, v12, v10
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m1, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v10
@@ -188,11 +188,11 @@ define <vscale x 4 x bfloat> @vfdiv_vf_nxv4bf16(<vscale x 4 x bfloat> %va, bfloa
 ; CHECK-LABEL: vfdiv_vf_nxv4bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m1, ta, ma
 ; CHECK-NEXT:    vmv.v.x v9, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
 ; CHECK-NEXT:    vfdiv.vv v10, v10, v12, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m1, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v10
@@ -207,11 +207,11 @@ define <vscale x 4 x bfloat> @vfdiv_vf_nxv4bf16_unmasked(<vscale x 4 x bfloat> %
 ; CHECK-LABEL: vfdiv_vf_nxv4bf16_unmasked:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m1, ta, ma
 ; CHECK-NEXT:    vmv.v.x v9, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
 ; CHECK-NEXT:    vfdiv.vv v10, v10, v12
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m1, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v10
@@ -227,10 +227,10 @@ declare <vscale x 8 x bfloat> @llvm.vp.fdiv.nxv8bf16(<vscale x 8 x bfloat>, <vsc
 define <vscale x 8 x bfloat> @vfdiv_vv_nxv8bf16(<vscale x 8 x bfloat> %va, <vscale x 8 x bfloat> %b, <vscale x 8 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: vfdiv_vv_nxv8bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v10
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vfdiv.vv v12, v16, v12, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m2, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v12
@@ -242,10 +242,10 @@ define <vscale x 8 x bfloat> @vfdiv_vv_nxv8bf16(<vscale x 8 x bfloat> %va, <vsca
 define <vscale x 8 x bfloat> @vfdiv_vv_nxv8bf16_unmasked(<vscale x 8 x bfloat> %va, <vscale x 8 x bfloat> %b, i32 zeroext %evl) {
 ; CHECK-LABEL: vfdiv_vv_nxv8bf16_unmasked:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v10
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vfdiv.vv v12, v16, v12
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m2, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v12
@@ -258,11 +258,11 @@ define <vscale x 8 x bfloat> @vfdiv_vf_nxv8bf16(<vscale x 8 x bfloat> %va, bfloa
 ; CHECK-LABEL: vfdiv_vf_nxv8bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vmv.v.x v10, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v10
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vfdiv.vv v12, v12, v16, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m2, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v12
@@ -277,11 +277,11 @@ define <vscale x 8 x bfloat> @vfdiv_vf_nxv8bf16_unmasked(<vscale x 8 x bfloat> %
 ; CHECK-LABEL: vfdiv_vf_nxv8bf16_unmasked:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vmv.v.x v10, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v10
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vfdiv.vv v12, v12, v16
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m2, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v12
@@ -297,10 +297,10 @@ declare <vscale x 16 x bfloat> @llvm.vp.fdiv.nxv16bf16(<vscale x 16 x bfloat>, <
 define <vscale x 16 x bfloat> @vfdiv_vv_nxv16bf16(<vscale x 16 x bfloat> %va, <vscale x 16 x bfloat> %b, <vscale x 16 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: vfdiv_vv_nxv16bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v12
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfdiv.vv v16, v24, v16, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v16
@@ -312,10 +312,10 @@ define <vscale x 16 x bfloat> @vfdiv_vv_nxv16bf16(<vscale x 16 x bfloat> %va, <v
 define <vscale x 16 x bfloat> @vfdiv_vv_nxv16bf16_unmasked(<vscale x 16 x bfloat> %va, <vscale x 16 x bfloat> %b, i32 zeroext %evl) {
 ; CHECK-LABEL: vfdiv_vv_nxv16bf16_unmasked:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v12
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfdiv.vv v16, v24, v16
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v16
@@ -328,11 +328,11 @@ define <vscale x 16 x bfloat> @vfdiv_vf_nxv16bf16(<vscale x 16 x bfloat> %va, bf
 ; CHECK-LABEL: vfdiv_vf_nxv16bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m4, ta, ma
 ; CHECK-NEXT:    vmv.v.x v12, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v12
-; CHECK-NEXT:    vsetvli zero, a0, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfdiv.vv v16, v16, v24, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v16
@@ -347,11 +347,11 @@ define <vscale x 16 x bfloat> @vfdiv_vf_nxv16bf16_unmasked(<vscale x 16 x bfloat
 ; CHECK-LABEL: vfdiv_vf_nxv16bf16_unmasked:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m4, ta, ma
 ; CHECK-NEXT:    vmv.v.x v12, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v12
-; CHECK-NEXT:    vsetvli zero, a0, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfdiv.vv v16, v16, v24
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v16
@@ -373,23 +373,22 @@ define <vscale x 32 x bfloat> @vfdiv_vv_nxv32bf16(<vscale x 32 x bfloat> %va, <v
 ; CHECK-NEXT:    slli a1, a1, 3
 ; CHECK-NEXT:    sub sp, sp, a1
 ; CHECK-NEXT:    .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x08, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 8 * vlenb
-; CHECK-NEXT:    vsetvli a1, zero, e16, m4, ta, ma
+; CHECK-NEXT:    vsetvli a1, zero, e8, mf2, ta, ma
 ; CHECK-NEXT:    vmv1r.v v7, v0
 ; CHECK-NEXT:    csrr a2, vlenb
-; CHECK-NEXT:    addi a1, sp, 16
-; CHECK-NEXT:    vs8r.v v16, (a1) # Unknown-size Folded Spill
-; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v20
 ; CHECK-NEXT:    slli a1, a2, 1
 ; CHECK-NEXT:    srli a2, a2, 2
 ; CHECK-NEXT:    sub a3, a0, a1
-; CHECK-NEXT:    vsetvli a4, zero, e8, mf2, ta, ma
 ; CHECK-NEXT:    vslidedown.vx v0, v0, a2
 ; CHECK-NEXT:    sltu a2, a0, a3
 ; CHECK-NEXT:    addi a2, a2, -1
 ; CHECK-NEXT:    and a2, a2, a3
-; CHECK-NEXT:    vsetvli a3, zero, e16, m4, ta, ma
+; CHECK-NEXT:    addi a3, sp, 16
+; CHECK-NEXT:    vs8r.v v16, (a3) # Unknown-size Folded Spill
+; CHECK-NEXT:    vsetvli zero, a2, e16, m4, ta, ma
+; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v20
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v12
-; CHECK-NEXT:    vsetvli zero, a2, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfdiv.vv v16, v16, v24, v0.t
 ; CHECK-NEXT:    vsetvli a2, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v12, v16
@@ -399,10 +398,11 @@ define <vscale x 32 x bfloat> @vfdiv_vv_nxv32bf16(<vscale x 32 x bfloat> %va, <v
 ; CHECK-NEXT:  .LBB20_2:
 ; CHECK-NEXT:    addi a1, sp, 16
 ; CHECK-NEXT:    vl8r.v v24, (a1) # Unknown-size Folded Reload
+; CHECK-NEXT:    vsetvli zero, a0, e16, m4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v24
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v8
 ; CHECK-NEXT:    vmv1r.v v0, v7
-; CHECK-NEXT:    vsetvli zero, a0, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfdiv.vv v16, v24, v16, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v16
@@ -428,22 +428,21 @@ define <vscale x 32 x bfloat> @vfdiv_vv_nxv32bf16_unmasked(<vscale x 32 x bfloat
 ; CHECK-NEXT:    .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x08, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 8 * vlenb
 ; CHECK-NEXT:    csrr a2, vlenb
 ; CHECK-NEXT:    vsetvli a1, zero, e8, m4, ta, ma
-; CHECK-NEXT:    vmset.m v7
-; CHECK-NEXT:    addi a1, sp, 16
-; CHECK-NEXT:    vs8r.v v16, (a1) # Unknown-size Folded Spill
-; CHECK-NEXT:    vsetvli a1, zero, e16, m4, ta, ma
-; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v20
+; CHECK-NEXT:    vmset.m v24
 ; CHECK-NEXT:    slli a1, a2, 1
 ; CHECK-NEXT:    srli a2, a2, 2
 ; CHECK-NEXT:    sub a3, a0, a1
 ; CHECK-NEXT:    vsetvli a4, zero, e8, mf2, ta, ma
-; CHECK-NEXT:    vslidedown.vx v0, v7, a2
+; CHECK-NEXT:    vslidedown.vx v0, v24, a2
 ; CHECK-NEXT:    sltu a2, a0, a3
 ; CHECK-NEXT:    addi a2, a2, -1
 ; CHECK-NEXT:    and a2, a2, a3
-; CHECK-NEXT:    vsetvli a3, zero, e16, m4, ta, ma
+; CHECK-NEXT:    addi a3, sp, 16
+; CHECK-NEXT:    vs8r.v v16, (a3) # Unknown-size Folded Spill
+; CHECK-NEXT:    vsetvli zero, a2, e16, m4, ta, ma
+; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v20
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v12
-; CHECK-NEXT:    vsetvli zero, a2, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfdiv.vv v16, v16, v24, v0.t
 ; CHECK-NEXT:    vsetvli a2, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v12, v16
@@ -453,9 +452,10 @@ define <vscale x 32 x bfloat> @vfdiv_vv_nxv32bf16_unmasked(<vscale x 32 x bfloat
 ; CHECK-NEXT:  .LBB21_2:
 ; CHECK-NEXT:    addi a1, sp, 16
 ; CHECK-NEXT:    vl8r.v v24, (a1) # Unknown-size Folded Reload
+; CHECK-NEXT:    vsetvli zero, a0, e16, m4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v24
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfdiv.vv v16, v24, v16
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v16
@@ -480,14 +480,10 @@ define <vscale x 32 x bfloat> @vfdiv_vf_nxv32bf16(<vscale x 32 x bfloat> %va, bf
 ; CHECK-NEXT:    add a1, a2, a1
 ; CHECK-NEXT:    sub sp, sp, a1
 ; CHECK-NEXT:    .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x11, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 17 * vlenb
-; CHECK-NEXT:    vsetvli a1, zero, e16, m4, ta, ma
-; CHECK-NEXT:    vmv8r.v v16, v8
+; CHECK-NEXT:    vsetvli a1, zero, e16, m8, ta, ma
+; CHECK-NEXT:    vmv8r.v v24, v8
 ; CHECK-NEXT:    fmv.x.h a1, fa0
 ; CHECK-NEXT:    csrr a2, vlenb
-; CHECK-NEXT:    addi a3, sp, 16
-; CHECK-NEXT:    vs8r.v v8, (a3) # Unknown-size Folded Spill
-; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v20
-; CHECK-NEXT:    vsetvli a3, zero, e16, m8, ta, ma
 ; CHECK-NEXT:    vmv.v.x v16, a1
 ; CHECK-NEXT:    csrr a1, vlenb
 ; CHECK-NEXT:    slli a3, a1, 3
