[llvm] [RISCV] Use predicated FP extension when promoting VP operands (PR #79328)
Simeon K via llvm-commits
llvm-commits at lists.llvm.org
Fri Feb 23 09:28:40 PST 2024
https://github.com/simeonkr updated https://github.com/llvm/llvm-project/pull/79328
>From 0ce6f3f06df123426d791c3e782fe18281e3a062 Mon Sep 17 00:00:00 2001
From: Simeon Krastnikov <simeon.krastnikov at imgtec.com>
Date: Thu, 18 Jan 2024 15:46:27 +0000
Subject: [PATCH 1/2] [RISCV] Use predicated FP extension when promoting VP
operands
This ensures that the VL used for the emitted vfwcvt matches the one in
the original operation. Otherwise, a larger VL may be used than is
necessary.
---
.../SelectionDAG/LegalizeVectorOps.cpp | 23 +-
.../RISCV/rvv/fixed-vectors-ceil-vp.ll | 40 +-
.../RISCV/rvv/fixed-vectors-floor-vp.ll | 40 +-
.../RISCV/rvv/fixed-vectors-fmaximum-vp.ll | 152 +++---
.../RISCV/rvv/fixed-vectors-fminimum-vp.ll | 152 +++---
.../RISCV/rvv/fixed-vectors-fp-buildvec.ll | 3 -
.../rvv/fixed-vectors-reduction-fp-vp.ll | 24 +-
.../RISCV/rvv/fixed-vectors-round-vp.ll | 40 +-
.../RISCV/rvv/fixed-vectors-roundeven-vp.ll | 40 +-
.../RISCV/rvv/fixed-vectors-roundtozero-vp.ll | 40 +-
.../RISCV/rvv/fixed-vectors-vfabs-vp.ll | 40 +-
.../RISCV/rvv/fixed-vectors-vfadd-vp.ll | 104 ++--
.../RISCV/rvv/fixed-vectors-vfdiv-vp.ll | 104 ++--
.../RISCV/rvv/fixed-vectors-vfma-vp.ll | 112 ++---
.../RISCV/rvv/fixed-vectors-vfmax-vp.ll | 48 +-
.../RISCV/rvv/fixed-vectors-vfmin-vp.ll | 48 +-
.../RISCV/rvv/fixed-vectors-vfmul-vp.ll | 104 ++--
.../RISCV/rvv/fixed-vectors-vfneg-vp.ll | 40 +-
.../RISCV/rvv/fixed-vectors-vfsqrt-vp.ll | 40 +-
.../RISCV/rvv/fixed-vectors-vfsub-vp.ll | 104 ++--
llvm/test/CodeGen/RISCV/rvv/fmaximum-vp.ll | 452 +++++++++---------
llvm/test/CodeGen/RISCV/rvv/fminimum-vp.ll | 452 +++++++++---------
llvm/test/CodeGen/RISCV/rvv/nearbyint-vp.ll | 103 ++--
llvm/test/CodeGen/RISCV/rvv/rint-vp.ll | 82 ++--
llvm/test/CodeGen/RISCV/rvv/round-vp.ll | 82 ++--
llvm/test/CodeGen/RISCV/rvv/roundeven-vp.ll | 82 ++--
llvm/test/CodeGen/RISCV/rvv/roundtozero-vp.ll | 82 ++--
llvm/test/CodeGen/RISCV/rvv/vcopysign-vp.ll | 134 ++++--
llvm/test/CodeGen/RISCV/rvv/vfabs-vp.ll | 70 +--
llvm/test/CodeGen/RISCV/rvv/vfadd-vp.ll | 255 ++++++----
llvm/test/CodeGen/RISCV/rvv/vfdiv-vp.ll | 245 ++++++----
llvm/test/CodeGen/RISCV/rvv/vfmax-vp.ll | 134 ++++--
llvm/test/CodeGen/RISCV/rvv/vfmin-vp.ll | 134 ++++--
llvm/test/CodeGen/RISCV/rvv/vfmul-vp.ll | 245 ++++++----
llvm/test/CodeGen/RISCV/rvv/vfneg-vp.ll | 70 +--
llvm/test/CodeGen/RISCV/rvv/vfsqrt-vp.ll | 70 +--
llvm/test/CodeGen/RISCV/rvv/vfsub-vp.ll | 245 ++++++----
.../CodeGen/RISCV/rvv/vreductions-fp-vp.ll | 278 ++++++-----
38 files changed, 2454 insertions(+), 2059 deletions(-)
diff --git a/llvm/lib/CodeGen/SelectionDAG/LegalizeVectorOps.cpp b/llvm/lib/CodeGen/SelectionDAG/LegalizeVectorOps.cpp
index 1fbd6322f9ed9c..937a6a0df65586 100644
--- a/llvm/lib/CodeGen/SelectionDAG/LegalizeVectorOps.cpp
+++ b/llvm/lib/CodeGen/SelectionDAG/LegalizeVectorOps.cpp
@@ -593,8 +593,16 @@ void VectorLegalizer::PromoteReduction(SDNode *Node,
ISD::getVPMaskIdx(Node->getOpcode()) == j)) // Skip mask operand.
// promote the vector operand.
if (Node->getOperand(j).getValueType().isFloatingPoint())
- Operands[j] =
- DAG.getNode(ISD::FP_EXTEND, DL, NewVecVT, Node->getOperand(j));
+ if (!ISD::isVPOpcode(Node->getOpcode()))
+ Operands[j] =
+ DAG.getNode(ISD::FP_EXTEND, DL, NewVecVT, Node->getOperand(j));
+ else {
+ auto MaskIdx = *ISD::getVPMaskIdx(Node->getOpcode());
+ auto EVLIdx = *ISD::getVPExplicitVectorLengthIdx(Node->getOpcode());
+ Operands[j] =
+ DAG.getNode(ISD::VP_FP_EXTEND, DL, NewVecVT, Node->getOperand(j),
+ Node->getOperand(MaskIdx), Node->getOperand(EVLIdx));
+ }
else
Operands[j] =
DAG.getNode(ISD::ANY_EXTEND, DL, NewVecVT, Node->getOperand(j));
@@ -756,7 +764,16 @@ void VectorLegalizer::Promote(SDNode *Node, SmallVectorImpl<SDValue> &Results) {
.getVectorElementType()
.isFloatingPoint() &&
NVT.isVector() && NVT.getVectorElementType().isFloatingPoint())
- Operands[j] = DAG.getNode(ISD::FP_EXTEND, dl, NVT, Node->getOperand(j));
+ if (!ISD::isVPOpcode(Node->getOpcode()))
+ Operands[j] =
+ DAG.getNode(ISD::FP_EXTEND, dl, NVT, Node->getOperand(j));
+ else {
+ auto MaskIdx = *ISD::getVPMaskIdx(Node->getOpcode());
+ auto EVLIdx = *ISD::getVPExplicitVectorLengthIdx(Node->getOpcode());
+ Operands[j] =
+ DAG.getNode(ISD::VP_FP_EXTEND, dl, NVT, Node->getOperand(j),
+ Node->getOperand(MaskIdx), Node->getOperand(EVLIdx));
+ }
else
Operands[j] = DAG.getNode(ISD::BITCAST, dl, NVT, Node->getOperand(j));
else
diff --git a/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-ceil-vp.ll b/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-ceil-vp.ll
index 194179f9f470e1..375ac3e29a71e6 100644
--- a/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-ceil-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-ceil-vp.ll
@@ -30,9 +30,9 @@ define <2 x half> @vp_ceil_v2f16(<2 x half> %va, <2 x i1> %m, i32 zeroext %evl)
;
; ZVFHMIN-LABEL: vp_ceil_v2f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v9, v0.t
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -70,9 +70,9 @@ define <2 x half> @vp_ceil_v2f16_unmasked(<2 x half> %va, i32 zeroext %evl) {
;
; ZVFHMIN-LABEL: vp_ceil_v2f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v9
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -114,9 +114,9 @@ define <4 x half> @vp_ceil_v4f16(<4 x half> %va, <4 x i1> %m, i32 zeroext %evl)
;
; ZVFHMIN-LABEL: vp_ceil_v4f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v9, v0.t
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -154,9 +154,9 @@ define <4 x half> @vp_ceil_v4f16_unmasked(<4 x half> %va, i32 zeroext %evl) {
;
; ZVFHMIN-LABEL: vp_ceil_v4f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v9
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -199,9 +199,9 @@ define <8 x half> @vp_ceil_v8f16(<8 x half> %va, <8 x i1> %m, i32 zeroext %evl)
; ZVFHMIN-LABEL: vp_ceil_v8f16:
; ZVFHMIN: # %bb.0:
; ZVFHMIN-NEXT: vmv1r.v v9, v0
-; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfabs.v v12, v10, v0.t
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -240,9 +240,9 @@ define <8 x half> @vp_ceil_v8f16_unmasked(<8 x half> %va, i32 zeroext %evl) {
;
; ZVFHMIN-LABEL: vp_ceil_v8f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v10
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -287,9 +287,9 @@ define <16 x half> @vp_ceil_v16f16(<16 x half> %va, <16 x i1> %m, i32 zeroext %e
; ZVFHMIN-LABEL: vp_ceil_v16f16:
; ZVFHMIN: # %bb.0:
; ZVFHMIN-NEXT: vmv1r.v v10, v0
-; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfabs.v v16, v12, v0.t
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -328,9 +328,9 @@ define <16 x half> @vp_ceil_v16f16_unmasked(<16 x half> %va, i32 zeroext %evl) {
;
; ZVFHMIN-LABEL: vp_ceil_v16f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v12
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
diff --git a/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-floor-vp.ll b/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-floor-vp.ll
index 583742224f8cf1..bdefb1525b71ea 100644
--- a/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-floor-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-floor-vp.ll
@@ -30,9 +30,9 @@ define <2 x half> @vp_floor_v2f16(<2 x half> %va, <2 x i1> %m, i32 zeroext %evl)
;
; ZVFHMIN-LABEL: vp_floor_v2f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v9, v0.t
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -70,9 +70,9 @@ define <2 x half> @vp_floor_v2f16_unmasked(<2 x half> %va, i32 zeroext %evl) {
;
; ZVFHMIN-LABEL: vp_floor_v2f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v9
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -114,9 +114,9 @@ define <4 x half> @vp_floor_v4f16(<4 x half> %va, <4 x i1> %m, i32 zeroext %evl)
;
; ZVFHMIN-LABEL: vp_floor_v4f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v9, v0.t
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -154,9 +154,9 @@ define <4 x half> @vp_floor_v4f16_unmasked(<4 x half> %va, i32 zeroext %evl) {
;
; ZVFHMIN-LABEL: vp_floor_v4f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v9
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -199,9 +199,9 @@ define <8 x half> @vp_floor_v8f16(<8 x half> %va, <8 x i1> %m, i32 zeroext %evl)
; ZVFHMIN-LABEL: vp_floor_v8f16:
; ZVFHMIN: # %bb.0:
; ZVFHMIN-NEXT: vmv1r.v v9, v0
-; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfabs.v v12, v10, v0.t
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -240,9 +240,9 @@ define <8 x half> @vp_floor_v8f16_unmasked(<8 x half> %va, i32 zeroext %evl) {
;
; ZVFHMIN-LABEL: vp_floor_v8f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v10
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -287,9 +287,9 @@ define <16 x half> @vp_floor_v16f16(<16 x half> %va, <16 x i1> %m, i32 zeroext %
; ZVFHMIN-LABEL: vp_floor_v16f16:
; ZVFHMIN: # %bb.0:
; ZVFHMIN-NEXT: vmv1r.v v10, v0
-; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfabs.v v16, v12, v0.t
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -328,9 +328,9 @@ define <16 x half> @vp_floor_v16f16_unmasked(<16 x half> %va, i32 zeroext %evl)
;
; ZVFHMIN-LABEL: vp_floor_v16f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v12
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
diff --git a/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-fmaximum-vp.ll b/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-fmaximum-vp.ll
index 4a5ef21efdb968..fe1a409288dbcf 100644
--- a/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-fmaximum-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-fmaximum-vp.ll
@@ -27,19 +27,17 @@ define <2 x half> @vfmax_vv_v2f16(<2 x half> %va, <2 x half> %vb, <2 x i1> %m, i
; ZVFHMIN-LABEL: vfmax_vv_v2f16:
; ZVFHMIN: # %bb.0:
; ZVFHMIN-NEXT: vmv1r.v v10, v0
-; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v11, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v0, v11, v11, v0.t
-; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v9
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
-; ZVFHMIN-NEXT: vmerge.vvm v9, v11, v8, v0
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v11, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vmfeq.vv v0, v9, v9, v0.t
+; ZVFHMIN-NEXT: vmerge.vvm v8, v9, v11, v0
; ZVFHMIN-NEXT: vmv1r.v v0, v10
-; ZVFHMIN-NEXT: vmfeq.vv v0, v8, v8, v0.t
-; ZVFHMIN-NEXT: vmerge.vvm v8, v8, v11, v0
+; ZVFHMIN-NEXT: vmfeq.vv v0, v11, v11, v0.t
+; ZVFHMIN-NEXT: vmerge.vvm v9, v11, v9, v0
; ZVFHMIN-NEXT: vmv1r.v v0, v10
-; ZVFHMIN-NEXT: vfmax.vv v9, v8, v9, v0.t
+; ZVFHMIN-NEXT: vfmax.vv v9, v9, v8, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
; ZVFHMIN-NEXT: ret
@@ -61,18 +59,16 @@ define <2 x half> @vfmax_vv_v2f16_unmasked(<2 x half> %va, <2 x half> %vb, i32 z
;
; ZVFHMIN-LABEL: vfmax_vv_v2f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v0, v10, v10
-; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v11, v9
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v8, v11, v11
-; ZVFHMIN-NEXT: vmerge.vvm v9, v10, v11, v0
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vmfeq.vv v0, v9, v9
+; ZVFHMIN-NEXT: vmfeq.vv v8, v10, v10
+; ZVFHMIN-NEXT: vmerge.vvm v11, v9, v10, v0
; ZVFHMIN-NEXT: vmv1r.v v0, v8
-; ZVFHMIN-NEXT: vmerge.vvm v8, v11, v10, v0
-; ZVFHMIN-NEXT: vfmax.vv v9, v8, v9
+; ZVFHMIN-NEXT: vmerge.vvm v8, v10, v9, v0
+; ZVFHMIN-NEXT: vfmax.vv v9, v8, v11
; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
; ZVFHMIN-NEXT: ret
@@ -101,19 +97,17 @@ define <4 x half> @vfmax_vv_v4f16(<4 x half> %va, <4 x half> %vb, <4 x i1> %m, i
; ZVFHMIN-LABEL: vfmax_vv_v4f16:
; ZVFHMIN: # %bb.0:
; ZVFHMIN-NEXT: vmv1r.v v10, v0
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v11, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v0, v11, v11, v0.t
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v9
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
-; ZVFHMIN-NEXT: vmerge.vvm v9, v11, v8, v0
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v11, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vmfeq.vv v0, v9, v9, v0.t
+; ZVFHMIN-NEXT: vmerge.vvm v8, v9, v11, v0
; ZVFHMIN-NEXT: vmv1r.v v0, v10
-; ZVFHMIN-NEXT: vmfeq.vv v0, v8, v8, v0.t
-; ZVFHMIN-NEXT: vmerge.vvm v8, v8, v11, v0
+; ZVFHMIN-NEXT: vmfeq.vv v0, v11, v11, v0.t
+; ZVFHMIN-NEXT: vmerge.vvm v9, v11, v9, v0
; ZVFHMIN-NEXT: vmv1r.v v0, v10
-; ZVFHMIN-NEXT: vfmax.vv v9, v8, v9, v0.t
+; ZVFHMIN-NEXT: vfmax.vv v9, v9, v8, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
; ZVFHMIN-NEXT: ret
@@ -135,18 +129,16 @@ define <4 x half> @vfmax_vv_v4f16_unmasked(<4 x half> %va, <4 x half> %vb, i32 z
;
; ZVFHMIN-LABEL: vfmax_vv_v4f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v0, v10, v10
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v11, v9
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v8, v11, v11
-; ZVFHMIN-NEXT: vmerge.vvm v9, v10, v11, v0
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vmfeq.vv v0, v9, v9
+; ZVFHMIN-NEXT: vmfeq.vv v8, v10, v10
+; ZVFHMIN-NEXT: vmerge.vvm v11, v9, v10, v0
; ZVFHMIN-NEXT: vmv.v.v v0, v8
-; ZVFHMIN-NEXT: vmerge.vvm v8, v11, v10, v0
-; ZVFHMIN-NEXT: vfmax.vv v9, v8, v9
+; ZVFHMIN-NEXT: vmerge.vvm v8, v10, v9, v0
+; ZVFHMIN-NEXT: vfmax.vv v9, v8, v11
; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
; ZVFHMIN-NEXT: ret
@@ -175,19 +167,17 @@ define <8 x half> @vfmax_vv_v8f16(<8 x half> %va, <8 x half> %vb, <8 x i1> %m, i
; ZVFHMIN-LABEL: vfmax_vv_v8f16:
; ZVFHMIN: # %bb.0:
; ZVFHMIN-NEXT: vmv1r.v v10, v0
-; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v8, v12, v12, v0.t
-; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v14, v9
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v14, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vmfeq.vv v8, v14, v14, v0.t
; ZVFHMIN-NEXT: vmv1r.v v0, v8
-; ZVFHMIN-NEXT: vmerge.vvm v16, v12, v14, v0
+; ZVFHMIN-NEXT: vmerge.vvm v16, v14, v12, v0
; ZVFHMIN-NEXT: vmv1r.v v0, v10
-; ZVFHMIN-NEXT: vmfeq.vv v8, v14, v14, v0.t
+; ZVFHMIN-NEXT: vmfeq.vv v8, v12, v12, v0.t
; ZVFHMIN-NEXT: vmv1r.v v0, v8
-; ZVFHMIN-NEXT: vmerge.vvm v8, v14, v12, v0
+; ZVFHMIN-NEXT: vmerge.vvm v8, v12, v14, v0
; ZVFHMIN-NEXT: vmv1r.v v0, v10
; ZVFHMIN-NEXT: vfmax.vv v10, v8, v16, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
@@ -211,17 +201,15 @@ define <8 x half> @vfmax_vv_v8f16_unmasked(<8 x half> %va, <8 x half> %vb, i32 z
;
; ZVFHMIN-LABEL: vfmax_vv_v8f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v0, v10, v10
-; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v9
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v8, v12, v12
-; ZVFHMIN-NEXT: vmerge.vvm v14, v10, v12, v0
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vmfeq.vv v0, v12, v12
+; ZVFHMIN-NEXT: vmfeq.vv v8, v10, v10
+; ZVFHMIN-NEXT: vmerge.vvm v14, v12, v10, v0
; ZVFHMIN-NEXT: vmv1r.v v0, v8
-; ZVFHMIN-NEXT: vmerge.vvm v8, v12, v10, v0
+; ZVFHMIN-NEXT: vmerge.vvm v8, v10, v12, v0
; ZVFHMIN-NEXT: vfmax.vv v10, v8, v14
; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -253,19 +241,17 @@ define <16 x half> @vfmax_vv_v16f16(<16 x half> %va, <16 x half> %vb, <16 x i1>
; ZVFHMIN-LABEL: vfmax_vv_v16f16:
; ZVFHMIN: # %bb.0:
; ZVFHMIN-NEXT: vmv1r.v v12, v0
-; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v8, v16, v16, v0.t
-; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v20, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v10, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v20, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vmfeq.vv v8, v20, v20, v0.t
; ZVFHMIN-NEXT: vmv1r.v v0, v8
-; ZVFHMIN-NEXT: vmerge.vvm v24, v16, v20, v0
+; ZVFHMIN-NEXT: vmerge.vvm v24, v20, v16, v0
; ZVFHMIN-NEXT: vmv1r.v v0, v12
-; ZVFHMIN-NEXT: vmfeq.vv v8, v20, v20, v0.t
+; ZVFHMIN-NEXT: vmfeq.vv v8, v16, v16, v0.t
; ZVFHMIN-NEXT: vmv1r.v v0, v8
-; ZVFHMIN-NEXT: vmerge.vvm v8, v20, v16, v0
+; ZVFHMIN-NEXT: vmerge.vvm v8, v16, v20, v0
; ZVFHMIN-NEXT: vmv1r.v v0, v12
; ZVFHMIN-NEXT: vfmax.vv v12, v8, v24, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
@@ -289,17 +275,15 @@ define <16 x half> @vfmax_vv_v16f16_unmasked(<16 x half> %va, <16 x half> %vb, i
;
; ZVFHMIN-LABEL: vfmax_vv_v16f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v0, v12, v12
-; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v8, v16, v16
-; ZVFHMIN-NEXT: vmerge.vvm v20, v12, v16, v0
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v10
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vmfeq.vv v0, v16, v16
+; ZVFHMIN-NEXT: vmfeq.vv v8, v12, v12
+; ZVFHMIN-NEXT: vmerge.vvm v20, v16, v12, v0
; ZVFHMIN-NEXT: vmv1r.v v0, v8
-; ZVFHMIN-NEXT: vmerge.vvm v8, v16, v12, v0
+; ZVFHMIN-NEXT: vmerge.vvm v8, v12, v16, v0
; ZVFHMIN-NEXT: vfmax.vv v12, v8, v20
; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
diff --git a/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-fminimum-vp.ll b/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-fminimum-vp.ll
index 8ea08bdedd6d34..cf21c95495eb86 100644
--- a/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-fminimum-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-fminimum-vp.ll
@@ -27,19 +27,17 @@ define <2 x half> @vfmin_vv_v2f16(<2 x half> %va, <2 x half> %vb, <2 x i1> %m, i
; ZVFHMIN-LABEL: vfmin_vv_v2f16:
; ZVFHMIN: # %bb.0:
; ZVFHMIN-NEXT: vmv1r.v v10, v0
-; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v11, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v0, v11, v11, v0.t
-; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v9
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
-; ZVFHMIN-NEXT: vmerge.vvm v9, v11, v8, v0
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v11, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vmfeq.vv v0, v9, v9, v0.t
+; ZVFHMIN-NEXT: vmerge.vvm v8, v9, v11, v0
; ZVFHMIN-NEXT: vmv1r.v v0, v10
-; ZVFHMIN-NEXT: vmfeq.vv v0, v8, v8, v0.t
-; ZVFHMIN-NEXT: vmerge.vvm v8, v8, v11, v0
+; ZVFHMIN-NEXT: vmfeq.vv v0, v11, v11, v0.t
+; ZVFHMIN-NEXT: vmerge.vvm v9, v11, v9, v0
; ZVFHMIN-NEXT: vmv1r.v v0, v10
-; ZVFHMIN-NEXT: vfmin.vv v9, v8, v9, v0.t
+; ZVFHMIN-NEXT: vfmin.vv v9, v9, v8, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
; ZVFHMIN-NEXT: ret
@@ -61,18 +59,16 @@ define <2 x half> @vfmin_vv_v2f16_unmasked(<2 x half> %va, <2 x half> %vb, i32 z
;
; ZVFHMIN-LABEL: vfmin_vv_v2f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v0, v10, v10
-; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v11, v9
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v8, v11, v11
-; ZVFHMIN-NEXT: vmerge.vvm v9, v10, v11, v0
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vmfeq.vv v0, v9, v9
+; ZVFHMIN-NEXT: vmfeq.vv v8, v10, v10
+; ZVFHMIN-NEXT: vmerge.vvm v11, v9, v10, v0
; ZVFHMIN-NEXT: vmv1r.v v0, v8
-; ZVFHMIN-NEXT: vmerge.vvm v8, v11, v10, v0
-; ZVFHMIN-NEXT: vfmin.vv v9, v8, v9
+; ZVFHMIN-NEXT: vmerge.vvm v8, v10, v9, v0
+; ZVFHMIN-NEXT: vfmin.vv v9, v8, v11
; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
; ZVFHMIN-NEXT: ret
@@ -101,19 +97,17 @@ define <4 x half> @vfmin_vv_v4f16(<4 x half> %va, <4 x half> %vb, <4 x i1> %m, i
; ZVFHMIN-LABEL: vfmin_vv_v4f16:
; ZVFHMIN: # %bb.0:
; ZVFHMIN-NEXT: vmv1r.v v10, v0
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v11, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v0, v11, v11, v0.t
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v9
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
-; ZVFHMIN-NEXT: vmerge.vvm v9, v11, v8, v0
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v11, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vmfeq.vv v0, v9, v9, v0.t
+; ZVFHMIN-NEXT: vmerge.vvm v8, v9, v11, v0
; ZVFHMIN-NEXT: vmv1r.v v0, v10
-; ZVFHMIN-NEXT: vmfeq.vv v0, v8, v8, v0.t
-; ZVFHMIN-NEXT: vmerge.vvm v8, v8, v11, v0
+; ZVFHMIN-NEXT: vmfeq.vv v0, v11, v11, v0.t
+; ZVFHMIN-NEXT: vmerge.vvm v9, v11, v9, v0
; ZVFHMIN-NEXT: vmv1r.v v0, v10
-; ZVFHMIN-NEXT: vfmin.vv v9, v8, v9, v0.t
+; ZVFHMIN-NEXT: vfmin.vv v9, v9, v8, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
; ZVFHMIN-NEXT: ret
@@ -135,18 +129,16 @@ define <4 x half> @vfmin_vv_v4f16_unmasked(<4 x half> %va, <4 x half> %vb, i32 z
;
; ZVFHMIN-LABEL: vfmin_vv_v4f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v0, v10, v10
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v11, v9
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v8, v11, v11
-; ZVFHMIN-NEXT: vmerge.vvm v9, v10, v11, v0
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vmfeq.vv v0, v9, v9
+; ZVFHMIN-NEXT: vmfeq.vv v8, v10, v10
+; ZVFHMIN-NEXT: vmerge.vvm v11, v9, v10, v0
; ZVFHMIN-NEXT: vmv.v.v v0, v8
-; ZVFHMIN-NEXT: vmerge.vvm v8, v11, v10, v0
-; ZVFHMIN-NEXT: vfmin.vv v9, v8, v9
+; ZVFHMIN-NEXT: vmerge.vvm v8, v10, v9, v0
+; ZVFHMIN-NEXT: vfmin.vv v9, v8, v11
; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
; ZVFHMIN-NEXT: ret
@@ -175,19 +167,17 @@ define <8 x half> @vfmin_vv_v8f16(<8 x half> %va, <8 x half> %vb, <8 x i1> %m, i
; ZVFHMIN-LABEL: vfmin_vv_v8f16:
; ZVFHMIN: # %bb.0:
; ZVFHMIN-NEXT: vmv1r.v v10, v0
-; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v8, v12, v12, v0.t
-; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v14, v9
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v14, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vmfeq.vv v8, v14, v14, v0.t
; ZVFHMIN-NEXT: vmv1r.v v0, v8
-; ZVFHMIN-NEXT: vmerge.vvm v16, v12, v14, v0
+; ZVFHMIN-NEXT: vmerge.vvm v16, v14, v12, v0
; ZVFHMIN-NEXT: vmv1r.v v0, v10
-; ZVFHMIN-NEXT: vmfeq.vv v8, v14, v14, v0.t
+; ZVFHMIN-NEXT: vmfeq.vv v8, v12, v12, v0.t
; ZVFHMIN-NEXT: vmv1r.v v0, v8
-; ZVFHMIN-NEXT: vmerge.vvm v8, v14, v12, v0
+; ZVFHMIN-NEXT: vmerge.vvm v8, v12, v14, v0
; ZVFHMIN-NEXT: vmv1r.v v0, v10
; ZVFHMIN-NEXT: vfmin.vv v10, v8, v16, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
@@ -211,17 +201,15 @@ define <8 x half> @vfmin_vv_v8f16_unmasked(<8 x half> %va, <8 x half> %vb, i32 z
;
; ZVFHMIN-LABEL: vfmin_vv_v8f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v0, v10, v10
-; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v9
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v8, v12, v12
-; ZVFHMIN-NEXT: vmerge.vvm v14, v10, v12, v0
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vmfeq.vv v0, v12, v12
+; ZVFHMIN-NEXT: vmfeq.vv v8, v10, v10
+; ZVFHMIN-NEXT: vmerge.vvm v14, v12, v10, v0
; ZVFHMIN-NEXT: vmv1r.v v0, v8
-; ZVFHMIN-NEXT: vmerge.vvm v8, v12, v10, v0
+; ZVFHMIN-NEXT: vmerge.vvm v8, v10, v12, v0
; ZVFHMIN-NEXT: vfmin.vv v10, v8, v14
; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -253,19 +241,17 @@ define <16 x half> @vfmin_vv_v16f16(<16 x half> %va, <16 x half> %vb, <16 x i1>
; ZVFHMIN-LABEL: vfmin_vv_v16f16:
; ZVFHMIN: # %bb.0:
; ZVFHMIN-NEXT: vmv1r.v v12, v0
-; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v8, v16, v16, v0.t
-; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v20, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v10, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v20, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vmfeq.vv v8, v20, v20, v0.t
; ZVFHMIN-NEXT: vmv1r.v v0, v8
-; ZVFHMIN-NEXT: vmerge.vvm v24, v16, v20, v0
+; ZVFHMIN-NEXT: vmerge.vvm v24, v20, v16, v0
; ZVFHMIN-NEXT: vmv1r.v v0, v12
-; ZVFHMIN-NEXT: vmfeq.vv v8, v20, v20, v0.t
+; ZVFHMIN-NEXT: vmfeq.vv v8, v16, v16, v0.t
; ZVFHMIN-NEXT: vmv1r.v v0, v8
-; ZVFHMIN-NEXT: vmerge.vvm v8, v20, v16, v0
+; ZVFHMIN-NEXT: vmerge.vvm v8, v16, v20, v0
; ZVFHMIN-NEXT: vmv1r.v v0, v12
; ZVFHMIN-NEXT: vfmin.vv v12, v8, v24, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
@@ -289,17 +275,15 @@ define <16 x half> @vfmin_vv_v16f16_unmasked(<16 x half> %va, <16 x half> %vb, i
;
; ZVFHMIN-LABEL: vfmin_vv_v16f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v0, v12, v12
-; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v8, v16, v16
-; ZVFHMIN-NEXT: vmerge.vvm v20, v12, v16, v0
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v10
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vmfeq.vv v0, v16, v16
+; ZVFHMIN-NEXT: vmfeq.vv v8, v12, v12
+; ZVFHMIN-NEXT: vmerge.vvm v20, v16, v12, v0
; ZVFHMIN-NEXT: vmv1r.v v0, v8
-; ZVFHMIN-NEXT: vmerge.vvm v8, v16, v12, v0
+; ZVFHMIN-NEXT: vmerge.vvm v8, v12, v16, v0
; ZVFHMIN-NEXT: vfmin.vv v12, v8, v20
; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
diff --git a/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-fp-buildvec.ll b/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-fp-buildvec.ll
index 4852850f234ba6..924094c00fe7b1 100644
--- a/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-fp-buildvec.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-fp-buildvec.ll
@@ -1394,6 +1394,3 @@ define <2 x double> @vid_step2_v2f64() {
; CHECK-NEXT: ret
ret <2 x double> <double 0.0, double 2.0>
}
-;; NOTE: These prefixes are unused and the list is autogenerated. Do not add tests below this line:
-; RV32: {{.*}}
-; RV64: {{.*}}
diff --git a/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-reduction-fp-vp.ll b/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-reduction-fp-vp.ll
index 9df160bf30f005..5cdc9315589eaf 100644
--- a/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-reduction-fp-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-reduction-fp-vp.ll
@@ -22,10 +22,10 @@ define half @vpreduce_fadd_v2f16(half %s, <2 x half> %v, <2 x i1> %m, i32 zeroex
;
; ZVFHMIN-LABEL: vpreduce_fadd_v2f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
; ZVFHMIN-NEXT: fcvt.s.h fa5, fa0
-; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetivli zero, 1, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfmv.s.f v8, fa5
; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfredusum.vs v8, v9, v8, v0.t
@@ -48,10 +48,10 @@ define half @vpreduce_ord_fadd_v2f16(half %s, <2 x half> %v, <2 x i1> %m, i32 ze
;
; ZVFHMIN-LABEL: vpreduce_ord_fadd_v2f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
; ZVFHMIN-NEXT: fcvt.s.h fa5, fa0
-; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetivli zero, 1, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfmv.s.f v8, fa5
; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfredosum.vs v8, v9, v8, v0.t
@@ -76,10 +76,10 @@ define half @vpreduce_fadd_v4f16(half %s, <4 x half> %v, <4 x i1> %m, i32 zeroex
;
; ZVFHMIN-LABEL: vpreduce_fadd_v4f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
; ZVFHMIN-NEXT: fcvt.s.h fa5, fa0
-; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetivli zero, 1, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfmv.s.f v8, fa5
; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfredusum.vs v8, v9, v8, v0.t
@@ -102,10 +102,10 @@ define half @vpreduce_ord_fadd_v4f16(half %s, <4 x half> %v, <4 x i1> %m, i32 ze
;
; ZVFHMIN-LABEL: vpreduce_ord_fadd_v4f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
; ZVFHMIN-NEXT: fcvt.s.h fa5, fa0
-; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetivli zero, 1, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfmv.s.f v8, fa5
; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfredosum.vs v8, v9, v8, v0.t
diff --git a/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-round-vp.ll b/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-round-vp.ll
index 96a72d0ddd18b8..b87b03a9f06a28 100644
--- a/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-round-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-round-vp.ll
@@ -30,9 +30,9 @@ define <2 x half> @vp_round_v2f16(<2 x half> %va, <2 x i1> %m, i32 zeroext %evl)
;
; ZVFHMIN-LABEL: vp_round_v2f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v9, v0.t
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -70,9 +70,9 @@ define <2 x half> @vp_round_v2f16_unmasked(<2 x half> %va, i32 zeroext %evl) {
;
; ZVFHMIN-LABEL: vp_round_v2f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v9
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -114,9 +114,9 @@ define <4 x half> @vp_round_v4f16(<4 x half> %va, <4 x i1> %m, i32 zeroext %evl)
;
; ZVFHMIN-LABEL: vp_round_v4f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v9, v0.t
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -154,9 +154,9 @@ define <4 x half> @vp_round_v4f16_unmasked(<4 x half> %va, i32 zeroext %evl) {
;
; ZVFHMIN-LABEL: vp_round_v4f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v9
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -199,9 +199,9 @@ define <8 x half> @vp_round_v8f16(<8 x half> %va, <8 x i1> %m, i32 zeroext %evl)
; ZVFHMIN-LABEL: vp_round_v8f16:
; ZVFHMIN: # %bb.0:
; ZVFHMIN-NEXT: vmv1r.v v9, v0
-; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfabs.v v12, v10, v0.t
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -240,9 +240,9 @@ define <8 x half> @vp_round_v8f16_unmasked(<8 x half> %va, i32 zeroext %evl) {
;
; ZVFHMIN-LABEL: vp_round_v8f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v10
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -287,9 +287,9 @@ define <16 x half> @vp_round_v16f16(<16 x half> %va, <16 x i1> %m, i32 zeroext %
; ZVFHMIN-LABEL: vp_round_v16f16:
; ZVFHMIN: # %bb.0:
; ZVFHMIN-NEXT: vmv1r.v v10, v0
-; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfabs.v v16, v12, v0.t
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -328,9 +328,9 @@ define <16 x half> @vp_round_v16f16_unmasked(<16 x half> %va, i32 zeroext %evl)
;
; ZVFHMIN-LABEL: vp_round_v16f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v12
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
diff --git a/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-roundeven-vp.ll b/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-roundeven-vp.ll
index 74a43f09542af4..4fb54c189b0854 100644
--- a/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-roundeven-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-roundeven-vp.ll
@@ -30,9 +30,9 @@ define <2 x half> @vp_roundeven_v2f16(<2 x half> %va, <2 x i1> %m, i32 zeroext %
;
; ZVFHMIN-LABEL: vp_roundeven_v2f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v9, v0.t