@@ -508,15 +504,18 @@ define <vscale x 32 x bfloat> @vfdiv_vf_nxv32bf16(<vscale x 32 x bfloat> %va, bf
 ; CHECK-NEXT:    sltu a2, a0, a3
 ; CHECK-NEXT:    addi a2, a2, -1
 ; CHECK-NEXT:    and a2, a2, a3
-; CHECK-NEXT:    csrr a3, vlenb
-; CHECK-NEXT:    slli a4, a3, 3
-; CHECK-NEXT:    add a3, a4, a3
-; CHECK-NEXT:    add a3, sp, a3
-; CHECK-NEXT:    addi a3, a3, 16
-; CHECK-NEXT:    vl8r.v v24, (a3) # Unknown-size Folded Reload
-; CHECK-NEXT:    vsetvli a3, zero, e16, m4, ta, ma
+; CHECK-NEXT:    addi a3, sp, 16
+; CHECK-NEXT:    vs8r.v v8, (a3) # Unknown-size Folded Spill
+; CHECK-NEXT:    vsetvli zero, a2, e16, m4, ta, ma
+; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v28
+; CHECK-NEXT:    csrr a2, vlenb
+; CHECK-NEXT:    slli a3, a2, 3
+; CHECK-NEXT:    add a2, a3, a2
+; CHECK-NEXT:    add a2, sp, a2
+; CHECK-NEXT:    addi a2, a2, 16
+; CHECK-NEXT:    vl8r.v v24, (a2) # Unknown-size Folded Reload
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v28
-; CHECK-NEXT:    vsetvli zero, a2, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfdiv.vv v16, v8, v16, v0.t
 ; CHECK-NEXT:    vsetvli a2, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v12, v16
@@ -526,20 +525,21 @@ define <vscale x 32 x bfloat> @vfdiv_vf_nxv32bf16(<vscale x 32 x bfloat> %va, bf
 ; CHECK-NEXT:  .LBB22_2:
 ; CHECK-NEXT:    addi a1, sp, 16
 ; CHECK-NEXT:    vl8r.v v24, (a1) # Unknown-size Folded Reload
+; CHECK-NEXT:    vsetvli zero, a0, e16, m4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v24
-; CHECK-NEXT:    csrr a1, vlenb
-; CHECK-NEXT:    slli a2, a1, 3
-; CHECK-NEXT:    add a1, a2, a1
-; CHECK-NEXT:    add a1, sp, a1
-; CHECK-NEXT:    addi a1, a1, 16
-; CHECK-NEXT:    vl8r.v v0, (a1) # Unknown-size Folded Reload
+; CHECK-NEXT:    csrr a0, vlenb
+; CHECK-NEXT:    slli a1, a0, 3
+; CHECK-NEXT:    add a0, a1, a0
+; CHECK-NEXT:    add a0, sp, a0
+; CHECK-NEXT:    addi a0, a0, 16
+; CHECK-NEXT:    vl8r.v v0, (a0) # Unknown-size Folded Reload
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v0
-; CHECK-NEXT:    csrr a1, vlenb
-; CHECK-NEXT:    slli a1, a1, 3
-; CHECK-NEXT:    add a1, sp, a1
-; CHECK-NEXT:    addi a1, a1, 16
-; CHECK-NEXT:    vl1r.v v0, (a1) # Unknown-size Folded Reload
-; CHECK-NEXT:    vsetvli zero, a0, e32, m8, ta, ma
+; CHECK-NEXT:    csrr a0, vlenb
+; CHECK-NEXT:    slli a0, a0, 3
+; CHECK-NEXT:    add a0, sp, a0
+; CHECK-NEXT:    addi a0, a0, 16
+; CHECK-NEXT:    vl1r.v v0, (a0) # Unknown-size Folded Reload
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfdiv.vv v16, v16, v24, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v16
@@ -566,16 +566,10 @@ define <vscale x 32 x bfloat> @vfdiv_vf_nxv32bf16_unmasked(<vscale x 32 x bfloat
 ; CHECK-NEXT:    slli a1, a1, 4
 ; CHECK-NEXT:    sub sp, sp, a1
 ; CHECK-NEXT:    .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x10, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 16 * vlenb
-; CHECK-NEXT:    vsetvli a1, zero, e8, m4, ta, ma
-; CHECK-NEXT:    vmv8r.v v16, v8
 ; CHECK-NEXT:    fmv.x.h a1, fa0
 ; CHECK-NEXT:    csrr a2, vlenb
-; CHECK-NEXT:    vmset.m v7
-; CHECK-NEXT:    addi a3, sp, 16
-; CHECK-NEXT:    vs8r.v v8, (a3) # Unknown-size Folded Spill
-; CHECK-NEXT:    vsetvli a3, zero, e16, m4, ta, ma
-; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v20
 ; CHECK-NEXT:    vsetvli a3, zero, e16, m8, ta, ma
+; CHECK-NEXT:    vmset.m v24
 ; CHECK-NEXT:    vmv.v.x v16, a1
 ; CHECK-NEXT:    csrr a1, vlenb
 ; CHECK-NEXT:    slli a1, a1, 3
@@ -586,18 +580,22 @@ define <vscale x 32 x bfloat> @vfdiv_vf_nxv32bf16_unmasked(<vscale x 32 x bfloat
 ; CHECK-NEXT:    srli a2, a2, 2
 ; CHECK-NEXT:    sub a3, a0, a1
 ; CHECK-NEXT:    vsetvli a4, zero, e8, mf2, ta, ma
-; CHECK-NEXT:    vslidedown.vx v0, v7, a2
+; CHECK-NEXT:    vslidedown.vx v0, v24, a2
 ; CHECK-NEXT:    sltu a2, a0, a3
 ; CHECK-NEXT:    addi a2, a2, -1
 ; CHECK-NEXT:    and a2, a2, a3
-; CHECK-NEXT:    csrr a3, vlenb
-; CHECK-NEXT:    slli a3, a3, 3
-; CHECK-NEXT:    add a3, sp, a3
-; CHECK-NEXT:    addi a3, a3, 16
-; CHECK-NEXT:    vl8r.v v24, (a3) # Unknown-size Folded Reload
-; CHECK-NEXT:    vsetvli a3, zero, e16, m4, ta, ma
+; CHECK-NEXT:    vmv8r.v v16, v8
+; CHECK-NEXT:    addi a3, sp, 16
+; CHECK-NEXT:    vs8r.v v8, (a3) # Unknown-size Folded Spill
+; CHECK-NEXT:    vsetvli zero, a2, e16, m4, ta, ma
+; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v20
+; CHECK-NEXT:    csrr a2, vlenb
+; CHECK-NEXT:    slli a2, a2, 3
+; CHECK-NEXT:    add a2, sp, a2
+; CHECK-NEXT:    addi a2, a2, 16
+; CHECK-NEXT:    vl8r.v v24, (a2) # Unknown-size Folded Reload
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v28
-; CHECK-NEXT:    vsetvli zero, a2, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfdiv.vv v16, v8, v16, v0.t
 ; CHECK-NEXT:    vsetvli a2, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v12, v16
@@ -607,14 +605,15 @@ define <vscale x 32 x bfloat> @vfdiv_vf_nxv32bf16_unmasked(<vscale x 32 x bfloat
 ; CHECK-NEXT:  .LBB23_2:
 ; CHECK-NEXT:    addi a1, sp, 16
 ; CHECK-NEXT:    vl8r.v v24, (a1) # Unknown-size Folded Reload
+; CHECK-NEXT:    vsetvli zero, a0, e16, m4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v24
-; CHECK-NEXT:    csrr a1, vlenb
-; CHECK-NEXT:    slli a1, a1, 3
-; CHECK-NEXT:    add a1, sp, a1
-; CHECK-NEXT:    addi a1, a1, 16
-; CHECK-NEXT:    vl8r.v v0, (a1) # Unknown-size Folded Reload
+; CHECK-NEXT:    csrr a0, vlenb
+; CHECK-NEXT:    slli a0, a0, 3
+; CHECK-NEXT:    add a0, sp, a0
+; CHECK-NEXT:    addi a0, a0, 16
+; CHECK-NEXT:    vl8r.v v0, (a0) # Unknown-size Folded Reload
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v0
-; CHECK-NEXT:    vsetvli zero, a0, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfdiv.vv v16, v16, v24
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v16
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfmax-vp.ll b/llvm/test/CodeGen/RISCV/rvv/vfmax-vp.ll
index 5d998c4e739d5..74d289951530a 100644
--- a/llvm/test/CodeGen/RISCV/rvv/vfmax-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/vfmax-vp.ll
@@ -17,10 +17,10 @@ declare <vscale x 1 x bfloat> @llvm.vp.maxnum.nxv1bf16(<vscale x 1 x bfloat>, <v
 define <vscale x 1 x bfloat> @vfmax_vv_nxv1bf16(<vscale x 1 x bfloat> %va, <vscale x 1 x bfloat> %vb, <vscale x 1 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: vfmax_vv_nxv1bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v9
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vfmax.vv v9, v9, v10, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v9
@@ -32,10 +32,10 @@ define <vscale x 1 x bfloat> @vfmax_vv_nxv1bf16(<vscale x 1 x bfloat> %va, <vsca
 define <vscale x 1 x bfloat> @vfmax_vv_nxv1bf16_unmasked(<vscale x 1 x bfloat> %va, <vscale x 1 x bfloat> %vb, i32 zeroext %evl) {
 ; CHECK-LABEL: vfmax_vv_nxv1bf16_unmasked:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v9
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vfmax.vv v9, v9, v10
 ; CHECK-NEXT:    vsetvli a0, zero, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v9
@@ -49,10 +49,10 @@ declare <vscale x 2 x bfloat> @llvm.vp.maxnum.nxv2bf16(<vscale x 2 x bfloat>, <v
 define <vscale x 2 x bfloat> @vfmax_vv_nxv2bf16(<vscale x 2 x bfloat> %va, <vscale x 2 x bfloat> %vb, <vscale x 2 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: vfmax_vv_nxv2bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v9
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m1, ta, ma
 ; CHECK-NEXT:    vfmax.vv v9, v9, v10, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, mf2, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v9
@@ -64,10 +64,10 @@ define <vscale x 2 x bfloat> @vfmax_vv_nxv2bf16(<vscale x 2 x bfloat> %va, <vsca
 define <vscale x 2 x bfloat> @vfmax_vv_nxv2bf16_unmasked(<vscale x 2 x bfloat> %va, <vscale x 2 x bfloat> %vb, i32 zeroext %evl) {
 ; CHECK-LABEL: vfmax_vv_nxv2bf16_unmasked:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v9
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m1, ta, ma
 ; CHECK-NEXT:    vfmax.vv v9, v9, v10
 ; CHECK-NEXT:    vsetvli a0, zero, e16, mf2, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v9
@@ -81,10 +81,10 @@ declare <vscale x 4 x bfloat> @llvm.vp.maxnum.nxv4bf16(<vscale x 4 x bfloat>, <v
 define <vscale x 4 x bfloat> @vfmax_vv_nxv4bf16(<vscale x 4 x bfloat> %va, <vscale x 4 x bfloat> %vb, <vscale x 4 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: vfmax_vv_nxv4bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m1, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v9
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
 ; CHECK-NEXT:    vfmax.vv v10, v12, v10, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m1, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v10
@@ -96,10 +96,10 @@ define <vscale x 4 x bfloat> @vfmax_vv_nxv4bf16(<vscale x 4 x bfloat> %va, <vsca
 define <vscale x 4 x bfloat> @vfmax_vv_nxv4bf16_unmasked(<vscale x 4 x bfloat> %va, <vscale x 4 x bfloat> %vb, i32 zeroext %evl) {
 ; CHECK-LABEL: vfmax_vv_nxv4bf16_unmasked:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m1, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v9
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
 ; CHECK-NEXT:    vfmax.vv v10, v12, v10
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m1, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v10
@@ -113,10 +113,10 @@ declare <vscale x 8 x bfloat> @llvm.vp.maxnum.nxv8bf16(<vscale x 8 x bfloat>, <v
 define <vscale x 8 x bfloat> @vfmax_vv_nxv8bf16(<vscale x 8 x bfloat> %va, <vscale x 8 x bfloat> %vb, <vscale x 8 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: vfmax_vv_nxv8bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v10