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -70,9 +70,9 @@ define <2 x half> @vp_roundeven_v2f16_unmasked(<2 x half> %va, i32 zeroext %evl)
;
; ZVFHMIN-LABEL: vp_roundeven_v2f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v9
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -114,9 +114,9 @@ define <4 x half> @vp_roundeven_v4f16(<4 x half> %va, <4 x i1> %m, i32 zeroext %
;
; ZVFHMIN-LABEL: vp_roundeven_v4f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v9, v0.t
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -154,9 +154,9 @@ define <4 x half> @vp_roundeven_v4f16_unmasked(<4 x half> %va, i32 zeroext %evl)
;
; ZVFHMIN-LABEL: vp_roundeven_v4f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v9
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -199,9 +199,9 @@ define <8 x half> @vp_roundeven_v8f16(<8 x half> %va, <8 x i1> %m, i32 zeroext %
; ZVFHMIN-LABEL: vp_roundeven_v8f16:
; ZVFHMIN: # %bb.0:
; ZVFHMIN-NEXT: vmv1r.v v9, v0
-; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfabs.v v12, v10, v0.t
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -240,9 +240,9 @@ define <8 x half> @vp_roundeven_v8f16_unmasked(<8 x half> %va, i32 zeroext %evl)
;
; ZVFHMIN-LABEL: vp_roundeven_v8f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v10
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -287,9 +287,9 @@ define <16 x half> @vp_roundeven_v16f16(<16 x half> %va, <16 x i1> %m, i32 zeroe
; ZVFHMIN-LABEL: vp_roundeven_v16f16:
; ZVFHMIN: # %bb.0:
; ZVFHMIN-NEXT: vmv1r.v v10, v0
-; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfabs.v v16, v12, v0.t
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -328,9 +328,9 @@ define <16 x half> @vp_roundeven_v16f16_unmasked(<16 x half> %va, i32 zeroext %e
;
; ZVFHMIN-LABEL: vp_roundeven_v16f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v12
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
diff --git a/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-roundtozero-vp.ll b/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-roundtozero-vp.ll
index 91de65c79bb705..190744e10315a7 100644
--- a/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-roundtozero-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-roundtozero-vp.ll
@@ -30,9 +30,9 @@ define <2 x half> @vp_roundtozero_v2f16(<2 x half> %va, <2 x i1> %m, i32 zeroext
;
; ZVFHMIN-LABEL: vp_roundtozero_v2f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v9, v0.t
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -70,9 +70,9 @@ define <2 x half> @vp_roundtozero_v2f16_unmasked(<2 x half> %va, i32 zeroext %ev
;
; ZVFHMIN-LABEL: vp_roundtozero_v2f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v9
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -114,9 +114,9 @@ define <4 x half> @vp_roundtozero_v4f16(<4 x half> %va, <4 x i1> %m, i32 zeroext
;
; ZVFHMIN-LABEL: vp_roundtozero_v4f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v9, v0.t
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -154,9 +154,9 @@ define <4 x half> @vp_roundtozero_v4f16_unmasked(<4 x half> %va, i32 zeroext %ev
;
; ZVFHMIN-LABEL: vp_roundtozero_v4f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v9
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -199,9 +199,9 @@ define <8 x half> @vp_roundtozero_v8f16(<8 x half> %va, <8 x i1> %m, i32 zeroext
; ZVFHMIN-LABEL: vp_roundtozero_v8f16:
; ZVFHMIN: # %bb.0:
; ZVFHMIN-NEXT: vmv1r.v v9, v0
-; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfabs.v v12, v10, v0.t
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -240,9 +240,9 @@ define <8 x half> @vp_roundtozero_v8f16_unmasked(<8 x half> %va, i32 zeroext %ev
;
; ZVFHMIN-LABEL: vp_roundtozero_v8f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v10
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -287,9 +287,9 @@ define <16 x half> @vp_roundtozero_v16f16(<16 x half> %va, <16 x i1> %m, i32 zer
; ZVFHMIN-LABEL: vp_roundtozero_v16f16:
; ZVFHMIN: # %bb.0:
; ZVFHMIN-NEXT: vmv1r.v v10, v0
-; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfabs.v v16, v12, v0.t
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -328,9 +328,9 @@ define <16 x half> @vp_roundtozero_v16f16_unmasked(<16 x half> %va, i32 zeroext
;
; ZVFHMIN-LABEL: vp_roundtozero_v16f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v12
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
diff --git a/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vfabs-vp.ll b/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vfabs-vp.ll
index a30f682d5cf1e2..9e650b77b74ebe 100644
--- a/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vfabs-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vfabs-vp.ll
@@ -19,9 +19,9 @@ define <2 x half> @vfabs_vv_v2f16(<2 x half> %va, <2 x i1> %m, i32 zeroext %evl)
;
; ZVFHMIN-LABEL: vfabs_vv_v2f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfabs.v v9, v9, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -39,9 +39,9 @@ define <2 x half> @vfabs_vv_v2f16_unmasked(<2 x half> %va, i32 zeroext %evl) {
;
; ZVFHMIN-LABEL: vfabs_vv_v2f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfabs.v v9, v9
; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -63,9 +63,9 @@ define <4 x half> @vfabs_vv_v4f16(<4 x half> %va, <4 x i1> %m, i32 zeroext %evl)
;
; ZVFHMIN-LABEL: vfabs_vv_v4f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfabs.v v9, v9, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -83,9 +83,9 @@ define <4 x half> @vfabs_vv_v4f16_unmasked(<4 x half> %va, i32 zeroext %evl) {
;
; ZVFHMIN-LABEL: vfabs_vv_v4f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfabs.v v9, v9
; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -107,9 +107,9 @@ define <8 x half> @vfabs_vv_v8f16(<8 x half> %va, <8 x i1> %m, i32 zeroext %evl)
;
; ZVFHMIN-LABEL: vfabs_vv_v8f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfabs.v v10, v10, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -127,9 +127,9 @@ define <8 x half> @vfabs_vv_v8f16_unmasked(<8 x half> %va, i32 zeroext %evl) {
;
; ZVFHMIN-LABEL: vfabs_vv_v8f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfabs.v v10, v10
; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -151,9 +151,9 @@ define <16 x half> @vfabs_vv_v16f16(<16 x half> %va, <16 x i1> %m, i32 zeroext %
;
; ZVFHMIN-LABEL: vfabs_vv_v16f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfabs.v v12, v12, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -171,9 +171,9 @@ define <16 x half> @vfabs_vv_v16f16_unmasked(<16 x half> %va, i32 zeroext %evl)
;
; ZVFHMIN-LABEL: vfabs_vv_v16f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfabs.v v12, v12
; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
diff --git a/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vfadd-vp.ll b/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vfadd-vp.ll
index fa01057b21203e..6e078d43a86c80 100644
--- a/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vfadd-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vfadd-vp.ll
@@ -19,10 +19,10 @@ define <2 x half> @vfadd_vv_v2f16(<2 x half> %va, <2 x half> %b, <2 x i1> %m, i3
;
; ZVFHMIN-LABEL: vfadd_vv_v2f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfadd.vv v9, v9, v10, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -40,10 +40,10 @@ define <2 x half> @vfadd_vv_v2f16_unmasked(<2 x half> %va, <2 x half> %b, i32 ze
;
; ZVFHMIN-LABEL: vfadd_vv_v2f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfadd.vv v9, v9, v10
; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -68,10 +68,10 @@ define <2 x half> @vfadd_vf_v2f16(<2 x half> %va, half %b, <2 x i1> %m, i32 zero
; ZVFHMIN-NEXT: vfmv.v.f v9, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v10, v9
-; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfadd.vv v9, v9, v8, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -96,10 +96,10 @@ define <2 x half> @vfadd_vf_v2f16_unmasked(<2 x half> %va, half %b, i32 zeroext
; ZVFHMIN-NEXT: vfmv.v.f v9, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v10, v9
-; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfadd.vv v9, v9, v8
; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -123,10 +123,10 @@ define <3 x half> @vfadd_vv_v3f16(<3 x half> %va, <3 x half> %b, <3 x i1> %m, i3
;
; ZVFHMIN-LABEL: vfadd_vv_v3f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfadd.vv v9, v9, v10, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -146,10 +146,10 @@ define <4 x half> @vfadd_vv_v4f16(<4 x half> %va, <4 x half> %b, <4 x i1> %m, i3
;
; ZVFHMIN-LABEL: vfadd_vv_v4f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfadd.vv v9, v9, v10, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -167,10 +167,10 @@ define <4 x half> @vfadd_vv_v4f16_unmasked(<4 x half> %va, <4 x half> %b, i32 ze
;
; ZVFHMIN-LABEL: vfadd_vv_v4f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfadd.vv v9, v9, v10
; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -195,10 +195,10 @@ define <4 x half> @vfadd_vf_v4f16(<4 x half> %va, half %b, <4 x i1> %m, i32 zero
; ZVFHMIN-NEXT: vfmv.v.f v9, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v10, v9
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfadd.vv v9, v9, v8, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -223,10 +223,10 @@ define <4 x half> @vfadd_vf_v4f16_unmasked(<4 x half> %va, half %b, i32 zeroext
; ZVFHMIN-NEXT: vfmv.v.f v9, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v10, v9
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfadd.vv v9, v9, v8
; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -250,10 +250,10 @@ define <8 x half> @vfadd_vv_v8f16(<8 x half> %va, <8 x half> %b, <8 x i1> %m, i3
;
; ZVFHMIN-LABEL: vfadd_vv_v8f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfadd.vv v10, v12, v10, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -271,10 +271,10 @@ define <8 x half> @vfadd_vv_v8f16_unmasked(<8 x half> %va, <8 x half> %b, i32 ze
;
; ZVFHMIN-LABEL: vfadd_vv_v8f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfadd.vv v10, v12, v10
; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -299,10 +299,10 @@ define <8 x half> @vfadd_vf_v8f16(<8 x half> %va, half %b, <8 x i1> %m, i32 zero
; ZVFHMIN-NEXT: vfmv.v.f v10, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v9, v10
-; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v9
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v9, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfadd.vv v10, v10, v12, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -327,10 +327,10 @@ define <8 x half> @vfadd_vf_v8f16_unmasked(<8 x half> %va, half %b, i32 zeroext
; ZVFHMIN-NEXT: vfmv.v.f v10, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v9, v10
-; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v9
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfadd.vv v10, v10, v12
; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -354,10 +354,10 @@ define <16 x half> @vfadd_vv_v16f16(<16 x half> %va, <16 x half> %b, <16 x i1> %
;
; ZVFHMIN-LABEL: vfadd_vv_v16f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v10
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v10, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfadd.vv v12, v16, v12, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -375,10 +375,10 @@ define <16 x half> @vfadd_vv_v16f16_unmasked(<16 x half> %va, <16 x half> %b, i3
;
; ZVFHMIN-LABEL: vfadd_vv_v16f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v10
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfadd.vv v12, v16, v12
; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -403,10 +403,10 @@ define <16 x half> @vfadd_vf_v16f16(<16 x half> %va, half %b, <16 x i1> %m, i32
; ZVFHMIN-NEXT: vfmv.v.f v12, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v10, v12
-; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v10, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfadd.vv v12, v12, v16, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -431,10 +431,10 @@ define <16 x half> @vfadd_vf_v16f16_unmasked(<16 x half> %va, half %b, i32 zeroe
; ZVFHMIN-NEXT: vfmv.v.f v12, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v10, v12
-; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfadd.vv v12, v12, v16
; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
diff --git a/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vfdiv-vp.ll b/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vfdiv-vp.ll
index 8c19e0bae81cfe..914581409dab9f 100644
--- a/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vfdiv-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vfdiv-vp.ll
@@ -19,10 +19,10 @@ define <2 x half> @vfdiv_vv_v2f16(<2 x half> %va, <2 x half> %b, <2 x i1> %m, i3
;
; ZVFHMIN-LABEL: vfdiv_vv_v2f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfdiv.vv v9, v9, v10, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -40,10 +40,10 @@ define <2 x half> @vfdiv_vv_v2f16_unmasked(<2 x half> %va, <2 x half> %b, i32 ze
;
; ZVFHMIN-LABEL: vfdiv_vv_v2f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfdiv.vv v9, v9, v10
; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -68,10 +68,10 @@ define <2 x half> @vfdiv_vf_v2f16(<2 x half> %va, half %b, <2 x i1> %m, i32 zero
; ZVFHMIN-NEXT: vfmv.v.f v9, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v10, v9
-; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfdiv.vv v9, v9, v8, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -96,10 +96,10 @@ define <2 x half> @vfdiv_vf_v2f16_unmasked(<2 x half> %va, half %b, i32 zeroext
; ZVFHMIN-NEXT: vfmv.v.f v9, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v10, v9
-; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfdiv.vv v9, v9, v8
; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -123,10 +123,10 @@ define <3 x half> @vfdiv_vv_v3f16(<3 x half> %va, <3 x half> %b, <3 x i1> %m, i3
;
; ZVFHMIN-LABEL: vfdiv_vv_v3f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfdiv.vv v9, v9, v10, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -146,10 +146,10 @@ define <4 x half> @vfdiv_vv_v4f16(<4 x half> %va, <4 x half> %b, <4 x i1> %m, i3
;
; ZVFHMIN-LABEL: vfdiv_vv_v4f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfdiv.vv v9, v9, v10, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -167,10 +167,10 @@ define <4 x half> @vfdiv_vv_v4f16_unmasked(<4 x half> %va, <4 x half> %b, i32 ze
;
; ZVFHMIN-LABEL: vfdiv_vv_v4f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfdiv.vv v9, v9, v10
; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -195,10 +195,10 @@ define <4 x half> @vfdiv_vf_v4f16(<4 x half> %va, half %b, <4 x i1> %m, i32 zero
; ZVFHMIN-NEXT: vfmv.v.f v9, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v10, v9
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfdiv.vv v9, v9, v8, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -223,10 +223,10 @@ define <4 x half> @vfdiv_vf_v4f16_unmasked(<4 x half> %va, half %b, i32 zeroext
; ZVFHMIN-NEXT: vfmv.v.f v9, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v10, v9
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfdiv.vv v9, v9, v8
; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -250,10 +250,10 @@ define <8 x half> @vfdiv_vv_v8f16(<8 x half> %va, <8 x half> %b, <8 x i1> %m, i3
;
; ZVFHMIN-LABEL: vfdiv_vv_v8f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfdiv.vv v10, v12, v10, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -271,10 +271,10 @@ define <8 x half> @vfdiv_vv_v8f16_unmasked(<8 x half> %va, <8 x half> %b, i32 ze
;
; ZVFHMIN-LABEL: vfdiv_vv_v8f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfdiv.vv v10, v12, v10
; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -299,10 +299,10 @@ define <8 x half> @vfdiv_vf_v8f16(<8 x half> %va, half %b, <8 x i1> %m, i32 zero
; ZVFHMIN-NEXT: vfmv.v.f v10, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v9, v10
-; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v9
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v9, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfdiv.vv v10, v10, v12, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -327,10 +327,10 @@ define <8 x half> @vfdiv_vf_v8f16_unmasked(<8 x half> %va, half %b, i32 zeroext
; ZVFHMIN-NEXT: vfmv.v.f v10, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v9, v10
-; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v9
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfdiv.vv v10, v10, v12
; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -354,10 +354,10 @@ define <16 x half> @vfdiv_vv_v16f16(<16 x half> %va, <16 x half> %b, <16 x i1> %
;
; ZVFHMIN-LABEL: vfdiv_vv_v16f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v10
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v10, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfdiv.vv v12, v16, v12, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -375,10 +375,10 @@ define <16 x half> @vfdiv_vv_v16f16_unmasked(<16 x half> %va, <16 x half> %b, i3
;
; ZVFHMIN-LABEL: vfdiv_vv_v16f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v10
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfdiv.vv v12, v16, v12
; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -403,10 +403,10 @@ define <16 x half> @vfdiv_vf_v16f16(<16 x half> %va, half %b, <16 x i1> %m, i32
; ZVFHMIN-NEXT: vfmv.v.f v12, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v10, v12
-; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v10, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfdiv.vv v12, v12, v16, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -431,10 +431,10 @@ define <16 x half> @vfdiv_vf_v16f16_unmasked(<16 x half> %va, half %b, i32 zeroe
; ZVFHMIN-NEXT: vfmv.v.f v12, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v10, v12
-; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfdiv.vv v12, v12, v16
; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
diff --git a/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vfma-vp.ll b/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vfma-vp.ll
index 65776339de07a5..66b365c21e09b8 100644
--- a/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vfma-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vfma-vp.ll
@@ -20,11 +20,11 @@ define <2 x half> @vfma_vv_v2f16(<2 x half> %va, <2 x half> %b, <2 x half> %c, <
;
; ZVFHMIN-LABEL: vfma_vv_v2f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v11, v10
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v9
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v11, v10, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v9, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfmadd.vv v12, v10, v11, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -42,11 +42,11 @@ define <2 x half> @vfma_vv_v2f16_unmasked(<2 x half> %va, <2 x half> %b, <2 x ha
;
; ZVFHMIN-LABEL: vfma_vv_v2f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v11, v10
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v9
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfmadd.vv v12, v10, v11
; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -71,11 +71,11 @@ define <2 x half> @vfma_vf_v2f16(<2 x half> %va, half %b, <2 x half> %vc, <2 x i
; ZVFHMIN-NEXT: vfmv.v.f v10, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v11, v10
-; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v11
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v11, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfmadd.vv v12, v9, v10, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -100,11 +100,11 @@ define <2 x half> @vfma_vf_v2f16_unmasked(<2 x half> %va, half %b, <2 x half> %v
; ZVFHMIN-NEXT: vfmv.v.f v10, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v11, v10
-; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v11
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfmadd.vv v12, v9, v10
; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -129,11 +129,11 @@ define <4 x half> @vfma_vv_v4f16(<4 x half> %va, <4 x half> %b, <4 x half> %c, <
;
; ZVFHMIN-LABEL: vfma_vv_v4f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v11, v10
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v9
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v11, v10, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v9, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfmadd.vv v12, v10, v11, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -151,11 +151,11 @@ define <4 x half> @vfma_vv_v4f16_unmasked(<4 x half> %va, <4 x half> %b, <4 x ha
;
; ZVFHMIN-LABEL: vfma_vv_v4f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v11, v10
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v9
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfmadd.vv v12, v10, v11
; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -180,11 +180,11 @@ define <4 x half> @vfma_vf_v4f16(<4 x half> %va, half %b, <4 x half> %vc, <4 x i
; ZVFHMIN-NEXT: vfmv.v.f v10, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v11, v10
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v11
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v11, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfmadd.vv v12, v9, v10, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -209,11 +209,11 @@ define <4 x half> @vfma_vf_v4f16_unmasked(<4 x half> %va, half %b, <4 x half> %v
; ZVFHMIN-NEXT: vfmv.v.f v10, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v11, v10
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v11
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfmadd.vv v12, v9, v10
; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -238,11 +238,11 @@ define <8 x half> @vfma_vv_v8f16(<8 x half> %va, <8 x half> %b, <8 x half> %c, <
;
; ZVFHMIN-LABEL: vfma_vv_v8f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v10
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v14, v9
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v10, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v14, v9, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfmadd.vv v14, v10, v12, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v14
@@ -260,11 +260,11 @@ define <8 x half> @vfma_vv_v8f16_unmasked(<8 x half> %va, <8 x half> %b, <8 x ha
;
; ZVFHMIN-LABEL: vfma_vv_v8f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v10
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
; ZVFHMIN-NEXT: vfwcvt.f.f.v v14, v9
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfmadd.vv v14, v10, v12
; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v14
@@ -289,11 +289,11 @@ define <8 x half> @vfma_vf_v8f16(<8 x half> %va, half %b, <8 x half> %vc, <8 x i
; ZVFHMIN-NEXT: vfmv.v.f v10, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v10
-; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v14, v8
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v14, v8, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfmadd.vv v16, v14, v10, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
@@ -318,11 +318,11 @@ define <8 x half> @vfma_vf_v8f16_unmasked(<8 x half> %va, half %b, <8 x half> %v
; ZVFHMIN-NEXT: vfmv.v.f v10, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v10
-; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
; ZVFHMIN-NEXT: vfwcvt.f.f.v v14, v8
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfmadd.vv v16, v14, v10
; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
@@ -347,11 +347,11 @@ define <16 x half> @vfma_vv_v16f16(<16 x half> %va, <16 x half> %b, <16 x half>
;
; ZVFHMIN-LABEL: vfma_vv_v16f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v20, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v20, v10, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfmadd.vv v20, v12, v16, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v20
@@ -369,11 +369,11 @@ define <16 x half> @vfma_vv_v16f16_unmasked(<16 x half> %va, <16 x half> %b, <16
;
; ZVFHMIN-LABEL: vfma_vv_v16f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
; ZVFHMIN-NEXT: vfwcvt.f.f.v v20, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfmadd.vv v20, v12, v16
; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v20
@@ -398,11 +398,11 @@ define <16 x half> @vfma_vf_v16f16(<16 x half> %va, half %b, <16 x half> %vc, <1
; ZVFHMIN-NEXT: vfmv.v.f v12, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v16, v12
-; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v10
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v20, v8
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v16
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v10, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v20, v8, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v16, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfmadd.vv v24, v20, v12, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v24
@@ -427,11 +427,11 @@ define <16 x half> @vfma_vf_v16f16_unmasked(<16 x half> %va, half %b, <16 x half
; ZVFHMIN-NEXT: vfmv.v.f v12, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v16, v12
-; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v10
; ZVFHMIN-NEXT: vfwcvt.f.f.v v20, v8
; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v16
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfmadd.vv v24, v20, v12
; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v24
diff --git a/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vfmax-vp.ll b/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vfmax-vp.ll
index ca033c26dba1f7..90aefd96d31af7 100644
--- a/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vfmax-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vfmax-vp.ll
@@ -19,10 +19,10 @@ define <2 x half> @vfmax_vv_v2f16(<2 x half> %va, <2 x half> %vb, <2 x i1> %m, i
;
; ZVFHMIN-LABEL: vfmax_vv_v2f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfmax.vv v9, v9, v10, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -40,10 +40,10 @@ define <2 x half> @vfmax_vv_v2f16_unmasked(<2 x half> %va, <2 x half> %vb, i32 z
;
; ZVFHMIN-LABEL: vfmax_vv_v2f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfmax.vv v9, v9, v10
; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -65,10 +65,10 @@ define <4 x half> @vfmax_vv_v4f16(<4 x half> %va, <4 x half> %vb, <4 x i1> %m, i
;
; ZVFHMIN-LABEL: vfmax_vv_v4f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfmax.vv v9, v9, v10, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -86,10 +86,10 @@ define <4 x half> @vfmax_vv_v4f16_unmasked(<4 x half> %va, <4 x half> %vb, i32 z
;
; ZVFHMIN-LABEL: vfmax_vv_v4f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfmax.vv v9, v9, v10
; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -111,10 +111,10 @@ define <8 x half> @vfmax_vv_v8f16(<8 x half> %va, <8 x half> %vb, <8 x i1> %m, i
;
; ZVFHMIN-LABEL: vfmax_vv_v8f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfmax.vv v10, v12, v10, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -132,10 +132,10 @@ define <8 x half> @vfmax_vv_v8f16_unmasked(<8 x half> %va, <8 x half> %vb, i32 z
;
; ZVFHMIN-LABEL: vfmax_vv_v8f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfmax.vv v10, v12, v10
; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -157,10 +157,10 @@ define <16 x half> @vfmax_vv_v16f16(<16 x half> %va, <16 x half> %vb, <16 x i1>
;
; ZVFHMIN-LABEL: vfmax_vv_v16f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v10
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v10, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfmax.vv v12, v16, v12, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -178,10 +178,10 @@ define <16 x half> @vfmax_vv_v16f16_unmasked(<16 x half> %va, <16 x half> %vb, i
;
; ZVFHMIN-LABEL: vfmax_vv_v16f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v10
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfmax.vv v12, v16, v12
; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
diff --git a/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vfmin-vp.ll b/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vfmin-vp.ll
index 1f1efdaf1ee5ac..3da54997a319ed 100644
--- a/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vfmin-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vfmin-vp.ll
@@ -19,10 +19,10 @@ define <2 x half> @vfmin_vv_v2f16(<2 x half> %va, <2 x half> %vb, <2 x i1> %m, i
;
; ZVFHMIN-LABEL: vfmin_vv_v2f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfmin.vv v9, v9, v10, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -40,10 +40,10 @@ define <2 x half> @vfmin_vv_v2f16_unmasked(<2 x half> %va, <2 x half> %vb, i32 z
;
; ZVFHMIN-LABEL: vfmin_vv_v2f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfmin.vv v9, v9, v10
; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -65,10 +65,10 @@ define <4 x half> @vfmin_vv_v4f16(<4 x half> %va, <4 x half> %vb, <4 x i1> %m, i
;
; ZVFHMIN-LABEL: vfmin_vv_v4f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfmin.vv v9, v9, v10, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -86,10 +86,10 @@ define <4 x half> @vfmin_vv_v4f16_unmasked(<4 x half> %va, <4 x half> %vb, i32 z
;
; ZVFHMIN-LABEL: vfmin_vv_v4f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfmin.vv v9, v9, v10
; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -111,10 +111,10 @@ define <8 x half> @vfmin_vv_v8f16(<8 x half> %va, <8 x half> %vb, <8 x i1> %m, i
;
; ZVFHMIN-LABEL: vfmin_vv_v8f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfmin.vv v10, v12, v10, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -132,10 +132,10 @@ define <8 x half> @vfmin_vv_v8f16_unmasked(<8 x half> %va, <8 x half> %vb, i32 z
;
; ZVFHMIN-LABEL: vfmin_vv_v8f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfmin.vv v10, v12, v10
; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -157,10 +157,10 @@ define <16 x half> @vfmin_vv_v16f16(<16 x half> %va, <16 x half> %vb, <16 x i1>
;
; ZVFHMIN-LABEL: vfmin_vv_v16f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v10