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vfmax.vv v12, v16, v12, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m2, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v12
@@ -128,10 +128,10 @@ define <vscale x 8 x bfloat> @vfmax_vv_nxv8bf16(<vscale x 8 x bfloat> %va, <vsca
 define <vscale x 8 x bfloat> @vfmax_vv_nxv8bf16_unmasked(<vscale x 8 x bfloat> %va, <vscale x 8 x bfloat> %vb, i32 zeroext %evl) {
 ; CHECK-LABEL: vfmax_vv_nxv8bf16_unmasked:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v10
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vfmax.vv v12, v16, v12
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m2, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v12
@@ -145,10 +145,10 @@ declare <vscale x 16 x bfloat> @llvm.vp.maxnum.nxv16bf16(<vscale x 16 x bfloat>,
 define <vscale x 16 x bfloat> @vfmax_vv_nxv16bf16(<vscale x 16 x bfloat> %va, <vscale x 16 x bfloat> %vb, <vscale x 16 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: vfmax_vv_nxv16bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v12
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfmax.vv v16, v24, v16, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v16
@@ -160,10 +160,10 @@ define <vscale x 16 x bfloat> @vfmax_vv_nxv16bf16(<vscale x 16 x bfloat> %va, <v
 define <vscale x 16 x bfloat> @vfmax_vv_nxv16bf16_unmasked(<vscale x 16 x bfloat> %va, <vscale x 16 x bfloat> %vb, i32 zeroext %evl) {
 ; CHECK-LABEL: vfmax_vv_nxv16bf16_unmasked:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v12
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfmax.vv v16, v24, v16
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v16
@@ -183,23 +183,22 @@ define <vscale x 32 x bfloat> @vfmax_vv_nxv32bf16(<vscale x 32 x bfloat> %va, <v
 ; CHECK-NEXT:    slli a1, a1, 3
 ; CHECK-NEXT:    sub sp, sp, a1
 ; CHECK-NEXT:    .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x08, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 8 * vlenb
-; CHECK-NEXT:    vsetvli a1, zero, e16, m4, ta, ma
+; CHECK-NEXT:    vsetvli a1, zero, e8, mf2, ta, ma
 ; CHECK-NEXT:    vmv1r.v v7, v0
 ; CHECK-NEXT:    csrr a2, vlenb
-; CHECK-NEXT:    addi a1, sp, 16
-; CHECK-NEXT:    vs8r.v v16, (a1) # Unknown-size Folded Spill
-; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v20
 ; CHECK-NEXT:    slli a1, a2, 1
 ; CHECK-NEXT:    srli a2, a2, 2
 ; CHECK-NEXT:    sub a3, a0, a1
-; CHECK-NEXT:    vsetvli a4, zero, e8, mf2, ta, ma
 ; CHECK-NEXT:    vslidedown.vx v0, v0, a2
 ; CHECK-NEXT:    sltu a2, a0, a3
 ; CHECK-NEXT:    addi a2, a2, -1
 ; CHECK-NEXT:    and a2, a2, a3
-; CHECK-NEXT:    vsetvli a3, zero, e16, m4, ta, ma
+; CHECK-NEXT:    addi a3, sp, 16
+; CHECK-NEXT:    vs8r.v v16, (a3) # Unknown-size Folded Spill
+; CHECK-NEXT:    vsetvli zero, a2, e16, m4, ta, ma
+; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v20
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v12
-; CHECK-NEXT:    vsetvli zero, a2, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfmax.vv v16, v16, v24, v0.t
 ; CHECK-NEXT:    vsetvli a2, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v12, v16
@@ -209,10 +208,11 @@ define <vscale x 32 x bfloat> @vfmax_vv_nxv32bf16(<vscale x 32 x bfloat> %va, <v
 ; CHECK-NEXT:  .LBB10_2:
 ; CHECK-NEXT:    addi a1, sp, 16
 ; CHECK-NEXT:    vl8r.v v24, (a1) # Unknown-size Folded Reload
+; CHECK-NEXT:    vsetvli zero, a0, e16, m4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v24
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v8
 ; CHECK-NEXT:    vmv1r.v v0, v7
-; CHECK-NEXT:    vsetvli zero, a0, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfmax.vv v16, v24, v16, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v16
@@ -238,22 +238,21 @@ define <vscale x 32 x bfloat> @vfmax_vv_nxv32bf16_unmasked(<vscale x 32 x bfloat
 ; CHECK-NEXT:    .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x08, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 8 * vlenb
 ; CHECK-NEXT:    csrr a2, vlenb
 ; CHECK-NEXT:    vsetvli a1, zero, e8, m4, ta, ma
-; CHECK-NEXT:    vmset.m v7
-; CHECK-NEXT:    addi a1, sp, 16
-; CHECK-NEXT:    vs8r.v v16, (a1) # Unknown-size Folded Spill
-; CHECK-NEXT:    vsetvli a1, zero, e16, m4, ta, ma
-; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v20
+; CHECK-NEXT:    vmset.m v24
 ; CHECK-NEXT:    slli a1, a2, 1
 ; CHECK-NEXT:    srli a2, a2, 2
 ; CHECK-NEXT:    sub a3, a0, a1
 ; CHECK-NEXT:    vsetvli a4, zero, e8, mf2, ta, ma
-; CHECK-NEXT:    vslidedown.vx v0, v7, a2
+; CHECK-NEXT:    vslidedown.vx v0, v24, a2
 ; CHECK-NEXT:    sltu a2, a0, a3
 ; CHECK-NEXT:    addi a2, a2, -1
 ; CHECK-NEXT:    and a2, a2, a3
-; CHECK-NEXT:    vsetvli a3, zero, e16, m4, ta, ma
+; CHECK-NEXT:    addi a3, sp, 16
+; CHECK-NEXT:    vs8r.v v16, (a3) # Unknown-size Folded Spill
+; CHECK-NEXT:    vsetvli zero, a2, e16, m4, ta, ma
+; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v20
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v12
-; CHECK-NEXT:    vsetvli zero, a2, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfmax.vv v16, v16, v24, v0.t
 ; CHECK-NEXT:    vsetvli a2, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v12, v16
@@ -263,9 +262,10 @@ define <vscale x 32 x bfloat> @vfmax_vv_nxv32bf16_unmasked(<vscale x 32 x bfloat
 ; CHECK-NEXT:  .LBB11_2:
 ; CHECK-NEXT:    addi a1, sp, 16
 ; CHECK-NEXT:    vl8r.v v24, (a1) # Unknown-size Folded Reload
+; CHECK-NEXT:    vsetvli zero, a0, e16, m4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v24
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfmax.vv v16, v24, v16
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v16
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfmin-vp.ll b/llvm/test/CodeGen/RISCV/rvv/vfmin-vp.ll
index 48a4c13869009..1aaddef39bc1c 100644
--- a/llvm/test/CodeGen/RISCV/rvv/vfmin-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/vfmin-vp.ll
@@ -17,10 +17,10 @@ declare <vscale x 1 x bfloat> @llvm.vp.minnum.nxv1bf16(<vscale x 1 x bfloat>, <v
 define <vscale x 1 x bfloat> @vfmin_vv_nxv1bf16(<vscale x 1 x bfloat> %va, <vscale x 1 x bfloat> %vb, <vscale x 1 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: vfmin_vv_nxv1bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v9
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vfmin.vv v9, v9, v10, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v9
@@ -32,10 +32,10 @@ define <vscale x 1 x bfloat> @vfmin_vv_nxv1bf16(<vscale x 1 x bfloat> %va, <vsca
 define <vscale x 1 x bfloat> @vfmin_vv_nxv1bf16_unmasked(<vscale x 1 x bfloat> %va, <vscale x 1 x bfloat> %vb, i32 zeroext %evl) {
 ; CHECK-LABEL: vfmin_vv_nxv1bf16_unmasked:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v9
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vfmin.vv v9, v9, v10
 ; CHECK-NEXT:    vsetvli a0, zero, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v9
@@ -49,10 +49,10 @@ declare <vscale x 2 x bfloat> @llvm.vp.minnum.nxv2bf16(<vscale x 2 x bfloat>, <v
 define <vscale x 2 x bfloat> @vfmin_vv_nxv2bf16(<vscale x 2 x bfloat> %va, <vscale x 2 x bfloat> %vb, <vscale x 2 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: vfmin_vv_nxv2bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v9
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m1, ta, ma
 ; CHECK-NEXT:    vfmin.vv v9, v9, v10, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, mf2, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v9
@@ -64,10 +64,10 @@ define <vscale x 2 x bfloat> @vfmin_vv_nxv2bf16(<vscale x 2 x bfloat> %va, <vsca
 define <vscale x 2 x bfloat> @vfmin_vv_nxv2bf16_unmasked(<vscale x 2 x bfloat> %va, <vscale x 2 x bfloat> %vb, i32 zeroext %evl) {
 ; CHECK-LABEL: vfmin_vv_nxv2bf16_unmasked:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v9
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m1, ta, ma
 ; CHECK-NEXT:    vfmin.vv v9, v9, v10
 ; CHECK-NEXT:    vsetvli a0, zero, e16, mf2, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v9
@@ -81,10 +81,10 @@ declare <vscale x 4 x bfloat> @llvm.vp.minnum.nxv4bf16(<vscale x 4 x bfloat>, <v
 define <vscale x 4 x bfloat> @vfmin_vv_nxv4bf16(<vscale x 4 x bfloat> %va, <vscale x 4 x bfloat> %vb, <vscale x 4 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: vfmin_vv_nxv4bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m1, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v9
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
 ; CHECK-NEXT:    vfmin.vv v10, v12, v10, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m1, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v10
@@ -96,10 +96,10 @@ define <vscale x 4 x bfloat> @vfmin_vv_nxv4bf16(<vscale x 4 x bfloat> %va, <vsca
 define <vscale x 4 x bfloat> @vfmin_vv_nxv4bf16_unmasked(<vscale x 4 x bfloat> %va, <vscale x 4 x bfloat> %vb, i32 zeroext %evl) {
 ; CHECK-LABEL: vfmin_vv_nxv4bf16_unmasked:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m1, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v9
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
 ; CHECK-NEXT:    vfmin.vv v10, v12, v10
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m1, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v10
@@ -113,10 +113,10 @@ declare <vscale x 8 x bfloat> @llvm.vp.minnum.nxv8bf16(<vscale x 8 x bfloat>, <v
 define <vscale x 8 x bfloat> @vfmin_vv_nxv8bf16(<vscale x 8 x bfloat> %va, <vscale x 8 x bfloat> %vb, <vscale x 8 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: vfmin_vv_nxv8bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v10
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vfmin.vv v12, v16, v12, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m2, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v12