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v10, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfmin.vv v12, v16, v12, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -178,10 +178,10 @@ define <16 x half> @vfmin_vv_v16f16_unmasked(<16 x half> %va, <16 x half> %vb, i
;
; ZVFHMIN-LABEL: vfmin_vv_v16f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v10
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfmin.vv v12, v16, v12
; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
diff --git a/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vfmul-vp.ll b/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vfmul-vp.ll
index 4e06263f4e8b70..91aa048fa5bafc 100644
--- a/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vfmul-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vfmul-vp.ll
@@ -19,10 +19,10 @@ define <2 x half> @vfmul_vv_v2f16(<2 x half> %va, <2 x half> %b, <2 x i1> %m, i3
;
; ZVFHMIN-LABEL: vfmul_vv_v2f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfmul.vv v9, v9, v10, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -40,10 +40,10 @@ define <2 x half> @vfmul_vv_v2f16_unmasked(<2 x half> %va, <2 x half> %b, i32 ze
;
; ZVFHMIN-LABEL: vfmul_vv_v2f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfmul.vv v9, v9, v10
; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -68,10 +68,10 @@ define <2 x half> @vfmul_vf_v2f16(<2 x half> %va, half %b, <2 x i1> %m, i32 zero
; ZVFHMIN-NEXT: vfmv.v.f v9, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v10, v9
-; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfmul.vv v9, v9, v8, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -96,10 +96,10 @@ define <2 x half> @vfmul_vf_v2f16_unmasked(<2 x half> %va, half %b, i32 zeroext
; ZVFHMIN-NEXT: vfmv.v.f v9, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v10, v9
-; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfmul.vv v9, v9, v8
; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -123,10 +123,10 @@ define <3 x half> @vfmul_vv_v3f16(<3 x half> %va, <3 x half> %b, <3 x i1> %m, i3
;
; ZVFHMIN-LABEL: vfmul_vv_v3f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfmul.vv v9, v9, v10, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -146,10 +146,10 @@ define <4 x half> @vfmul_vv_v4f16(<4 x half> %va, <4 x half> %b, <4 x i1> %m, i3
;
; ZVFHMIN-LABEL: vfmul_vv_v4f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfmul.vv v9, v9, v10, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -167,10 +167,10 @@ define <4 x half> @vfmul_vv_v4f16_unmasked(<4 x half> %va, <4 x half> %b, i32 ze
;
; ZVFHMIN-LABEL: vfmul_vv_v4f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfmul.vv v9, v9, v10
; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -195,10 +195,10 @@ define <4 x half> @vfmul_vf_v4f16(<4 x half> %va, half %b, <4 x i1> %m, i32 zero
; ZVFHMIN-NEXT: vfmv.v.f v9, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v10, v9
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfmul.vv v9, v9, v8, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -223,10 +223,10 @@ define <4 x half> @vfmul_vf_v4f16_unmasked(<4 x half> %va, half %b, i32 zeroext
; ZVFHMIN-NEXT: vfmv.v.f v9, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v10, v9
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfmul.vv v9, v9, v8
; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -250,10 +250,10 @@ define <8 x half> @vfmul_vv_v8f16(<8 x half> %va, <8 x half> %b, <8 x i1> %m, i3
;
; ZVFHMIN-LABEL: vfmul_vv_v8f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfmul.vv v10, v12, v10, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -271,10 +271,10 @@ define <8 x half> @vfmul_vv_v8f16_unmasked(<8 x half> %va, <8 x half> %b, i32 ze
;
; ZVFHMIN-LABEL: vfmul_vv_v8f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfmul.vv v10, v12, v10
; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -299,10 +299,10 @@ define <8 x half> @vfmul_vf_v8f16(<8 x half> %va, half %b, <8 x i1> %m, i32 zero
; ZVFHMIN-NEXT: vfmv.v.f v10, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v9, v10
-; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v9
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v9, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfmul.vv v10, v10, v12, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -327,10 +327,10 @@ define <8 x half> @vfmul_vf_v8f16_unmasked(<8 x half> %va, half %b, i32 zeroext
; ZVFHMIN-NEXT: vfmv.v.f v10, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v9, v10
-; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v9
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfmul.vv v10, v10, v12
; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -354,10 +354,10 @@ define <16 x half> @vfmul_vv_v16f16(<16 x half> %va, <16 x half> %b, <16 x i1> %
;
; ZVFHMIN-LABEL: vfmul_vv_v16f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v10
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v10, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfmul.vv v12, v16, v12, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -375,10 +375,10 @@ define <16 x half> @vfmul_vv_v16f16_unmasked(<16 x half> %va, <16 x half> %b, i3
;
; ZVFHMIN-LABEL: vfmul_vv_v16f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v10
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfmul.vv v12, v16, v12
; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -403,10 +403,10 @@ define <16 x half> @vfmul_vf_v16f16(<16 x half> %va, half %b, <16 x i1> %m, i32
; ZVFHMIN-NEXT: vfmv.v.f v12, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v10, v12
-; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v10, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfmul.vv v12, v12, v16, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -431,10 +431,10 @@ define <16 x half> @vfmul_vf_v16f16_unmasked(<16 x half> %va, half %b, i32 zeroe
; ZVFHMIN-NEXT: vfmv.v.f v12, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v10, v12
-; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfmul.vv v12, v12, v16
; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
diff --git a/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vfneg-vp.ll b/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vfneg-vp.ll
index 3d037a5589a1d1..0c230b8b2f5f94 100644
--- a/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vfneg-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vfneg-vp.ll
@@ -19,9 +19,9 @@ define <2 x half> @vfneg_vv_v2f16(<2 x half> %va, <2 x i1> %m, i32 zeroext %evl)
;
; ZVFHMIN-LABEL: vfneg_vv_v2f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfneg.v v9, v9, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -39,9 +39,9 @@ define <2 x half> @vfneg_vv_v2f16_unmasked(<2 x half> %va, i32 zeroext %evl) {
;
; ZVFHMIN-LABEL: vfneg_vv_v2f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfneg.v v9, v9
; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -63,9 +63,9 @@ define <4 x half> @vfneg_vv_v4f16(<4 x half> %va, <4 x i1> %m, i32 zeroext %evl)
;
; ZVFHMIN-LABEL: vfneg_vv_v4f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfneg.v v9, v9, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -83,9 +83,9 @@ define <4 x half> @vfneg_vv_v4f16_unmasked(<4 x half> %va, i32 zeroext %evl) {
;
; ZVFHMIN-LABEL: vfneg_vv_v4f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfneg.v v9, v9
; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -107,9 +107,9 @@ define <8 x half> @vfneg_vv_v8f16(<8 x half> %va, <8 x i1> %m, i32 zeroext %evl)
;
; ZVFHMIN-LABEL: vfneg_vv_v8f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfneg.v v10, v10, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -127,9 +127,9 @@ define <8 x half> @vfneg_vv_v8f16_unmasked(<8 x half> %va, i32 zeroext %evl) {
;
; ZVFHMIN-LABEL: vfneg_vv_v8f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfneg.v v10, v10
; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -151,9 +151,9 @@ define <16 x half> @vfneg_vv_v16f16(<16 x half> %va, <16 x i1> %m, i32 zeroext %
;
; ZVFHMIN-LABEL: vfneg_vv_v16f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfneg.v v12, v12, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -171,9 +171,9 @@ define <16 x half> @vfneg_vv_v16f16_unmasked(<16 x half> %va, i32 zeroext %evl)
;
; ZVFHMIN-LABEL: vfneg_vv_v16f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfneg.v v12, v12
; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
diff --git a/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vfsqrt-vp.ll b/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vfsqrt-vp.ll
index 60022644e5ab68..fbf9e50d844664 100644
--- a/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vfsqrt-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vfsqrt-vp.ll
@@ -19,9 +19,9 @@ define <2 x half> @vfsqrt_vv_v2f16(<2 x half> %va, <2 x i1> %m, i32 zeroext %evl
;
; ZVFHMIN-LABEL: vfsqrt_vv_v2f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfsqrt.v v9, v9, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -39,9 +39,9 @@ define <2 x half> @vfsqrt_vv_v2f16_unmasked(<2 x half> %va, i32 zeroext %evl) {
;
; ZVFHMIN-LABEL: vfsqrt_vv_v2f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfsqrt.v v9, v9
; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -63,9 +63,9 @@ define <4 x half> @vfsqrt_vv_v4f16(<4 x half> %va, <4 x i1> %m, i32 zeroext %evl
;
; ZVFHMIN-LABEL: vfsqrt_vv_v4f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfsqrt.v v9, v9, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -83,9 +83,9 @@ define <4 x half> @vfsqrt_vv_v4f16_unmasked(<4 x half> %va, i32 zeroext %evl) {
;
; ZVFHMIN-LABEL: vfsqrt_vv_v4f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfsqrt.v v9, v9
; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -107,9 +107,9 @@ define <8 x half> @vfsqrt_vv_v8f16(<8 x half> %va, <8 x i1> %m, i32 zeroext %evl
;
; ZVFHMIN-LABEL: vfsqrt_vv_v8f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfsqrt.v v10, v10, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -127,9 +127,9 @@ define <8 x half> @vfsqrt_vv_v8f16_unmasked(<8 x half> %va, i32 zeroext %evl) {
;
; ZVFHMIN-LABEL: vfsqrt_vv_v8f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfsqrt.v v10, v10
; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -151,9 +151,9 @@ define <16 x half> @vfsqrt_vv_v16f16(<16 x half> %va, <16 x i1> %m, i32 zeroext
;
; ZVFHMIN-LABEL: vfsqrt_vv_v16f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfsqrt.v v12, v12, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -171,9 +171,9 @@ define <16 x half> @vfsqrt_vv_v16f16_unmasked(<16 x half> %va, i32 zeroext %evl)
;
; ZVFHMIN-LABEL: vfsqrt_vv_v16f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfsqrt.v v12, v12
; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
diff --git a/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vfsub-vp.ll b/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vfsub-vp.ll
index 76ca7d971eb49a..09d24e9cff655a 100644
--- a/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vfsub-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vfsub-vp.ll
@@ -19,10 +19,10 @@ define <2 x half> @vfsub_vv_v2f16(<2 x half> %va, <2 x half> %b, <2 x i1> %m, i3
;
; ZVFHMIN-LABEL: vfsub_vv_v2f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfsub.vv v9, v9, v10, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -40,10 +40,10 @@ define <2 x half> @vfsub_vv_v2f16_unmasked(<2 x half> %va, <2 x half> %b, i32 ze
;
; ZVFHMIN-LABEL: vfsub_vv_v2f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfsub.vv v9, v9, v10
; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -68,10 +68,10 @@ define <2 x half> @vfsub_vf_v2f16(<2 x half> %va, half %b, <2 x i1> %m, i32 zero
; ZVFHMIN-NEXT: vfmv.v.f v9, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v10, v9
-; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfsub.vv v9, v9, v8, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -96,10 +96,10 @@ define <2 x half> @vfsub_vf_v2f16_unmasked(<2 x half> %va, half %b, i32 zeroext
; ZVFHMIN-NEXT: vfmv.v.f v9, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v10, v9
-; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfsub.vv v9, v9, v8
; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -123,10 +123,10 @@ define <3 x half> @vfsub_vv_v3f16(<3 x half> %va, <3 x half> %b, <3 x i1> %m, i3
;
; ZVFHMIN-LABEL: vfsub_vv_v3f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfsub.vv v9, v9, v10, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -146,10 +146,10 @@ define <4 x half> @vfsub_vv_v4f16(<4 x half> %va, <4 x half> %b, <4 x i1> %m, i3
;
; ZVFHMIN-LABEL: vfsub_vv_v4f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfsub.vv v9, v9, v10, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -167,10 +167,10 @@ define <4 x half> @vfsub_vv_v4f16_unmasked(<4 x half> %va, <4 x half> %b, i32 ze
;
; ZVFHMIN-LABEL: vfsub_vv_v4f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfsub.vv v9, v9, v10
; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -195,10 +195,10 @@ define <4 x half> @vfsub_vf_v4f16(<4 x half> %va, half %b, <4 x i1> %m, i32 zero
; ZVFHMIN-NEXT: vfmv.v.f v9, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v10, v9
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfsub.vv v9, v9, v8, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -223,10 +223,10 @@ define <4 x half> @vfsub_vf_v4f16_unmasked(<4 x half> %va, half %b, i32 zeroext
; ZVFHMIN-NEXT: vfmv.v.f v9, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v10, v9
-; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfsub.vv v9, v9, v8
; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -250,10 +250,10 @@ define <8 x half> @vfsub_vv_v8f16(<8 x half> %va, <8 x half> %b, <8 x i1> %m, i3
;
; ZVFHMIN-LABEL: vfsub_vv_v8f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfsub.vv v10, v12, v10, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -271,10 +271,10 @@ define <8 x half> @vfsub_vv_v8f16_unmasked(<8 x half> %va, <8 x half> %b, i32 ze
;
; ZVFHMIN-LABEL: vfsub_vv_v8f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfsub.vv v10, v12, v10
; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -299,10 +299,10 @@ define <8 x half> @vfsub_vf_v8f16(<8 x half> %va, half %b, <8 x i1> %m, i32 zero
; ZVFHMIN-NEXT: vfmv.v.f v10, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v9, v10
-; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v9
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v9, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfsub.vv v10, v10, v12, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -327,10 +327,10 @@ define <8 x half> @vfsub_vf_v8f16_unmasked(<8 x half> %va, half %b, i32 zeroext
; ZVFHMIN-NEXT: vfmv.v.f v10, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v9, v10
-; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v9
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfsub.vv v10, v10, v12
; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -354,10 +354,10 @@ define <16 x half> @vfsub_vv_v16f16(<16 x half> %va, <16 x half> %b, <16 x i1> %
;
; ZVFHMIN-LABEL: vfsub_vv_v16f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v10
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v10, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfsub.vv v12, v16, v12, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -375,10 +375,10 @@ define <16 x half> @vfsub_vv_v16f16_unmasked(<16 x half> %va, <16 x half> %b, i3
;
; ZVFHMIN-LABEL: vfsub_vv_v16f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v10
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfsub.vv v12, v16, v12
; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -403,10 +403,10 @@ define <16 x half> @vfsub_vf_v16f16(<16 x half> %va, half %b, <16 x i1> %m, i32
; ZVFHMIN-NEXT: vfmv.v.f v12, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v10, v12
-; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v10, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfsub.vv v12, v12, v16, v0.t
; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -431,10 +431,10 @@ define <16 x half> @vfsub_vf_v16f16_unmasked(<16 x half> %va, half %b, i32 zeroe
; ZVFHMIN-NEXT: vfmv.v.f v12, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v10, v12
-; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfsub.vv v12, v12, v16
; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
diff --git a/llvm/test/CodeGen/RISCV/rvv/fmaximum-vp.ll b/llvm/test/CodeGen/RISCV/rvv/fmaximum-vp.ll
index f4350a1e1ced31..ec2de0c69a8890 100644
--- a/llvm/test/CodeGen/RISCV/rvv/fmaximum-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/fmaximum-vp.ll
@@ -27,19 +27,17 @@ define <vscale x 1 x half> @vfmax_vv_nxv1f16(<vscale x 1 x half> %va, <vscale x
; ZVFHMIN-LABEL: vfmax_vv_nxv1f16:
; ZVFHMIN: # %bb.0:
; ZVFHMIN-NEXT: vmv1r.v v10, v0
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v11, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v0, v11, v11, v0.t
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v9
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
-; ZVFHMIN-NEXT: vmerge.vvm v9, v11, v8, v0
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v11, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vmfeq.vv v0, v9, v9, v0.t
+; ZVFHMIN-NEXT: vmerge.vvm v8, v9, v11, v0
; ZVFHMIN-NEXT: vmv1r.v v0, v10
-; ZVFHMIN-NEXT: vmfeq.vv v0, v8, v8, v0.t
-; ZVFHMIN-NEXT: vmerge.vvm v8, v8, v11, v0
+; ZVFHMIN-NEXT: vmfeq.vv v0, v11, v11, v0.t
+; ZVFHMIN-NEXT: vmerge.vvm v9, v11, v9, v0
; ZVFHMIN-NEXT: vmv1r.v v0, v10
-; ZVFHMIN-NEXT: vfmax.vv v9, v8, v9, v0.t
+; ZVFHMIN-NEXT: vfmax.vv v9, v9, v8, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
; ZVFHMIN-NEXT: ret
@@ -61,18 +59,16 @@ define <vscale x 1 x half> @vfmax_vv_nxv1f16_unmasked(<vscale x 1 x half> %va, <
;
; ZVFHMIN-LABEL: vfmax_vv_nxv1f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v0, v10, v10
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v11, v9
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v8, v11, v11
-; ZVFHMIN-NEXT: vmerge.vvm v9, v10, v11, v0
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vmfeq.vv v0, v9, v9
+; ZVFHMIN-NEXT: vmfeq.vv v8, v10, v10
+; ZVFHMIN-NEXT: vmerge.vvm v11, v9, v10, v0
; ZVFHMIN-NEXT: vmv1r.v v0, v8
-; ZVFHMIN-NEXT: vmerge.vvm v8, v11, v10, v0
-; ZVFHMIN-NEXT: vfmax.vv v9, v8, v9
+; ZVFHMIN-NEXT: vmerge.vvm v8, v10, v9, v0
+; ZVFHMIN-NEXT: vfmax.vv v9, v8, v11
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
; ZVFHMIN-NEXT: ret
@@ -101,19 +97,17 @@ define <vscale x 2 x half> @vfmax_vv_nxv2f16(<vscale x 2 x half> %va, <vscale x
; ZVFHMIN-LABEL: vfmax_vv_nxv2f16:
; ZVFHMIN: # %bb.0:
; ZVFHMIN-NEXT: vmv1r.v v10, v0
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v11, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v0, v11, v11, v0.t
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v9
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
-; ZVFHMIN-NEXT: vmerge.vvm v9, v11, v8, v0
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v11, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vmfeq.vv v0, v9, v9, v0.t
+; ZVFHMIN-NEXT: vmerge.vvm v8, v9, v11, v0
; ZVFHMIN-NEXT: vmv1r.v v0, v10
-; ZVFHMIN-NEXT: vmfeq.vv v0, v8, v8, v0.t
-; ZVFHMIN-NEXT: vmerge.vvm v8, v8, v11, v0
+; ZVFHMIN-NEXT: vmfeq.vv v0, v11, v11, v0.t
+; ZVFHMIN-NEXT: vmerge.vvm v9, v11, v9, v0
; ZVFHMIN-NEXT: vmv1r.v v0, v10
-; ZVFHMIN-NEXT: vfmax.vv v9, v8, v9, v0.t
+; ZVFHMIN-NEXT: vfmax.vv v9, v9, v8, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
; ZVFHMIN-NEXT: ret
@@ -135,18 +129,16 @@ define <vscale x 2 x half> @vfmax_vv_nxv2f16_unmasked(<vscale x 2 x half> %va, <
;
; ZVFHMIN-LABEL: vfmax_vv_nxv2f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v0, v10, v10
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v11, v9
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v8, v11, v11
-; ZVFHMIN-NEXT: vmerge.vvm v9, v10, v11, v0
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vmfeq.vv v0, v9, v9
+; ZVFHMIN-NEXT: vmfeq.vv v8, v10, v10
+; ZVFHMIN-NEXT: vmerge.vvm v11, v9, v10, v0
; ZVFHMIN-NEXT: vmv.v.v v0, v8
-; ZVFHMIN-NEXT: vmerge.vvm v8, v11, v10, v0
-; ZVFHMIN-NEXT: vfmax.vv v9, v8, v9
+; ZVFHMIN-NEXT: vmerge.vvm v8, v10, v9, v0
+; ZVFHMIN-NEXT: vfmax.vv v9, v8, v11
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
; ZVFHMIN-NEXT: ret
@@ -175,19 +167,17 @@ define <vscale x 4 x half> @vfmax_vv_nxv4f16(<vscale x 4 x half> %va, <vscale x
; ZVFHMIN-LABEL: vfmax_vv_nxv4f16:
; ZVFHMIN: # %bb.0:
; ZVFHMIN-NEXT: vmv1r.v v10, v0
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v8, v12, v12, v0.t
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v14, v9
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v14, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vmfeq.vv v8, v14, v14, v0.t
; ZVFHMIN-NEXT: vmv1r.v v0, v8
-; ZVFHMIN-NEXT: vmerge.vvm v16, v12, v14, v0
+; ZVFHMIN-NEXT: vmerge.vvm v16, v14, v12, v0
; ZVFHMIN-NEXT: vmv1r.v v0, v10
-; ZVFHMIN-NEXT: vmfeq.vv v8, v14, v14, v0.t
+; ZVFHMIN-NEXT: vmfeq.vv v8, v12, v12, v0.t
; ZVFHMIN-NEXT: vmv1r.v v0, v8
-; ZVFHMIN-NEXT: vmerge.vvm v8, v14, v12, v0
+; ZVFHMIN-NEXT: vmerge.vvm v8, v12, v14, v0
; ZVFHMIN-NEXT: vmv1r.v v0, v10
; ZVFHMIN-NEXT: vfmax.vv v10, v8, v16, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m1, ta, ma
@@ -211,17 +201,15 @@ define <vscale x 4 x half> @vfmax_vv_nxv4f16_unmasked(<vscale x 4 x half> %va, <
;
; ZVFHMIN-LABEL: vfmax_vv_nxv4f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v0, v10, v10
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v9
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v8, v12, v12
-; ZVFHMIN-NEXT: vmerge.vvm v14, v10, v12, v0
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vmfeq.vv v0, v12, v12
+; ZVFHMIN-NEXT: vmfeq.vv v8, v10, v10
+; ZVFHMIN-NEXT: vmerge.vvm v14, v12, v10, v0
; ZVFHMIN-NEXT: vmv1r.v v0, v8
-; ZVFHMIN-NEXT: vmerge.vvm v8, v12, v10, v0
+; ZVFHMIN-NEXT: vmerge.vvm v8, v10, v12, v0
; ZVFHMIN-NEXT: vfmax.vv v10, v8, v14
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -253,19 +241,17 @@ define <vscale x 8 x half> @vfmax_vv_nxv8f16(<vscale x 8 x half> %va, <vscale x
; ZVFHMIN-LABEL: vfmax_vv_nxv8f16:
; ZVFHMIN: # %bb.0:
; ZVFHMIN-NEXT: vmv1r.v v12, v0
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v8, v16, v16, v0.t
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v20, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v10, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v20, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vmfeq.vv v8, v20, v20, v0.t
; ZVFHMIN-NEXT: vmv1r.v v0, v8
-; ZVFHMIN-NEXT: vmerge.vvm v24, v16, v20, v0
+; ZVFHMIN-NEXT: vmerge.vvm v24, v20, v16, v0
; ZVFHMIN-NEXT: vmv1r.v v0, v12
-; ZVFHMIN-NEXT: vmfeq.vv v8, v20, v20, v0.t
+; ZVFHMIN-NEXT: vmfeq.vv v8, v16, v16, v0.t
; ZVFHMIN-NEXT: vmv1r.v v0, v8
-; ZVFHMIN-NEXT: vmerge.vvm v8, v20, v16, v0
+; ZVFHMIN-NEXT: vmerge.vvm v8, v16, v20, v0
; ZVFHMIN-NEXT: vmv1r.v v0, v12
; ZVFHMIN-NEXT: vfmax.vv v12, v8, v24, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m2, ta, ma
@@ -289,17 +275,15 @@ define <vscale x 8 x half> @vfmax_vv_nxv8f16_unmasked(<vscale x 8 x half> %va, <
;
; ZVFHMIN-LABEL: vfmax_vv_nxv8f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v0, v12, v12
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v8, v16, v16
-; ZVFHMIN-NEXT: vmerge.vvm v20, v12, v16, v0
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v10
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vmfeq.vv v0, v16, v16
+; ZVFHMIN-NEXT: vmfeq.vv v8, v12, v12
+; ZVFHMIN-NEXT: vmerge.vvm v20, v16, v12, v0
; ZVFHMIN-NEXT: vmv1r.v v0, v8
-; ZVFHMIN-NEXT: vmerge.vvm v8, v16, v12, v0
+; ZVFHMIN-NEXT: vmerge.vvm v8, v12, v16, v0
; ZVFHMIN-NEXT: vfmax.vv v12, v8, v20
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -337,13 +321,11 @@ define <vscale x 16 x half> @vfmax_vv_nxv16f16(<vscale x 16 x half> %va, <vscale
; ZVFHMIN-NEXT: sub sp, sp, a1
; ZVFHMIN-NEXT: .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x08, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 8 * vlenb
; ZVFHMIN-NEXT: vmv1r.v v1, v0
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vmfeq.vv v8, v24, v24, v0.t
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
; ZVFHMIN-NEXT: vmv1r.v v0, v8
; ZVFHMIN-NEXT: vmerge.vvm v8, v24, v16, v0
; ZVFHMIN-NEXT: addi a0, sp, 16
@@ -386,19 +368,17 @@ define <vscale x 16 x half> @vfmax_vv_nxv16f16_unmasked(<vscale x 16 x half> %va
; ZVFHMIN-NEXT: slli a1, a1, 3
; ZVFHMIN-NEXT: sub sp, sp, a1
; ZVFHMIN-NEXT: .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x08, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 8 * vlenb
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v0, v16, v16
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v12
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v1, v24, v24
-; ZVFHMIN-NEXT: vmerge.vvm v8, v16, v24, v0
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vmfeq.vv v0, v24, v24
+; ZVFHMIN-NEXT: vmfeq.vv v1, v16, v16
+; ZVFHMIN-NEXT: vmerge.vvm v8, v24, v16, v0
; ZVFHMIN-NEXT: addi a0, sp, 16
; ZVFHMIN-NEXT: vs8r.v v8, (a0) # Unknown-size Folded Spill
; ZVFHMIN-NEXT: vmv1r.v v0, v1
-; ZVFHMIN-NEXT: vmerge.vvm v8, v24, v16, v0
+; ZVFHMIN-NEXT: vmerge.vvm v8, v16, v24, v0
; ZVFHMIN-NEXT: vl8r.v v16, (a0) # Unknown-size Folded Reload
; ZVFHMIN-NEXT: vfmax.vv v16, v8, v16
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
@@ -450,12 +430,12 @@ define <vscale x 32 x half> @vfmax_vv_nxv32f16(<vscale x 32 x half> %va, <vscale
; ZVFHMIN-NEXT: addi sp, sp, -16
; ZVFHMIN-NEXT: .cfi_def_cfa_offset 16
; ZVFHMIN-NEXT: csrr a1, vlenb
-; ZVFHMIN-NEXT: li a2, 34
-; ZVFHMIN-NEXT: mul a1, a1, a2
+; ZVFHMIN-NEXT: slli a1, a1, 5
; ZVFHMIN-NEXT: sub sp, sp, a1
-; ZVFHMIN-NEXT: .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x22, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 34 * vlenb
+; ZVFHMIN-NEXT: .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x20, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 32 * vlenb
+; ZVFHMIN-NEXT: vmv1r.v v1, v0
; ZVFHMIN-NEXT: csrr a1, vlenb
-; ZVFHMIN-NEXT: li a2, 25
+; ZVFHMIN-NEXT: li a2, 24
; ZVFHMIN-NEXT: mul a1, a1, a2
; ZVFHMIN-NEXT: add a1, sp, a1
; ZVFHMIN-NEXT: addi a1, a1, 16
@@ -469,136 +449,117 @@ define <vscale x 32 x half> @vfmax_vv_nxv32f16(<vscale x 32 x half> %va, <vscale
; ZVFHMIN-NEXT: and a3, a4, a3
; ZVFHMIN-NEXT: srli a2, a2, 2
; ZVFHMIN-NEXT: vsetvli a4, zero, e8, mf2, ta, ma
-; ZVFHMIN-NEXT: csrr a4, vlenb
-; ZVFHMIN-NEXT: li a5, 24
-; ZVFHMIN-NEXT: mul a4, a4, a5
-; ZVFHMIN-NEXT: add a4, sp, a4
-; ZVFHMIN-NEXT: addi a4, a4, 16
-; ZVFHMIN-NEXT: vs1r.v v0, (a4) # Unknown-size Folded Spill
-; ZVFHMIN-NEXT: vslidedown.vx v0, v0, a2
-; ZVFHMIN-NEXT: addi a2, sp, 16
-; ZVFHMIN-NEXT: vs1r.v v0, (a2) # Unknown-size Folded Spill
-; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vslidedown.vx v25, v0, a2
+; ZVFHMIN-NEXT: vsetvli zero, a3, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vmv1r.v v0, v25
+; ZVFHMIN-NEXT: csrr a2, vlenb
+; ZVFHMIN-NEXT: slli a2, a2, 3
+; ZVFHMIN-NEXT: add a2, sp, a2
+; ZVFHMIN-NEXT: addi a2, a2, 16
+; ZVFHMIN-NEXT: vs8r.v v8, (a2) # Unknown-size Folded Spill
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v20, v0.t
; ZVFHMIN-NEXT: csrr a2, vlenb
; ZVFHMIN-NEXT: slli a2, a2, 4
; ZVFHMIN-NEXT: add a2, sp, a2
; ZVFHMIN-NEXT: addi a2, a2, 16
; ZVFHMIN-NEXT: vs8r.v v8, (a2) # Unknown-size Folded Spill
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v20
-; ZVFHMIN-NEXT: vsetvli zero, a3, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v12, v24, v24, v0.t
-; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vmfeq.vv v24, v8, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: csrr a2, vlenb
-; ZVFHMIN-NEXT: li a4, 25
-; ZVFHMIN-NEXT: mul a2, a2, a4
+; ZVFHMIN-NEXT: li a3, 24
+; ZVFHMIN-NEXT: mul a2, a2, a3
; ZVFHMIN-NEXT: add a2, sp, a2
; ZVFHMIN-NEXT: addi a2, a2, 16
-; ZVFHMIN-NEXT: vl8r.v v0, (a2) # Unknown-size Folded Reload
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v4
-; ZVFHMIN-NEXT: vsetvli zero, a3, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vmv1r.v v0, v12
-; ZVFHMIN-NEXT: vmerge.vvm v8, v24, v16, v0
+; ZVFHMIN-NEXT: vl8r.v v8, (a2) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vmv1r.v v0, v24
; ZVFHMIN-NEXT: csrr a2, vlenb
-; ZVFHMIN-NEXT: slli a2, a2, 3
+; ZVFHMIN-NEXT: slli a2, a2, 4
; ZVFHMIN-NEXT: add a2, sp, a2
; ZVFHMIN-NEXT: addi a2, a2, 16
-; ZVFHMIN-NEXT: vs8r.v v8, (a2) # Unknown-size Folded Spill
+; ZVFHMIN-NEXT: vl8r.v v8, (a2) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vmerge.vvm v8, v8, v16, v0
; ZVFHMIN-NEXT: addi a2, sp, 16
-; ZVFHMIN-NEXT: vl1r.v v8, (a2) # Unknown-size Folded Reload
-; ZVFHMIN-NEXT: vmv1r.v v0, v8
+; ZVFHMIN-NEXT: vs8r.v v8, (a2) # Unknown-size Folded Spill
+; ZVFHMIN-NEXT: vmv1r.v v0, v25
; ZVFHMIN-NEXT: vmfeq.vv v12, v16, v16, v0.t
; ZVFHMIN-NEXT: vmv1r.v v0, v12
-; ZVFHMIN-NEXT: vmerge.vvm v16, v16, v24, v0
-; ZVFHMIN-NEXT: vmv1r.v v0, v8
; ZVFHMIN-NEXT: csrr a2, vlenb
-; ZVFHMIN-NEXT: slli a2, a2, 3
+; ZVFHMIN-NEXT: slli a2, a2, 4
; ZVFHMIN-NEXT: add a2, sp, a2
; ZVFHMIN-NEXT: addi a2, a2, 16
-; ZVFHMIN-NEXT: vl8r.v v24, (a2) # Unknown-size Folded Reload
-; ZVFHMIN-NEXT: vfmax.vv v16, v16, v24, v0.t
+; ZVFHMIN-NEXT: vl8r.v v8, (a2) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vmerge.vvm v16, v16, v8, v0
+; ZVFHMIN-NEXT: vmv1r.v v0, v25
+; ZVFHMIN-NEXT: addi a2, sp, 16
+; ZVFHMIN-NEXT: vl8r.v v8, (a2) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfmax.vv v16, v16, v8, v0.t
; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v16
-; ZVFHMIN-NEXT: csrr a2, vlenb
-; ZVFHMIN-NEXT: slli a2, a2, 3
-; ZVFHMIN-NEXT: add a2, sp, a2
-; ZVFHMIN-NEXT: addi a2, a2, 16
-; ZVFHMIN-NEXT: vs8r.v v8, (a2) # Unknown-size Folded Spill
; ZVFHMIN-NEXT: bltu a0, a1, .LBB10_2
; ZVFHMIN-NEXT: # %bb.1:
; ZVFHMIN-NEXT: mv a0, a1
; ZVFHMIN-NEXT: .LBB10_2:
-; ZVFHMIN-NEXT: csrr a1, vlenb
-; ZVFHMIN-NEXT: slli a1, a1, 4
-; ZVFHMIN-NEXT: add a1, sp, a1
-; ZVFHMIN-NEXT: addi a1, a1, 16
-; ZVFHMIN-NEXT: vl8r.v v16, (a1) # Unknown-size Folded Reload