@@ -128,10 +128,10 @@ define <vscale x 8 x bfloat> @vfmin_vv_nxv8bf16(<vscale x 8 x bfloat> %va, <vsca
 define <vscale x 8 x bfloat> @vfmin_vv_nxv8bf16_unmasked(<vscale x 8 x bfloat> %va, <vscale x 8 x bfloat> %vb, i32 zeroext %evl) {
 ; CHECK-LABEL: vfmin_vv_nxv8bf16_unmasked:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v10
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vfmin.vv v12, v16, v12
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m2, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v12
@@ -145,10 +145,10 @@ declare <vscale x 16 x bfloat> @llvm.vp.minnum.nxv16bf16(<vscale x 16 x bfloat>,
 define <vscale x 16 x bfloat> @vfmin_vv_nxv16bf16(<vscale x 16 x bfloat> %va, <vscale x 16 x bfloat> %vb, <vscale x 16 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: vfmin_vv_nxv16bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v12
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfmin.vv v16, v24, v16, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v16
@@ -160,10 +160,10 @@ define <vscale x 16 x bfloat> @vfmin_vv_nxv16bf16(<vscale x 16 x bfloat> %va, <v
 define <vscale x 16 x bfloat> @vfmin_vv_nxv16bf16_unmasked(<vscale x 16 x bfloat> %va, <vscale x 16 x bfloat> %vb, i32 zeroext %evl) {
 ; CHECK-LABEL: vfmin_vv_nxv16bf16_unmasked:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v12
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfmin.vv v16, v24, v16
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v16
@@ -183,23 +183,22 @@ define <vscale x 32 x bfloat> @vfmin_vv_nxv32bf16(<vscale x 32 x bfloat> %va, <v
 ; CHECK-NEXT:    slli a1, a1, 3
 ; CHECK-NEXT:    sub sp, sp, a1
 ; CHECK-NEXT:    .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x08, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 8 * vlenb
-; CHECK-NEXT:    vsetvli a1, zero, e16, m4, ta, ma
+; CHECK-NEXT:    vsetvli a1, zero, e8, mf2, ta, ma
 ; CHECK-NEXT:    vmv1r.v v7, v0
 ; CHECK-NEXT:    csrr a2, vlenb
-; CHECK-NEXT:    addi a1, sp, 16
-; CHECK-NEXT:    vs8r.v v16, (a1) # Unknown-size Folded Spill
-; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v20
 ; CHECK-NEXT:    slli a1, a2, 1
 ; CHECK-NEXT:    srli a2, a2, 2
 ; CHECK-NEXT:    sub a3, a0, a1
-; CHECK-NEXT:    vsetvli a4, zero, e8, mf2, ta, ma
 ; CHECK-NEXT:    vslidedown.vx v0, v0, a2
 ; CHECK-NEXT:    sltu a2, a0, a3
 ; CHECK-NEXT:    addi a2, a2, -1
 ; CHECK-NEXT:    and a2, a2, a3
-; CHECK-NEXT:    vsetvli a3, zero, e16, m4, ta, ma
+; CHECK-NEXT:    addi a3, sp, 16
+; CHECK-NEXT:    vs8r.v v16, (a3) # Unknown-size Folded Spill
+; CHECK-NEXT:    vsetvli zero, a2, e16, m4, ta, ma
+; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v20
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v12
-; CHECK-NEXT:    vsetvli zero, a2, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfmin.vv v16, v16, v24, v0.t
 ; CHECK-NEXT:    vsetvli a2, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v12, v16
@@ -209,10 +208,11 @@ define <vscale x 32 x bfloat> @vfmin_vv_nxv32bf16(<vscale x 32 x bfloat> %va, <v
 ; CHECK-NEXT:  .LBB10_2:
 ; CHECK-NEXT:    addi a1, sp, 16
 ; CHECK-NEXT:    vl8r.v v24, (a1) # Unknown-size Folded Reload
+; CHECK-NEXT:    vsetvli zero, a0, e16, m4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v24
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v8
 ; CHECK-NEXT:    vmv1r.v v0, v7
-; CHECK-NEXT:    vsetvli zero, a0, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfmin.vv v16, v24, v16, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v16
@@ -238,22 +238,21 @@ define <vscale x 32 x bfloat> @vfmin_vv_nxv32bf16_unmasked(<vscale x 32 x bfloat
 ; CHECK-NEXT:    .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x08, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 8 * vlenb
 ; CHECK-NEXT:    csrr a2, vlenb
 ; CHECK-NEXT:    vsetvli a1, zero, e8, m4, ta, ma
-; CHECK-NEXT:    vmset.m v7
-; CHECK-NEXT:    addi a1, sp, 16
-; CHECK-NEXT:    vs8r.v v16, (a1) # Unknown-size Folded Spill
-; CHECK-NEXT:    vsetvli a1, zero, e16, m4, ta, ma
-; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v20
+; CHECK-NEXT:    vmset.m v24
 ; CHECK-NEXT:    slli a1, a2, 1
 ; CHECK-NEXT:    srli a2, a2, 2
 ; CHECK-NEXT:    sub a3, a0, a1
 ; CHECK-NEXT:    vsetvli a4, zero, e8, mf2, ta, ma
-; CHECK-NEXT:    vslidedown.vx v0, v7, a2
+; CHECK-NEXT:    vslidedown.vx v0, v24, a2
 ; CHECK-NEXT:    sltu a2, a0, a3
 ; CHECK-NEXT:    addi a2, a2, -1
 ; CHECK-NEXT:    and a2, a2, a3
-; CHECK-NEXT:    vsetvli a3, zero, e16, m4, ta, ma
+; CHECK-NEXT:    addi a3, sp, 16
+; CHECK-NEXT:    vs8r.v v16, (a3) # Unknown-size Folded Spill
+; CHECK-NEXT:    vsetvli zero, a2, e16, m4, ta, ma
+; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v20
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v12
-; CHECK-NEXT:    vsetvli zero, a2, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfmin.vv v16, v16, v24, v0.t
 ; CHECK-NEXT:    vsetvli a2, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v12, v16
@@ -263,9 +262,10 @@ define <vscale x 32 x bfloat> @vfmin_vv_nxv32bf16_unmasked(<vscale x 32 x bfloat
 ; CHECK-NEXT:  .LBB11_2:
 ; CHECK-NEXT:    addi a1, sp, 16
 ; CHECK-NEXT:    vl8r.v v24, (a1) # Unknown-size Folded Reload
+; CHECK-NEXT:    vsetvli zero, a0, e16, m4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v24
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfmin.vv v16, v24, v16
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v16
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfptosi-vp-mask.ll b/llvm/test/CodeGen/RISCV/rvv/vfptosi-vp-mask.ll
index 575d50d11f0ac..33decd8aa1b91 100644
--- a/llvm/test/CodeGen/RISCV/rvv/vfptosi-vp-mask.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/vfptosi-vp-mask.ll
@@ -7,9 +7,9 @@
 define <vscale x 2 x i1> @vfptosi_nxv2i1_nxv2bf16(<vscale x 2 x bfloat> %va, <vscale x 2 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: vfptosi_nxv2i1_nxv2bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m1, ta, ma
 ; CHECK-NEXT:    vfcvt.rtz.x.f.v v8, v9, v0.t
 ; CHECK-NEXT:    vmsne.vi v0, v8, 0, v0.t
 ; CHECK-NEXT:    ret
@@ -20,9 +20,9 @@ define <vscale x 2 x i1> @vfptosi_nxv2i1_nxv2bf16(<vscale x 2 x bfloat> %va, <vs
 define <vscale x 2 x i1> @vfptosi_nxv2i1_nxv2bf16_unmasked(<vscale x 2 x bfloat> %va, i32 zeroext %evl) {
 ; CHECK-LABEL: vfptosi_nxv2i1_nxv2bf16_unmasked:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m1, ta, ma
 ; CHECK-NEXT:    vfcvt.rtz.x.f.v v8, v9
 ; CHECK-NEXT:    vmsne.vi v0, v8, 0
 ; CHECK-NEXT:    ret
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfptosi-vp.ll b/llvm/test/CodeGen/RISCV/rvv/vfptosi-vp.ll
index e33ab98c0f85d..ed0023e21497e 100644
--- a/llvm/test/CodeGen/RISCV/rvv/vfptosi-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/vfptosi-vp.ll
@@ -7,9 +7,8 @@
 define <vscale x 2 x i7> @vfptosi_v4i7_v4bf16(<vscale x 2 x bfloat> %va, <vscale x 2 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: vfptosi_v4i7_v4bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf2, ta, ma
-; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
 ; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, ma
+; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
 ; CHECK-NEXT:    vfncvt.rtz.x.f.w v8, v9, v0.t
 ; CHECK-NEXT:    vsetvli zero, zero, e8, mf4, ta, ma
 ; CHECK-NEXT:    vnsrl.wi v8, v8, 0, v0.t
@@ -21,9 +20,8 @@ define <vscale x 2 x i7> @vfptosi_v4i7_v4bf16(<vscale x 2 x bfloat> %va, <vscale
 define <vscale x 2 x i8> @vfptosi_nxv2i8_nxv2bf16(<vscale x 2 x bfloat> %va, <vscale x 2 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: vfptosi_nxv2i8_nxv2bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf2, ta, ma
-; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
 ; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, ma
+; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
 ; CHECK-NEXT:    vfncvt.rtz.x.f.w v8, v9, v0.t
 ; CHECK-NEXT:    vsetvli zero, zero, e8, mf4, ta, ma
 ; CHECK-NEXT:    vnsrl.wi v8, v8, 0, v0.t
@@ -35,9 +33,8 @@ define <vscale x 2 x i8> @vfptosi_nxv2i8_nxv2bf16(<vscale x 2 x bfloat> %va, <vs
 define <vscale x 2 x i8> @vfptosi_nxv2i8_nxv2bf16_unmasked(<vscale x 2 x bfloat> %va, i32 zeroext %evl) {
 ; CHECK-LABEL: vfptosi_nxv2i8_nxv2bf16_unmasked:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf2, ta, ma
-; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
 ; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, ma
+; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
 ; CHECK-NEXT:    vfncvt.rtz.x.f.w v8, v9
 ; CHECK-NEXT:    vsetvli zero, zero, e8, mf4, ta, ma
 ; CHECK-NEXT:    vnsrl.wi v8, v8, 0
@@ -49,9 +46,8 @@ define <vscale x 2 x i8> @vfptosi_nxv2i8_nxv2bf16_unmasked(<vscale x 2 x bfloat>
 define <vscale x 2 x i16> @vfptosi_nxv2i16_nxv2bf16(<vscale x 2 x bfloat> %va, <vscale x 2 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: vfptosi_nxv2i16_nxv2bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf2, ta, ma
-; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
 ; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, ma
+; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
 ; CHECK-NEXT:    vfncvt.rtz.x.f.w v8, v9, v0.t
 ; CHECK-NEXT:    ret
   %v = call <vscale x 2 x i16> @llvm.vp.fptosi.nxv2i16.nxv2bf16(<vscale x 2 x bfloat> %va, <vscale x 2 x i1> %m, i32 %evl)
@@ -61,9 +57,8 @@ define <vscale x 2 x i16> @vfptosi_nxv2i16_nxv2bf16(<vscale x 2 x bfloat> %va, <
 define <vscale x 2 x i16> @vfptosi_nxv2i16_nxv2bf16_unmasked(<vscale x 2 x bfloat> %va, i32 zeroext %evl) {
 ; CHECK-LABEL: vfptosi_nxv2i16_nxv2bf16_unmasked:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf2, ta, ma
-; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
 ; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, ma
+; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
 ; CHECK-NEXT:    vfncvt.rtz.x.f.w v8, v9
 ; CHECK-NEXT:    ret
   %v = call <vscale x 2 x i16> @llvm.vp.fptosi.nxv2i16.nxv2bf16(<vscale x 2 x bfloat> %va, <vscale x 2 x i1> splat (i1 true), i32 %evl)
@@ -73,9 +68,9 @@ define <vscale x 2 x i16> @vfptosi_nxv2i16_nxv2bf16_unmasked(<vscale x 2 x bfloa
 define <vscale x 2 x i32> @vfptosi_nxv2i32_nxv2bf16(<vscale x 2 x bfloat> %va, <vscale x 2 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: vfptosi_nxv2i32_nxv2bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m1, ta, ma
 ; CHECK-NEXT:    vfcvt.rtz.x.f.v v8, v9, v0.t
 ; CHECK-NEXT:    ret
   %v = call <vscale x 2 x i32> @llvm.vp.fptosi.nxv2i32.nxv2bf16(<vscale x 2 x bfloat> %va, <vscale x 2 x i1> %m, i32 %evl)
@@ -85,9 +80,9 @@ define <vscale x 2 x i32> @vfptosi_nxv2i32_nxv2bf16(<vscale x 2 x bfloat> %va, <
 define <vscale x 2 x i32> @vfptosi_nxv2i32_nxv2bf16_unmasked(<vscale x 2 x bfloat> %va, i32 zeroext %evl) {
 ; CHECK-LABEL: vfptosi_nxv2i32_nxv2bf16_unmasked:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m1, ta, ma
 ; CHECK-NEXT:    vfcvt.rtz.x.f.v v8, v9
 ; CHECK-NEXT:    ret
   %v = call <vscale x 2 x i32> @llvm.vp.fptosi.nxv2i32.nxv2bf16(<vscale x 2 x bfloat> %va, <vscale x 2 x i1> splat (i1 true), i32 %evl)
@@ -97,9 +92,9 @@ define <vscale x 2 x i32> @vfptosi_nxv2i32_nxv2bf16_unmasked(<vscale x 2 x bfloa
 define <vscale x 2 x i64> @vfptosi_nxv2i64_nxv2bf16(<vscale x 2 x bfloat> %va, <vscale x 2 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: vfptosi_nxv2i64_nxv2bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m1, ta, ma
 ; CHECK-NEXT:    vfwcvt.rtz.x.f.v v8, v10, v0.t
 ; CHECK-NEXT:    ret
   %v = call <vscale x 2 x i64> @llvm.vp.fptosi.nxv2i64.nxv2bf16(<vscale x 2 x bfloat> %va, <vscale x 2 x i1> %m, i32 %evl)
@@ -109,9 +104,9 @@ define <vscale x 2 x i64> @vfptosi_nxv2i64_nxv2bf16(<vscale x 2 x bfloat> %va, <
 define <vscale x 2 x i64> @vfptosi_nxv2i64_nxv2bf16_unmasked(<vscale x 2 x bfloat> %va, i32 zeroext %evl) {
 ; CHECK-LABEL: vfptosi_nxv2i64_nxv2bf16_unmasked:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m1, ta, ma
 ; CHECK-NEXT:    vfwcvt.rtz.x.f.v v8, v10
 ; CHECK-NEXT:    ret
   %v = call <vscale x 2 x i64> @llvm.vp.fptosi.nxv2i64.nxv2bf16(<vscale x 2 x bfloat> %va, <vscale x 2 x i1> splat (i1 true), i32 %evl)
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfptoui-vp-mask.ll b/llvm/test/CodeGen/RISCV/rvv/vfptoui-vp-mask.ll
index e1d0ad4758586..59c6791c12f79 100644
--- a/llvm/test/CodeGen/RISCV/rvv/vfptoui-vp-mask.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/vfptoui-vp-mask.ll
@@ -7,9 +7,9 @@
 define <vscale x 2 x i1> @vfptoui_nxv2i1_nxv2bf16(<vscale x 2 x bfloat> %va, <vscale x 2 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: vfptoui_nxv2i1_nxv2bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m1, ta, ma
 ; CHECK-NEXT:    vfcvt.rtz.xu.f.v v8, v9, v0.t
 ; CHECK-NEXT:    vmsne.vi v0, v8, 0, v0.t
 ; CHECK-NEXT:    ret
@@ -20,9 +20,9 @@ define <vscale x 2 x i1> @vfptoui_nxv2i1_nxv2bf16(<vscale x 2 x bfloat> %va, <vs
 define <vscale x 2 x i1> @vfptoui_nxv2i1_nxv2bf16_unmasked(<vscale x 2 x bfloat> %va, i32 zeroext %evl) {
 ; CHECK-LABEL: vfptoui_nxv2i1_nxv2bf16_unmasked:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m1, ta, ma
 ; CHECK-NEXT:    vfcvt.rtz.xu.f.v v8, v9
 ; CHECK-NEXT:    vmsne.vi v0, v8, 0
 ; CHECK-NEXT:    ret
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfptoui-vp.ll b/llvm/test/CodeGen/RISCV/rvv/vfptoui-vp.ll
index 86222ecfadfea..a11139fea9e5b 100644
--- a/llvm/test/CodeGen/RISCV/rvv/vfptoui-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/vfptoui-vp.ll
@@ -7,9 +7,8 @@
 define <vscale x 2 x i7> @vfptoui_v4i7_v4bf16(<vscale x 2 x bfloat> %va, <vscale x 2 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: vfptoui_v4i7_v4bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf2, ta, ma
-; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
 ; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, ma
+; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
 ; CHECK-NEXT:    vfncvt.rtz.x.f.w v8, v9, v0.t
 ; CHECK-NEXT:    vsetvli zero, zero, e8, mf4, ta, ma
 ; CHECK-NEXT:    vnsrl.wi v8, v8, 0, v0.t
@@ -21,9 +20,8 @@ define <vscale x 2 x i7> @vfptoui_v4i7_v4bf16(<vscale x 2 x bfloat> %va, <vscale
 define <vscale x 2 x i8> @vfptoui_nxv2i8_nxv2bf16(<vscale x 2 x bfloat> %va, <vscale x 2 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: vfptoui_nxv2i8_nxv2bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf2, ta, ma
-; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
 ; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, ma
+; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
 ; CHECK-NEXT:    vfncvt.rtz.xu.f.w v8, v9, v0.t
 ; CHECK-NEXT:    vsetvli zero, zero, e8, mf4, ta, ma
 ; CHECK-NEXT:    vnsrl.wi v8, v8, 0, v0.t
@@ -35,9 +33,8 @@ define <vscale x 2 x i8> @vfptoui_nxv2i8_nxv2bf16(<vscale x 2 x bfloat> %va, <vs
 define <vscale x 2 x i8> @vfptoui_nxv2i8_nxv2bf16_unmasked(<vscale x 2 x bfloat> %va, i32 zeroext %evl) {
 ; CHECK-LABEL: vfptoui_nxv2i8_nxv2bf16_unmasked:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf2, ta, ma
-; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
 ; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, ma
+; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
 ; CHECK-NEXT:    vfncvt.rtz.xu.f.w v8, v9
 ; CHECK-NEXT:    vsetvli zero, zero, e8, mf4, ta, ma
 ; CHECK-NEXT:    vnsrl.wi v8, v8, 0
@@ -49,9 +46,8 @@ define <vscale x 2 x i8> @vfptoui_nxv2i8_nxv2bf16_unmasked(<vscale x 2 x bfloat>
 define <vscale x 2 x i16> @vfptoui_nxv2i16_nxv2bf16(<vscale x 2 x bfloat> %va, <vscale x 2 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: vfptoui_nxv2i16_nxv2bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf2, ta, ma
-; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
 ; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, ma
+; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
 ; CHECK-NEXT:    vfncvt.rtz.xu.f.w v8, v9, v0.t
 ; CHECK-NEXT:    ret
   %v = call <vscale x 2 x i16> @llvm.vp.fptoui.nxv2i16.nxv2bf16(<vscale x 2 x bfloat> %va, <vscale x 2 x i1> %m, i32 %evl)
@@ -61,9 +57,8 @@ define <vscale x 2 x i16> @vfptoui_nxv2i16_nxv2bf16(<vscale x 2 x bfloat> %va, <
 define <vscale x 2 x i16> @vfptoui_nxv2i16_nxv2bf16_unmasked(<vscale x 2 x bfloat> %va, i32 zeroext %evl) {
 ; CHECK-LABEL: vfptoui_nxv2i16_nxv2bf16_unmasked:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf2, ta, ma
-; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
 ; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, ma
+; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
 ; CHECK-NEXT:    vfncvt.rtz.xu.f.w v8, v9
 ; CHECK-NEXT:    ret
   %v = call <vscale x 2 x i16> @llvm.vp.fptoui.nxv2i16.nxv2bf16(<vscale x 2 x bfloat> %va, <vscale x 2 x i1> splat (i1 true), i32 %evl)
@@ -73,9 +68,9 @@ define <vscale x 2 x i16> @vfptoui_nxv2i16_nxv2bf16_unmasked(<vscale x 2 x bfloa
 define <vscale x 2 x i32> @vfptoui_nxv2i32_nxv2bf16(<vscale x 2 x bfloat> %va, <vscale x 2 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: vfptoui_nxv2i32_nxv2bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m1, ta, ma
 ; CHECK-NEXT:    vfcvt.rtz.xu.f.v v8, v9, v0.t
 ; CHECK-NEXT:    ret
   %v = call <vscale x 2 x i32> @llvm.vp.fptoui.nxv2i32.nxv2bf16(<vscale x 2 x bfloat> %va, <vscale x 2 x i1> %m, i32 %evl)
@@ -85,9 +80,9 @@ define <vscale x 2 x i32> @vfptoui_nxv2i32_nxv2bf16(<vscale x 2 x bfloat> %va, <
 define <vscale x 2 x i32> @vfptoui_nxv2i32_nxv2bf16_unmasked(<vscale x 2 x bfloat> %va, i32 zeroext %evl) {
 ; CHECK-LABEL: vfptoui_nxv2i32_nxv2bf16_unmasked:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m1, ta, ma
 ; CHECK-NEXT:    vfcvt.rtz.xu.f.v v8, v9
 ; CHECK-NEXT:    ret
   %v = call <vscale x 2 x i32> @llvm.vp.fptoui.nxv2i32.nxv2bf16(<vscale x 2 x bfloat> %va, <vscale x 2 x i1> splat (i1 true), i32 %evl)
@@ -97,9 +92,9 @@ define <vscale x 2 x i32> @vfptoui_nxv2i32_nxv2bf16_unmasked(<vscale x 2 x bfloa
 define <vscale x 2 x i64> @vfptoui_nxv2i64_nxv2bf16(<vscale x 2 x bfloat> %va, <vscale x 2 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: vfptoui_nxv2i64_nxv2bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m1, ta, ma
 ; CHECK-NEXT:    vfwcvt.rtz.xu.f.v v8, v10, v0.t
 ; CHECK-NEXT:    ret
   %v = call <vscale x 2 x i64> @llvm.vp.fptoui.nxv2i64.nxv2bf16(<vscale x 2 x bfloat> %va, <vscale x 2 x i1> %m, i32 %evl)
@@ -109,9 +104,9 @@ define <vscale x 2 x i64> @vfptoui_nxv2i64_nxv2bf16(<vscale x 2 x bfloat> %va, <
 define <vscale x 2 x i64> @vfptoui_nxv2i64_nxv2bf16_unmasked(<vscale x 2 x bfloat> %va, i32 zeroext %evl) {
 ; CHECK-LABEL: vfptoui_nxv2i64_nxv2bf16_unmasked:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m1, ta, ma
 ; CHECK-NEXT:    vfwcvt.rtz.xu.f.v v8, v10
 ; CHECK-NEXT:    ret
   %v = call <vscale x 2 x i64> @llvm.vp.fptoui.nxv2i64.nxv2bf16(<vscale x 2 x bfloat> %va, <vscale x 2 x i1> splat (i1 true), i32 %evl)
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfsqrt-vp.ll b/llvm/test/CodeGen/RISCV/rvv/vfsqrt-vp.ll
index e94d0a60bbfc7..6193fdd38a642 100644
--- a/llvm/test/CodeGen/RISCV/rvv/vfsqrt-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/vfsqrt-vp.ll
@@ -17,9 +17,9 @@ declare <vscale x 1 x bfloat> @llvm.vp.sqrt.nxv1bf16(<vscale x 1 x bfloat>, <vsc
 define <vscale x 1 x bfloat> @vfsqrt_vv_nxv1bf16(<vscale x 1 x bfloat> %va, <vscale x 1 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: vfsqrt_vv_nxv1bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vfsqrt.v v9, v9, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v9