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v16
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
-; ZVFHMIN-NEXT: csrr a1, vlenb
-; ZVFHMIN-NEXT: li a2, 24
-; ZVFHMIN-NEXT: mul a1, a1, a2
-; ZVFHMIN-NEXT: add a1, sp, a1
-; ZVFHMIN-NEXT: addi a1, a1, 16
-; ZVFHMIN-NEXT: vl1r.v v0, (a1) # Unknown-size Folded Reload
-; ZVFHMIN-NEXT: vmfeq.vv v8, v24, v24, v0.t
-; ZVFHMIN-NEXT: csrr a1, vlenb
-; ZVFHMIN-NEXT: slli a1, a1, 4
-; ZVFHMIN-NEXT: add a1, sp, a1
-; ZVFHMIN-NEXT: addi a1, a1, 16
-; ZVFHMIN-NEXT: vs1r.v v8, (a1) # Unknown-size Folded Spill
-; ZVFHMIN-NEXT: addi a1, sp, 16
-; ZVFHMIN-NEXT: vs8r.v v24, (a1) # Unknown-size Folded Spill
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: csrr a1, vlenb
-; ZVFHMIN-NEXT: li a2, 25
-; ZVFHMIN-NEXT: mul a1, a1, a2
-; ZVFHMIN-NEXT: add a1, sp, a1
-; ZVFHMIN-NEXT: addi a1, a1, 16
-; ZVFHMIN-NEXT: vl8r.v v0, (a1) # Unknown-size Folded Reload
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v0
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vmv1r.v v0, v1
; ZVFHMIN-NEXT: csrr a0, vlenb
-; ZVFHMIN-NEXT: slli a0, a0, 4
+; ZVFHMIN-NEXT: slli a0, a0, 3
; ZVFHMIN-NEXT: add a0, sp, a0
; ZVFHMIN-NEXT: addi a0, a0, 16
-; ZVFHMIN-NEXT: vl1r.v v0, (a0) # Unknown-size Folded Reload
-; ZVFHMIN-NEXT: vmerge.vvm v24, v24, v16, v0
+; ZVFHMIN-NEXT: vl8r.v v16, (a0) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v16, v0.t
; ZVFHMIN-NEXT: csrr a0, vlenb
-; ZVFHMIN-NEXT: li a1, 25
-; ZVFHMIN-NEXT: mul a0, a0, a1
+; ZVFHMIN-NEXT: slli a0, a0, 4
; ZVFHMIN-NEXT: add a0, sp, a0
; ZVFHMIN-NEXT: addi a0, a0, 16
; ZVFHMIN-NEXT: vs8r.v v24, (a0) # Unknown-size Folded Spill
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vmfeq.vv v8, v24, v24, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: csrr a0, vlenb
; ZVFHMIN-NEXT: li a1, 24
; ZVFHMIN-NEXT: mul a0, a0, a1
; ZVFHMIN-NEXT: add a0, sp, a0
; ZVFHMIN-NEXT: addi a0, a0, 16
-; ZVFHMIN-NEXT: vl1r.v v9, (a0) # Unknown-size Folded Reload
-; ZVFHMIN-NEXT: vmv1r.v v0, v9
-; ZVFHMIN-NEXT: vmfeq.vv v8, v16, v16, v0.t
+; ZVFHMIN-NEXT: vl8r.v v16, (a0) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v16, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vmv1r.v v0, v8
-; ZVFHMIN-NEXT: addi a0, sp, 16
-; ZVFHMIN-NEXT: vl8r.v v24, (a0) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: csrr a0, vlenb
+; ZVFHMIN-NEXT: slli a0, a0, 4
+; ZVFHMIN-NEXT: add a0, sp, a0
+; ZVFHMIN-NEXT: addi a0, a0, 16
+; ZVFHMIN-NEXT: vl8r.v v16, (a0) # Unknown-size Folded Reload
; ZVFHMIN-NEXT: vmerge.vvm v16, v16, v24, v0
-; ZVFHMIN-NEXT: vmv1r.v v0, v9
; ZVFHMIN-NEXT: csrr a0, vlenb
-; ZVFHMIN-NEXT: li a1, 25
+; ZVFHMIN-NEXT: li a1, 24
; ZVFHMIN-NEXT: mul a0, a0, a1
; ZVFHMIN-NEXT: add a0, sp, a0
; ZVFHMIN-NEXT: addi a0, a0, 16
-; ZVFHMIN-NEXT: vl8r.v v24, (a0) # Unknown-size Folded Reload
-; ZVFHMIN-NEXT: vfmax.vv v16, v16, v24, v0.t
-; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vs8r.v v16, (a0) # Unknown-size Folded Spill
+; ZVFHMIN-NEXT: vmv1r.v v0, v1
+; ZVFHMIN-NEXT: vmfeq.vv v8, v24, v24, v0.t
+; ZVFHMIN-NEXT: vmv1r.v v0, v8
; ZVFHMIN-NEXT: csrr a0, vlenb
-; ZVFHMIN-NEXT: slli a0, a0, 3
+; ZVFHMIN-NEXT: slli a0, a0, 4
; ZVFHMIN-NEXT: add a0, sp, a0
; ZVFHMIN-NEXT: addi a0, a0, 16
-; ZVFHMIN-NEXT: vl8r.v v8, (a0) # Unknown-size Folded Reload
-; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
+; ZVFHMIN-NEXT: vl8r.v v16, (a0) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vmerge.vvm v16, v24, v16, v0
+; ZVFHMIN-NEXT: vmv1r.v v0, v1
; ZVFHMIN-NEXT: csrr a0, vlenb
-; ZVFHMIN-NEXT: li a1, 34
+; ZVFHMIN-NEXT: li a1, 24
; ZVFHMIN-NEXT: mul a0, a0, a1
+; ZVFHMIN-NEXT: add a0, sp, a0
+; ZVFHMIN-NEXT: addi a0, a0, 16
+; ZVFHMIN-NEXT: vl8r.v v24, (a0) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfmax.vv v16, v16, v24, v0.t
+; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
+; ZVFHMIN-NEXT: csrr a0, vlenb
+; ZVFHMIN-NEXT: slli a0, a0, 5
; ZVFHMIN-NEXT: add sp, sp, a0
; ZVFHMIN-NEXT: addi sp, sp, 16
; ZVFHMIN-NEXT: ret
@@ -626,8 +587,14 @@ define <vscale x 32 x half> @vfmax_vv_nxv32f16_unmasked(<vscale x 32 x half> %va
; ZVFHMIN-NEXT: slli a1, a1, 5
; ZVFHMIN-NEXT: sub sp, sp, a1
; ZVFHMIN-NEXT: .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x20, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 32 * vlenb
+; ZVFHMIN-NEXT: csrr a1, vlenb
+; ZVFHMIN-NEXT: li a2, 24
+; ZVFHMIN-NEXT: mul a1, a1, a2
+; ZVFHMIN-NEXT: add a1, sp, a1
+; ZVFHMIN-NEXT: addi a1, a1, 16
+; ZVFHMIN-NEXT: vs8r.v v16, (a1) # Unknown-size Folded Spill
; ZVFHMIN-NEXT: vsetvli a1, zero, e8, m4, ta, ma
-; ZVFHMIN-NEXT: vmset.m v24
+; ZVFHMIN-NEXT: vmset.m v16
; ZVFHMIN-NEXT: csrr a2, vlenb
; ZVFHMIN-NEXT: slli a1, a2, 1
; ZVFHMIN-NEXT: sub a3, a0, a1
@@ -636,48 +603,52 @@ define <vscale x 32 x half> @vfmax_vv_nxv32f16_unmasked(<vscale x 32 x half> %va
; ZVFHMIN-NEXT: and a3, a4, a3
; ZVFHMIN-NEXT: srli a2, a2, 2
; ZVFHMIN-NEXT: vsetvli a4, zero, e8, mf2, ta, ma
-; ZVFHMIN-NEXT: vslidedown.vx v1, v24, a2
-; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vslidedown.vx v24, v16, a2
+; ZVFHMIN-NEXT: vsetvli zero, a3, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vmv1r.v v0, v24
; ZVFHMIN-NEXT: csrr a2, vlenb
-; ZVFHMIN-NEXT: slli a2, a2, 4
+; ZVFHMIN-NEXT: slli a2, a2, 3
; ZVFHMIN-NEXT: add a2, sp, a2
; ZVFHMIN-NEXT: addi a2, a2, 16
; ZVFHMIN-NEXT: vs8r.v v8, (a2) # Unknown-size Folded Spill
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v12
-; ZVFHMIN-NEXT: addi a2, sp, 16
-; ZVFHMIN-NEXT: vs8r.v v24, (a2) # Unknown-size Folded Spill
-; ZVFHMIN-NEXT: vsetvli zero, a3, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vmv1r.v v0, v1
-; ZVFHMIN-NEXT: vmfeq.vv v12, v24, v24, v0.t
-; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vmv4r.v v8, v16
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12, v0.t
; ZVFHMIN-NEXT: csrr a2, vlenb
-; ZVFHMIN-NEXT: li a4, 24
-; ZVFHMIN-NEXT: mul a2, a2, a4
+; ZVFHMIN-NEXT: slli a2, a2, 4
; ZVFHMIN-NEXT: add a2, sp, a2
; ZVFHMIN-NEXT: addi a2, a2, 16
-; ZVFHMIN-NEXT: vs8r.v v8, (a2) # Unknown-size Folded Spill
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v20
-; ZVFHMIN-NEXT: vsetvli zero, a3, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vmv1r.v v0, v12
-; ZVFHMIN-NEXT: addi a2, sp, 16
-; ZVFHMIN-NEXT: vl8r.v v16, (a2) # Unknown-size Folded Reload
-; ZVFHMIN-NEXT: vmerge.vvm v8, v16, v24, v0
+; ZVFHMIN-NEXT: vs8r.v v16, (a2) # Unknown-size Folded Spill
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vmfeq.vv v25, v16, v16, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: csrr a2, vlenb
-; ZVFHMIN-NEXT: slli a2, a2, 3
+; ZVFHMIN-NEXT: li a3, 24
+; ZVFHMIN-NEXT: mul a2, a2, a3
; ZVFHMIN-NEXT: add a2, sp, a2
; ZVFHMIN-NEXT: addi a2, a2, 16
+; ZVFHMIN-NEXT: vl8r.v v8, (a2) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vmv1r.v v0, v25
+; ZVFHMIN-NEXT: csrr a2, vlenb
+; ZVFHMIN-NEXT: slli a2, a2, 4
+; ZVFHMIN-NEXT: add a2, sp, a2
+; ZVFHMIN-NEXT: addi a2, a2, 16
+; ZVFHMIN-NEXT: vl8r.v v8, (a2) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vmerge.vvm v8, v8, v16, v0
+; ZVFHMIN-NEXT: addi a2, sp, 16
; ZVFHMIN-NEXT: vs8r.v v8, (a2) # Unknown-size Folded Spill
-; ZVFHMIN-NEXT: vmv1r.v v0, v1
-; ZVFHMIN-NEXT: vmfeq.vv v12, v24, v24, v0.t
+; ZVFHMIN-NEXT: vmv1r.v v0, v24
+; ZVFHMIN-NEXT: vmfeq.vv v12, v16, v16, v0.t
; ZVFHMIN-NEXT: vmv1r.v v0, v12
-; ZVFHMIN-NEXT: vmerge.vvm v16, v24, v16, v0
-; ZVFHMIN-NEXT: vmv1r.v v0, v1
; ZVFHMIN-NEXT: csrr a2, vlenb
-; ZVFHMIN-NEXT: slli a2, a2, 3
+; ZVFHMIN-NEXT: slli a2, a2, 4
; ZVFHMIN-NEXT: add a2, sp, a2
; ZVFHMIN-NEXT: addi a2, a2, 16
; ZVFHMIN-NEXT: vl8r.v v8, (a2) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vmerge.vvm v16, v16, v8, v0
+; ZVFHMIN-NEXT: vmv1r.v v0, v24
+; ZVFHMIN-NEXT: addi a2, sp, 16
+; ZVFHMIN-NEXT: vl8r.v v8, (a2) # Unknown-size Folded Reload
; ZVFHMIN-NEXT: vfmax.vv v16, v16, v8, v0.t
; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v4, v16
@@ -685,28 +656,41 @@ define <vscale x 32 x half> @vfmax_vv_nxv32f16_unmasked(<vscale x 32 x half> %va
; ZVFHMIN-NEXT: # %bb.1:
; ZVFHMIN-NEXT: mv a0, a1
; ZVFHMIN-NEXT: .LBB11_2:
-; ZVFHMIN-NEXT: csrr a1, vlenb
-; ZVFHMIN-NEXT: slli a1, a1, 4
-; ZVFHMIN-NEXT: add a1, sp, a1
-; ZVFHMIN-NEXT: addi a1, a1, 16
-; ZVFHMIN-NEXT: vl8r.v v16, (a1) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; ZVFHMIN-NEXT: csrr a0, vlenb
+; ZVFHMIN-NEXT: slli a0, a0, 3
+; ZVFHMIN-NEXT: add a0, sp, a0
+; ZVFHMIN-NEXT: addi a0, a0, 16
+; ZVFHMIN-NEXT: vl8r.v v16, (a0) # Unknown-size Folded Reload
; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v16
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vmfeq.vv v0, v8, v8
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: csrr a1, vlenb
-; ZVFHMIN-NEXT: li a2, 24
-; ZVFHMIN-NEXT: mul a1, a1, a2
-; ZVFHMIN-NEXT: add a1, sp, a1
-; ZVFHMIN-NEXT: addi a1, a1, 16
-; ZVFHMIN-NEXT: vl8r.v v24, (a1) # Unknown-size Folded Reload
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v24
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v1, v16, v16
-; ZVFHMIN-NEXT: vmerge.vvm v24, v8, v16, v0
+; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: csrr a0, vlenb
+; ZVFHMIN-NEXT: li a1, 24
+; ZVFHMIN-NEXT: mul a0, a0, a1
+; ZVFHMIN-NEXT: add a0, sp, a0
+; ZVFHMIN-NEXT: addi a0, a0, 16
+; ZVFHMIN-NEXT: vl8r.v v16, (a0) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v16
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vmfeq.vv v1, v24, v24
+; ZVFHMIN-NEXT: vmerge.vvm v16, v8, v24, v0
+; ZVFHMIN-NEXT: csrr a0, vlenb
+; ZVFHMIN-NEXT: li a1, 24
+; ZVFHMIN-NEXT: mul a0, a0, a1
+; ZVFHMIN-NEXT: add a0, sp, a0
+; ZVFHMIN-NEXT: addi a0, a0, 16
+; ZVFHMIN-NEXT: vs8r.v v16, (a0) # Unknown-size Folded Spill
; ZVFHMIN-NEXT: vmv1r.v v0, v1
-; ZVFHMIN-NEXT: vmerge.vvm v16, v16, v8, v0
-; ZVFHMIN-NEXT: vfmax.vv v16, v16, v24
+; ZVFHMIN-NEXT: vmerge.vvm v16, v24, v8, v0
+; ZVFHMIN-NEXT: csrr a0, vlenb
+; ZVFHMIN-NEXT: li a1, 24
+; ZVFHMIN-NEXT: mul a0, a0, a1
+; ZVFHMIN-NEXT: add a0, sp, a0
+; ZVFHMIN-NEXT: addi a0, a0, 16
+; ZVFHMIN-NEXT: vl8r.v v8, (a0) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfmax.vv v16, v16, v8
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v0, v16
; ZVFHMIN-NEXT: vmv8r.v v8, v0
diff --git a/llvm/test/CodeGen/RISCV/rvv/fminimum-vp.ll b/llvm/test/CodeGen/RISCV/rvv/fminimum-vp.ll
index 1f8af732e3f995..a4a043e8e71706 100644
--- a/llvm/test/CodeGen/RISCV/rvv/fminimum-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/fminimum-vp.ll
@@ -27,19 +27,17 @@ define <vscale x 1 x half> @vfmin_vv_nxv1f16(<vscale x 1 x half> %va, <vscale x
; ZVFHMIN-LABEL: vfmin_vv_nxv1f16:
; ZVFHMIN: # %bb.0:
; ZVFHMIN-NEXT: vmv1r.v v10, v0
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v11, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v0, v11, v11, v0.t
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v9
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
-; ZVFHMIN-NEXT: vmerge.vvm v9, v11, v8, v0
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v11, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vmfeq.vv v0, v9, v9, v0.t
+; ZVFHMIN-NEXT: vmerge.vvm v8, v9, v11, v0
; ZVFHMIN-NEXT: vmv1r.v v0, v10
-; ZVFHMIN-NEXT: vmfeq.vv v0, v8, v8, v0.t
-; ZVFHMIN-NEXT: vmerge.vvm v8, v8, v11, v0
+; ZVFHMIN-NEXT: vmfeq.vv v0, v11, v11, v0.t
+; ZVFHMIN-NEXT: vmerge.vvm v9, v11, v9, v0
; ZVFHMIN-NEXT: vmv1r.v v0, v10
-; ZVFHMIN-NEXT: vfmin.vv v9, v8, v9, v0.t
+; ZVFHMIN-NEXT: vfmin.vv v9, v9, v8, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
; ZVFHMIN-NEXT: ret
@@ -61,18 +59,16 @@ define <vscale x 1 x half> @vfmin_vv_nxv1f16_unmasked(<vscale x 1 x half> %va, <
;
; ZVFHMIN-LABEL: vfmin_vv_nxv1f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v0, v10, v10
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v11, v9
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v8, v11, v11
-; ZVFHMIN-NEXT: vmerge.vvm v9, v10, v11, v0
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vmfeq.vv v0, v9, v9
+; ZVFHMIN-NEXT: vmfeq.vv v8, v10, v10
+; ZVFHMIN-NEXT: vmerge.vvm v11, v9, v10, v0
; ZVFHMIN-NEXT: vmv1r.v v0, v8
-; ZVFHMIN-NEXT: vmerge.vvm v8, v11, v10, v0
-; ZVFHMIN-NEXT: vfmin.vv v9, v8, v9
+; ZVFHMIN-NEXT: vmerge.vvm v8, v10, v9, v0
+; ZVFHMIN-NEXT: vfmin.vv v9, v8, v11
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
; ZVFHMIN-NEXT: ret
@@ -101,19 +97,17 @@ define <vscale x 2 x half> @vfmin_vv_nxv2f16(<vscale x 2 x half> %va, <vscale x
; ZVFHMIN-LABEL: vfmin_vv_nxv2f16:
; ZVFHMIN: # %bb.0:
; ZVFHMIN-NEXT: vmv1r.v v10, v0
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v11, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v0, v11, v11, v0.t
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v9
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
-; ZVFHMIN-NEXT: vmerge.vvm v9, v11, v8, v0
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v11, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vmfeq.vv v0, v9, v9, v0.t
+; ZVFHMIN-NEXT: vmerge.vvm v8, v9, v11, v0
; ZVFHMIN-NEXT: vmv1r.v v0, v10
-; ZVFHMIN-NEXT: vmfeq.vv v0, v8, v8, v0.t
-; ZVFHMIN-NEXT: vmerge.vvm v8, v8, v11, v0
+; ZVFHMIN-NEXT: vmfeq.vv v0, v11, v11, v0.t
+; ZVFHMIN-NEXT: vmerge.vvm v9, v11, v9, v0
; ZVFHMIN-NEXT: vmv1r.v v0, v10
-; ZVFHMIN-NEXT: vfmin.vv v9, v8, v9, v0.t
+; ZVFHMIN-NEXT: vfmin.vv v9, v9, v8, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
; ZVFHMIN-NEXT: ret
@@ -135,18 +129,16 @@ define <vscale x 2 x half> @vfmin_vv_nxv2f16_unmasked(<vscale x 2 x half> %va, <
;
; ZVFHMIN-LABEL: vfmin_vv_nxv2f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v0, v10, v10
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v11, v9
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v8, v11, v11
-; ZVFHMIN-NEXT: vmerge.vvm v9, v10, v11, v0
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vmfeq.vv v0, v9, v9
+; ZVFHMIN-NEXT: vmfeq.vv v8, v10, v10
+; ZVFHMIN-NEXT: vmerge.vvm v11, v9, v10, v0
; ZVFHMIN-NEXT: vmv.v.v v0, v8
-; ZVFHMIN-NEXT: vmerge.vvm v8, v11, v10, v0
-; ZVFHMIN-NEXT: vfmin.vv v9, v8, v9
+; ZVFHMIN-NEXT: vmerge.vvm v8, v10, v9, v0
+; ZVFHMIN-NEXT: vfmin.vv v9, v8, v11
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
; ZVFHMIN-NEXT: ret
@@ -175,19 +167,17 @@ define <vscale x 4 x half> @vfmin_vv_nxv4f16(<vscale x 4 x half> %va, <vscale x
; ZVFHMIN-LABEL: vfmin_vv_nxv4f16:
; ZVFHMIN: # %bb.0:
; ZVFHMIN-NEXT: vmv1r.v v10, v0
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v8, v12, v12, v0.t
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v14, v9
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v14, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vmfeq.vv v8, v14, v14, v0.t
; ZVFHMIN-NEXT: vmv1r.v v0, v8
-; ZVFHMIN-NEXT: vmerge.vvm v16, v12, v14, v0
+; ZVFHMIN-NEXT: vmerge.vvm v16, v14, v12, v0
; ZVFHMIN-NEXT: vmv1r.v v0, v10
-; ZVFHMIN-NEXT: vmfeq.vv v8, v14, v14, v0.t
+; ZVFHMIN-NEXT: vmfeq.vv v8, v12, v12, v0.t
; ZVFHMIN-NEXT: vmv1r.v v0, v8
-; ZVFHMIN-NEXT: vmerge.vvm v8, v14, v12, v0
+; ZVFHMIN-NEXT: vmerge.vvm v8, v12, v14, v0
; ZVFHMIN-NEXT: vmv1r.v v0, v10
; ZVFHMIN-NEXT: vfmin.vv v10, v8, v16, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m1, ta, ma
@@ -211,17 +201,15 @@ define <vscale x 4 x half> @vfmin_vv_nxv4f16_unmasked(<vscale x 4 x half> %va, <
;
; ZVFHMIN-LABEL: vfmin_vv_nxv4f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v0, v10, v10
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v9
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v8, v12, v12
-; ZVFHMIN-NEXT: vmerge.vvm v14, v10, v12, v0
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vmfeq.vv v0, v12, v12
+; ZVFHMIN-NEXT: vmfeq.vv v8, v10, v10
+; ZVFHMIN-NEXT: vmerge.vvm v14, v12, v10, v0
; ZVFHMIN-NEXT: vmv1r.v v0, v8
-; ZVFHMIN-NEXT: vmerge.vvm v8, v12, v10, v0
+; ZVFHMIN-NEXT: vmerge.vvm v8, v10, v12, v0
; ZVFHMIN-NEXT: vfmin.vv v10, v8, v14
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -253,19 +241,17 @@ define <vscale x 8 x half> @vfmin_vv_nxv8f16(<vscale x 8 x half> %va, <vscale x
; ZVFHMIN-LABEL: vfmin_vv_nxv8f16:
; ZVFHMIN: # %bb.0:
; ZVFHMIN-NEXT: vmv1r.v v12, v0
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v8, v16, v16, v0.t
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v20, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v10, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v20, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vmfeq.vv v8, v20, v20, v0.t
; ZVFHMIN-NEXT: vmv1r.v v0, v8
-; ZVFHMIN-NEXT: vmerge.vvm v24, v16, v20, v0
+; ZVFHMIN-NEXT: vmerge.vvm v24, v20, v16, v0
; ZVFHMIN-NEXT: vmv1r.v v0, v12
-; ZVFHMIN-NEXT: vmfeq.vv v8, v20, v20, v0.t
+; ZVFHMIN-NEXT: vmfeq.vv v8, v16, v16, v0.t
; ZVFHMIN-NEXT: vmv1r.v v0, v8
-; ZVFHMIN-NEXT: vmerge.vvm v8, v20, v16, v0
+; ZVFHMIN-NEXT: vmerge.vvm v8, v16, v20, v0
; ZVFHMIN-NEXT: vmv1r.v v0, v12
; ZVFHMIN-NEXT: vfmin.vv v12, v8, v24, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m2, ta, ma
@@ -289,17 +275,15 @@ define <vscale x 8 x half> @vfmin_vv_nxv8f16_unmasked(<vscale x 8 x half> %va, <
;
; ZVFHMIN-LABEL: vfmin_vv_nxv8f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v0, v12, v12
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v8, v16, v16
-; ZVFHMIN-NEXT: vmerge.vvm v20, v12, v16, v0
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v10
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vmfeq.vv v0, v16, v16
+; ZVFHMIN-NEXT: vmfeq.vv v8, v12, v12
+; ZVFHMIN-NEXT: vmerge.vvm v20, v16, v12, v0
; ZVFHMIN-NEXT: vmv1r.v v0, v8
-; ZVFHMIN-NEXT: vmerge.vvm v8, v16, v12, v0
+; ZVFHMIN-NEXT: vmerge.vvm v8, v12, v16, v0
; ZVFHMIN-NEXT: vfmin.vv v12, v8, v20
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -337,13 +321,11 @@ define <vscale x 16 x half> @vfmin_vv_nxv16f16(<vscale x 16 x half> %va, <vscale
; ZVFHMIN-NEXT: sub sp, sp, a1
; ZVFHMIN-NEXT: .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x08, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 8 * vlenb
; ZVFHMIN-NEXT: vmv1r.v v1, v0
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vmfeq.vv v8, v24, v24, v0.t
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
; ZVFHMIN-NEXT: vmv1r.v v0, v8
; ZVFHMIN-NEXT: vmerge.vvm v8, v24, v16, v0
; ZVFHMIN-NEXT: addi a0, sp, 16
@@ -386,19 +368,17 @@ define <vscale x 16 x half> @vfmin_vv_nxv16f16_unmasked(<vscale x 16 x half> %va
; ZVFHMIN-NEXT: slli a1, a1, 3
; ZVFHMIN-NEXT: sub sp, sp, a1
; ZVFHMIN-NEXT: .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x08, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 8 * vlenb
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v0, v16, v16
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v12
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v1, v24, v24
-; ZVFHMIN-NEXT: vmerge.vvm v8, v16, v24, v0
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vmfeq.vv v0, v24, v24
+; ZVFHMIN-NEXT: vmfeq.vv v1, v16, v16
+; ZVFHMIN-NEXT: vmerge.vvm v8, v24, v16, v0
; ZVFHMIN-NEXT: addi a0, sp, 16
; ZVFHMIN-NEXT: vs8r.v v8, (a0) # Unknown-size Folded Spill
; ZVFHMIN-NEXT: vmv1r.v v0, v1
-; ZVFHMIN-NEXT: vmerge.vvm v8, v24, v16, v0
+; ZVFHMIN-NEXT: vmerge.vvm v8, v16, v24, v0
; ZVFHMIN-NEXT: vl8r.v v16, (a0) # Unknown-size Folded Reload
; ZVFHMIN-NEXT: vfmin.vv v16, v8, v16
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
@@ -450,12 +430,12 @@ define <vscale x 32 x half> @vfmin_vv_nxv32f16(<vscale x 32 x half> %va, <vscale
; ZVFHMIN-NEXT: addi sp, sp, -16
; ZVFHMIN-NEXT: .cfi_def_cfa_offset 16
; ZVFHMIN-NEXT: csrr a1, vlenb
-; ZVFHMIN-NEXT: li a2, 34
-; ZVFHMIN-NEXT: mul a1, a1, a2
+; ZVFHMIN-NEXT: slli a1, a1, 5
; ZVFHMIN-NEXT: sub sp, sp, a1
-; ZVFHMIN-NEXT: .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x22, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 34 * vlenb
+; ZVFHMIN-NEXT: .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x20, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 32 * vlenb
+; ZVFHMIN-NEXT: vmv1r.v v1, v0
; ZVFHMIN-NEXT: csrr a1, vlenb
-; ZVFHMIN-NEXT: li a2, 25
+; ZVFHMIN-NEXT: li a2, 24
; ZVFHMIN-NEXT: mul a1, a1, a2
; ZVFHMIN-NEXT: add a1, sp, a1
; ZVFHMIN-NEXT: addi a1, a1, 16
@@ -469,136 +449,117 @@ define <vscale x 32 x half> @vfmin_vv_nxv32f16(<vscale x 32 x half> %va, <vscale
; ZVFHMIN-NEXT: and a3, a4, a3
; ZVFHMIN-NEXT: srli a2, a2, 2
; ZVFHMIN-NEXT: vsetvli a4, zero, e8, mf2, ta, ma
-; ZVFHMIN-NEXT: csrr a4, vlenb
-; ZVFHMIN-NEXT: li a5, 24
-; ZVFHMIN-NEXT: mul a4, a4, a5
-; ZVFHMIN-NEXT: add a4, sp, a4
-; ZVFHMIN-NEXT: addi a4, a4, 16
-; ZVFHMIN-NEXT: vs1r.v v0, (a4) # Unknown-size Folded Spill
-; ZVFHMIN-NEXT: vslidedown.vx v0, v0, a2
-; ZVFHMIN-NEXT: addi a2, sp, 16
-; ZVFHMIN-NEXT: vs1r.v v0, (a2) # Unknown-size Folded Spill
-; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vslidedown.vx v25, v0, a2
+; ZVFHMIN-NEXT: vsetvli zero, a3, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vmv1r.v v0, v25
+; ZVFHMIN-NEXT: csrr a2, vlenb
+; ZVFHMIN-NEXT: slli a2, a2, 3
+; ZVFHMIN-NEXT: add a2, sp, a2
+; ZVFHMIN-NEXT: addi a2, a2, 16
+; ZVFHMIN-NEXT: vs8r.v v8, (a2) # Unknown-size Folded Spill
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v20, v0.t
; ZVFHMIN-NEXT: csrr a2, vlenb
; ZVFHMIN-NEXT: slli a2, a2, 4
; ZVFHMIN-NEXT: add a2, sp, a2
; ZVFHMIN-NEXT: addi a2, a2, 16
; ZVFHMIN-NEXT: vs8r.v v8, (a2) # Unknown-size Folded Spill
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v20
-; ZVFHMIN-NEXT: vsetvli zero, a3, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v12, v24, v24, v0.t
-; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vmfeq.vv v24, v8, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: csrr a2, vlenb
-; ZVFHMIN-NEXT: li a4, 25
-; ZVFHMIN-NEXT: mul a2, a2, a4
+; ZVFHMIN-NEXT: li a3, 24
+; ZVFHMIN-NEXT: mul a2, a2, a3
; ZVFHMIN-NEXT: add a2, sp, a2
; ZVFHMIN-NEXT: addi a2, a2, 16
-; ZVFHMIN-NEXT: vl8r.v v0, (a2) # Unknown-size Folded Reload
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v4
-; ZVFHMIN-NEXT: vsetvli zero, a3, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vmv1r.v v0, v12
-; ZVFHMIN-NEXT: vmerge.vvm v8, v24, v16, v0
+; ZVFHMIN-NEXT: vl8r.v v8, (a2) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vmv1r.v v0, v24
; ZVFHMIN-NEXT: csrr a2, vlenb
-; ZVFHMIN-NEXT: slli a2, a2, 3
+; ZVFHMIN-NEXT: slli a2, a2, 4
; ZVFHMIN-NEXT: add a2, sp, a2
; ZVFHMIN-NEXT: addi a2, a2, 16
-; ZVFHMIN-NEXT: vs8r.v v8, (a2) # Unknown-size Folded Spill
+; ZVFHMIN-NEXT: vl8r.v v8, (a2) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vmerge.vvm v8, v8, v16, v0
; ZVFHMIN-NEXT: addi a2, sp, 16
-; ZVFHMIN-NEXT: vl1r.v v8, (a2) # Unknown-size Folded Reload
-; ZVFHMIN-NEXT: vmv1r.v v0, v8
+; ZVFHMIN-NEXT: vs8r.v v8, (a2) # Unknown-size Folded Spill
+; ZVFHMIN-NEXT: vmv1r.v v0, v25
; ZVFHMIN-NEXT: vmfeq.vv v12, v16, v16, v0.t
; ZVFHMIN-NEXT: vmv1r.v v0, v12
-; ZVFHMIN-NEXT: vmerge.vvm v16, v16, v24, v0
-; ZVFHMIN-NEXT: vmv1r.v v0, v8
; ZVFHMIN-NEXT: csrr a2, vlenb
-; ZVFHMIN-NEXT: slli a2, a2, 3
+; ZVFHMIN-NEXT: slli a2, a2, 4
; ZVFHMIN-NEXT: add a2, sp, a2
; ZVFHMIN-NEXT: addi a2, a2, 16
-; ZVFHMIN-NEXT: vl8r.v v24, (a2) # Unknown-size Folded Reload
-; ZVFHMIN-NEXT: vfmin.vv v16, v16, v24, v0.t
+; ZVFHMIN-NEXT: vl8r.v v8, (a2) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vmerge.vvm v16, v16, v8, v0
+; ZVFHMIN-NEXT: vmv1r.v v0, v25
+; ZVFHMIN-NEXT: addi a2, sp, 16
+; ZVFHMIN-NEXT: vl8r.v v8, (a2) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfmin.vv v16, v16, v8, v0.t
; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v16
-; ZVFHMIN-NEXT: csrr a2, vlenb
-; ZVFHMIN-NEXT: slli a2, a2, 3
-; ZVFHMIN-NEXT: add a2, sp, a2
-; ZVFHMIN-NEXT: addi a2, a2, 16
-; ZVFHMIN-NEXT: vs8r.v v8, (a2) # Unknown-size Folded Spill
; ZVFHMIN-NEXT: bltu a0, a1, .LBB10_2
; ZVFHMIN-NEXT: # %bb.1:
; ZVFHMIN-NEXT: mv a0, a1
; ZVFHMIN-NEXT: .LBB10_2:
-; ZVFHMIN-NEXT: csrr a1, vlenb
-; ZVFHMIN-NEXT: slli a1, a1, 4
-; ZVFHMIN-NEXT: add a1, sp, a1
-; ZVFHMIN-NEXT: addi a1, a1, 16
-; ZVFHMIN-NEXT: vl8r.v v16, (a1) # Unknown-size Folded Reload
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v16
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
-; ZVFHMIN-NEXT: csrr a1, vlenb
-; ZVFHMIN-NEXT: li a2, 24
-; ZVFHMIN-NEXT: mul a1, a1, a2
-; ZVFHMIN-NEXT: add a1, sp, a1
-; ZVFHMIN-NEXT: addi a1, a1, 16
-; ZVFHMIN-NEXT: vl1r.v v0, (a1) # Unknown-size Folded Reload
-; ZVFHMIN-NEXT: vmfeq.vv v8, v24, v24, v0.t
-; ZVFHMIN-NEXT: csrr a1, vlenb
-; ZVFHMIN-NEXT: slli a1, a1, 4
-; ZVFHMIN-NEXT: add a1, sp, a1
-; ZVFHMIN-NEXT: addi a1, a1, 16
-; ZVFHMIN-NEXT: vs1r.v v8, (a1) # Unknown-size Folded Spill
-; ZVFHMIN-NEXT: addi a1, sp, 16
-; ZVFHMIN-NEXT: vs8r.v v24, (a1) # Unknown-size Folded Spill
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: csrr a1, vlenb
-; ZVFHMIN-NEXT: li a2, 25
-; ZVFHMIN-NEXT: mul a1, a1, a2
-; ZVFHMIN-NEXT: add a1, sp, a1
-; ZVFHMIN-NEXT: addi a1, a1, 16
-; ZVFHMIN-NEXT: vl8r.v v0, (a1) # Unknown-size Folded Reload
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v0
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vmv1r.v v0, v1
; ZVFHMIN-NEXT: csrr a0, vlenb
-; ZVFHMIN-NEXT: slli a0, a0, 4
+; ZVFHMIN-NEXT: slli a0, a0, 3
; ZVFHMIN-NEXT: add a0, sp, a0
; ZVFHMIN-NEXT: addi a0, a0, 16
-; ZVFHMIN-NEXT: vl1r.v v0, (a0) # Unknown-size Folded Reload
-; ZVFHMIN-NEXT: vmerge.vvm v24, v24, v16, v0
+; ZVFHMIN-NEXT: vl8r.v v16, (a0) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v16, v0.t
; ZVFHMIN-NEXT: csrr a0, vlenb
-; ZVFHMIN-NEXT: li a1, 25
-; ZVFHMIN-NEXT: mul a0, a0, a1
+; ZVFHMIN-NEXT: slli a0, a0, 4
; ZVFHMIN-NEXT: add a0, sp, a0
; ZVFHMIN-NEXT: addi a0, a0, 16
; ZVFHMIN-NEXT: vs8r.v v24, (a0) # Unknown-size Folded Spill
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vmfeq.vv v8, v24, v24, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: csrr a0, vlenb
; ZVFHMIN-NEXT: li a1, 24
; ZVFHMIN-NEXT: mul a0, a0, a1
; ZVFHMIN-NEXT: add a0, sp, a0
; ZVFHMIN-NEXT: addi a0, a0, 16
-; ZVFHMIN-NEXT: vl1r.v v9, (a0) # Unknown-size Folded Reload
-; ZVFHMIN-NEXT: vmv1r.v v0, v9
-; ZVFHMIN-NEXT: vmfeq.vv v8, v16, v16, v0.t
+; ZVFHMIN-NEXT: vl8r.v v16, (a0) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v16, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vmv1r.v v0, v8
-; ZVFHMIN-NEXT: addi a0, sp, 16
-; ZVFHMIN-NEXT: vl8r.v v24, (a0) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: csrr a0, vlenb
+; ZVFHMIN-NEXT: slli a0, a0, 4
+; ZVFHMIN-NEXT: add a0, sp, a0
+; ZVFHMIN-NEXT: addi a0, a0, 16
+; ZVFHMIN-NEXT: vl8r.v v16, (a0) # Unknown-size Folded Reload
; ZVFHMIN-NEXT: vmerge.vvm v16, v16, v24, v0
-; ZVFHMIN-NEXT: vmv1r.v v0, v9
; ZVFHMIN-NEXT: csrr a0, vlenb
-; ZVFHMIN-NEXT: li a1, 25
+; ZVFHMIN-NEXT: li a1, 24
; ZVFHMIN-NEXT: mul a0, a0, a1
; ZVFHMIN-NEXT: add a0, sp, a0
; ZVFHMIN-NEXT: addi a0, a0, 16
-; ZVFHMIN-NEXT: vl8r.v v24, (a0) # Unknown-size Folded Reload
-; ZVFHMIN-NEXT: vfmin.vv v16, v16, v24, v0.t
-; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vs8r.v v16, (a0) # Unknown-size Folded Spill
+; ZVFHMIN-NEXT: vmv1r.v v0, v1
+; ZVFHMIN-NEXT: vmfeq.vv v8, v24, v24, v0.t
+; ZVFHMIN-NEXT: vmv1r.v v0, v8
; ZVFHMIN-NEXT: csrr a0, vlenb
-; ZVFHMIN-NEXT: slli a0, a0, 3
+; ZVFHMIN-NEXT: slli a0, a0, 4
; ZVFHMIN-NEXT: add a0, sp, a0
; ZVFHMIN-NEXT: addi a0, a0, 16
-; ZVFHMIN-NEXT: vl8r.v v8, (a0) # Unknown-size Folded Reload
-; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
+; ZVFHMIN-NEXT: vl8r.v v16, (a0) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vmerge.vvm v16, v24, v16, v0
+; ZVFHMIN-NEXT: vmv1r.v v0, v1
; ZVFHMIN-NEXT: csrr a0, vlenb
-; ZVFHMIN-NEXT: li a1, 34
+; ZVFHMIN-NEXT: li a1, 24
; ZVFHMIN-NEXT: mul a0, a0, a1
+; ZVFHMIN-NEXT: add a0, sp, a0
+; ZVFHMIN-NEXT: addi a0, a0, 16
+; ZVFHMIN-NEXT: vl8r.v v24, (a0) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfmin.vv v16, v16, v24, v0.t
+; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
+; ZVFHMIN-NEXT: csrr a0, vlenb
+; ZVFHMIN-NEXT: slli a0, a0, 5
; ZVFHMIN-NEXT: add sp, sp, a0
; ZVFHMIN-NEXT: addi sp, sp, 16
; ZVFHMIN-NEXT: ret
@@ -626,8 +587,14 @@ define <vscale x 32 x half> @vfmin_vv_nxv32f16_unmasked(<vscale x 32 x half> %va
; ZVFHMIN-NEXT: slli a1, a1, 5
; ZVFHMIN-NEXT: sub sp, sp, a1
; ZVFHMIN-NEXT: .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x20, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 32 * vlenb
+; ZVFHMIN-NEXT: csrr a1, vlenb
+; ZVFHMIN-NEXT: li a2, 24
+; ZVFHMIN-NEXT: mul a1, a1, a2
+; ZVFHMIN-NEXT: add a1, sp, a1
+; ZVFHMIN-NEXT: addi a1, a1, 16
+; ZVFHMIN-NEXT: vs8r.v v16, (a1) # Unknown-size Folded Spill
; ZVFHMIN-NEXT: vsetvli a1, zero, e8, m4, ta, ma
-; ZVFHMIN-NEXT: vmset.m v24
+; ZVFHMIN-NEXT: vmset.m v16
; ZVFHMIN-NEXT: csrr a2, vlenb
; ZVFHMIN-NEXT: slli a1, a2, 1
; ZVFHMIN-NEXT: sub a3, a0, a1
@@ -636,48 +603,52 @@ define <vscale x 32 x half> @vfmin_vv_nxv32f16_unmasked(<vscale x 32 x half> %va
; ZVFHMIN-NEXT: and a3, a4, a3
; ZVFHMIN-NEXT: srli a2, a2, 2
; ZVFHMIN-NEXT: vsetvli a4, zero, e8, mf2, ta, ma
-; ZVFHMIN-NEXT: vslidedown.vx v1, v24, a2
-; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vslidedown.vx v24, v16, a2
+; ZVFHMIN-NEXT: vsetvli zero, a3, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vmv1r.v v0, v24
; ZVFHMIN-NEXT: csrr a2, vlenb
-; ZVFHMIN-NEXT: slli a2, a2, 4
+; ZVFHMIN-NEXT: slli a2, a2, 3
; ZVFHMIN-NEXT: add a2, sp, a2
; ZVFHMIN-NEXT: addi a2, a2, 16
; ZVFHMIN-NEXT: vs8r.v v8, (a2) # Unknown-size Folded Spill
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v12
-; ZVFHMIN-NEXT: addi a2, sp, 16
-; ZVFHMIN-NEXT: vs8r.v v24, (a2) # Unknown-size Folded Spill
-; ZVFHMIN-NEXT: vsetvli zero, a3, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vmv1r.v v0, v1
-; ZVFHMIN-NEXT: vmfeq.vv v12, v24, v24, v0.t
-; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vmv4r.v v8, v16
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12, v0.t
; ZVFHMIN-NEXT: csrr a2, vlenb
-; ZVFHMIN-NEXT: li a4, 24
-; ZVFHMIN-NEXT: mul a2, a2, a4
+; ZVFHMIN-NEXT: slli a2, a2, 4
; ZVFHMIN-NEXT: add a2, sp, a2
; ZVFHMIN-NEXT: addi a2, a2, 16
-; ZVFHMIN-NEXT: vs8r.v v8, (a2) # Unknown-size Folded Spill
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v20
-; ZVFHMIN-NEXT: vsetvli zero, a3, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vmv1r.v v0, v12
-; ZVFHMIN-NEXT: addi a2, sp, 16
-; ZVFHMIN-NEXT: vl8r.v v16, (a2) # Unknown-size Folded Reload
-; ZVFHMIN-NEXT: vmerge.vvm v8, v16, v24, v0
+; ZVFHMIN-NEXT: vs8r.v v16, (a2) # Unknown-size Folded Spill
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vmfeq.vv v25, v16, v16, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: csrr a2, vlenb
-; ZVFHMIN-NEXT: slli a2, a2, 3
+; ZVFHMIN-NEXT: li a3, 24
+; ZVFHMIN-NEXT: mul a2, a2, a3
; ZVFHMIN-NEXT: add a2, sp, a2
; ZVFHMIN-NEXT: addi a2, a2, 16
+; ZVFHMIN-NEXT: vl8r.v v8, (a2) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vmv1r.v v0, v25
+; ZVFHMIN-NEXT: csrr a2, vlenb
+; ZVFHMIN-NEXT: slli a2, a2, 4
+; ZVFHMIN-NEXT: add a2, sp, a2
+; ZVFHMIN-NEXT: addi a2, a2, 16
+; ZVFHMIN-NEXT: vl8r.v v8, (a2) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vmerge.vvm v8, v8, v16, v0
+; ZVFHMIN-NEXT: addi a2, sp, 16
; ZVFHMIN-NEXT: vs8r.v v8, (a2) # Unknown-size Folded Spill
-; ZVFHMIN-NEXT: vmv1r.v v0, v1
-; ZVFHMIN-NEXT: vmfeq.vv v12, v24, v24, v0.t
+; ZVFHMIN-NEXT: vmv1r.v v0, v24
+; ZVFHMIN-NEXT: vmfeq.vv v12, v16, v16, v0.t
; ZVFHMIN-NEXT: vmv1r.v v0, v12
-; ZVFHMIN-NEXT: vmerge.vvm v16, v24, v16, v0
-; ZVFHMIN-NEXT: vmv1r.v v0, v1
; ZVFHMIN-NEXT: csrr a2, vlenb
-; ZVFHMIN-NEXT: slli a2, a2, 3
+; ZVFHMIN-NEXT: slli a2, a2, 4
; ZVFHMIN-NEXT: add a2, sp, a2
; ZVFHMIN-NEXT: addi a2, a2, 16
; ZVFHMIN-NEXT: vl8r.v v8, (a2) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vmerge.vvm v16, v16, v8, v0
+; ZVFHMIN-NEXT: vmv1r.v v0, v24
+; ZVFHMIN-NEXT: addi a2, sp, 16
+; ZVFHMIN-NEXT: vl8r.v v8, (a2) # Unknown-size Folded Reload
; ZVFHMIN-NEXT: vfmin.vv v16, v16, v8, v0.t
; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v4, v16
@@ -685,28 +656,41 @@ define <vscale x 32 x half> @vfmin_vv_nxv32f16_unmasked(<vscale x 32 x half> %va
; ZVFHMIN-NEXT: # %bb.1:
; ZVFHMIN-NEXT: mv a0, a1
; ZVFHMIN-NEXT: .LBB11_2:
-; ZVFHMIN-NEXT: csrr a1, vlenb
-; ZVFHMIN-NEXT: slli a1, a1, 4
-; ZVFHMIN-NEXT: add a1, sp, a1
-; ZVFHMIN-NEXT: addi a1, a1, 16
-; ZVFHMIN-NEXT: vl8r.v v16, (a1) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; ZVFHMIN-NEXT: csrr a0, vlenb
+; ZVFHMIN-NEXT: slli a0, a0, 3
+; ZVFHMIN-NEXT: add a0, sp, a0
+; ZVFHMIN-NEXT: addi a0, a0, 16
+; ZVFHMIN-NEXT: vl8r.v v16, (a0) # Unknown-size Folded Reload
; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v16
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vmfeq.vv v0, v8, v8
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: csrr a1, vlenb
-; ZVFHMIN-NEXT: li a2, 24
-; ZVFHMIN-NEXT: mul a1, a1, a2
-; ZVFHMIN-NEXT: add a1, sp, a1
-; ZVFHMIN-NEXT: addi a1, a1, 16
-; ZVFHMIN-NEXT: vl8r.v v24, (a1) # Unknown-size Folded Reload
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v24
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vmfeq.vv v1, v16, v16
-; ZVFHMIN-NEXT: vmerge.vvm v24, v8, v16, v0
+; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: csrr a0, vlenb
+; ZVFHMIN-NEXT: li a1, 24
+; ZVFHMIN-NEXT: mul a0, a0, a1
+; ZVFHMIN-NEXT: add a0, sp, a0
+; ZVFHMIN-NEXT: addi a0, a0, 16
+; ZVFHMIN-NEXT: vl8r.v v16, (a0) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v16
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vmfeq.vv v1, v24, v24
+; ZVFHMIN-NEXT: vmerge.vvm v16, v8, v24, v0
+; ZVFHMIN-NEXT: csrr a0, vlenb
+; ZVFHMIN-NEXT: li a1, 24
+; ZVFHMIN-NEXT: mul a0, a0, a1
+; ZVFHMIN-NEXT: add a0, sp, a0
+; ZVFHMIN-NEXT: addi a0, a0, 16
+; ZVFHMIN-NEXT: vs8r.v v16, (a0) # Unknown-size Folded Spill
; ZVFHMIN-NEXT: vmv1r.v v0, v1
-; ZVFHMIN-NEXT: vmerge.vvm v16, v16, v8, v0
-; ZVFHMIN-NEXT: vfmin.vv v16, v16, v24
+; ZVFHMIN-NEXT: vmerge.vvm v16, v24, v8, v0
+; ZVFHMIN-NEXT: csrr a0, vlenb
+; ZVFHMIN-NEXT: li a1, 24
+; ZVFHMIN-NEXT: mul a0, a0, a1
+; ZVFHMIN-NEXT: add a0, sp, a0
+; ZVFHMIN-NEXT: addi a0, a0, 16
+; ZVFHMIN-NEXT: vl8r.v v8, (a0) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfmin.vv v16, v16, v8
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v0, v16
; ZVFHMIN-NEXT: vmv8r.v v8, v0
diff --git a/llvm/test/CodeGen/RISCV/rvv/nearbyint-vp.ll b/llvm/test/CodeGen/RISCV/rvv/nearbyint-vp.ll
index 7c354c3714c6f4..284329f2f22d91 100644
--- a/llvm/test/CodeGen/RISCV/rvv/nearbyint-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/nearbyint-vp.ll
@@ -30,9 +30,9 @@ define <vscale x 1 x half> @vp_nearbyint_nxv1f16(<vscale x 1 x half> %va, <vscal
;
; ZVFHMIN-LABEL: vp_nearbyint_nxv1f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v9, v0.t
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -70,9 +70,9 @@ define <vscale x 1 x half> @vp_nearbyint_nxv1f16_unmasked(<vscale x 1 x half> %v
;
; ZVFHMIN-LABEL: vp_nearbyint_nxv1f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v9
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -114,9 +114,9 @@ define <vscale x 2 x half> @vp_nearbyint_nxv2f16(<vscale x 2 x half> %va, <vscal
;
; ZVFHMIN-LABEL: vp_nearbyint_nxv2f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v9, v0.t
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -154,9 +154,9 @@ define <vscale x 2 x half> @vp_nearbyint_nxv2f16_unmasked(<vscale x 2 x half> %v
;
; ZVFHMIN-LABEL: vp_nearbyint_nxv2f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v9
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -199,9 +199,9 @@ define <vscale x 4 x half> @vp_nearbyint_nxv4f16(<vscale x 4 x half> %va, <vscal
; ZVFHMIN-LABEL: vp_nearbyint_nxv4f16:
; ZVFHMIN: # %bb.0:
; ZVFHMIN-NEXT: vmv1r.v v9, v0
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfabs.v v12, v10, v0.t
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -240,9 +240,9 @@ define <vscale x 4 x half> @vp_nearbyint_nxv4f16_unmasked(<vscale x 4 x half> %v
;
; ZVFHMIN-LABEL: vp_nearbyint_nxv4f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v10
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -287,9 +287,9 @@ define <vscale x 8 x half> @vp_nearbyint_nxv8f16(<vscale x 8 x half> %va, <vscal
; ZVFHMIN-LABEL: vp_nearbyint_nxv8f16:
; ZVFHMIN: # %bb.0:
; ZVFHMIN-NEXT: vmv1r.v v10, v0
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfabs.v v16, v12, v0.t
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -328,9 +328,9 @@ define <vscale x 8 x half> @vp_nearbyint_nxv8f16_unmasked(<vscale x 8 x half> %v
;
; ZVFHMIN-LABEL: vp_nearbyint_nxv8f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v12
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -375,9 +375,9 @@ define <vscale x 16 x half> @vp_nearbyint_nxv16f16(<vscale x 16 x half> %va, <vs
; ZVFHMIN-LABEL: vp_nearbyint_nxv16f16:
; ZVFHMIN: # %bb.0:
; ZVFHMIN-NEXT: vmv1r.v v12, v0
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfabs.v v24, v16, v0.t
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -416,9 +416,9 @@ define <vscale x 16 x half> @vp_nearbyint_nxv16f16_unmasked(<vscale x 16 x half>
;
; ZVFHMIN-LABEL: vp_nearbyint_nxv16f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v16
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -468,7 +468,7 @@ define <vscale x 32 x half> @vp_nearbyint_nxv32f16(<vscale x 32 x half> %va, <vs
; ZVFHMIN-NEXT: slli a1, a1, 3
; ZVFHMIN-NEXT: sub sp, sp, a1
; ZVFHMIN-NEXT: .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x08, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 8 * vlenb
-; ZVFHMIN-NEXT: vmv1r.v v1, v0
+; ZVFHMIN-NEXT: vmv1r.v v16, v0
; ZVFHMIN-NEXT: csrr a2, vlenb
; ZVFHMIN-NEXT: slli a1, a2, 1
; ZVFHMIN-NEXT: sub a3, a0, a1
@@ -478,12 +478,12 @@ define <vscale x 32 x half> @vp_nearbyint_nxv32f16(<vscale x 32 x half> %va, <vs
; ZVFHMIN-NEXT: srli a2, a2, 2
; ZVFHMIN-NEXT: vsetvli a4, zero, e8, mf2, ta, ma
; ZVFHMIN-NEXT: vslidedown.vx v17, v0, a2
-; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a3, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vmv1r.v v0, v17
; ZVFHMIN-NEXT: addi a2, sp, 16
; ZVFHMIN-NEXT: vs8r.v v8, (a2) # Unknown-size Folded Spill
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v12
-; ZVFHMIN-NEXT: vsetvli zero, a3, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vmv1r.v v0, v17
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v12, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v24, v0.t
; ZVFHMIN-NEXT: lui a2, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a2
@@ -498,29 +498,31 @@ define <vscale x 32 x half> @vp_nearbyint_nxv32f16(<vscale x 32 x half> %va, <vs
; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, mu
; ZVFHMIN-NEXT: vfsgnj.vv v24, v8, v24, v0.t
; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v24
+; ZVFHMIN-NEXT: vfncvt.f.f.w v20, v24
; ZVFHMIN-NEXT: bltu a0, a1, .LBB10_2
; ZVFHMIN-NEXT: # %bb.1:
; ZVFHMIN-NEXT: mv a0, a1
; ZVFHMIN-NEXT: .LBB10_2:
-; ZVFHMIN-NEXT: addi a1, sp, 16
-; ZVFHMIN-NEXT: vl8r.v v16, (a1) # Unknown-size Folded Reload
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v16
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vmv1r.v v0, v1
-; ZVFHMIN-NEXT: vfabs.v v16, v24, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vmv1r.v v0, v16
+; ZVFHMIN-NEXT: addi a0, sp, 16
+; ZVFHMIN-NEXT: vl8r.v v8, (a0) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vfabs.v v8, v24, v0.t
; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, mu
-; ZVFHMIN-NEXT: vmflt.vf v1, v16, fa5, v0.t
+; ZVFHMIN-NEXT: vmflt.vf v16, v8, fa5, v0.t
; ZVFHMIN-NEXT: frflags a0
; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vmv1r.v v0, v1
-; ZVFHMIN-NEXT: vfcvt.x.f.v v16, v24, v0.t
-; ZVFHMIN-NEXT: vfcvt.f.x.v v16, v16, v0.t
+; ZVFHMIN-NEXT: vmv1r.v v0, v16
+; ZVFHMIN-NEXT: vfcvt.x.f.v v8, v24, v0.t
+; ZVFHMIN-NEXT: vfcvt.f.x.v v8, v8, v0.t
; ZVFHMIN-NEXT: fsflags a0
; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, mu
-; ZVFHMIN-NEXT: vfsgnj.vv v24, v16, v24, v0.t
+; ZVFHMIN-NEXT: vfsgnj.vv v24, v8, v24, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v24
+; ZVFHMIN-NEXT: vfncvt.f.f.w v16, v24
+; ZVFHMIN-NEXT: vmv8r.v v8, v16
; ZVFHMIN-NEXT: csrr a0, vlenb
; ZVFHMIN-NEXT: slli a0, a0, 3
; ZVFHMIN-NEXT: add sp, sp, a0
@@ -565,12 +567,12 @@ define <vscale x 32 x half> @vp_nearbyint_nxv32f16_unmasked(<vscale x 32 x half>
; ZVFHMIN-NEXT: srli a2, a2, 2
; ZVFHMIN-NEXT: vsetvli a4, zero, e8, mf2, ta, ma
; ZVFHMIN-NEXT: vslidedown.vx v16, v16, a2
-; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a3, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vmv1r.v v0, v16
; ZVFHMIN-NEXT: addi a2, sp, 16
; ZVFHMIN-NEXT: vs8r.v v8, (a2) # Unknown-size Folded Spill
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v12
-; ZVFHMIN-NEXT: vsetvli zero, a3, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vmv1r.v v0, v16
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v12, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v24, v0.t
; ZVFHMIN-NEXT: lui a2, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a2
@@ -590,10 +592,11 @@ define <vscale x 32 x half> @vp_nearbyint_nxv32f16_unmasked(<vscale x 32 x half>
; ZVFHMIN-NEXT: # %bb.1:
; ZVFHMIN-NEXT: mv a0, a1
; ZVFHMIN-NEXT: .LBB11_2:
-; ZVFHMIN-NEXT: addi a1, sp, 16
-; ZVFHMIN-NEXT: vl8r.v v24, (a1) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; ZVFHMIN-NEXT: addi a0, sp, 16
+; ZVFHMIN-NEXT: vl8r.v v24, (a0) # Unknown-size Folded Reload
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v24
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfabs.v v24, v16
; ZVFHMIN-NEXT: vmflt.vf v0, v24, fa5
; ZVFHMIN-NEXT: frflags a0
diff --git a/llvm/test/CodeGen/RISCV/rvv/rint-vp.ll b/llvm/test/CodeGen/RISCV/rvv/rint-vp.ll
index 04761d4e7bfc4a..0ca9d2a4d69817 100644
--- a/llvm/test/CodeGen/RISCV/rvv/rint-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/rint-vp.ll
@@ -28,9 +28,9 @@ define <vscale x 1 x half> @vp_rint_nxv1f16(<vscale x 1 x half> %va, <vscale x 1
;
; ZVFHMIN-LABEL: vp_rint_nxv1f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v9, v0.t
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -64,9 +64,9 @@ define <vscale x 1 x half> @vp_rint_nxv1f16_unmasked(<vscale x 1 x half> %va, i3
;
; ZVFHMIN-LABEL: vp_rint_nxv1f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v9
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -104,9 +104,9 @@ define <vscale x 2 x half> @vp_rint_nxv2f16(<vscale x 2 x half> %va, <vscale x 2
;
; ZVFHMIN-LABEL: vp_rint_nxv2f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v9, v0.t
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -140,9 +140,9 @@ define <vscale x 2 x half> @vp_rint_nxv2f16_unmasked(<vscale x 2 x half> %va, i3
;
; ZVFHMIN-LABEL: vp_rint_nxv2f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v9
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -181,9 +181,9 @@ define <vscale x 4 x half> @vp_rint_nxv4f16(<vscale x 4 x half> %va, <vscale x 4
; ZVFHMIN-LABEL: vp_rint_nxv4f16:
; ZVFHMIN: # %bb.0:
; ZVFHMIN-NEXT: vmv1r.v v9, v0
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfabs.v v12, v10, v0.t
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -218,9 +218,9 @@ define <vscale x 4 x half> @vp_rint_nxv4f16_unmasked(<vscale x 4 x half> %va, i3
;
; ZVFHMIN-LABEL: vp_rint_nxv4f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v10
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -261,9 +261,9 @@ define <vscale x 8 x half> @vp_rint_nxv8f16(<vscale x 8 x half> %va, <vscale x 8
; ZVFHMIN-LABEL: vp_rint_nxv8f16:
; ZVFHMIN: # %bb.0:
; ZVFHMIN-NEXT: vmv1r.v v10, v0
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfabs.v v16, v12, v0.t
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -298,9 +298,9 @@ define <vscale x 8 x half> @vp_rint_nxv8f16_unmasked(<vscale x 8 x half> %va, i3
;
; ZVFHMIN-LABEL: vp_rint_nxv8f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v12
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -341,9 +341,9 @@ define <vscale x 16 x half> @vp_rint_nxv16f16(<vscale x 16 x half> %va, <vscale
; ZVFHMIN-LABEL: vp_rint_nxv16f16:
; ZVFHMIN: # %bb.0:
; ZVFHMIN-NEXT: vmv1r.v v12, v0
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfabs.v v24, v16, v0.t
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -378,9 +378,9 @@ define <vscale x 16 x half> @vp_rint_nxv16f16_unmasked(<vscale x 16 x half> %va,
;
; ZVFHMIN-LABEL: vp_rint_nxv16f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v16
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -436,12 +436,12 @@ define <vscale x 32 x half> @vp_rint_nxv32f16(<vscale x 32 x half> %va, <vscale
; ZVFHMIN-NEXT: srli a2, a2, 2
; ZVFHMIN-NEXT: vsetvli a4, zero, e8, mf2, ta, ma
; ZVFHMIN-NEXT: vslidedown.vx v17, v0, a2
-; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a3, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vmv1r.v v0, v17
; ZVFHMIN-NEXT: addi a2, sp, 16
; ZVFHMIN-NEXT: vs8r.v v8, (a2) # Unknown-size Folded Spill
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v12
-; ZVFHMIN-NEXT: vsetvli zero, a3, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vmv1r.v v0, v17
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v12, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v24, v0.t
; ZVFHMIN-NEXT: lui a2, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a2
@@ -459,11 +459,12 @@ define <vscale x 32 x half> @vp_rint_nxv32f16(<vscale x 32 x half> %va, <vscale
; ZVFHMIN-NEXT: # %bb.1:
; ZVFHMIN-NEXT: mv a0, a1
; ZVFHMIN-NEXT: .LBB10_2:
-; ZVFHMIN-NEXT: addi a1, sp, 16
-; ZVFHMIN-NEXT: vl8r.v v8, (a1) # Unknown-size Folded Reload
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
; ZVFHMIN-NEXT: vmv1r.v v0, v16
+; ZVFHMIN-NEXT: addi a0, sp, 16
+; ZVFHMIN-NEXT: vl8r.v v8, (a0) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v24, v0.t
; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, mu
; ZVFHMIN-NEXT: vmflt.vf v16, v8, fa5, v0.t
@@ -518,12 +519,12 @@ define <vscale x 32 x half> @vp_rint_nxv32f16_unmasked(<vscale x 32 x half> %va,
; ZVFHMIN-NEXT: srli a2, a2, 2
; ZVFHMIN-NEXT: vsetvli a4, zero, e8, mf2, ta, ma
; ZVFHMIN-NEXT: vslidedown.vx v16, v16, a2
-; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a3, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vmv1r.v v0, v16
; ZVFHMIN-NEXT: addi a2, sp, 16
; ZVFHMIN-NEXT: vs8r.v v8, (a2) # Unknown-size Folded Spill
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v12
-; ZVFHMIN-NEXT: vsetvli zero, a3, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vmv1r.v v0, v16
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v12, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v24, v0.t
; ZVFHMIN-NEXT: lui a2, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a2
@@ -541,10 +542,11 @@ define <vscale x 32 x half> @vp_rint_nxv32f16_unmasked(<vscale x 32 x half> %va,
; ZVFHMIN-NEXT: # %bb.1:
; ZVFHMIN-NEXT: mv a0, a1
; ZVFHMIN-NEXT: .LBB11_2:
-; ZVFHMIN-NEXT: addi a1, sp, 16
-; ZVFHMIN-NEXT: vl8r.v v24, (a1) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; ZVFHMIN-NEXT: addi a0, sp, 16
+; ZVFHMIN-NEXT: vl8r.v v24, (a0) # Unknown-size Folded Reload
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v24
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfabs.v v24, v16
; ZVFHMIN-NEXT: vmflt.vf v0, v24, fa5
; ZVFHMIN-NEXT: vfcvt.x.f.v v24, v16, v0.t
diff --git a/llvm/test/CodeGen/RISCV/rvv/round-vp.ll b/llvm/test/CodeGen/RISCV/rvv/round-vp.ll
index 16bd665dd0de98..cbff0fc6855de9 100644
--- a/llvm/test/CodeGen/RISCV/rvv/round-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/round-vp.ll
@@ -30,9 +30,9 @@ define <vscale x 1 x half> @vp_round_nxv1f16(<vscale x 1 x half> %va, <vscale x
;
; ZVFHMIN-LABEL: vp_round_nxv1f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v9, v0.t
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -70,9 +70,9 @@ define <vscale x 1 x half> @vp_round_nxv1f16_unmasked(<vscale x 1 x half> %va, i
;
; ZVFHMIN-LABEL: vp_round_nxv1f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v9
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -114,9 +114,9 @@ define <vscale x 2 x half> @vp_round_nxv2f16(<vscale x 2 x half> %va, <vscale x
;
; ZVFHMIN-LABEL: vp_round_nxv2f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v9, v0.t
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -154,9 +154,9 @@ define <vscale x 2 x half> @vp_round_nxv2f16_unmasked(<vscale x 2 x half> %va, i
;
; ZVFHMIN-LABEL: vp_round_nxv2f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v9
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -199,9 +199,9 @@ define <vscale x 4 x half> @vp_round_nxv4f16(<vscale x 4 x half> %va, <vscale x
; ZVFHMIN-LABEL: vp_round_nxv4f16:
; ZVFHMIN: # %bb.0:
; ZVFHMIN-NEXT: vmv1r.v v9, v0
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfabs.v v12, v10, v0.t
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -240,9 +240,9 @@ define <vscale x 4 x half> @vp_round_nxv4f16_unmasked(<vscale x 4 x half> %va, i
;
; ZVFHMIN-LABEL: vp_round_nxv4f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v10
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -287,9 +287,9 @@ define <vscale x 8 x half> @vp_round_nxv8f16(<vscale x 8 x half> %va, <vscale x
; ZVFHMIN-LABEL: vp_round_nxv8f16:
; ZVFHMIN: # %bb.0:
; ZVFHMIN-NEXT: vmv1r.v v10, v0
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfabs.v v16, v12, v0.t
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -328,9 +328,9 @@ define <vscale x 8 x half> @vp_round_nxv8f16_unmasked(<vscale x 8 x half> %va, i
;
; ZVFHMIN-LABEL: vp_round_nxv8f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v12
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -375,9 +375,9 @@ define <vscale x 16 x half> @vp_round_nxv16f16(<vscale x 16 x half> %va, <vscale
; ZVFHMIN-LABEL: vp_round_nxv16f16:
; ZVFHMIN: # %bb.0:
; ZVFHMIN-NEXT: vmv1r.v v12, v0
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfabs.v v24, v16, v0.t
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -416,9 +416,9 @@ define <vscale x 16 x half> @vp_round_nxv16f16_unmasked(<vscale x 16 x half> %va
;
; ZVFHMIN-LABEL: vp_round_nxv16f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v16
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -478,12 +478,12 @@ define <vscale x 32 x half> @vp_round_nxv32f16(<vscale x 32 x half> %va, <vscale
; ZVFHMIN-NEXT: srli a2, a2, 2
; ZVFHMIN-NEXT: vsetvli a4, zero, e8, mf2, ta, ma
; ZVFHMIN-NEXT: vslidedown.vx v17, v0, a2
-; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a3, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vmv1r.v v0, v17
; ZVFHMIN-NEXT: addi a2, sp, 16
; ZVFHMIN-NEXT: vs8r.v v8, (a2) # Unknown-size Folded Spill
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v12
-; ZVFHMIN-NEXT: vsetvli zero, a3, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vmv1r.v v0, v17
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v12, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v24, v0.t
; ZVFHMIN-NEXT: lui a2, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a2
@@ -503,11 +503,12 @@ define <vscale x 32 x half> @vp_round_nxv32f16(<vscale x 32 x half> %va, <vscale
; ZVFHMIN-NEXT: # %bb.1:
; ZVFHMIN-NEXT: mv a0, a1
; ZVFHMIN-NEXT: .LBB10_2:
-; ZVFHMIN-NEXT: addi a1, sp, 16
-; ZVFHMIN-NEXT: vl8r.v v8, (a1) # Unknown-size Folded Reload
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
; ZVFHMIN-NEXT: vmv1r.v v0, v16
+; ZVFHMIN-NEXT: addi a0, sp, 16
+; ZVFHMIN-NEXT: vl8r.v v8, (a0) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v24, v0.t
; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, mu
; ZVFHMIN-NEXT: vmflt.vf v16, v8, fa5, v0.t
@@ -566,12 +567,12 @@ define <vscale x 32 x half> @vp_round_nxv32f16_unmasked(<vscale x 32 x half> %va
; ZVFHMIN-NEXT: srli a2, a2, 2
; ZVFHMIN-NEXT: vsetvli a4, zero, e8, mf2, ta, ma
; ZVFHMIN-NEXT: vslidedown.vx v16, v16, a2
-; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a3, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vmv1r.v v0, v16
; ZVFHMIN-NEXT: addi a2, sp, 16
; ZVFHMIN-NEXT: vs8r.v v8, (a2) # Unknown-size Folded Spill
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v12
-; ZVFHMIN-NEXT: vsetvli zero, a3, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vmv1r.v v0, v16
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v12, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v24, v0.t
; ZVFHMIN-NEXT: lui a2, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a2
@@ -591,10 +592,11 @@ define <vscale x 32 x half> @vp_round_nxv32f16_unmasked(<vscale x 32 x half> %va
; ZVFHMIN-NEXT: # %bb.1:
; ZVFHMIN-NEXT: mv a0, a1
; ZVFHMIN-NEXT: .LBB11_2:
-; ZVFHMIN-NEXT: addi a1, sp, 16
-; ZVFHMIN-NEXT: vl8r.v v24, (a1) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; ZVFHMIN-NEXT: addi a0, sp, 16
+; ZVFHMIN-NEXT: vl8r.v v24, (a0) # Unknown-size Folded Reload
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v24
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfabs.v v24, v16
; ZVFHMIN-NEXT: vmflt.vf v0, v24, fa5
; ZVFHMIN-NEXT: fsrmi a0, 4
diff --git a/llvm/test/CodeGen/RISCV/rvv/roundeven-vp.ll b/llvm/test/CodeGen/RISCV/rvv/roundeven-vp.ll
index 429ddb6c71be34..39d8e1023aadba 100644
--- a/llvm/test/CodeGen/RISCV/rvv/roundeven-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/roundeven-vp.ll
@@ -30,9 +30,9 @@ define <vscale x 1 x half> @vp_roundeven_nxv1f16(<vscale x 1 x half> %va, <vscal
;
; ZVFHMIN-LABEL: vp_roundeven_nxv1f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v9, v0.t
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -70,9 +70,9 @@ define <vscale x 1 x half> @vp_roundeven_nxv1f16_unmasked(<vscale x 1 x half> %v
;
; ZVFHMIN-LABEL: vp_roundeven_nxv1f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v9
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -114,9 +114,9 @@ define <vscale x 2 x half> @vp_roundeven_nxv2f16(<vscale x 2 x half> %va, <vscal
;
; ZVFHMIN-LABEL: vp_roundeven_nxv2f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v9, v0.t
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -154,9 +154,9 @@ define <vscale x 2 x half> @vp_roundeven_nxv2f16_unmasked(<vscale x 2 x half> %v
;
; ZVFHMIN-LABEL: vp_roundeven_nxv2f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v9
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -199,9 +199,9 @@ define <vscale x 4 x half> @vp_roundeven_nxv4f16(<vscale x 4 x half> %va, <vscal
; ZVFHMIN-LABEL: vp_roundeven_nxv4f16:
; ZVFHMIN: # %bb.0:
; ZVFHMIN-NEXT: vmv1r.v v9, v0
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfabs.v v12, v10, v0.t
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -240,9 +240,9 @@ define <vscale x 4 x half> @vp_roundeven_nxv4f16_unmasked(<vscale x 4 x half> %v
;
; ZVFHMIN-LABEL: vp_roundeven_nxv4f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v10
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -287,9 +287,9 @@ define <vscale x 8 x half> @vp_roundeven_nxv8f16(<vscale x 8 x half> %va, <vscal
; ZVFHMIN-LABEL: vp_roundeven_nxv8f16:
; ZVFHMIN: # %bb.0:
; ZVFHMIN-NEXT: vmv1r.v v10, v0
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfabs.v v16, v12, v0.t
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -328,9 +328,9 @@ define <vscale x 8 x half> @vp_roundeven_nxv8f16_unmasked(<vscale x 8 x half> %v
;
; ZVFHMIN-LABEL: vp_roundeven_nxv8f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v12
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -375,9 +375,9 @@ define <vscale x 16 x half> @vp_roundeven_nxv16f16(<vscale x 16 x half> %va, <vs
; ZVFHMIN-LABEL: vp_roundeven_nxv16f16:
; ZVFHMIN: # %bb.0:
; ZVFHMIN-NEXT: vmv1r.v v12, v0
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfabs.v v24, v16, v0.t
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -416,9 +416,9 @@ define <vscale x 16 x half> @vp_roundeven_nxv16f16_unmasked(<vscale x 16 x half>
;
; ZVFHMIN-LABEL: vp_roundeven_nxv16f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v16
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -478,12 +478,12 @@ define <vscale x 32 x half> @vp_roundeven_nxv32f16(<vscale x 32 x half> %va, <vs
; ZVFHMIN-NEXT: srli a2, a2, 2
; ZVFHMIN-NEXT: vsetvli a4, zero, e8, mf2, ta, ma
; ZVFHMIN-NEXT: vslidedown.vx v17, v0, a2
-; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a3, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vmv1r.v v0, v17
; ZVFHMIN-NEXT: addi a2, sp, 16
; ZVFHMIN-NEXT: vs8r.v v8, (a2) # Unknown-size Folded Spill
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v12
-; ZVFHMIN-NEXT: vsetvli zero, a3, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vmv1r.v v0, v17
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v12, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v24, v0.t
; ZVFHMIN-NEXT: lui a2, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a2
@@ -503,11 +503,12 @@ define <vscale x 32 x half> @vp_roundeven_nxv32f16(<vscale x 32 x half> %va, <vs
; ZVFHMIN-NEXT: # %bb.1:
; ZVFHMIN-NEXT: mv a0, a1
; ZVFHMIN-NEXT: .LBB10_2:
-; ZVFHMIN-NEXT: addi a1, sp, 16
-; ZVFHMIN-NEXT: vl8r.v v8, (a1) # Unknown-size Folded Reload
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
; ZVFHMIN-NEXT: vmv1r.v v0, v16
+; ZVFHMIN-NEXT: addi a0, sp, 16
+; ZVFHMIN-NEXT: vl8r.v v8, (a0) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v24, v0.t
; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, mu
; ZVFHMIN-NEXT: vmflt.vf v16, v8, fa5, v0.t
@@ -566,12 +567,12 @@ define <vscale x 32 x half> @vp_roundeven_nxv32f16_unmasked(<vscale x 32 x half>
; ZVFHMIN-NEXT: srli a2, a2, 2
; ZVFHMIN-NEXT: vsetvli a4, zero, e8, mf2, ta, ma
; ZVFHMIN-NEXT: vslidedown.vx v16, v16, a2
-; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a3, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vmv1r.v v0, v16
; ZVFHMIN-NEXT: addi a2, sp, 16
; ZVFHMIN-NEXT: vs8r.v v8, (a2) # Unknown-size Folded Spill
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v12
-; ZVFHMIN-NEXT: vsetvli zero, a3, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vmv1r.v v0, v16
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v12, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v24, v0.t
; ZVFHMIN-NEXT: lui a2, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a2
@@ -591,10 +592,11 @@ define <vscale x 32 x half> @vp_roundeven_nxv32f16_unmasked(<vscale x 32 x half>
; ZVFHMIN-NEXT: # %bb.1:
; ZVFHMIN-NEXT: mv a0, a1
; ZVFHMIN-NEXT: .LBB11_2:
-; ZVFHMIN-NEXT: addi a1, sp, 16
-; ZVFHMIN-NEXT: vl8r.v v24, (a1) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; ZVFHMIN-NEXT: addi a0, sp, 16
+; ZVFHMIN-NEXT: vl8r.v v24, (a0) # Unknown-size Folded Reload
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v24
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfabs.v v24, v16
; ZVFHMIN-NEXT: vmflt.vf v0, v24, fa5
; ZVFHMIN-NEXT: fsrmi a0, 0
diff --git a/llvm/test/CodeGen/RISCV/rvv/roundtozero-vp.ll b/llvm/test/CodeGen/RISCV/rvv/roundtozero-vp.ll
index c854e0fb8a05d9..98972fbe537d77 100644
--- a/llvm/test/CodeGen/RISCV/rvv/roundtozero-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/roundtozero-vp.ll
@@ -30,9 +30,9 @@ define <vscale x 1 x half> @vp_roundtozero_nxv1f16(<vscale x 1 x half> %va, <vsc
;
; ZVFHMIN-LABEL: vp_roundtozero_nxv1f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v9, v0.t
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -70,9 +70,9 @@ define <vscale x 1 x half> @vp_roundtozero_nxv1f16_unmasked(<vscale x 1 x half>
;
; ZVFHMIN-LABEL: vp_roundtozero_nxv1f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v9
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -114,9 +114,9 @@ define <vscale x 2 x half> @vp_roundtozero_nxv2f16(<vscale x 2 x half> %va, <vsc
;
; ZVFHMIN-LABEL: vp_roundtozero_nxv2f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v9, v0.t
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -154,9 +154,9 @@ define <vscale x 2 x half> @vp_roundtozero_nxv2f16_unmasked(<vscale x 2 x half>
;
; ZVFHMIN-LABEL: vp_roundtozero_nxv2f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v9
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -199,9 +199,9 @@ define <vscale x 4 x half> @vp_roundtozero_nxv4f16(<vscale x 4 x half> %va, <vsc