@@ -31,9 +31,9 @@ define <vscale x 1 x bfloat> @vfsqrt_vv_nxv1bf16(<vscale x 1 x bfloat> %va, <vsc
 define <vscale x 1 x bfloat> @vfsqrt_vv_nxv1bf16_unmasked(<vscale x 1 x bfloat> %va, i32 zeroext %evl) {
 ; CHECK-LABEL: vfsqrt_vv_nxv1bf16_unmasked:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vfsqrt.v v9, v9
 ; CHECK-NEXT:    vsetvli a0, zero, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v9
@@ -47,9 +47,9 @@ declare <vscale x 2 x bfloat> @llvm.vp.sqrt.nxv2bf16(<vscale x 2 x bfloat>, <vsc
 define <vscale x 2 x bfloat> @vfsqrt_vv_nxv2bf16(<vscale x 2 x bfloat> %va, <vscale x 2 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: vfsqrt_vv_nxv2bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m1, ta, ma
 ; CHECK-NEXT:    vfsqrt.v v9, v9, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, mf2, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v9
@@ -61,9 +61,9 @@ define <vscale x 2 x bfloat> @vfsqrt_vv_nxv2bf16(<vscale x 2 x bfloat> %va, <vsc
 define <vscale x 2 x bfloat> @vfsqrt_vv_nxv2bf16_unmasked(<vscale x 2 x bfloat> %va, i32 zeroext %evl) {
 ; CHECK-LABEL: vfsqrt_vv_nxv2bf16_unmasked:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m1, ta, ma
 ; CHECK-NEXT:    vfsqrt.v v9, v9
 ; CHECK-NEXT:    vsetvli a0, zero, e16, mf2, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v9
@@ -77,9 +77,9 @@ declare <vscale x 4 x bfloat> @llvm.vp.sqrt.nxv4bf16(<vscale x 4 x bfloat>, <vsc
 define <vscale x 4 x bfloat> @vfsqrt_vv_nxv4bf16(<vscale x 4 x bfloat> %va, <vscale x 4 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: vfsqrt_vv_nxv4bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m1, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
 ; CHECK-NEXT:    vfsqrt.v v10, v10, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m1, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v10
@@ -91,9 +91,9 @@ define <vscale x 4 x bfloat> @vfsqrt_vv_nxv4bf16(<vscale x 4 x bfloat> %va, <vsc
 define <vscale x 4 x bfloat> @vfsqrt_vv_nxv4bf16_unmasked(<vscale x 4 x bfloat> %va, i32 zeroext %evl) {
 ; CHECK-LABEL: vfsqrt_vv_nxv4bf16_unmasked:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m1, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
 ; CHECK-NEXT:    vfsqrt.v v10, v10
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m1, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v10
@@ -107,9 +107,9 @@ declare <vscale x 8 x bfloat> @llvm.vp.sqrt.nxv8bf16(<vscale x 8 x bfloat>, <vsc
 define <vscale x 8 x bfloat> @vfsqrt_vv_nxv8bf16(<vscale x 8 x bfloat> %va, <vscale x 8 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: vfsqrt_vv_nxv8bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vfsqrt.v v12, v12, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m2, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v12
@@ -121,9 +121,9 @@ define <vscale x 8 x bfloat> @vfsqrt_vv_nxv8bf16(<vscale x 8 x bfloat> %va, <vsc
 define <vscale x 8 x bfloat> @vfsqrt_vv_nxv8bf16_unmasked(<vscale x 8 x bfloat> %va, i32 zeroext %evl) {
 ; CHECK-LABEL: vfsqrt_vv_nxv8bf16_unmasked:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vfsqrt.v v12, v12
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m2, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v12
@@ -137,9 +137,9 @@ declare <vscale x 16 x bfloat> @llvm.vp.sqrt.nxv16bf16(<vscale x 16 x bfloat>, <
 define <vscale x 16 x bfloat> @vfsqrt_vv_nxv16bf16(<vscale x 16 x bfloat> %va, <vscale x 16 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: vfsqrt_vv_nxv16bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfsqrt.v v16, v16, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v16
@@ -151,9 +151,9 @@ define <vscale x 16 x bfloat> @vfsqrt_vv_nxv16bf16(<vscale x 16 x bfloat> %va, <
 define <vscale x 16 x bfloat> @vfsqrt_vv_nxv16bf16_unmasked(<vscale x 16 x bfloat> %va, i32 zeroext %evl) {
 ; CHECK-LABEL: vfsqrt_vv_nxv16bf16_unmasked:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfsqrt.v v16, v16
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v16
@@ -173,13 +173,13 @@ define <vscale x 32 x bfloat> @vfsqrt_vv_nxv32bf16(<vscale x 32 x bfloat> %va, <
 ; CHECK-NEXT:    slli a1, a2, 1
 ; CHECK-NEXT:    srli a2, a2, 2
 ; CHECK-NEXT:    sub a3, a0, a1
+; CHECK-NEXT:    sltu a4, a0, a3
+; CHECK-NEXT:    addi a4, a4, -1
+; CHECK-NEXT:    and a3, a4, a3
 ; CHECK-NEXT:    vslidedown.vx v0, v0, a2
-; CHECK-NEXT:    sltu a2, a0, a3
-; CHECK-NEXT:    addi a2, a2, -1
-; CHECK-NEXT:    and a2, a2, a3
-; CHECK-NEXT:    vsetvli a3, zero, e16, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a3, e16, m4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v12
-; CHECK-NEXT:    vsetvli zero, a2, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfsqrt.v v24, v24, v0.t
 ; CHECK-NEXT:    vsetvli a2, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v12, v24
@@ -187,9 +187,10 @@ define <vscale x 32 x bfloat> @vfsqrt_vv_nxv32bf16(<vscale x 32 x bfloat> %va, <
 ; CHECK-NEXT:  # %bb.1:
 ; CHECK-NEXT:    mv a0, a1
 ; CHECK-NEXT:  .LBB10_2:
+; CHECK-NEXT:    vsetvli zero, a0, e16, m4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v8
 ; CHECK-NEXT:    vmv1r.v v0, v16
-; CHECK-NEXT:    vsetvli zero, a0, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfsqrt.v v16, v24, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v16
@@ -207,14 +208,14 @@ define <vscale x 32 x bfloat> @vfsqrt_vv_nxv32bf16_unmasked(<vscale x 32 x bfloa
 ; CHECK-NEXT:    slli a1, a2, 1
 ; CHECK-NEXT:    srli a2, a2, 2
 ; CHECK-NEXT:    sub a3, a0, a1
+; CHECK-NEXT:    sltu a4, a0, a3
+; CHECK-NEXT:    addi a4, a4, -1
+; CHECK-NEXT:    and a3, a4, a3
 ; CHECK-NEXT:    vsetvli a4, zero, e8, mf2, ta, ma
 ; CHECK-NEXT:    vslidedown.vx v0, v16, a2
-; CHECK-NEXT:    sltu a2, a0, a3
-; CHECK-NEXT:    addi a2, a2, -1
-; CHECK-NEXT:    and a2, a2, a3
-; CHECK-NEXT:    vsetvli a3, zero, e16, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a3, e16, m4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v12
-; CHECK-NEXT:    vsetvli zero, a2, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfsqrt.v v16, v16, v0.t
 ; CHECK-NEXT:    vsetvli a2, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v12, v16
@@ -222,8 +223,9 @@ define <vscale x 32 x bfloat> @vfsqrt_vv_nxv32bf16_unmasked(<vscale x 32 x bfloa
 ; CHECK-NEXT:  # %bb.1:
 ; CHECK-NEXT:    mv a0, a1
 ; CHECK-NEXT:  .LBB11_2:
+; CHECK-NEXT:    vsetvli zero, a0, e16, m4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfsqrt.v v16, v16
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v16
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfsub-vp.ll b/llvm/test/CodeGen/RISCV/rvv/vfsub-vp.ll
index 56ed560f9ec93..08664fd1c4819 100644
--- a/llvm/test/CodeGen/RISCV/rvv/vfsub-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/vfsub-vp.ll
@@ -17,10 +17,10 @@ declare <vscale x 1 x bfloat> @llvm.vp.fsub.nxv1bf16(<vscale x 1 x bfloat>, <vsc
 define <vscale x 1 x bfloat> @vfsub_vv_nxv1bf16(<vscale x 1 x bfloat> %va, <vscale x 1 x bfloat> %b, <vscale x 1 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: vfsub_vv_nxv1bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v9
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vfsub.vv v9, v9, v10, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v9
@@ -32,10 +32,10 @@ define <vscale x 1 x bfloat> @vfsub_vv_nxv1bf16(<vscale x 1 x bfloat> %va, <vsca
 define <vscale x 1 x bfloat> @vfsub_vv_nxv1bf16_unmasked(<vscale x 1 x bfloat> %va, <vscale x 1 x bfloat> %b, i32 zeroext %evl) {
 ; CHECK-LABEL: vfsub_vv_nxv1bf16_unmasked:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v9
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vfsub.vv v9, v9, v10
 ; CHECK-NEXT:    vsetvli a0, zero, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v9
@@ -48,11 +48,11 @@ define <vscale x 1 x bfloat> @vfsub_vf_nxv1bf16(<vscale x 1 x bfloat> %va, bfloa
 ; CHECK-LABEL: vfsub_vf_nxv1bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vmv.v.x v9, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vfsub.vv v9, v10, v8, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v9
@@ -67,11 +67,11 @@ define <vscale x 1 x bfloat> @vfsub_vf_nxv1bf16_unmasked(<vscale x 1 x bfloat> %
 ; CHECK-LABEL: vfsub_vf_nxv1bf16_unmasked:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
 ; CHECK-NEXT:    vmv.v.x v9, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
 ; CHECK-NEXT:    vfsub.vv v9, v10, v8
 ; CHECK-NEXT:    vsetvli a0, zero, e16, mf4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v9
@@ -87,10 +87,10 @@ declare <vscale x 2 x bfloat> @llvm.vp.fsub.nxv2bf16(<vscale x 2 x bfloat>, <vsc
 define <vscale x 2 x bfloat> @vfsub_vv_nxv2bf16(<vscale x 2 x bfloat> %va, <vscale x 2 x bfloat> %b, <vscale x 2 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: vfsub_vv_nxv2bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v9
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m1, ta, ma
 ; CHECK-NEXT:    vfsub.vv v9, v9, v10, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, mf2, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v9
@@ -102,10 +102,10 @@ define <vscale x 2 x bfloat> @vfsub_vv_nxv2bf16(<vscale x 2 x bfloat> %va, <vsca
 define <vscale x 2 x bfloat> @vfsub_vv_nxv2bf16_unmasked(<vscale x 2 x bfloat> %va, <vscale x 2 x bfloat> %b, i32 zeroext %evl) {
 ; CHECK-LABEL: vfsub_vv_nxv2bf16_unmasked:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v9
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v9, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m1, ta, ma
 ; CHECK-NEXT:    vfsub.vv v9, v9, v10
 ; CHECK-NEXT:    vsetvli a0, zero, e16, mf2, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v9
@@ -118,11 +118,11 @@ define <vscale x 2 x bfloat> @vfsub_vf_nxv2bf16(<vscale x 2 x bfloat> %va, bfloa