; ZVFHMIN-LABEL: vp_roundtozero_nxv4f16:
; ZVFHMIN: # %bb.0:
; ZVFHMIN-NEXT: vmv1r.v v9, v0
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfabs.v v12, v10, v0.t
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -240,9 +240,9 @@ define <vscale x 4 x half> @vp_roundtozero_nxv4f16_unmasked(<vscale x 4 x half>
;
; ZVFHMIN-LABEL: vp_roundtozero_nxv4f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v10
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -287,9 +287,9 @@ define <vscale x 8 x half> @vp_roundtozero_nxv8f16(<vscale x 8 x half> %va, <vsc
; ZVFHMIN-LABEL: vp_roundtozero_nxv8f16:
; ZVFHMIN: # %bb.0:
; ZVFHMIN-NEXT: vmv1r.v v10, v0
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfabs.v v16, v12, v0.t
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -328,9 +328,9 @@ define <vscale x 8 x half> @vp_roundtozero_nxv8f16_unmasked(<vscale x 8 x half>
;
; ZVFHMIN-LABEL: vp_roundtozero_nxv8f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v12
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -375,9 +375,9 @@ define <vscale x 16 x half> @vp_roundtozero_nxv16f16(<vscale x 16 x half> %va, <
; ZVFHMIN-LABEL: vp_roundtozero_nxv16f16:
; ZVFHMIN: # %bb.0:
; ZVFHMIN-NEXT: vmv1r.v v12, v0
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfabs.v v24, v16, v0.t
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -416,9 +416,9 @@ define <vscale x 16 x half> @vp_roundtozero_nxv16f16_unmasked(<vscale x 16 x hal
;
; ZVFHMIN-LABEL: vp_roundtozero_nxv16f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v16
; ZVFHMIN-NEXT: lui a0, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a0
@@ -478,12 +478,12 @@ define <vscale x 32 x half> @vp_roundtozero_nxv32f16(<vscale x 32 x half> %va, <
; ZVFHMIN-NEXT: srli a2, a2, 2
; ZVFHMIN-NEXT: vsetvli a4, zero, e8, mf2, ta, ma
; ZVFHMIN-NEXT: vslidedown.vx v17, v0, a2
-; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a3, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vmv1r.v v0, v17
; ZVFHMIN-NEXT: addi a2, sp, 16
; ZVFHMIN-NEXT: vs8r.v v8, (a2) # Unknown-size Folded Spill
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v12
-; ZVFHMIN-NEXT: vsetvli zero, a3, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vmv1r.v v0, v17
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v12, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v24, v0.t
; ZVFHMIN-NEXT: lui a2, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a2
@@ -503,11 +503,12 @@ define <vscale x 32 x half> @vp_roundtozero_nxv32f16(<vscale x 32 x half> %va, <
; ZVFHMIN-NEXT: # %bb.1:
; ZVFHMIN-NEXT: mv a0, a1
; ZVFHMIN-NEXT: .LBB10_2:
-; ZVFHMIN-NEXT: addi a1, sp, 16
-; ZVFHMIN-NEXT: vl8r.v v8, (a1) # Unknown-size Folded Reload
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
; ZVFHMIN-NEXT: vmv1r.v v0, v16
+; ZVFHMIN-NEXT: addi a0, sp, 16
+; ZVFHMIN-NEXT: vl8r.v v8, (a0) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v24, v0.t
; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, mu
; ZVFHMIN-NEXT: vmflt.vf v16, v8, fa5, v0.t
@@ -566,12 +567,12 @@ define <vscale x 32 x half> @vp_roundtozero_nxv32f16_unmasked(<vscale x 32 x hal
; ZVFHMIN-NEXT: srli a2, a2, 2
; ZVFHMIN-NEXT: vsetvli a4, zero, e8, mf2, ta, ma
; ZVFHMIN-NEXT: vslidedown.vx v16, v16, a2
-; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a3, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vmv1r.v v0, v16
; ZVFHMIN-NEXT: addi a2, sp, 16
; ZVFHMIN-NEXT: vs8r.v v8, (a2) # Unknown-size Folded Spill
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v12
-; ZVFHMIN-NEXT: vsetvli zero, a3, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vmv1r.v v0, v16
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v12, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfabs.v v8, v24, v0.t
; ZVFHMIN-NEXT: lui a2, 307200
; ZVFHMIN-NEXT: fmv.w.x fa5, a2
@@ -591,10 +592,11 @@ define <vscale x 32 x half> @vp_roundtozero_nxv32f16_unmasked(<vscale x 32 x hal
; ZVFHMIN-NEXT: # %bb.1:
; ZVFHMIN-NEXT: mv a0, a1
; ZVFHMIN-NEXT: .LBB11_2:
-; ZVFHMIN-NEXT: addi a1, sp, 16
-; ZVFHMIN-NEXT: vl8r.v v24, (a1) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; ZVFHMIN-NEXT: addi a0, sp, 16
+; ZVFHMIN-NEXT: vl8r.v v24, (a0) # Unknown-size Folded Reload
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v24
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfabs.v v24, v16
; ZVFHMIN-NEXT: vmflt.vf v0, v24, fa5
; ZVFHMIN-NEXT: fsrmi a0, 1
diff --git a/llvm/test/CodeGen/RISCV/rvv/vcopysign-vp.ll b/llvm/test/CodeGen/RISCV/rvv/vcopysign-vp.ll
index f3574200054fd4..6251ea6c3a0da8 100644
--- a/llvm/test/CodeGen/RISCV/rvv/vcopysign-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/vcopysign-vp.ll
@@ -19,10 +19,10 @@ define <vscale x 1 x half> @vfsgnj_vv_nxv1f16(<vscale x 1 x half> %va, <vscale x
;
; ZVFHMIN-LABEL: vfsgnj_vv_nxv1f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfsgnj.vv v9, v9, v10, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -40,10 +40,10 @@ define <vscale x 1 x half> @vfsgnj_vv_nxv1f16_unmasked(<vscale x 1 x half> %va,
;
; ZVFHMIN-LABEL: vfsgnj_vv_nxv1f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfsgnj.vv v9, v9, v10
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -65,10 +65,10 @@ define <vscale x 2 x half> @vfsgnj_vv_nxv2f16(<vscale x 2 x half> %va, <vscale x
;
; ZVFHMIN-LABEL: vfsgnj_vv_nxv2f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfsgnj.vv v9, v9, v10, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -86,10 +86,10 @@ define <vscale x 2 x half> @vfsgnj_vv_nxv2f16_unmasked(<vscale x 2 x half> %va,
;
; ZVFHMIN-LABEL: vfsgnj_vv_nxv2f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfsgnj.vv v9, v9, v10
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -111,10 +111,10 @@ define <vscale x 4 x half> @vfsgnj_vv_nxv4f16(<vscale x 4 x half> %va, <vscale x
;
; ZVFHMIN-LABEL: vfsgnj_vv_nxv4f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfsgnj.vv v10, v12, v10, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -132,10 +132,10 @@ define <vscale x 4 x half> @vfsgnj_vv_nxv4f16_unmasked(<vscale x 4 x half> %va,
;
; ZVFHMIN-LABEL: vfsgnj_vv_nxv4f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfsgnj.vv v10, v12, v10
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -157,10 +157,10 @@ define <vscale x 8 x half> @vfsgnj_vv_nxv8f16(<vscale x 8 x half> %va, <vscale x
;
; ZVFHMIN-LABEL: vfsgnj_vv_nxv8f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v10
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v10, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfsgnj.vv v12, v16, v12, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -178,10 +178,10 @@ define <vscale x 8 x half> @vfsgnj_vv_nxv8f16_unmasked(<vscale x 8 x half> %va,
;
; ZVFHMIN-LABEL: vfsgnj_vv_nxv8f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v10
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfsgnj.vv v12, v16, v12
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -203,10 +203,10 @@ define <vscale x 16 x half> @vfsgnj_vv_nxv16f16(<vscale x 16 x half> %va, <vscal
;
; ZVFHMIN-LABEL: vfsgnj_vv_nxv16f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfsgnj.vv v16, v24, v16, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
@@ -224,10 +224,10 @@ define <vscale x 16 x half> @vfsgnj_vv_nxv16f16_unmasked(<vscale x 16 x half> %v
;
; ZVFHMIN-LABEL: vfsgnj_vv_nxv16f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12
; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfsgnj.vv v16, v24, v16
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
@@ -252,9 +252,13 @@ define <vscale x 32 x half> @vfsgnj_vv_nxv32f16(<vscale x 32 x half> %va, <vscal
; ZVFHMIN-NEXT: addi sp, sp, -16
; ZVFHMIN-NEXT: .cfi_def_cfa_offset 16
; ZVFHMIN-NEXT: csrr a1, vlenb
+; ZVFHMIN-NEXT: li a2, 0
; ZVFHMIN-NEXT: slli a1, a1, 3
+; ZVFHMIN-NEXT: add a2, a2, a1
+; ZVFHMIN-NEXT: slli a1, a1, 1
+; ZVFHMIN-NEXT: add a1, a1, a2
; ZVFHMIN-NEXT: sub sp, sp, a1
-; ZVFHMIN-NEXT: .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x08, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 8 * vlenb
+; ZVFHMIN-NEXT: .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x18, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 24 * vlenb
; ZVFHMIN-NEXT: vmv1r.v v1, v0
; ZVFHMIN-NEXT: csrr a2, vlenb
; ZVFHMIN-NEXT: slli a1, a2, 1
@@ -265,30 +269,61 @@ define <vscale x 32 x half> @vfsgnj_vv_nxv32f16(<vscale x 32 x half> %va, <vscal
; ZVFHMIN-NEXT: srli a2, a2, 2
; ZVFHMIN-NEXT: vsetvli a4, zero, e8, mf2, ta, ma
; ZVFHMIN-NEXT: vslidedown.vx v0, v0, a2
-; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a3, e16, m4, ta, ma
+; ZVFHMIN-NEXT: csrr a2, vlenb
+; ZVFHMIN-NEXT: slli a2, a2, 3
+; ZVFHMIN-NEXT: add a2, sp, a2
+; ZVFHMIN-NEXT: addi a2, a2, 16
+; ZVFHMIN-NEXT: vs8r.v v16, (a2) # Unknown-size Folded Spill
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v20, v0.t
; ZVFHMIN-NEXT: addi a2, sp, 16
+; ZVFHMIN-NEXT: vs8r.v v24, (a2) # Unknown-size Folded Spill
+; ZVFHMIN-NEXT: vmv4r.v v16, v8
+; ZVFHMIN-NEXT: csrr a2, vlenb
+; ZVFHMIN-NEXT: slli a2, a2, 4
+; ZVFHMIN-NEXT: add a2, sp, a2
+; ZVFHMIN-NEXT: addi a2, a2, 16
; ZVFHMIN-NEXT: vs8r.v v16, (a2) # Unknown-size Folded Spill
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v20
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12
-; ZVFHMIN-NEXT: vsetvli zero, a3, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vfsgnj.vv v16, v16, v24, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; ZVFHMIN-NEXT: addi a2, sp, 16
+; ZVFHMIN-NEXT: vl8r.v v8, (a2) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfsgnj.vv v16, v16, v8, v0.t
; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v16
; ZVFHMIN-NEXT: bltu a0, a1, .LBB10_2
; ZVFHMIN-NEXT: # %bb.1:
; ZVFHMIN-NEXT: mv a0, a1
; ZVFHMIN-NEXT: .LBB10_2:
-; ZVFHMIN-NEXT: addi a1, sp, 16
-; ZVFHMIN-NEXT: vl8r.v v24, (a1) # Unknown-size Folded Reload
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v24
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
; ZVFHMIN-NEXT: vmv1r.v v0, v1
-; ZVFHMIN-NEXT: vfsgnj.vv v16, v24, v16, v0.t
+; ZVFHMIN-NEXT: csrr a0, vlenb
+; ZVFHMIN-NEXT: slli a0, a0, 3
+; ZVFHMIN-NEXT: add a0, sp, a0
+; ZVFHMIN-NEXT: addi a0, a0, 16
+; ZVFHMIN-NEXT: vl8r.v v16, (a0) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v16, v0.t
+; ZVFHMIN-NEXT: addi a0, sp, 16
+; ZVFHMIN-NEXT: vs8r.v v24, (a0) # Unknown-size Folded Spill
+; ZVFHMIN-NEXT: csrr a0, vlenb
+; ZVFHMIN-NEXT: slli a0, a0, 4
+; ZVFHMIN-NEXT: add a0, sp, a0
+; ZVFHMIN-NEXT: addi a0, a0, 16
+; ZVFHMIN-NEXT: vl8r.v v16, (a0) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v16, v0.t
+; ZVFHMIN-NEXT: vmv8r.v v16, v24
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; ZVFHMIN-NEXT: addi a0, sp, 16
+; ZVFHMIN-NEXT: vl8r.v v24, (a0) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfsgnj.vv v16, v16, v24, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
; ZVFHMIN-NEXT: csrr a0, vlenb
+; ZVFHMIN-NEXT: li a1, 0
; ZVFHMIN-NEXT: slli a0, a0, 3
+; ZVFHMIN-NEXT: add a1, a1, a0
+; ZVFHMIN-NEXT: slli a0, a0, 1
+; ZVFHMIN-NEXT: add a0, a0, a1
; ZVFHMIN-NEXT: add sp, sp, a0
; ZVFHMIN-NEXT: addi sp, sp, 16
; ZVFHMIN-NEXT: ret
@@ -322,12 +357,12 @@ define <vscale x 32 x half> @vfsgnj_vv_nxv32f16_unmasked(<vscale x 32 x half> %v
; ZVFHMIN-NEXT: srli a2, a2, 2
; ZVFHMIN-NEXT: vsetvli a4, zero, e8, mf2, ta, ma
; ZVFHMIN-NEXT: vslidedown.vx v0, v24, a2
-; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a3, e16, m4, ta, ma
; ZVFHMIN-NEXT: addi a2, sp, 16
; ZVFHMIN-NEXT: vs8r.v v16, (a2) # Unknown-size Folded Spill
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v20
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12
-; ZVFHMIN-NEXT: vsetvli zero, a3, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v20, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfsgnj.vv v16, v16, v24, v0.t
; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v16
@@ -335,11 +370,12 @@ define <vscale x 32 x half> @vfsgnj_vv_nxv32f16_unmasked(<vscale x 32 x half> %v
; ZVFHMIN-NEXT: # %bb.1:
; ZVFHMIN-NEXT: mv a0, a1
; ZVFHMIN-NEXT: .LBB11_2:
-; ZVFHMIN-NEXT: addi a1, sp, 16
-; ZVFHMIN-NEXT: vl8r.v v24, (a1) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; ZVFHMIN-NEXT: addi a0, sp, 16
+; ZVFHMIN-NEXT: vl8r.v v24, (a0) # Unknown-size Folded Reload
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v24
; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfsgnj.vv v16, v24, v16
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfabs-vp.ll b/llvm/test/CodeGen/RISCV/rvv/vfabs-vp.ll
index 22f0d8bba2e629..20d17510b36300 100644
--- a/llvm/test/CodeGen/RISCV/rvv/vfabs-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/vfabs-vp.ll
@@ -19,9 +19,9 @@ define <vscale x 1 x half> @vfabs_vv_nxv1f16(<vscale x 1 x half> %va, <vscale x
;
; ZVFHMIN-LABEL: vfabs_vv_nxv1f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfabs.v v9, v9, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -39,9 +39,9 @@ define <vscale x 1 x half> @vfabs_vv_nxv1f16_unmasked(<vscale x 1 x half> %va, i
;
; ZVFHMIN-LABEL: vfabs_vv_nxv1f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfabs.v v9, v9
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -63,9 +63,9 @@ define <vscale x 2 x half> @vfabs_vv_nxv2f16(<vscale x 2 x half> %va, <vscale x
;
; ZVFHMIN-LABEL: vfabs_vv_nxv2f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfabs.v v9, v9, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -83,9 +83,9 @@ define <vscale x 2 x half> @vfabs_vv_nxv2f16_unmasked(<vscale x 2 x half> %va, i
;
; ZVFHMIN-LABEL: vfabs_vv_nxv2f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfabs.v v9, v9
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -107,9 +107,9 @@ define <vscale x 4 x half> @vfabs_vv_nxv4f16(<vscale x 4 x half> %va, <vscale x
;
; ZVFHMIN-LABEL: vfabs_vv_nxv4f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfabs.v v10, v10, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -127,9 +127,9 @@ define <vscale x 4 x half> @vfabs_vv_nxv4f16_unmasked(<vscale x 4 x half> %va, i
;
; ZVFHMIN-LABEL: vfabs_vv_nxv4f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfabs.v v10, v10
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -151,9 +151,9 @@ define <vscale x 8 x half> @vfabs_vv_nxv8f16(<vscale x 8 x half> %va, <vscale x
;
; ZVFHMIN-LABEL: vfabs_vv_nxv8f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfabs.v v12, v12, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -171,9 +171,9 @@ define <vscale x 8 x half> @vfabs_vv_nxv8f16_unmasked(<vscale x 8 x half> %va, i
;
; ZVFHMIN-LABEL: vfabs_vv_nxv8f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfabs.v v12, v12
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -195,9 +195,9 @@ define <vscale x 16 x half> @vfabs_vv_nxv16f16(<vscale x 16 x half> %va, <vscale
;
; ZVFHMIN-LABEL: vfabs_vv_nxv16f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfabs.v v16, v16, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
@@ -215,9 +215,9 @@ define <vscale x 16 x half> @vfabs_vv_nxv16f16_unmasked(<vscale x 16 x half> %va
;
; ZVFHMIN-LABEL: vfabs_vv_nxv16f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfabs.v v16, v16
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
@@ -249,9 +249,9 @@ define <vscale x 32 x half> @vfabs_vv_nxv32f16(<vscale x 32 x half> %va, <vscale
; ZVFHMIN-NEXT: srli a2, a2, 2
; ZVFHMIN-NEXT: vsetvli a4, zero, e8, mf2, ta, ma
; ZVFHMIN-NEXT: vslidedown.vx v0, v0, a2
-; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v12
-; ZVFHMIN-NEXT: vsetvli zero, a3, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a3, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v12, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfabs.v v24, v24, v0.t
; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v24
@@ -259,9 +259,10 @@ define <vscale x 32 x half> @vfabs_vv_nxv32f16(<vscale x 32 x half> %va, <vscale
; ZVFHMIN-NEXT: # %bb.1:
; ZVFHMIN-NEXT: mv a0, a1
; ZVFHMIN-NEXT: .LBB10_2:
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
; ZVFHMIN-NEXT: vmv1r.v v0, v16
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfabs.v v16, v24, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
@@ -290,9 +291,9 @@ define <vscale x 32 x half> @vfabs_vv_nxv32f16_unmasked(<vscale x 32 x half> %va
; ZVFHMIN-NEXT: srli a2, a2, 2
; ZVFHMIN-NEXT: vsetvli a4, zero, e8, mf2, ta, ma
; ZVFHMIN-NEXT: vslidedown.vx v0, v16, a2
-; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12
-; ZVFHMIN-NEXT: vsetvli zero, a3, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a3, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfabs.v v16, v16, v0.t
; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v16
@@ -300,8 +301,9 @@ define <vscale x 32 x half> @vfabs_vv_nxv32f16_unmasked(<vscale x 32 x half> %va
; ZVFHMIN-NEXT: # %bb.1:
; ZVFHMIN-NEXT: mv a0, a1
; ZVFHMIN-NEXT: .LBB11_2:
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfabs.v v16, v16
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfadd-vp.ll b/llvm/test/CodeGen/RISCV/rvv/vfadd-vp.ll
index 00cb54c61a7a94..c4ff9f6b0a5b67 100644
--- a/llvm/test/CodeGen/RISCV/rvv/vfadd-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/vfadd-vp.ll
@@ -19,10 +19,10 @@ define <vscale x 1 x half> @vfadd_vv_nxv1f16(<vscale x 1 x half> %va, <vscale x
;
; ZVFHMIN-LABEL: vfadd_vv_nxv1f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfadd.vv v9, v9, v10, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -40,10 +40,10 @@ define <vscale x 1 x half> @vfadd_vv_nxv1f16_unmasked(<vscale x 1 x half> %va, <
;
; ZVFHMIN-LABEL: vfadd_vv_nxv1f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfadd.vv v9, v9, v10
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -68,9 +68,10 @@ define <vscale x 1 x half> @vfadd_vf_nxv1f16(<vscale x 1 x half> %va, half %b, <
; ZVFHMIN-NEXT: vfmv.v.f v9, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfadd.vv v9, v9, v8, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -95,9 +96,10 @@ define <vscale x 1 x half> @vfadd_vf_nxv1f16_commute(<vscale x 1 x half> %va, ha
; ZVFHMIN-NEXT: vfmv.v.f v9, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfadd.vv v9, v8, v9, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -122,9 +124,10 @@ define <vscale x 1 x half> @vfadd_vf_nxv1f16_unmasked(<vscale x 1 x half> %va, h
; ZVFHMIN-NEXT: vfmv.v.f v9, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v10, v9
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfadd.vv v9, v9, v8
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -151,9 +154,10 @@ define <vscale x 1 x half> @vfadd_vf_nxv1f16_unmasked_commute(<vscale x 1 x half
; ZVFHMIN-NEXT: vfmv.v.f v9, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v10, v9
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfadd.vv v9, v8, v9
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -177,10 +181,10 @@ define <vscale x 2 x half> @vfadd_vv_nxv2f16(<vscale x 2 x half> %va, <vscale x
;
; ZVFHMIN-LABEL: vfadd_vv_nxv2f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfadd.vv v9, v9, v10, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -198,10 +202,10 @@ define <vscale x 2 x half> @vfadd_vv_nxv2f16_unmasked(<vscale x 2 x half> %va, <
;
; ZVFHMIN-LABEL: vfadd_vv_nxv2f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfadd.vv v9, v9, v10
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -226,9 +230,10 @@ define <vscale x 2 x half> @vfadd_vf_nxv2f16(<vscale x 2 x half> %va, half %b, <
; ZVFHMIN-NEXT: vfmv.v.f v9, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfadd.vv v9, v9, v8, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -253,9 +258,10 @@ define <vscale x 2 x half> @vfadd_vf_nxv2f16_unmasked(<vscale x 2 x half> %va, h
; ZVFHMIN-NEXT: vfmv.v.f v9, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v10, v9
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfadd.vv v9, v9, v8
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -279,10 +285,10 @@ define <vscale x 4 x half> @vfadd_vv_nxv4f16(<vscale x 4 x half> %va, <vscale x
;
; ZVFHMIN-LABEL: vfadd_vv_nxv4f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfadd.vv v10, v12, v10, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -300,10 +306,10 @@ define <vscale x 4 x half> @vfadd_vv_nxv4f16_unmasked(<vscale x 4 x half> %va, <
;
; ZVFHMIN-LABEL: vfadd_vv_nxv4f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfadd.vv v10, v12, v10
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -328,9 +334,10 @@ define <vscale x 4 x half> @vfadd_vf_nxv4f16(<vscale x 4 x half> %va, half %b, <
; ZVFHMIN-NEXT: vfmv.v.f v10, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v9, v10
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v9
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v9, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfadd.vv v10, v10, v12, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -355,9 +362,10 @@ define <vscale x 4 x half> @vfadd_vf_nxv4f16_unmasked(<vscale x 4 x half> %va, h
; ZVFHMIN-NEXT: vfmv.v.f v10, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v9, v10
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v9
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfadd.vv v10, v10, v12
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -381,10 +389,10 @@ define <vscale x 8 x half> @vfadd_vv_nxv8f16(<vscale x 8 x half> %va, <vscale x
;
; ZVFHMIN-LABEL: vfadd_vv_nxv8f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v10
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v10, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfadd.vv v12, v16, v12, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -402,10 +410,10 @@ define <vscale x 8 x half> @vfadd_vv_nxv8f16_unmasked(<vscale x 8 x half> %va, <
;
; ZVFHMIN-LABEL: vfadd_vv_nxv8f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v10
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfadd.vv v12, v16, v12
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -430,9 +438,10 @@ define <vscale x 8 x half> @vfadd_vf_nxv8f16(<vscale x 8 x half> %va, half %b, <
; ZVFHMIN-NEXT: vfmv.v.f v12, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v10, v12
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v10, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfadd.vv v12, v12, v16, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -457,9 +466,10 @@ define <vscale x 8 x half> @vfadd_vf_nxv8f16_unmasked(<vscale x 8 x half> %va, h
; ZVFHMIN-NEXT: vfmv.v.f v12, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v10, v12
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfadd.vv v12, v12, v16
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -483,10 +493,10 @@ define <vscale x 16 x half> @vfadd_vv_nxv16f16(<vscale x 16 x half> %va, <vscale
;
; ZVFHMIN-LABEL: vfadd_vv_nxv16f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfadd.vv v16, v24, v16, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
@@ -504,10 +514,10 @@ define <vscale x 16 x half> @vfadd_vv_nxv16f16_unmasked(<vscale x 16 x half> %va
;
; ZVFHMIN-LABEL: vfadd_vv_nxv16f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12
; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfadd.vv v16, v24, v16
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
@@ -532,9 +542,10 @@ define <vscale x 16 x half> @vfadd_vf_nxv16f16(<vscale x 16 x half> %va, half %b
; ZVFHMIN-NEXT: vfmv.v.f v16, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v16
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v12
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v12, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfadd.vv v16, v16, v24, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
@@ -559,9 +570,10 @@ define <vscale x 16 x half> @vfadd_vf_nxv16f16_unmasked(<vscale x 16 x half> %va
; ZVFHMIN-NEXT: vfmv.v.f v16, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v16
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v12
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfadd.vv v16, v16, v24
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
@@ -588,9 +600,13 @@ define <vscale x 32 x half> @vfadd_vv_nxv32f16(<vscale x 32 x half> %va, <vscale
; ZVFHMIN-NEXT: addi sp, sp, -16
; ZVFHMIN-NEXT: .cfi_def_cfa_offset 16
; ZVFHMIN-NEXT: csrr a1, vlenb
+; ZVFHMIN-NEXT: li a2, 0
; ZVFHMIN-NEXT: slli a1, a1, 3
+; ZVFHMIN-NEXT: add a2, a2, a1
+; ZVFHMIN-NEXT: slli a1, a1, 1
+; ZVFHMIN-NEXT: add a1, a1, a2
; ZVFHMIN-NEXT: sub sp, sp, a1
-; ZVFHMIN-NEXT: .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x08, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 8 * vlenb
+; ZVFHMIN-NEXT: .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x18, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 24 * vlenb
; ZVFHMIN-NEXT: vmv1r.v v1, v0
; ZVFHMIN-NEXT: csrr a2, vlenb
; ZVFHMIN-NEXT: slli a1, a2, 1
@@ -601,30 +617,61 @@ define <vscale x 32 x half> @vfadd_vv_nxv32f16(<vscale x 32 x half> %va, <vscale
; ZVFHMIN-NEXT: srli a2, a2, 2
; ZVFHMIN-NEXT: vsetvli a4, zero, e8, mf2, ta, ma
; ZVFHMIN-NEXT: vslidedown.vx v0, v0, a2
-; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a3, e16, m4, ta, ma
+; ZVFHMIN-NEXT: csrr a2, vlenb
+; ZVFHMIN-NEXT: slli a2, a2, 3
+; ZVFHMIN-NEXT: add a2, sp, a2
+; ZVFHMIN-NEXT: addi a2, a2, 16
+; ZVFHMIN-NEXT: vs8r.v v16, (a2) # Unknown-size Folded Spill
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v20, v0.t
; ZVFHMIN-NEXT: addi a2, sp, 16
+; ZVFHMIN-NEXT: vs8r.v v24, (a2) # Unknown-size Folded Spill
+; ZVFHMIN-NEXT: vmv4r.v v16, v8
+; ZVFHMIN-NEXT: csrr a2, vlenb
+; ZVFHMIN-NEXT: slli a2, a2, 4
+; ZVFHMIN-NEXT: add a2, sp, a2
+; ZVFHMIN-NEXT: addi a2, a2, 16
; ZVFHMIN-NEXT: vs8r.v v16, (a2) # Unknown-size Folded Spill
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v20
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12
-; ZVFHMIN-NEXT: vsetvli zero, a3, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vfadd.vv v16, v16, v24, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; ZVFHMIN-NEXT: addi a2, sp, 16
+; ZVFHMIN-NEXT: vl8r.v v8, (a2) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfadd.vv v16, v16, v8, v0.t
; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v16
; ZVFHMIN-NEXT: bltu a0, a1, .LBB22_2
; ZVFHMIN-NEXT: # %bb.1:
; ZVFHMIN-NEXT: mv a0, a1
; ZVFHMIN-NEXT: .LBB22_2:
-; ZVFHMIN-NEXT: addi a1, sp, 16
-; ZVFHMIN-NEXT: vl8r.v v24, (a1) # Unknown-size Folded Reload
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v24
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
; ZVFHMIN-NEXT: vmv1r.v v0, v1
-; ZVFHMIN-NEXT: vfadd.vv v16, v24, v16, v0.t
+; ZVFHMIN-NEXT: csrr a0, vlenb
+; ZVFHMIN-NEXT: slli a0, a0, 3
+; ZVFHMIN-NEXT: add a0, sp, a0
+; ZVFHMIN-NEXT: addi a0, a0, 16
+; ZVFHMIN-NEXT: vl8r.v v16, (a0) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v16, v0.t
+; ZVFHMIN-NEXT: addi a0, sp, 16
+; ZVFHMIN-NEXT: vs8r.v v24, (a0) # Unknown-size Folded Spill
+; ZVFHMIN-NEXT: csrr a0, vlenb
+; ZVFHMIN-NEXT: slli a0, a0, 4
+; ZVFHMIN-NEXT: add a0, sp, a0
+; ZVFHMIN-NEXT: addi a0, a0, 16
+; ZVFHMIN-NEXT: vl8r.v v16, (a0) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v16, v0.t
+; ZVFHMIN-NEXT: vmv8r.v v16, v24
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; ZVFHMIN-NEXT: addi a0, sp, 16
+; ZVFHMIN-NEXT: vl8r.v v24, (a0) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfadd.vv v16, v16, v24, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
; ZVFHMIN-NEXT: csrr a0, vlenb
+; ZVFHMIN-NEXT: li a1, 0
; ZVFHMIN-NEXT: slli a0, a0, 3
+; ZVFHMIN-NEXT: add a1, a1, a0
+; ZVFHMIN-NEXT: slli a0, a0, 1
+; ZVFHMIN-NEXT: add a0, a0, a1
; ZVFHMIN-NEXT: add sp, sp, a0
; ZVFHMIN-NEXT: addi sp, sp, 16
; ZVFHMIN-NEXT: ret
@@ -658,12 +705,12 @@ define <vscale x 32 x half> @vfadd_vv_nxv32f16_unmasked(<vscale x 32 x half> %va
; ZVFHMIN-NEXT: srli a2, a2, 2
; ZVFHMIN-NEXT: vsetvli a4, zero, e8, mf2, ta, ma
; ZVFHMIN-NEXT: vslidedown.vx v0, v24, a2
-; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a3, e16, m4, ta, ma
; ZVFHMIN-NEXT: addi a2, sp, 16
; ZVFHMIN-NEXT: vs8r.v v16, (a2) # Unknown-size Folded Spill
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v20
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12
-; ZVFHMIN-NEXT: vsetvli zero, a3, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v20, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfadd.vv v16, v16, v24, v0.t
; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v16
@@ -671,11 +718,12 @@ define <vscale x 32 x half> @vfadd_vv_nxv32f16_unmasked(<vscale x 32 x half> %va
; ZVFHMIN-NEXT: # %bb.1:
; ZVFHMIN-NEXT: mv a0, a1
; ZVFHMIN-NEXT: .LBB23_2:
-; ZVFHMIN-NEXT: addi a1, sp, 16
-; ZVFHMIN-NEXT: vl8r.v v24, (a1) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; ZVFHMIN-NEXT: addi a0, sp, 16
+; ZVFHMIN-NEXT: vl8r.v v24, (a0) # Unknown-size Folded Reload
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v24
; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfadd.vv v16, v24, v16
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
@@ -699,7 +747,14 @@ define <vscale x 32 x half> @vfadd_vf_nxv32f16(<vscale x 32 x half> %va, half %b
;
; ZVFHMIN-LABEL: vfadd_vf_nxv32f16:
; ZVFHMIN: # %bb.0:
+; ZVFHMIN-NEXT: addi sp, sp, -16
+; ZVFHMIN-NEXT: .cfi_def_cfa_offset 16
+; ZVFHMIN-NEXT: csrr a1, vlenb
+; ZVFHMIN-NEXT: slli a1, a1, 3
+; ZVFHMIN-NEXT: sub sp, sp, a1
+; ZVFHMIN-NEXT: .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x08, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 8 * vlenb
; ZVFHMIN-NEXT: vmv1r.v v1, v0
+; ZVFHMIN-NEXT: vmv8r.v v16, v8
; ZVFHMIN-NEXT: fcvt.s.h fa5, fa0
; ZVFHMIN-NEXT: vsetvli a1, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfmv.v.f v24, fa5
@@ -714,23 +769,33 @@ define <vscale x 32 x half> @vfadd_vf_nxv32f16(<vscale x 32 x half> %va, half %b
; ZVFHMIN-NEXT: srli a2, a2, 2
; ZVFHMIN-NEXT: vsetvli a4, zero, e8, mf2, ta, ma
; ZVFHMIN-NEXT: vslidedown.vx v0, v0, a2
+; ZVFHMIN-NEXT: vsetvli zero, a3, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v4, v0.t