 ; CHECK-LABEL: vfsub_vf_nxv2bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, ma
 ; CHECK-NEXT:    vmv.v.x v9, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m1, ta, ma
 ; CHECK-NEXT:    vfsub.vv v9, v10, v8, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, mf2, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v9
@@ -137,11 +137,11 @@ define <vscale x 2 x bfloat> @vfsub_vf_nxv2bf16_unmasked(<vscale x 2 x bfloat> %
 ; CHECK-LABEL: vfsub_vf_nxv2bf16_unmasked:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, mf2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, ma
 ; CHECK-NEXT:    vmv.v.x v9, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m1, ta, ma
 ; CHECK-NEXT:    vfsub.vv v9, v10, v8
 ; CHECK-NEXT:    vsetvli a0, zero, e16, mf2, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v9
@@ -157,10 +157,10 @@ declare <vscale x 4 x bfloat> @llvm.vp.fsub.nxv4bf16(<vscale x 4 x bfloat>, <vsc
 define <vscale x 4 x bfloat> @vfsub_vv_nxv4bf16(<vscale x 4 x bfloat> %va, <vscale x 4 x bfloat> %b, <vscale x 4 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: vfsub_vv_nxv4bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m1, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v9
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
 ; CHECK-NEXT:    vfsub.vv v10, v12, v10, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m1, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v10
@@ -172,10 +172,10 @@ define <vscale x 4 x bfloat> @vfsub_vv_nxv4bf16(<vscale x 4 x bfloat> %va, <vsca
 define <vscale x 4 x bfloat> @vfsub_vv_nxv4bf16_unmasked(<vscale x 4 x bfloat> %va, <vscale x 4 x bfloat> %b, i32 zeroext %evl) {
 ; CHECK-LABEL: vfsub_vv_nxv4bf16_unmasked:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m1, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v9
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
 ; CHECK-NEXT:    vfsub.vv v10, v12, v10
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m1, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v10
@@ -188,11 +188,11 @@ define <vscale x 4 x bfloat> @vfsub_vf_nxv4bf16(<vscale x 4 x bfloat> %va, bfloa
 ; CHECK-LABEL: vfsub_vf_nxv4bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m1, ta, ma
 ; CHECK-NEXT:    vmv.v.x v9, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
 ; CHECK-NEXT:    vfsub.vv v10, v10, v12, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m1, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v10
@@ -207,11 +207,11 @@ define <vscale x 4 x bfloat> @vfsub_vf_nxv4bf16_unmasked(<vscale x 4 x bfloat> %
 ; CHECK-LABEL: vfsub_vf_nxv4bf16_unmasked:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m1, ta, ma
 ; CHECK-NEXT:    vmv.v.x v9, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v9
-; CHECK-NEXT:    vsetvli zero, a0, e32, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
 ; CHECK-NEXT:    vfsub.vv v10, v10, v12
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m1, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v10
@@ -227,10 +227,10 @@ declare <vscale x 8 x bfloat> @llvm.vp.fsub.nxv8bf16(<vscale x 8 x bfloat>, <vsc
 define <vscale x 8 x bfloat> @vfsub_vv_nxv8bf16(<vscale x 8 x bfloat> %va, <vscale x 8 x bfloat> %b, <vscale x 8 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: vfsub_vv_nxv8bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v10
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vfsub.vv v12, v16, v12, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m2, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v12
@@ -242,10 +242,10 @@ define <vscale x 8 x bfloat> @vfsub_vv_nxv8bf16(<vscale x 8 x bfloat> %va, <vsca
 define <vscale x 8 x bfloat> @vfsub_vv_nxv8bf16_unmasked(<vscale x 8 x bfloat> %va, <vscale x 8 x bfloat> %b, i32 zeroext %evl) {
 ; CHECK-LABEL: vfsub_vv_nxv8bf16_unmasked:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v10
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vfsub.vv v12, v16, v12
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m2, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v12
@@ -258,11 +258,11 @@ define <vscale x 8 x bfloat> @vfsub_vf_nxv8bf16(<vscale x 8 x bfloat> %va, bfloa
 ; CHECK-LABEL: vfsub_vf_nxv8bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vmv.v.x v10, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v10
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vfsub.vv v12, v12, v16, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m2, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v12
@@ -277,11 +277,11 @@ define <vscale x 8 x bfloat> @vfsub_vf_nxv8bf16_unmasked(<vscale x 8 x bfloat> %
 ; CHECK-LABEL: vfsub_vf_nxv8bf16_unmasked:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, m2, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
 ; CHECK-NEXT:    vmv.v.x v10, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v12, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v10
-; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
 ; CHECK-NEXT:    vfsub.vv v12, v12, v16
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m2, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v12
@@ -297,10 +297,10 @@ declare <vscale x 16 x bfloat> @llvm.vp.fsub.nxv16bf16(<vscale x 16 x bfloat>, <
 define <vscale x 16 x bfloat> @vfsub_vv_nxv16bf16(<vscale x 16 x bfloat> %va, <vscale x 16 x bfloat> %b, <vscale x 16 x i1> %m, i32 zeroext %evl) {
 ; CHECK-LABEL: vfsub_vv_nxv16bf16:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v12
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfsub.vv v16, v24, v16, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v16
@@ -312,10 +312,10 @@ define <vscale x 16 x bfloat> @vfsub_vv_nxv16bf16(<vscale x 16 x bfloat> %va, <v
 define <vscale x 16 x bfloat> @vfsub_vv_nxv16bf16_unmasked(<vscale x 16 x bfloat> %va, <vscale x 16 x bfloat> %b, i32 zeroext %evl) {
 ; CHECK-LABEL: vfsub_vv_nxv16bf16_unmasked:
 ; CHECK:       # %bb.0:
-; CHECK-NEXT:    vsetvli a1, zero, e16, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v12
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfsub.vv v16, v24, v16
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v16
@@ -328,11 +328,11 @@ define <vscale x 16 x bfloat> @vfsub_vf_nxv16bf16(<vscale x 16 x bfloat> %va, bf
 ; CHECK-LABEL: vfsub_vf_nxv16bf16:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m4, ta, ma
 ; CHECK-NEXT:    vmv.v.x v12, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v12
-; CHECK-NEXT:    vsetvli zero, a0, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfsub.vv v16, v16, v24, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v16
@@ -347,11 +347,11 @@ define <vscale x 16 x bfloat> @vfsub_vf_nxv16bf16_unmasked(<vscale x 16 x bfloat
 ; CHECK-LABEL: vfsub_vf_nxv16bf16_unmasked:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    fmv.x.h a1, fa0
-; CHECK-NEXT:    vsetvli a2, zero, e16, m4, ta, ma
+; CHECK-NEXT:    vsetvli zero, a0, e16, m4, ta, ma
 ; CHECK-NEXT:    vmv.v.x v12, a1
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v8
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v12
-; CHECK-NEXT:    vsetvli zero, a0, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfsub.vv v16, v16, v24
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v16
@@ -373,23 +373,22 @@ define <vscale x 32 x bfloat> @vfsub_vv_nxv32bf16(<vscale x 32 x bfloat> %va, <v
 ; CHECK-NEXT:    slli a1, a1, 3
 ; CHECK-NEXT:    sub sp, sp, a1
 ; CHECK-NEXT:    .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x08, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 8 * vlenb
-; CHECK-NEXT:    vsetvli a1, zero, e16, m4, ta, ma
+; CHECK-NEXT:    vsetvli a1, zero, e8, mf2, ta, ma
 ; CHECK-NEXT:    vmv1r.v v7, v0
 ; CHECK-NEXT:    csrr a2, vlenb
-; CHECK-NEXT:    addi a1, sp, 16
-; CHECK-NEXT:    vs8r.v v16, (a1) # Unknown-size Folded Spill
-; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v20
 ; CHECK-NEXT:    slli a1, a2, 1
 ; CHECK-NEXT:    srli a2, a2, 2
 ; CHECK-NEXT:    sub a3, a0, a1
-; CHECK-NEXT:    vsetvli a4, zero, e8, mf2, ta, ma
 ; CHECK-NEXT:    vslidedown.vx v0, v0, a2
 ; CHECK-NEXT:    sltu a2, a0, a3
 ; CHECK-NEXT:    addi a2, a2, -1
 ; CHECK-NEXT:    and a2, a2, a3
-; CHECK-NEXT:    vsetvli a3, zero, e16, m4, ta, ma
+; CHECK-NEXT:    addi a3, sp, 16
+; CHECK-NEXT:    vs8r.v v16, (a3) # Unknown-size Folded Spill
+; CHECK-NEXT:    vsetvli zero, a2, e16, m4, ta, ma
+; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v20
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v12
-; CHECK-NEXT:    vsetvli zero, a2, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfsub.vv v16, v16, v24, v0.t
 ; CHECK-NEXT:    vsetvli a2, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v12, v16
@@ -399,10 +398,11 @@ define <vscale x 32 x bfloat> @vfsub_vv_nxv32bf16(<vscale x 32 x bfloat> %va, <v
 ; CHECK-NEXT:  .LBB20_2:
 ; CHECK-NEXT:    addi a1, sp, 16
 ; CHECK-NEXT:    vl8r.v v24, (a1) # Unknown-size Folded Reload
+; CHECK-NEXT:    vsetvli zero, a0, e16, m4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v24
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v8
 ; CHECK-NEXT:    vmv1r.v v0, v7
-; CHECK-NEXT:    vsetvli zero, a0, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfsub.vv v16, v24, v16, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v16
@@ -428,22 +428,21 @@ define <vscale x 32 x bfloat> @vfsub_vv_nxv32bf16_unmasked(<vscale x 32 x bfloat
 ; CHECK-NEXT:    .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x08, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 8 * vlenb
 ; CHECK-NEXT:    csrr a2, vlenb
 ; CHECK-NEXT:    vsetvli a1, zero, e8, m4, ta, ma
-; CHECK-NEXT:    vmset.m v7