+; ZVFHMIN-NEXT: addi a2, sp, 16
+; ZVFHMIN-NEXT: vs8r.v v8, (a2) # Unknown-size Folded Spill
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v20, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vfadd.vv v24, v8, v24, v0.t
; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v4
-; ZVFHMIN-NEXT: vsetvli zero, a3, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vfadd.vv v16, v16, v24, v0.t
-; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v16
+; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v24
; ZVFHMIN-NEXT: bltu a0, a1, .LBB24_2
; ZVFHMIN-NEXT: # %bb.1:
; ZVFHMIN-NEXT: mv a0, a1
; ZVFHMIN-NEXT: .LBB24_2:
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
; ZVFHMIN-NEXT: vmv1r.v v0, v1
+; ZVFHMIN-NEXT: addi a0, sp, 16
+; ZVFHMIN-NEXT: vl8r.v v24, (a0) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v24, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v4, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfadd.vv v16, v16, v24, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
+; ZVFHMIN-NEXT: csrr a0, vlenb
+; ZVFHMIN-NEXT: slli a0, a0, 3
+; ZVFHMIN-NEXT: add sp, sp, a0
+; ZVFHMIN-NEXT: addi sp, sp, 16
; ZVFHMIN-NEXT: ret
%elt.head = insertelement <vscale x 32 x half> poison, half %b, i32 0
%vb = shufflevector <vscale x 32 x half> %elt.head, <vscale x 32 x half> poison, <vscale x 32 x i32> zeroinitializer
@@ -749,11 +814,11 @@ define <vscale x 32 x half> @vfadd_vf_nxv32f16_unmasked(<vscale x 32 x half> %va
; ZVFHMIN: # %bb.0:
; ZVFHMIN-NEXT: fcvt.s.h fa5, fa0
; ZVFHMIN-NEXT: vsetvli a1, zero, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vfmv.v.f v16, fa5
+; ZVFHMIN-NEXT: vfmv.v.f v24, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfncvt.f.f.w v4, v16
+; ZVFHMIN-NEXT: vfncvt.f.f.w v4, v24
; ZVFHMIN-NEXT: vsetvli a1, zero, e8, m4, ta, ma
-; ZVFHMIN-NEXT: vmset.m v16
+; ZVFHMIN-NEXT: vmset.m v20
; ZVFHMIN-NEXT: csrr a2, vlenb
; ZVFHMIN-NEXT: slli a1, a2, 1
; ZVFHMIN-NEXT: sub a3, a0, a1
@@ -762,21 +827,23 @@ define <vscale x 32 x half> @vfadd_vf_nxv32f16_unmasked(<vscale x 32 x half> %va
; ZVFHMIN-NEXT: and a3, a4, a3
; ZVFHMIN-NEXT: srli a2, a2, 2
; ZVFHMIN-NEXT: vsetvli a4, zero, e8, mf2, ta, ma
-; ZVFHMIN-NEXT: vslidedown.vx v0, v16, a2
-; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v12
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v4
-; ZVFHMIN-NEXT: vsetvli zero, a3, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vfadd.vv v24, v24, v16, v0.t
+; ZVFHMIN-NEXT: vslidedown.vx v0, v20, a2
+; ZVFHMIN-NEXT: vsetvli zero, a3, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v4, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vfadd.vv v16, v16, v24, v0.t
; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v24
+; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v16
; ZVFHMIN-NEXT: bltu a0, a1, .LBB25_2
; ZVFHMIN-NEXT: # %bb.1:
; ZVFHMIN-NEXT: mv a0, a1
; ZVFHMIN-NEXT: .LBB25_2:
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vfadd.vv v16, v24, v16
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v4
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vfadd.vv v16, v16, v24
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
; ZVFHMIN-NEXT: ret
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfdiv-vp.ll b/llvm/test/CodeGen/RISCV/rvv/vfdiv-vp.ll
index a49c0fd08ffe46..b20bfd73c41fbe 100644
--- a/llvm/test/CodeGen/RISCV/rvv/vfdiv-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/vfdiv-vp.ll
@@ -19,10 +19,10 @@ define <vscale x 1 x half> @vfdiv_vv_nxv1f16(<vscale x 1 x half> %va, <vscale x
;
; ZVFHMIN-LABEL: vfdiv_vv_nxv1f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfdiv.vv v9, v9, v10, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -40,10 +40,10 @@ define <vscale x 1 x half> @vfdiv_vv_nxv1f16_unmasked(<vscale x 1 x half> %va, <
;
; ZVFHMIN-LABEL: vfdiv_vv_nxv1f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfdiv.vv v9, v9, v10
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -68,9 +68,10 @@ define <vscale x 1 x half> @vfdiv_vf_nxv1f16(<vscale x 1 x half> %va, half %b, <
; ZVFHMIN-NEXT: vfmv.v.f v9, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfdiv.vv v9, v9, v8, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -95,9 +96,10 @@ define <vscale x 1 x half> @vfdiv_vf_nxv1f16_unmasked(<vscale x 1 x half> %va, h
; ZVFHMIN-NEXT: vfmv.v.f v9, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v10, v9
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfdiv.vv v9, v9, v8
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -121,10 +123,10 @@ define <vscale x 2 x half> @vfdiv_vv_nxv2f16(<vscale x 2 x half> %va, <vscale x
;
; ZVFHMIN-LABEL: vfdiv_vv_nxv2f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfdiv.vv v9, v9, v10, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -142,10 +144,10 @@ define <vscale x 2 x half> @vfdiv_vv_nxv2f16_unmasked(<vscale x 2 x half> %va, <
;
; ZVFHMIN-LABEL: vfdiv_vv_nxv2f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfdiv.vv v9, v9, v10
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -170,9 +172,10 @@ define <vscale x 2 x half> @vfdiv_vf_nxv2f16(<vscale x 2 x half> %va, half %b, <
; ZVFHMIN-NEXT: vfmv.v.f v9, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfdiv.vv v9, v9, v8, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -197,9 +200,10 @@ define <vscale x 2 x half> @vfdiv_vf_nxv2f16_unmasked(<vscale x 2 x half> %va, h
; ZVFHMIN-NEXT: vfmv.v.f v9, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v10, v9
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfdiv.vv v9, v9, v8
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -223,10 +227,10 @@ define <vscale x 4 x half> @vfdiv_vv_nxv4f16(<vscale x 4 x half> %va, <vscale x
;
; ZVFHMIN-LABEL: vfdiv_vv_nxv4f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfdiv.vv v10, v12, v10, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -244,10 +248,10 @@ define <vscale x 4 x half> @vfdiv_vv_nxv4f16_unmasked(<vscale x 4 x half> %va, <
;
; ZVFHMIN-LABEL: vfdiv_vv_nxv4f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfdiv.vv v10, v12, v10
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -272,9 +276,10 @@ define <vscale x 4 x half> @vfdiv_vf_nxv4f16(<vscale x 4 x half> %va, half %b, <
; ZVFHMIN-NEXT: vfmv.v.f v10, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v9, v10
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v9
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v9, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfdiv.vv v10, v10, v12, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -299,9 +304,10 @@ define <vscale x 4 x half> @vfdiv_vf_nxv4f16_unmasked(<vscale x 4 x half> %va, h
; ZVFHMIN-NEXT: vfmv.v.f v10, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v9, v10
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v9
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfdiv.vv v10, v10, v12
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -325,10 +331,10 @@ define <vscale x 8 x half> @vfdiv_vv_nxv8f16(<vscale x 8 x half> %va, <vscale x
;
; ZVFHMIN-LABEL: vfdiv_vv_nxv8f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v10
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v10, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfdiv.vv v12, v16, v12, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -346,10 +352,10 @@ define <vscale x 8 x half> @vfdiv_vv_nxv8f16_unmasked(<vscale x 8 x half> %va, <
;
; ZVFHMIN-LABEL: vfdiv_vv_nxv8f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v10
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfdiv.vv v12, v16, v12
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -374,9 +380,10 @@ define <vscale x 8 x half> @vfdiv_vf_nxv8f16(<vscale x 8 x half> %va, half %b, <
; ZVFHMIN-NEXT: vfmv.v.f v12, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v10, v12
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v10, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfdiv.vv v12, v12, v16, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -401,9 +408,10 @@ define <vscale x 8 x half> @vfdiv_vf_nxv8f16_unmasked(<vscale x 8 x half> %va, h
; ZVFHMIN-NEXT: vfmv.v.f v12, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v10, v12
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfdiv.vv v12, v12, v16
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -427,10 +435,10 @@ define <vscale x 16 x half> @vfdiv_vv_nxv16f16(<vscale x 16 x half> %va, <vscale
;
; ZVFHMIN-LABEL: vfdiv_vv_nxv16f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfdiv.vv v16, v24, v16, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
@@ -448,10 +456,10 @@ define <vscale x 16 x half> @vfdiv_vv_nxv16f16_unmasked(<vscale x 16 x half> %va
;
; ZVFHMIN-LABEL: vfdiv_vv_nxv16f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12
; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfdiv.vv v16, v24, v16
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
@@ -476,9 +484,10 @@ define <vscale x 16 x half> @vfdiv_vf_nxv16f16(<vscale x 16 x half> %va, half %b
; ZVFHMIN-NEXT: vfmv.v.f v16, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v16
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v12
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v12, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfdiv.vv v16, v16, v24, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
@@ -503,9 +512,10 @@ define <vscale x 16 x half> @vfdiv_vf_nxv16f16_unmasked(<vscale x 16 x half> %va
; ZVFHMIN-NEXT: vfmv.v.f v16, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v16
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v12
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfdiv.vv v16, v16, v24
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
@@ -532,9 +542,13 @@ define <vscale x 32 x half> @vfdiv_vv_nxv32f16(<vscale x 32 x half> %va, <vscale
; ZVFHMIN-NEXT: addi sp, sp, -16
; ZVFHMIN-NEXT: .cfi_def_cfa_offset 16
; ZVFHMIN-NEXT: csrr a1, vlenb
+; ZVFHMIN-NEXT: li a2, 0
; ZVFHMIN-NEXT: slli a1, a1, 3
+; ZVFHMIN-NEXT: add a2, a2, a1
+; ZVFHMIN-NEXT: slli a1, a1, 1
+; ZVFHMIN-NEXT: add a1, a1, a2
; ZVFHMIN-NEXT: sub sp, sp, a1
-; ZVFHMIN-NEXT: .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x08, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 8 * vlenb
+; ZVFHMIN-NEXT: .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x18, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 24 * vlenb
; ZVFHMIN-NEXT: vmv1r.v v1, v0
; ZVFHMIN-NEXT: csrr a2, vlenb
; ZVFHMIN-NEXT: slli a1, a2, 1
@@ -545,30 +559,61 @@ define <vscale x 32 x half> @vfdiv_vv_nxv32f16(<vscale x 32 x half> %va, <vscale
; ZVFHMIN-NEXT: srli a2, a2, 2
; ZVFHMIN-NEXT: vsetvli a4, zero, e8, mf2, ta, ma
; ZVFHMIN-NEXT: vslidedown.vx v0, v0, a2
-; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a3, e16, m4, ta, ma
+; ZVFHMIN-NEXT: csrr a2, vlenb
+; ZVFHMIN-NEXT: slli a2, a2, 3
+; ZVFHMIN-NEXT: add a2, sp, a2
+; ZVFHMIN-NEXT: addi a2, a2, 16
+; ZVFHMIN-NEXT: vs8r.v v16, (a2) # Unknown-size Folded Spill
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v20, v0.t
; ZVFHMIN-NEXT: addi a2, sp, 16
+; ZVFHMIN-NEXT: vs8r.v v24, (a2) # Unknown-size Folded Spill
+; ZVFHMIN-NEXT: vmv4r.v v16, v8
+; ZVFHMIN-NEXT: csrr a2, vlenb
+; ZVFHMIN-NEXT: slli a2, a2, 4
+; ZVFHMIN-NEXT: add a2, sp, a2
+; ZVFHMIN-NEXT: addi a2, a2, 16
; ZVFHMIN-NEXT: vs8r.v v16, (a2) # Unknown-size Folded Spill
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v20
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12
-; ZVFHMIN-NEXT: vsetvli zero, a3, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vfdiv.vv v16, v16, v24, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; ZVFHMIN-NEXT: addi a2, sp, 16
+; ZVFHMIN-NEXT: vl8r.v v8, (a2) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfdiv.vv v16, v16, v8, v0.t
; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v16
; ZVFHMIN-NEXT: bltu a0, a1, .LBB20_2
; ZVFHMIN-NEXT: # %bb.1:
; ZVFHMIN-NEXT: mv a0, a1
; ZVFHMIN-NEXT: .LBB20_2:
-; ZVFHMIN-NEXT: addi a1, sp, 16
-; ZVFHMIN-NEXT: vl8r.v v24, (a1) # Unknown-size Folded Reload
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v24
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
; ZVFHMIN-NEXT: vmv1r.v v0, v1
-; ZVFHMIN-NEXT: vfdiv.vv v16, v24, v16, v0.t
+; ZVFHMIN-NEXT: csrr a0, vlenb
+; ZVFHMIN-NEXT: slli a0, a0, 3
+; ZVFHMIN-NEXT: add a0, sp, a0
+; ZVFHMIN-NEXT: addi a0, a0, 16
+; ZVFHMIN-NEXT: vl8r.v v16, (a0) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v16, v0.t
+; ZVFHMIN-NEXT: addi a0, sp, 16
+; ZVFHMIN-NEXT: vs8r.v v24, (a0) # Unknown-size Folded Spill
+; ZVFHMIN-NEXT: csrr a0, vlenb
+; ZVFHMIN-NEXT: slli a0, a0, 4
+; ZVFHMIN-NEXT: add a0, sp, a0
+; ZVFHMIN-NEXT: addi a0, a0, 16
+; ZVFHMIN-NEXT: vl8r.v v16, (a0) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v16, v0.t
+; ZVFHMIN-NEXT: vmv8r.v v16, v24
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; ZVFHMIN-NEXT: addi a0, sp, 16
+; ZVFHMIN-NEXT: vl8r.v v24, (a0) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfdiv.vv v16, v16, v24, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
; ZVFHMIN-NEXT: csrr a0, vlenb
+; ZVFHMIN-NEXT: li a1, 0
; ZVFHMIN-NEXT: slli a0, a0, 3
+; ZVFHMIN-NEXT: add a1, a1, a0
+; ZVFHMIN-NEXT: slli a0, a0, 1
+; ZVFHMIN-NEXT: add a0, a0, a1
; ZVFHMIN-NEXT: add sp, sp, a0
; ZVFHMIN-NEXT: addi sp, sp, 16
; ZVFHMIN-NEXT: ret
@@ -602,12 +647,12 @@ define <vscale x 32 x half> @vfdiv_vv_nxv32f16_unmasked(<vscale x 32 x half> %va
; ZVFHMIN-NEXT: srli a2, a2, 2
; ZVFHMIN-NEXT: vsetvli a4, zero, e8, mf2, ta, ma
; ZVFHMIN-NEXT: vslidedown.vx v0, v24, a2
-; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a3, e16, m4, ta, ma
; ZVFHMIN-NEXT: addi a2, sp, 16
; ZVFHMIN-NEXT: vs8r.v v16, (a2) # Unknown-size Folded Spill
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v20
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12
-; ZVFHMIN-NEXT: vsetvli zero, a3, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v20, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfdiv.vv v16, v16, v24, v0.t
; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v16
@@ -615,11 +660,12 @@ define <vscale x 32 x half> @vfdiv_vv_nxv32f16_unmasked(<vscale x 32 x half> %va
; ZVFHMIN-NEXT: # %bb.1:
; ZVFHMIN-NEXT: mv a0, a1
; ZVFHMIN-NEXT: .LBB21_2:
-; ZVFHMIN-NEXT: addi a1, sp, 16
-; ZVFHMIN-NEXT: vl8r.v v24, (a1) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; ZVFHMIN-NEXT: addi a0, sp, 16
+; ZVFHMIN-NEXT: vl8r.v v24, (a0) # Unknown-size Folded Reload
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v24
; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfdiv.vv v16, v24, v16
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
@@ -643,7 +689,14 @@ define <vscale x 32 x half> @vfdiv_vf_nxv32f16(<vscale x 32 x half> %va, half %b
;
; ZVFHMIN-LABEL: vfdiv_vf_nxv32f16:
; ZVFHMIN: # %bb.0:
+; ZVFHMIN-NEXT: addi sp, sp, -16
+; ZVFHMIN-NEXT: .cfi_def_cfa_offset 16
+; ZVFHMIN-NEXT: csrr a1, vlenb
+; ZVFHMIN-NEXT: slli a1, a1, 3
+; ZVFHMIN-NEXT: sub sp, sp, a1
+; ZVFHMIN-NEXT: .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x08, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 8 * vlenb
; ZVFHMIN-NEXT: vmv1r.v v1, v0
+; ZVFHMIN-NEXT: vmv8r.v v16, v8
; ZVFHMIN-NEXT: fcvt.s.h fa5, fa0
; ZVFHMIN-NEXT: vsetvli a1, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfmv.v.f v24, fa5
@@ -658,23 +711,33 @@ define <vscale x 32 x half> @vfdiv_vf_nxv32f16(<vscale x 32 x half> %va, half %b
; ZVFHMIN-NEXT: srli a2, a2, 2
; ZVFHMIN-NEXT: vsetvli a4, zero, e8, mf2, ta, ma
; ZVFHMIN-NEXT: vslidedown.vx v0, v0, a2
+; ZVFHMIN-NEXT: vsetvli zero, a3, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v4, v0.t
+; ZVFHMIN-NEXT: addi a2, sp, 16
+; ZVFHMIN-NEXT: vs8r.v v8, (a2) # Unknown-size Folded Spill
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v20, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vfdiv.vv v24, v8, v24, v0.t
; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v4
-; ZVFHMIN-NEXT: vsetvli zero, a3, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vfdiv.vv v16, v16, v24, v0.t
-; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v16
+; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v24
; ZVFHMIN-NEXT: bltu a0, a1, .LBB22_2
; ZVFHMIN-NEXT: # %bb.1:
; ZVFHMIN-NEXT: mv a0, a1
; ZVFHMIN-NEXT: .LBB22_2:
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
; ZVFHMIN-NEXT: vmv1r.v v0, v1
+; ZVFHMIN-NEXT: addi a0, sp, 16
+; ZVFHMIN-NEXT: vl8r.v v24, (a0) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v24, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v4, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfdiv.vv v16, v16, v24, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
+; ZVFHMIN-NEXT: csrr a0, vlenb
+; ZVFHMIN-NEXT: slli a0, a0, 3
+; ZVFHMIN-NEXT: add sp, sp, a0
+; ZVFHMIN-NEXT: addi sp, sp, 16
; ZVFHMIN-NEXT: ret
%elt.head = insertelement <vscale x 32 x half> poison, half %b, i32 0
%vb = shufflevector <vscale x 32 x half> %elt.head, <vscale x 32 x half> poison, <vscale x 32 x i32> zeroinitializer
@@ -693,11 +756,11 @@ define <vscale x 32 x half> @vfdiv_vf_nxv32f16_unmasked(<vscale x 32 x half> %va
; ZVFHMIN: # %bb.0:
; ZVFHMIN-NEXT: fcvt.s.h fa5, fa0
; ZVFHMIN-NEXT: vsetvli a1, zero, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vfmv.v.f v16, fa5
+; ZVFHMIN-NEXT: vfmv.v.f v24, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfncvt.f.f.w v4, v16
+; ZVFHMIN-NEXT: vfncvt.f.f.w v4, v24
; ZVFHMIN-NEXT: vsetvli a1, zero, e8, m4, ta, ma
-; ZVFHMIN-NEXT: vmset.m v16
+; ZVFHMIN-NEXT: vmset.m v20
; ZVFHMIN-NEXT: csrr a2, vlenb
; ZVFHMIN-NEXT: slli a1, a2, 1
; ZVFHMIN-NEXT: sub a3, a0, a1
@@ -706,21 +769,23 @@ define <vscale x 32 x half> @vfdiv_vf_nxv32f16_unmasked(<vscale x 32 x half> %va
; ZVFHMIN-NEXT: and a3, a4, a3
; ZVFHMIN-NEXT: srli a2, a2, 2
; ZVFHMIN-NEXT: vsetvli a4, zero, e8, mf2, ta, ma
-; ZVFHMIN-NEXT: vslidedown.vx v0, v16, a2
-; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v12
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v4
-; ZVFHMIN-NEXT: vsetvli zero, a3, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vfdiv.vv v24, v24, v16, v0.t
+; ZVFHMIN-NEXT: vslidedown.vx v0, v20, a2
+; ZVFHMIN-NEXT: vsetvli zero, a3, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v4, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vfdiv.vv v16, v16, v24, v0.t
; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v24
+; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v16
; ZVFHMIN-NEXT: bltu a0, a1, .LBB23_2
; ZVFHMIN-NEXT: # %bb.1:
; ZVFHMIN-NEXT: mv a0, a1
; ZVFHMIN-NEXT: .LBB23_2:
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vfdiv.vv v16, v24, v16
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v4
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vfdiv.vv v16, v16, v24
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
; ZVFHMIN-NEXT: ret
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfmax-vp.ll b/llvm/test/CodeGen/RISCV/rvv/vfmax-vp.ll
index c3aaf743af170b..fcab080af93087 100644
--- a/llvm/test/CodeGen/RISCV/rvv/vfmax-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/vfmax-vp.ll
@@ -19,10 +19,10 @@ define <vscale x 1 x half> @vfmax_vv_nxv1f16(<vscale x 1 x half> %va, <vscale x
;
; ZVFHMIN-LABEL: vfmax_vv_nxv1f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfmax.vv v9, v9, v10, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -40,10 +40,10 @@ define <vscale x 1 x half> @vfmax_vv_nxv1f16_unmasked(<vscale x 1 x half> %va, <
;
; ZVFHMIN-LABEL: vfmax_vv_nxv1f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfmax.vv v9, v9, v10
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -65,10 +65,10 @@ define <vscale x 2 x half> @vfmax_vv_nxv2f16(<vscale x 2 x half> %va, <vscale x
;
; ZVFHMIN-LABEL: vfmax_vv_nxv2f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfmax.vv v9, v9, v10, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -86,10 +86,10 @@ define <vscale x 2 x half> @vfmax_vv_nxv2f16_unmasked(<vscale x 2 x half> %va, <
;
; ZVFHMIN-LABEL: vfmax_vv_nxv2f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfmax.vv v9, v9, v10
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -111,10 +111,10 @@ define <vscale x 4 x half> @vfmax_vv_nxv4f16(<vscale x 4 x half> %va, <vscale x
;
; ZVFHMIN-LABEL: vfmax_vv_nxv4f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfmax.vv v10, v12, v10, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -132,10 +132,10 @@ define <vscale x 4 x half> @vfmax_vv_nxv4f16_unmasked(<vscale x 4 x half> %va, <
;
; ZVFHMIN-LABEL: vfmax_vv_nxv4f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfmax.vv v10, v12, v10
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -157,10 +157,10 @@ define <vscale x 8 x half> @vfmax_vv_nxv8f16(<vscale x 8 x half> %va, <vscale x
;
; ZVFHMIN-LABEL: vfmax_vv_nxv8f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v10
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v10, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfmax.vv v12, v16, v12, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -178,10 +178,10 @@ define <vscale x 8 x half> @vfmax_vv_nxv8f16_unmasked(<vscale x 8 x half> %va, <
;
; ZVFHMIN-LABEL: vfmax_vv_nxv8f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v10
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfmax.vv v12, v16, v12
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -203,10 +203,10 @@ define <vscale x 16 x half> @vfmax_vv_nxv16f16(<vscale x 16 x half> %va, <vscale
;
; ZVFHMIN-LABEL: vfmax_vv_nxv16f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfmax.vv v16, v24, v16, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
@@ -224,10 +224,10 @@ define <vscale x 16 x half> @vfmax_vv_nxv16f16_unmasked(<vscale x 16 x half> %va
;
; ZVFHMIN-LABEL: vfmax_vv_nxv16f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12
; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfmax.vv v16, v24, v16
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
@@ -252,9 +252,13 @@ define <vscale x 32 x half> @vfmax_vv_nxv32f16(<vscale x 32 x half> %va, <vscale
; ZVFHMIN-NEXT: addi sp, sp, -16
; ZVFHMIN-NEXT: .cfi_def_cfa_offset 16
; ZVFHMIN-NEXT: csrr a1, vlenb
+; ZVFHMIN-NEXT: li a2, 0
; ZVFHMIN-NEXT: slli a1, a1, 3
+; ZVFHMIN-NEXT: add a2, a2, a1
+; ZVFHMIN-NEXT: slli a1, a1, 1
+; ZVFHMIN-NEXT: add a1, a1, a2
; ZVFHMIN-NEXT: sub sp, sp, a1
-; ZVFHMIN-NEXT: .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x08, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 8 * vlenb
+; ZVFHMIN-NEXT: .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x18, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 24 * vlenb
; ZVFHMIN-NEXT: vmv1r.v v1, v0
; ZVFHMIN-NEXT: csrr a2, vlenb
; ZVFHMIN-NEXT: slli a1, a2, 1
@@ -265,30 +269,61 @@ define <vscale x 32 x half> @vfmax_vv_nxv32f16(<vscale x 32 x half> %va, <vscale
; ZVFHMIN-NEXT: srli a2, a2, 2
; ZVFHMIN-NEXT: vsetvli a4, zero, e8, mf2, ta, ma
; ZVFHMIN-NEXT: vslidedown.vx v0, v0, a2
-; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a3, e16, m4, ta, ma
+; ZVFHMIN-NEXT: csrr a2, vlenb
+; ZVFHMIN-NEXT: slli a2, a2, 3
+; ZVFHMIN-NEXT: add a2, sp, a2
+; ZVFHMIN-NEXT: addi a2, a2, 16
+; ZVFHMIN-NEXT: vs8r.v v16, (a2) # Unknown-size Folded Spill
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v20, v0.t
; ZVFHMIN-NEXT: addi a2, sp, 16
+; ZVFHMIN-NEXT: vs8r.v v24, (a2) # Unknown-size Folded Spill
+; ZVFHMIN-NEXT: vmv4r.v v16, v8
+; ZVFHMIN-NEXT: csrr a2, vlenb
+; ZVFHMIN-NEXT: slli a2, a2, 4
+; ZVFHMIN-NEXT: add a2, sp, a2
+; ZVFHMIN-NEXT: addi a2, a2, 16
; ZVFHMIN-NEXT: vs8r.v v16, (a2) # Unknown-size Folded Spill
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v20
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12
-; ZVFHMIN-NEXT: vsetvli zero, a3, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vfmax.vv v16, v16, v24, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; ZVFHMIN-NEXT: addi a2, sp, 16
+; ZVFHMIN-NEXT: vl8r.v v8, (a2) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfmax.vv v16, v16, v8, v0.t
; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v16
; ZVFHMIN-NEXT: bltu a0, a1, .LBB10_2
; ZVFHMIN-NEXT: # %bb.1:
; ZVFHMIN-NEXT: mv a0, a1
; ZVFHMIN-NEXT: .LBB10_2:
-; ZVFHMIN-NEXT: addi a1, sp, 16
-; ZVFHMIN-NEXT: vl8r.v v24, (a1) # Unknown-size Folded Reload
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v24
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
; ZVFHMIN-NEXT: vmv1r.v v0, v1
-; ZVFHMIN-NEXT: vfmax.vv v16, v24, v16, v0.t
+; ZVFHMIN-NEXT: csrr a0, vlenb
+; ZVFHMIN-NEXT: slli a0, a0, 3
+; ZVFHMIN-NEXT: add a0, sp, a0
+; ZVFHMIN-NEXT: addi a0, a0, 16
+; ZVFHMIN-NEXT: vl8r.v v16, (a0) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v16, v0.t
+; ZVFHMIN-NEXT: addi a0, sp, 16
+; ZVFHMIN-NEXT: vs8r.v v24, (a0) # Unknown-size Folded Spill
+; ZVFHMIN-NEXT: csrr a0, vlenb
+; ZVFHMIN-NEXT: slli a0, a0, 4
+; ZVFHMIN-NEXT: add a0, sp, a0
+; ZVFHMIN-NEXT: addi a0, a0, 16
+; ZVFHMIN-NEXT: vl8r.v v16, (a0) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v16, v0.t
+; ZVFHMIN-NEXT: vmv8r.v v16, v24
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; ZVFHMIN-NEXT: addi a0, sp, 16
+; ZVFHMIN-NEXT: vl8r.v v24, (a0) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfmax.vv v16, v16, v24, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
; ZVFHMIN-NEXT: csrr a0, vlenb
+; ZVFHMIN-NEXT: li a1, 0
; ZVFHMIN-NEXT: slli a0, a0, 3
+; ZVFHMIN-NEXT: add a1, a1, a0
+; ZVFHMIN-NEXT: slli a0, a0, 1
+; ZVFHMIN-NEXT: add a0, a0, a1
; ZVFHMIN-NEXT: add sp, sp, a0
; ZVFHMIN-NEXT: addi sp, sp, 16
; ZVFHMIN-NEXT: ret
@@ -322,12 +357,12 @@ define <vscale x 32 x half> @vfmax_vv_nxv32f16_unmasked(<vscale x 32 x half> %va
; ZVFHMIN-NEXT: srli a2, a2, 2
; ZVFHMIN-NEXT: vsetvli a4, zero, e8, mf2, ta, ma
; ZVFHMIN-NEXT: vslidedown.vx v0, v24, a2
-; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a3, e16, m4, ta, ma
; ZVFHMIN-NEXT: addi a2, sp, 16
; ZVFHMIN-NEXT: vs8r.v v16, (a2) # Unknown-size Folded Spill
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v20
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12
-; ZVFHMIN-NEXT: vsetvli zero, a3, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v20, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfmax.vv v16, v16, v24, v0.t
; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v16
@@ -335,11 +370,12 @@ define <vscale x 32 x half> @vfmax_vv_nxv32f16_unmasked(<vscale x 32 x half> %va
; ZVFHMIN-NEXT: # %bb.1:
; ZVFHMIN-NEXT: mv a0, a1
; ZVFHMIN-NEXT: .LBB11_2:
-; ZVFHMIN-NEXT: addi a1, sp, 16
-; ZVFHMIN-NEXT: vl8r.v v24, (a1) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; ZVFHMIN-NEXT: addi a0, sp, 16
+; ZVFHMIN-NEXT: vl8r.v v24, (a0) # Unknown-size Folded Reload
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v24
; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfmax.vv v16, v24, v16
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.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 f18fa85e68d1be..104dbb13d511d7 100644
--- a/llvm/test/CodeGen/RISCV/rvv/vfmin-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/vfmin-vp.ll
@@ -19,10 +19,10 @@ define <vscale x 1 x half> @vfmin_vv_nxv1f16(<vscale x 1 x half> %va, <vscale x
;
; ZVFHMIN-LABEL: vfmin_vv_nxv1f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfmin.vv v9, v9, v10, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -40,10 +40,10 @@ define <vscale x 1 x half> @vfmin_vv_nxv1f16_unmasked(<vscale x 1 x half> %va, <
;
; ZVFHMIN-LABEL: vfmin_vv_nxv1f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfmin.vv v9, v9, v10
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -65,10 +65,10 @@ define <vscale x 2 x half> @vfmin_vv_nxv2f16(<vscale x 2 x half> %va, <vscale x
;
; ZVFHMIN-LABEL: vfmin_vv_nxv2f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfmin.vv v9, v9, v10, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -86,10 +86,10 @@ define <vscale x 2 x half> @vfmin_vv_nxv2f16_unmasked(<vscale x 2 x half> %va, <
;
; ZVFHMIN-LABEL: vfmin_vv_nxv2f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfmin.vv v9, v9, v10
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -111,10 +111,10 @@ define <vscale x 4 x half> @vfmin_vv_nxv4f16(<vscale x 4 x half> %va, <vscale x
;
; ZVFHMIN-LABEL: vfmin_vv_nxv4f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfmin.vv v10, v12, v10, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -132,10 +132,10 @@ define <vscale x 4 x half> @vfmin_vv_nxv4f16_unmasked(<vscale x 4 x half> %va, <
;
; ZVFHMIN-LABEL: vfmin_vv_nxv4f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfmin.vv v10, v12, v10
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -157,10 +157,10 @@ define <vscale x 8 x half> @vfmin_vv_nxv8f16(<vscale x 8 x half> %va, <vscale x
;
; ZVFHMIN-LABEL: vfmin_vv_nxv8f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v10
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v10, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfmin.vv v12, v16, v12, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -178,10 +178,10 @@ define <vscale x 8 x half> @vfmin_vv_nxv8f16_unmasked(<vscale x 8 x half> %va, <
;
; ZVFHMIN-LABEL: vfmin_vv_nxv8f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v10
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfmin.vv v12, v16, v12
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -203,10 +203,10 @@ define <vscale x 16 x half> @vfmin_vv_nxv16f16(<vscale x 16 x half> %va, <vscale
;
; ZVFHMIN-LABEL: vfmin_vv_nxv16f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfmin.vv v16, v24, v16, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
@@ -224,10 +224,10 @@ define <vscale x 16 x half> @vfmin_vv_nxv16f16_unmasked(<vscale x 16 x half> %va
;
; ZVFHMIN-LABEL: vfmin_vv_nxv16f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12
; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfmin.vv v16, v24, v16
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
@@ -252,9 +252,13 @@ define <vscale x 32 x half> @vfmin_vv_nxv32f16(<vscale x 32 x half> %va, <vscale