-; CHECK-NEXT:    addi a1, sp, 16
-; CHECK-NEXT:    vs8r.v v16, (a1) # Unknown-size Folded Spill
-; CHECK-NEXT:    vsetvli a1, zero, e16, m4, ta, ma
-; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v20
+; CHECK-NEXT:    vmset.m v24
 ; CHECK-NEXT:    slli a1, a2, 1
 ; CHECK-NEXT:    srli a2, a2, 2
 ; CHECK-NEXT:    sub a3, a0, a1
 ; CHECK-NEXT:    vsetvli a4, zero, e8, mf2, ta, ma
-; CHECK-NEXT:    vslidedown.vx v0, v7, a2
+; CHECK-NEXT:    vslidedown.vx v0, v24, a2
 ; CHECK-NEXT:    sltu a2, a0, a3
 ; CHECK-NEXT:    addi a2, a2, -1
 ; CHECK-NEXT:    and a2, a2, a3
-; CHECK-NEXT:    vsetvli a3, zero, e16, m4, ta, ma
+; CHECK-NEXT:    addi a3, sp, 16
+; CHECK-NEXT:    vs8r.v v16, (a3) # Unknown-size Folded Spill
+; CHECK-NEXT:    vsetvli zero, a2, e16, m4, ta, ma
+; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v20
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v12
-; CHECK-NEXT:    vsetvli zero, a2, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfsub.vv v16, v16, v24, v0.t
 ; CHECK-NEXT:    vsetvli a2, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v12, v16
@@ -453,9 +452,10 @@ define <vscale x 32 x bfloat> @vfsub_vv_nxv32bf16_unmasked(<vscale x 32 x bfloat
 ; CHECK-NEXT:  .LBB21_2:
 ; CHECK-NEXT:    addi a1, sp, 16
 ; CHECK-NEXT:    vl8r.v v24, (a1) # Unknown-size Folded Reload
+; CHECK-NEXT:    vsetvli zero, a0, e16, m4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v24
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v8
-; CHECK-NEXT:    vsetvli zero, a0, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfsub.vv v16, v24, v16
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v16
@@ -480,14 +480,10 @@ define <vscale x 32 x bfloat> @vfsub_vf_nxv32bf16(<vscale x 32 x bfloat> %va, bf
 ; CHECK-NEXT:    add a1, a2, a1
 ; CHECK-NEXT:    sub sp, sp, a1
 ; CHECK-NEXT:    .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x11, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 17 * vlenb
-; CHECK-NEXT:    vsetvli a1, zero, e16, m4, ta, ma
-; CHECK-NEXT:    vmv8r.v v16, v8
+; CHECK-NEXT:    vsetvli a1, zero, e16, m8, ta, ma
+; CHECK-NEXT:    vmv8r.v v24, v8
 ; CHECK-NEXT:    fmv.x.h a1, fa0
 ; CHECK-NEXT:    csrr a2, vlenb
-; CHECK-NEXT:    addi a3, sp, 16
-; CHECK-NEXT:    vs8r.v v8, (a3) # Unknown-size Folded Spill
-; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v20
-; CHECK-NEXT:    vsetvli a3, zero, e16, m8, ta, ma
 ; CHECK-NEXT:    vmv.v.x v16, a1
 ; CHECK-NEXT:    csrr a1, vlenb
 ; CHECK-NEXT:    slli a3, a1, 3
@@ -508,15 +504,18 @@ define <vscale x 32 x bfloat> @vfsub_vf_nxv32bf16(<vscale x 32 x bfloat> %va, bf
 ; CHECK-NEXT:    sltu a2, a0, a3
 ; CHECK-NEXT:    addi a2, a2, -1
 ; CHECK-NEXT:    and a2, a2, a3
-; CHECK-NEXT:    csrr a3, vlenb
-; CHECK-NEXT:    slli a4, a3, 3
-; CHECK-NEXT:    add a3, a4, a3
-; CHECK-NEXT:    add a3, sp, a3
-; CHECK-NEXT:    addi a3, a3, 16
-; CHECK-NEXT:    vl8r.v v24, (a3) # Unknown-size Folded Reload
-; CHECK-NEXT:    vsetvli a3, zero, e16, m4, ta, ma
+; CHECK-NEXT:    addi a3, sp, 16
+; CHECK-NEXT:    vs8r.v v8, (a3) # Unknown-size Folded Spill
+; CHECK-NEXT:    vsetvli zero, a2, e16, m4, ta, ma
+; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v28
+; CHECK-NEXT:    csrr a2, vlenb
+; CHECK-NEXT:    slli a3, a2, 3
+; CHECK-NEXT:    add a2, a3, a2
+; CHECK-NEXT:    add a2, sp, a2
+; CHECK-NEXT:    addi a2, a2, 16
+; CHECK-NEXT:    vl8r.v v24, (a2) # Unknown-size Folded Reload
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v28
-; CHECK-NEXT:    vsetvli zero, a2, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfsub.vv v16, v8, v16, v0.t
 ; CHECK-NEXT:    vsetvli a2, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v12, v16
@@ -526,20 +525,21 @@ define <vscale x 32 x bfloat> @vfsub_vf_nxv32bf16(<vscale x 32 x bfloat> %va, bf
 ; CHECK-NEXT:  .LBB22_2:
 ; CHECK-NEXT:    addi a1, sp, 16
 ; CHECK-NEXT:    vl8r.v v24, (a1) # Unknown-size Folded Reload
+; CHECK-NEXT:    vsetvli zero, a0, e16, m4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v24
-; CHECK-NEXT:    csrr a1, vlenb
-; CHECK-NEXT:    slli a2, a1, 3
-; CHECK-NEXT:    add a1, a2, a1
-; CHECK-NEXT:    add a1, sp, a1
-; CHECK-NEXT:    addi a1, a1, 16
-; CHECK-NEXT:    vl8r.v v0, (a1) # Unknown-size Folded Reload
+; CHECK-NEXT:    csrr a0, vlenb
+; CHECK-NEXT:    slli a1, a0, 3
+; CHECK-NEXT:    add a0, a1, a0
+; CHECK-NEXT:    add a0, sp, a0
+; CHECK-NEXT:    addi a0, a0, 16
+; CHECK-NEXT:    vl8r.v v0, (a0) # Unknown-size Folded Reload
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v0
-; CHECK-NEXT:    csrr a1, vlenb
-; CHECK-NEXT:    slli a1, a1, 3
-; CHECK-NEXT:    add a1, sp, a1
-; CHECK-NEXT:    addi a1, a1, 16
-; CHECK-NEXT:    vl1r.v v0, (a1) # Unknown-size Folded Reload
-; CHECK-NEXT:    vsetvli zero, a0, e32, m8, ta, ma
+; CHECK-NEXT:    csrr a0, vlenb
+; CHECK-NEXT:    slli a0, a0, 3
+; CHECK-NEXT:    add a0, sp, a0
+; CHECK-NEXT:    addi a0, a0, 16
+; CHECK-NEXT:    vl1r.v v0, (a0) # Unknown-size Folded Reload
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfsub.vv v16, v16, v24, v0.t
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v16
@@ -566,16 +566,10 @@ define <vscale x 32 x bfloat> @vfsub_vf_nxv32bf16_unmasked(<vscale x 32 x bfloat
 ; CHECK-NEXT:    slli a1, a1, 4
 ; CHECK-NEXT:    sub sp, sp, a1
 ; CHECK-NEXT:    .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x10, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 16 * vlenb
-; CHECK-NEXT:    vsetvli a1, zero, e8, m4, ta, ma
-; CHECK-NEXT:    vmv8r.v v16, v8
 ; CHECK-NEXT:    fmv.x.h a1, fa0
 ; CHECK-NEXT:    csrr a2, vlenb
-; CHECK-NEXT:    vmset.m v7
-; CHECK-NEXT:    addi a3, sp, 16
-; CHECK-NEXT:    vs8r.v v8, (a3) # Unknown-size Folded Spill
-; CHECK-NEXT:    vsetvli a3, zero, e16, m4, ta, ma
-; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v20
 ; CHECK-NEXT:    vsetvli a3, zero, e16, m8, ta, ma
+; CHECK-NEXT:    vmset.m v24
 ; CHECK-NEXT:    vmv.v.x v16, a1
 ; CHECK-NEXT:    csrr a1, vlenb
 ; CHECK-NEXT:    slli a1, a1, 3
@@ -586,18 +580,22 @@ define <vscale x 32 x bfloat> @vfsub_vf_nxv32bf16_unmasked(<vscale x 32 x bfloat
 ; CHECK-NEXT:    srli a2, a2, 2
 ; CHECK-NEXT:    sub a3, a0, a1
 ; CHECK-NEXT:    vsetvli a4, zero, e8, mf2, ta, ma
-; CHECK-NEXT:    vslidedown.vx v0, v7, a2
+; CHECK-NEXT:    vslidedown.vx v0, v24, a2
 ; CHECK-NEXT:    sltu a2, a0, a3
 ; CHECK-NEXT:    addi a2, a2, -1
 ; CHECK-NEXT:    and a2, a2, a3
-; CHECK-NEXT:    csrr a3, vlenb
-; CHECK-NEXT:    slli a3, a3, 3
-; CHECK-NEXT:    add a3, sp, a3
-; CHECK-NEXT:    addi a3, a3, 16
-; CHECK-NEXT:    vl8r.v v24, (a3) # Unknown-size Folded Reload
-; CHECK-NEXT:    vsetvli a3, zero, e16, m4, ta, ma
+; CHECK-NEXT:    vmv8r.v v16, v8
+; CHECK-NEXT:    addi a3, sp, 16
+; CHECK-NEXT:    vs8r.v v8, (a3) # Unknown-size Folded Spill
+; CHECK-NEXT:    vsetvli zero, a2, e16, m4, ta, ma
+; CHECK-NEXT:    vfwcvtbf16.f.f.v v8, v20
+; CHECK-NEXT:    csrr a2, vlenb
+; CHECK-NEXT:    slli a2, a2, 3
+; CHECK-NEXT:    add a2, sp, a2
+; CHECK-NEXT:    addi a2, a2, 16
+; CHECK-NEXT:    vl8r.v v24, (a2) # Unknown-size Folded Reload
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v28
-; CHECK-NEXT:    vsetvli zero, a2, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfsub.vv v16, v8, v16, v0.t
 ; CHECK-NEXT:    vsetvli a2, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v12, v16
@@ -607,14 +605,15 @@ define <vscale x 32 x bfloat> @vfsub_vf_nxv32bf16_unmasked(<vscale x 32 x bfloat
 ; CHECK-NEXT:  .LBB23_2:
 ; CHECK-NEXT:    addi a1, sp, 16
 ; CHECK-NEXT:    vl8r.v v24, (a1) # Unknown-size Folded Reload
+; CHECK-NEXT:    vsetvli zero, a0, e16, m4, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v16, v24
-; CHECK-NEXT:    csrr a1, vlenb
-; CHECK-NEXT:    slli a1, a1, 3
-; CHECK-NEXT:    add a1, sp, a1
-; CHECK-NEXT:    addi a1, a1, 16
-; CHECK-NEXT:    vl8r.v v0, (a1) # Unknown-size Folded Reload
+; CHECK-NEXT:    csrr a0, vlenb
+; CHECK-NEXT:    slli a0, a0, 3
+; CHECK-NEXT:    add a0, sp, a0
+; CHECK-NEXT:    addi a0, a0, 16
+; CHECK-NEXT:    vl8r.v v0, (a0) # Unknown-size Folded Reload
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v24, v0
-; CHECK-NEXT:    vsetvli zero, a0, e32, m8, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
 ; CHECK-NEXT:    vfsub.vv v16, v16, v24
 ; CHECK-NEXT:    vsetvli a0, zero, e16, m4, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v8, v16
diff --git a/llvm/test/CodeGen/RISCV/rvv/vl-opt.mir b/llvm/test/CodeGen/RISCV/rvv/vl-opt.mir
index 0a366f4fd89cf..56bfe0fd3eb93 100644
--- a/llvm/test/CodeGen/RISCV/rvv/vl-opt.mir
+++ b/llvm/test/CodeGen/RISCV/rvv/vl-opt.mir
@@ -1,5 +1,5 @@
 # NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py UTC_ARGS: --version 5
-# RUN: llc %s -o - -mtriple=riscv64 -mattr=+v -run-pass=riscv-vl-optimizer -verify-machineinstrs | FileCheck %s
+# RUN: llc %s -o - -mtriple=riscv64 -mattr=+v,+zvfbfmin -run-pass=riscv-vl-optimizer -verify-machineinstrs | FileCheck %s
 
 ---
 name: vnsrl_wv_user
@@ -130,4 +130,14 @@ body: |
     %x:vr = PseudoVFCVT_X_F_V_M1 $noreg, $noreg, 0, -1, 3 /* e32 */, 0
     %y:vr = PseudoVADD_VV_M1 $noreg, %x, $noreg, 1, 3 /* e8 */, 0
 ...
+---
+name: vfncvtbf16_f_f_w_nofpexcept
+body: |
+  bb.0:
+    ; CHECK-LABEL: name: vfncvtbf16_f_f_w_nofpexcept
+    ; CHECK: early-clobber %x:vr = nofpexcept PseudoVFNCVTBF16_F_F_W_M1_E16 $noreg, $noreg, 7, 1, 4 /* e16 */, 0 /* tu, mu */, implicit $frm
+    ; CHECK-NEXT: %y:vr = PseudoVADD_VV_M1 $noreg, %x, $noreg, 1, 4 /* e16 */, 0 /* tu, mu */
+    %x:vr = nofpexcept PseudoVFNCVTBF16_F_F_W_M1_E16 $noreg, $noreg, 7, -1, 4 /* e16 */, 0 /* tu, mu */, implicit $frm
+    %y:vr = PseudoVADD_VV_M1 $noreg, %x, $noreg, 1, 4 /* e16 */, 0
+...