; ZVFHMIN-NEXT: addi sp, sp, -16
; ZVFHMIN-NEXT: .cfi_def_cfa_offset 16
; ZVFHMIN-NEXT: csrr a1, vlenb
+; ZVFHMIN-NEXT: li a2, 0
; ZVFHMIN-NEXT: slli a1, a1, 3
+; ZVFHMIN-NEXT: add a2, a2, a1
+; ZVFHMIN-NEXT: slli a1, a1, 1
+; ZVFHMIN-NEXT: add a1, a1, a2
; ZVFHMIN-NEXT: sub sp, sp, a1
-; ZVFHMIN-NEXT: .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x08, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 8 * vlenb
+; ZVFHMIN-NEXT: .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x18, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 24 * vlenb
; ZVFHMIN-NEXT: vmv1r.v v1, v0
; ZVFHMIN-NEXT: csrr a2, vlenb
; ZVFHMIN-NEXT: slli a1, a2, 1
@@ -265,30 +269,61 @@ define <vscale x 32 x half> @vfmin_vv_nxv32f16(<vscale x 32 x half> %va, <vscale
; ZVFHMIN-NEXT: srli a2, a2, 2
; ZVFHMIN-NEXT: vsetvli a4, zero, e8, mf2, ta, ma
; ZVFHMIN-NEXT: vslidedown.vx v0, v0, a2
-; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a3, e16, m4, ta, ma
+; ZVFHMIN-NEXT: csrr a2, vlenb
+; ZVFHMIN-NEXT: slli a2, a2, 3
+; ZVFHMIN-NEXT: add a2, sp, a2
+; ZVFHMIN-NEXT: addi a2, a2, 16
+; ZVFHMIN-NEXT: vs8r.v v16, (a2) # Unknown-size Folded Spill
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v20, v0.t
; ZVFHMIN-NEXT: addi a2, sp, 16
+; ZVFHMIN-NEXT: vs8r.v v24, (a2) # Unknown-size Folded Spill
+; ZVFHMIN-NEXT: vmv4r.v v16, v8
+; ZVFHMIN-NEXT: csrr a2, vlenb
+; ZVFHMIN-NEXT: slli a2, a2, 4
+; ZVFHMIN-NEXT: add a2, sp, a2
+; ZVFHMIN-NEXT: addi a2, a2, 16
; ZVFHMIN-NEXT: vs8r.v v16, (a2) # Unknown-size Folded Spill
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v20
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12
-; ZVFHMIN-NEXT: vsetvli zero, a3, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vfmin.vv v16, v16, v24, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; ZVFHMIN-NEXT: addi a2, sp, 16
+; ZVFHMIN-NEXT: vl8r.v v8, (a2) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfmin.vv v16, v16, v8, v0.t
; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v16
; ZVFHMIN-NEXT: bltu a0, a1, .LBB10_2
; ZVFHMIN-NEXT: # %bb.1:
; ZVFHMIN-NEXT: mv a0, a1
; ZVFHMIN-NEXT: .LBB10_2:
-; ZVFHMIN-NEXT: addi a1, sp, 16
-; ZVFHMIN-NEXT: vl8r.v v24, (a1) # Unknown-size Folded Reload
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v24
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
; ZVFHMIN-NEXT: vmv1r.v v0, v1
-; ZVFHMIN-NEXT: vfmin.vv v16, v24, v16, v0.t
+; ZVFHMIN-NEXT: csrr a0, vlenb
+; ZVFHMIN-NEXT: slli a0, a0, 3
+; ZVFHMIN-NEXT: add a0, sp, a0
+; ZVFHMIN-NEXT: addi a0, a0, 16
+; ZVFHMIN-NEXT: vl8r.v v16, (a0) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v16, v0.t
+; ZVFHMIN-NEXT: addi a0, sp, 16
+; ZVFHMIN-NEXT: vs8r.v v24, (a0) # Unknown-size Folded Spill
+; ZVFHMIN-NEXT: csrr a0, vlenb
+; ZVFHMIN-NEXT: slli a0, a0, 4
+; ZVFHMIN-NEXT: add a0, sp, a0
+; ZVFHMIN-NEXT: addi a0, a0, 16
+; ZVFHMIN-NEXT: vl8r.v v16, (a0) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v16, v0.t
+; ZVFHMIN-NEXT: vmv8r.v v16, v24
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; ZVFHMIN-NEXT: addi a0, sp, 16
+; ZVFHMIN-NEXT: vl8r.v v24, (a0) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfmin.vv v16, v16, v24, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
; ZVFHMIN-NEXT: csrr a0, vlenb
+; ZVFHMIN-NEXT: li a1, 0
; ZVFHMIN-NEXT: slli a0, a0, 3
+; ZVFHMIN-NEXT: add a1, a1, a0
+; ZVFHMIN-NEXT: slli a0, a0, 1
+; ZVFHMIN-NEXT: add a0, a0, a1
; ZVFHMIN-NEXT: add sp, sp, a0
; ZVFHMIN-NEXT: addi sp, sp, 16
; ZVFHMIN-NEXT: ret
@@ -322,12 +357,12 @@ define <vscale x 32 x half> @vfmin_vv_nxv32f16_unmasked(<vscale x 32 x half> %va
; ZVFHMIN-NEXT: srli a2, a2, 2
; ZVFHMIN-NEXT: vsetvli a4, zero, e8, mf2, ta, ma
; ZVFHMIN-NEXT: vslidedown.vx v0, v24, a2
-; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a3, e16, m4, ta, ma
; ZVFHMIN-NEXT: addi a2, sp, 16
; ZVFHMIN-NEXT: vs8r.v v16, (a2) # Unknown-size Folded Spill
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v20
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12
-; ZVFHMIN-NEXT: vsetvli zero, a3, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v20, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfmin.vv v16, v16, v24, v0.t
; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v16
@@ -335,11 +370,12 @@ define <vscale x 32 x half> @vfmin_vv_nxv32f16_unmasked(<vscale x 32 x half> %va
; ZVFHMIN-NEXT: # %bb.1:
; ZVFHMIN-NEXT: mv a0, a1
; ZVFHMIN-NEXT: .LBB11_2:
-; ZVFHMIN-NEXT: addi a1, sp, 16
-; ZVFHMIN-NEXT: vl8r.v v24, (a1) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; ZVFHMIN-NEXT: addi a0, sp, 16
+; ZVFHMIN-NEXT: vl8r.v v24, (a0) # Unknown-size Folded Reload
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v24
; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfmin.vv v16, v24, v16
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfmul-vp.ll b/llvm/test/CodeGen/RISCV/rvv/vfmul-vp.ll
index 46b14153447cf7..4e99c90391ccae 100644
--- a/llvm/test/CodeGen/RISCV/rvv/vfmul-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/vfmul-vp.ll
@@ -19,10 +19,10 @@ define <vscale x 1 x half> @vfmul_vv_nxv1f16(<vscale x 1 x half> %va, <vscale x
;
; ZVFHMIN-LABEL: vfmul_vv_nxv1f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfmul.vv v9, v9, v10, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -40,10 +40,10 @@ define <vscale x 1 x half> @vfmul_vv_nxv1f16_unmasked(<vscale x 1 x half> %va, <
;
; ZVFHMIN-LABEL: vfmul_vv_nxv1f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfmul.vv v9, v9, v10
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -68,9 +68,10 @@ define <vscale x 1 x half> @vfmul_vf_nxv1f16(<vscale x 1 x half> %va, half %b, <
; ZVFHMIN-NEXT: vfmv.v.f v9, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfmul.vv v9, v9, v8, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -95,9 +96,10 @@ define <vscale x 1 x half> @vfmul_vf_nxv1f16_unmasked(<vscale x 1 x half> %va, h
; ZVFHMIN-NEXT: vfmv.v.f v9, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v10, v9
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfmul.vv v9, v9, v8
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -121,10 +123,10 @@ define <vscale x 2 x half> @vfmul_vv_nxv2f16(<vscale x 2 x half> %va, <vscale x
;
; ZVFHMIN-LABEL: vfmul_vv_nxv2f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfmul.vv v9, v9, v10, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -142,10 +144,10 @@ define <vscale x 2 x half> @vfmul_vv_nxv2f16_unmasked(<vscale x 2 x half> %va, <
;
; ZVFHMIN-LABEL: vfmul_vv_nxv2f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfmul.vv v9, v9, v10
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -170,9 +172,10 @@ define <vscale x 2 x half> @vfmul_vf_nxv2f16(<vscale x 2 x half> %va, half %b, <
; ZVFHMIN-NEXT: vfmv.v.f v9, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfmul.vv v9, v9, v8, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -197,9 +200,10 @@ define <vscale x 2 x half> @vfmul_vf_nxv2f16_unmasked(<vscale x 2 x half> %va, h
; ZVFHMIN-NEXT: vfmv.v.f v9, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v10, v9
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfmul.vv v9, v9, v8
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -223,10 +227,10 @@ define <vscale x 4 x half> @vfmul_vv_nxv4f16(<vscale x 4 x half> %va, <vscale x
;
; ZVFHMIN-LABEL: vfmul_vv_nxv4f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfmul.vv v10, v12, v10, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -244,10 +248,10 @@ define <vscale x 4 x half> @vfmul_vv_nxv4f16_unmasked(<vscale x 4 x half> %va, <
;
; ZVFHMIN-LABEL: vfmul_vv_nxv4f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfmul.vv v10, v12, v10
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -272,9 +276,10 @@ define <vscale x 4 x half> @vfmul_vf_nxv4f16(<vscale x 4 x half> %va, half %b, <
; ZVFHMIN-NEXT: vfmv.v.f v10, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v9, v10
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v9
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v9, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfmul.vv v10, v10, v12, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -299,9 +304,10 @@ define <vscale x 4 x half> @vfmul_vf_nxv4f16_unmasked(<vscale x 4 x half> %va, h
; ZVFHMIN-NEXT: vfmv.v.f v10, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v9, v10
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v9
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfmul.vv v10, v10, v12
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -325,10 +331,10 @@ define <vscale x 8 x half> @vfmul_vv_nxv8f16(<vscale x 8 x half> %va, <vscale x
;
; ZVFHMIN-LABEL: vfmul_vv_nxv8f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v10
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v10, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfmul.vv v12, v16, v12, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -346,10 +352,10 @@ define <vscale x 8 x half> @vfmul_vv_nxv8f16_unmasked(<vscale x 8 x half> %va, <
;
; ZVFHMIN-LABEL: vfmul_vv_nxv8f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v10
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfmul.vv v12, v16, v12
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -374,9 +380,10 @@ define <vscale x 8 x half> @vfmul_vf_nxv8f16(<vscale x 8 x half> %va, half %b, <
; ZVFHMIN-NEXT: vfmv.v.f v12, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v10, v12
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v10, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfmul.vv v12, v12, v16, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -401,9 +408,10 @@ define <vscale x 8 x half> @vfmul_vf_nxv8f16_unmasked(<vscale x 8 x half> %va, h
; ZVFHMIN-NEXT: vfmv.v.f v12, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v10, v12
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfmul.vv v12, v12, v16
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -427,10 +435,10 @@ define <vscale x 16 x half> @vfmul_vv_nxv16f16(<vscale x 16 x half> %va, <vscale
;
; ZVFHMIN-LABEL: vfmul_vv_nxv16f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfmul.vv v16, v24, v16, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
@@ -448,10 +456,10 @@ define <vscale x 16 x half> @vfmul_vv_nxv16f16_unmasked(<vscale x 16 x half> %va
;
; ZVFHMIN-LABEL: vfmul_vv_nxv16f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12
; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfmul.vv v16, v24, v16
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
@@ -476,9 +484,10 @@ define <vscale x 16 x half> @vfmul_vf_nxv16f16(<vscale x 16 x half> %va, half %b
; ZVFHMIN-NEXT: vfmv.v.f v16, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v16
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v12
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v12, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfmul.vv v16, v16, v24, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
@@ -503,9 +512,10 @@ define <vscale x 16 x half> @vfmul_vf_nxv16f16_unmasked(<vscale x 16 x half> %va
; ZVFHMIN-NEXT: vfmv.v.f v16, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v16
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v12
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfmul.vv v16, v16, v24
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
@@ -532,9 +542,13 @@ define <vscale x 32 x half> @vfmul_vv_nxv32f16(<vscale x 32 x half> %va, <vscale
; ZVFHMIN-NEXT: addi sp, sp, -16
; ZVFHMIN-NEXT: .cfi_def_cfa_offset 16
; ZVFHMIN-NEXT: csrr a1, vlenb
+; ZVFHMIN-NEXT: li a2, 0
; ZVFHMIN-NEXT: slli a1, a1, 3
+; ZVFHMIN-NEXT: add a2, a2, a1
+; ZVFHMIN-NEXT: slli a1, a1, 1
+; ZVFHMIN-NEXT: add a1, a1, a2
; ZVFHMIN-NEXT: sub sp, sp, a1
-; ZVFHMIN-NEXT: .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x08, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 8 * vlenb
+; ZVFHMIN-NEXT: .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x18, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 24 * vlenb
; ZVFHMIN-NEXT: vmv1r.v v1, v0
; ZVFHMIN-NEXT: csrr a2, vlenb
; ZVFHMIN-NEXT: slli a1, a2, 1
@@ -545,30 +559,61 @@ define <vscale x 32 x half> @vfmul_vv_nxv32f16(<vscale x 32 x half> %va, <vscale
; ZVFHMIN-NEXT: srli a2, a2, 2
; ZVFHMIN-NEXT: vsetvli a4, zero, e8, mf2, ta, ma
; ZVFHMIN-NEXT: vslidedown.vx v0, v0, a2
-; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a3, e16, m4, ta, ma
+; ZVFHMIN-NEXT: csrr a2, vlenb
+; ZVFHMIN-NEXT: slli a2, a2, 3
+; ZVFHMIN-NEXT: add a2, sp, a2
+; ZVFHMIN-NEXT: addi a2, a2, 16
+; ZVFHMIN-NEXT: vs8r.v v16, (a2) # Unknown-size Folded Spill
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v20, v0.t
; ZVFHMIN-NEXT: addi a2, sp, 16
+; ZVFHMIN-NEXT: vs8r.v v24, (a2) # Unknown-size Folded Spill
+; ZVFHMIN-NEXT: vmv4r.v v16, v8
+; ZVFHMIN-NEXT: csrr a2, vlenb
+; ZVFHMIN-NEXT: slli a2, a2, 4
+; ZVFHMIN-NEXT: add a2, sp, a2
+; ZVFHMIN-NEXT: addi a2, a2, 16
; ZVFHMIN-NEXT: vs8r.v v16, (a2) # Unknown-size Folded Spill
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v20
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12
-; ZVFHMIN-NEXT: vsetvli zero, a3, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vfmul.vv v16, v16, v24, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; ZVFHMIN-NEXT: addi a2, sp, 16
+; ZVFHMIN-NEXT: vl8r.v v8, (a2) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfmul.vv v16, v16, v8, v0.t
; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v16
; ZVFHMIN-NEXT: bltu a0, a1, .LBB20_2
; ZVFHMIN-NEXT: # %bb.1:
; ZVFHMIN-NEXT: mv a0, a1
; ZVFHMIN-NEXT: .LBB20_2:
-; ZVFHMIN-NEXT: addi a1, sp, 16
-; ZVFHMIN-NEXT: vl8r.v v24, (a1) # Unknown-size Folded Reload
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v24
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
; ZVFHMIN-NEXT: vmv1r.v v0, v1
-; ZVFHMIN-NEXT: vfmul.vv v16, v24, v16, v0.t
+; ZVFHMIN-NEXT: csrr a0, vlenb
+; ZVFHMIN-NEXT: slli a0, a0, 3
+; ZVFHMIN-NEXT: add a0, sp, a0
+; ZVFHMIN-NEXT: addi a0, a0, 16
+; ZVFHMIN-NEXT: vl8r.v v16, (a0) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v16, v0.t
+; ZVFHMIN-NEXT: addi a0, sp, 16
+; ZVFHMIN-NEXT: vs8r.v v24, (a0) # Unknown-size Folded Spill
+; ZVFHMIN-NEXT: csrr a0, vlenb
+; ZVFHMIN-NEXT: slli a0, a0, 4
+; ZVFHMIN-NEXT: add a0, sp, a0
+; ZVFHMIN-NEXT: addi a0, a0, 16
+; ZVFHMIN-NEXT: vl8r.v v16, (a0) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v16, v0.t
+; ZVFHMIN-NEXT: vmv8r.v v16, v24
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; ZVFHMIN-NEXT: addi a0, sp, 16
+; ZVFHMIN-NEXT: vl8r.v v24, (a0) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfmul.vv v16, v16, v24, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
; ZVFHMIN-NEXT: csrr a0, vlenb
+; ZVFHMIN-NEXT: li a1, 0
; ZVFHMIN-NEXT: slli a0, a0, 3
+; ZVFHMIN-NEXT: add a1, a1, a0
+; ZVFHMIN-NEXT: slli a0, a0, 1
+; ZVFHMIN-NEXT: add a0, a0, a1
; ZVFHMIN-NEXT: add sp, sp, a0
; ZVFHMIN-NEXT: addi sp, sp, 16
; ZVFHMIN-NEXT: ret
@@ -602,12 +647,12 @@ define <vscale x 32 x half> @vfmul_vv_nxv32f16_unmasked(<vscale x 32 x half> %va
; ZVFHMIN-NEXT: srli a2, a2, 2
; ZVFHMIN-NEXT: vsetvli a4, zero, e8, mf2, ta, ma
; ZVFHMIN-NEXT: vslidedown.vx v0, v24, a2
-; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a3, e16, m4, ta, ma
; ZVFHMIN-NEXT: addi a2, sp, 16
; ZVFHMIN-NEXT: vs8r.v v16, (a2) # Unknown-size Folded Spill
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v20
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12
-; ZVFHMIN-NEXT: vsetvli zero, a3, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v20, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfmul.vv v16, v16, v24, v0.t
; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v16
@@ -615,11 +660,12 @@ define <vscale x 32 x half> @vfmul_vv_nxv32f16_unmasked(<vscale x 32 x half> %va
; ZVFHMIN-NEXT: # %bb.1:
; ZVFHMIN-NEXT: mv a0, a1
; ZVFHMIN-NEXT: .LBB21_2:
-; ZVFHMIN-NEXT: addi a1, sp, 16
-; ZVFHMIN-NEXT: vl8r.v v24, (a1) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; ZVFHMIN-NEXT: addi a0, sp, 16
+; ZVFHMIN-NEXT: vl8r.v v24, (a0) # Unknown-size Folded Reload
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v24
; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfmul.vv v16, v24, v16
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
@@ -643,7 +689,14 @@ define <vscale x 32 x half> @vfmul_vf_nxv32f16(<vscale x 32 x half> %va, half %b
;
; ZVFHMIN-LABEL: vfmul_vf_nxv32f16:
; ZVFHMIN: # %bb.0:
+; ZVFHMIN-NEXT: addi sp, sp, -16
+; ZVFHMIN-NEXT: .cfi_def_cfa_offset 16
+; ZVFHMIN-NEXT: csrr a1, vlenb
+; ZVFHMIN-NEXT: slli a1, a1, 3
+; ZVFHMIN-NEXT: sub sp, sp, a1
+; ZVFHMIN-NEXT: .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x08, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 8 * vlenb
; ZVFHMIN-NEXT: vmv1r.v v1, v0
+; ZVFHMIN-NEXT: vmv8r.v v16, v8
; ZVFHMIN-NEXT: fcvt.s.h fa5, fa0
; ZVFHMIN-NEXT: vsetvli a1, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfmv.v.f v24, fa5
@@ -658,23 +711,33 @@ define <vscale x 32 x half> @vfmul_vf_nxv32f16(<vscale x 32 x half> %va, half %b
; ZVFHMIN-NEXT: srli a2, a2, 2
; ZVFHMIN-NEXT: vsetvli a4, zero, e8, mf2, ta, ma
; ZVFHMIN-NEXT: vslidedown.vx v0, v0, a2
+; ZVFHMIN-NEXT: vsetvli zero, a3, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v4, v0.t
+; ZVFHMIN-NEXT: addi a2, sp, 16
+; ZVFHMIN-NEXT: vs8r.v v8, (a2) # Unknown-size Folded Spill
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v20, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vfmul.vv v24, v8, v24, v0.t
; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v4
-; ZVFHMIN-NEXT: vsetvli zero, a3, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vfmul.vv v16, v16, v24, v0.t
-; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v16
+; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v24
; ZVFHMIN-NEXT: bltu a0, a1, .LBB22_2
; ZVFHMIN-NEXT: # %bb.1:
; ZVFHMIN-NEXT: mv a0, a1
; ZVFHMIN-NEXT: .LBB22_2:
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
; ZVFHMIN-NEXT: vmv1r.v v0, v1
+; ZVFHMIN-NEXT: addi a0, sp, 16
+; ZVFHMIN-NEXT: vl8r.v v24, (a0) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v24, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v4, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfmul.vv v16, v16, v24, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
+; ZVFHMIN-NEXT: csrr a0, vlenb
+; ZVFHMIN-NEXT: slli a0, a0, 3
+; ZVFHMIN-NEXT: add sp, sp, a0
+; ZVFHMIN-NEXT: addi sp, sp, 16
; ZVFHMIN-NEXT: ret
%elt.head = insertelement <vscale x 32 x half> poison, half %b, i32 0
%vb = shufflevector <vscale x 32 x half> %elt.head, <vscale x 32 x half> poison, <vscale x 32 x i32> zeroinitializer
@@ -693,11 +756,11 @@ define <vscale x 32 x half> @vfmul_vf_nxv32f16_unmasked(<vscale x 32 x half> %va
; ZVFHMIN: # %bb.0:
; ZVFHMIN-NEXT: fcvt.s.h fa5, fa0
; ZVFHMIN-NEXT: vsetvli a1, zero, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vfmv.v.f v16, fa5
+; ZVFHMIN-NEXT: vfmv.v.f v24, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfncvt.f.f.w v4, v16
+; ZVFHMIN-NEXT: vfncvt.f.f.w v4, v24
; ZVFHMIN-NEXT: vsetvli a1, zero, e8, m4, ta, ma
-; ZVFHMIN-NEXT: vmset.m v16
+; ZVFHMIN-NEXT: vmset.m v20
; ZVFHMIN-NEXT: csrr a2, vlenb
; ZVFHMIN-NEXT: slli a1, a2, 1
; ZVFHMIN-NEXT: sub a3, a0, a1
@@ -706,21 +769,23 @@ define <vscale x 32 x half> @vfmul_vf_nxv32f16_unmasked(<vscale x 32 x half> %va
; ZVFHMIN-NEXT: and a3, a4, a3
; ZVFHMIN-NEXT: srli a2, a2, 2
; ZVFHMIN-NEXT: vsetvli a4, zero, e8, mf2, ta, ma
-; ZVFHMIN-NEXT: vslidedown.vx v0, v16, a2
-; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v12
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v4
-; ZVFHMIN-NEXT: vsetvli zero, a3, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vfmul.vv v24, v24, v16, v0.t
+; ZVFHMIN-NEXT: vslidedown.vx v0, v20, a2
+; ZVFHMIN-NEXT: vsetvli zero, a3, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v4, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vfmul.vv v16, v16, v24, v0.t
; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v24
+; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v16
; ZVFHMIN-NEXT: bltu a0, a1, .LBB23_2
; ZVFHMIN-NEXT: # %bb.1:
; ZVFHMIN-NEXT: mv a0, a1
; ZVFHMIN-NEXT: .LBB23_2:
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vfmul.vv v16, v24, v16
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v4
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vfmul.vv v16, v16, v24
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
; ZVFHMIN-NEXT: ret
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfneg-vp.ll b/llvm/test/CodeGen/RISCV/rvv/vfneg-vp.ll
index ef08865100f14b..b8ab320736516a 100644
--- a/llvm/test/CodeGen/RISCV/rvv/vfneg-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/vfneg-vp.ll
@@ -19,9 +19,9 @@ define <vscale x 1 x half> @vfneg_vv_nxv1f16(<vscale x 1 x half> %va, <vscale x
;
; ZVFHMIN-LABEL: vfneg_vv_nxv1f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfneg.v v9, v9, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -39,9 +39,9 @@ define <vscale x 1 x half> @vfneg_vv_nxv1f16_unmasked(<vscale x 1 x half> %va, i
;
; ZVFHMIN-LABEL: vfneg_vv_nxv1f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfneg.v v9, v9
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -63,9 +63,9 @@ define <vscale x 2 x half> @vfneg_vv_nxv2f16(<vscale x 2 x half> %va, <vscale x
;
; ZVFHMIN-LABEL: vfneg_vv_nxv2f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfneg.v v9, v9, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -83,9 +83,9 @@ define <vscale x 2 x half> @vfneg_vv_nxv2f16_unmasked(<vscale x 2 x half> %va, i
;
; ZVFHMIN-LABEL: vfneg_vv_nxv2f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfneg.v v9, v9
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -107,9 +107,9 @@ define <vscale x 4 x half> @vfneg_vv_nxv4f16(<vscale x 4 x half> %va, <vscale x
;
; ZVFHMIN-LABEL: vfneg_vv_nxv4f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfneg.v v10, v10, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -127,9 +127,9 @@ define <vscale x 4 x half> @vfneg_vv_nxv4f16_unmasked(<vscale x 4 x half> %va, i
;
; ZVFHMIN-LABEL: vfneg_vv_nxv4f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfneg.v v10, v10
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -151,9 +151,9 @@ define <vscale x 8 x half> @vfneg_vv_nxv8f16(<vscale x 8 x half> %va, <vscale x
;
; ZVFHMIN-LABEL: vfneg_vv_nxv8f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfneg.v v12, v12, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -171,9 +171,9 @@ define <vscale x 8 x half> @vfneg_vv_nxv8f16_unmasked(<vscale x 8 x half> %va, i
;
; ZVFHMIN-LABEL: vfneg_vv_nxv8f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfneg.v v12, v12
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -195,9 +195,9 @@ define <vscale x 16 x half> @vfneg_vv_nxv16f16(<vscale x 16 x half> %va, <vscale
;
; ZVFHMIN-LABEL: vfneg_vv_nxv16f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfneg.v v16, v16, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
@@ -215,9 +215,9 @@ define <vscale x 16 x half> @vfneg_vv_nxv16f16_unmasked(<vscale x 16 x half> %va
;
; ZVFHMIN-LABEL: vfneg_vv_nxv16f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfneg.v v16, v16
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
@@ -249,9 +249,9 @@ define <vscale x 32 x half> @vfneg_vv_nxv32f16(<vscale x 32 x half> %va, <vscale
; ZVFHMIN-NEXT: srli a2, a2, 2
; ZVFHMIN-NEXT: vsetvli a4, zero, e8, mf2, ta, ma
; ZVFHMIN-NEXT: vslidedown.vx v0, v0, a2
-; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v12
-; ZVFHMIN-NEXT: vsetvli zero, a3, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a3, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v12, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfneg.v v24, v24, v0.t
; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v24
@@ -259,9 +259,10 @@ define <vscale x 32 x half> @vfneg_vv_nxv32f16(<vscale x 32 x half> %va, <vscale
; ZVFHMIN-NEXT: # %bb.1:
; ZVFHMIN-NEXT: mv a0, a1
; ZVFHMIN-NEXT: .LBB10_2:
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
; ZVFHMIN-NEXT: vmv1r.v v0, v16
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfneg.v v16, v24, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
@@ -290,9 +291,9 @@ define <vscale x 32 x half> @vfneg_vv_nxv32f16_unmasked(<vscale x 32 x half> %va
; ZVFHMIN-NEXT: srli a2, a2, 2
; ZVFHMIN-NEXT: vsetvli a4, zero, e8, mf2, ta, ma
; ZVFHMIN-NEXT: vslidedown.vx v0, v16, a2
-; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12
-; ZVFHMIN-NEXT: vsetvli zero, a3, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a3, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfneg.v v16, v16, v0.t
; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v16
@@ -300,8 +301,9 @@ define <vscale x 32 x half> @vfneg_vv_nxv32f16_unmasked(<vscale x 32 x half> %va
; ZVFHMIN-NEXT: # %bb.1:
; ZVFHMIN-NEXT: mv a0, a1
; ZVFHMIN-NEXT: .LBB11_2:
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfneg.v v16, v16
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfsqrt-vp.ll b/llvm/test/CodeGen/RISCV/rvv/vfsqrt-vp.ll
index b13d221e00e634..6bbbd1813b0dd1 100644
--- a/llvm/test/CodeGen/RISCV/rvv/vfsqrt-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/vfsqrt-vp.ll
@@ -19,9 +19,9 @@ define <vscale x 1 x half> @vfsqrt_vv_nxv1f16(<vscale x 1 x half> %va, <vscale x
;
; ZVFHMIN-LABEL: vfsqrt_vv_nxv1f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfsqrt.v v9, v9, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -39,9 +39,9 @@ define <vscale x 1 x half> @vfsqrt_vv_nxv1f16_unmasked(<vscale x 1 x half> %va,
;
; ZVFHMIN-LABEL: vfsqrt_vv_nxv1f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfsqrt.v v9, v9
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -63,9 +63,9 @@ define <vscale x 2 x half> @vfsqrt_vv_nxv2f16(<vscale x 2 x half> %va, <vscale x
;
; ZVFHMIN-LABEL: vfsqrt_vv_nxv2f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfsqrt.v v9, v9, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -83,9 +83,9 @@ define <vscale x 2 x half> @vfsqrt_vv_nxv2f16_unmasked(<vscale x 2 x half> %va,
;
; ZVFHMIN-LABEL: vfsqrt_vv_nxv2f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfsqrt.v v9, v9
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -107,9 +107,9 @@ define <vscale x 4 x half> @vfsqrt_vv_nxv4f16(<vscale x 4 x half> %va, <vscale x
;
; ZVFHMIN-LABEL: vfsqrt_vv_nxv4f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfsqrt.v v10, v10, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -127,9 +127,9 @@ define <vscale x 4 x half> @vfsqrt_vv_nxv4f16_unmasked(<vscale x 4 x half> %va,
;
; ZVFHMIN-LABEL: vfsqrt_vv_nxv4f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfsqrt.v v10, v10
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -151,9 +151,9 @@ define <vscale x 8 x half> @vfsqrt_vv_nxv8f16(<vscale x 8 x half> %va, <vscale x
;
; ZVFHMIN-LABEL: vfsqrt_vv_nxv8f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfsqrt.v v12, v12, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -171,9 +171,9 @@ define <vscale x 8 x half> @vfsqrt_vv_nxv8f16_unmasked(<vscale x 8 x half> %va,
;
; ZVFHMIN-LABEL: vfsqrt_vv_nxv8f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfsqrt.v v12, v12
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -195,9 +195,9 @@ define <vscale x 16 x half> @vfsqrt_vv_nxv16f16(<vscale x 16 x half> %va, <vscal
;
; ZVFHMIN-LABEL: vfsqrt_vv_nxv16f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfsqrt.v v16, v16, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
@@ -215,9 +215,9 @@ define <vscale x 16 x half> @vfsqrt_vv_nxv16f16_unmasked(<vscale x 16 x half> %v
;
; ZVFHMIN-LABEL: vfsqrt_vv_nxv16f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfsqrt.v v16, v16
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
@@ -249,9 +249,9 @@ define <vscale x 32 x half> @vfsqrt_vv_nxv32f16(<vscale x 32 x half> %va, <vscal
; ZVFHMIN-NEXT: srli a2, a2, 2
; ZVFHMIN-NEXT: vsetvli a4, zero, e8, mf2, ta, ma
; ZVFHMIN-NEXT: vslidedown.vx v0, v0, a2
-; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v12
-; ZVFHMIN-NEXT: vsetvli zero, a3, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a3, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v12, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfsqrt.v v24, v24, v0.t
; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v24
@@ -259,9 +259,10 @@ define <vscale x 32 x half> @vfsqrt_vv_nxv32f16(<vscale x 32 x half> %va, <vscal
; ZVFHMIN-NEXT: # %bb.1:
; ZVFHMIN-NEXT: mv a0, a1
; ZVFHMIN-NEXT: .LBB10_2:
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
; ZVFHMIN-NEXT: vmv1r.v v0, v16
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfsqrt.v v16, v24, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
@@ -290,9 +291,9 @@ define <vscale x 32 x half> @vfsqrt_vv_nxv32f16_unmasked(<vscale x 32 x half> %v
; ZVFHMIN-NEXT: srli a2, a2, 2
; ZVFHMIN-NEXT: vsetvli a4, zero, e8, mf2, ta, ma
; ZVFHMIN-NEXT: vslidedown.vx v0, v16, a2
-; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12
-; ZVFHMIN-NEXT: vsetvli zero, a3, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a3, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfsqrt.v v16, v16, v0.t
; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v16
@@ -300,8 +301,9 @@ define <vscale x 32 x half> @vfsqrt_vv_nxv32f16_unmasked(<vscale x 32 x half> %v
; ZVFHMIN-NEXT: # %bb.1:
; ZVFHMIN-NEXT: mv a0, a1
; ZVFHMIN-NEXT: .LBB11_2:
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfsqrt.v v16, v16
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.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 0df7b2ce1978da..231c19cdd64d80 100644
--- a/llvm/test/CodeGen/RISCV/rvv/vfsub-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/vfsub-vp.ll
@@ -19,10 +19,10 @@ define <vscale x 1 x half> @vfsub_vv_nxv1f16(<vscale x 1 x half> %va, <vscale x
;
; ZVFHMIN-LABEL: vfsub_vv_nxv1f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfsub.vv v9, v9, v10, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -40,10 +40,10 @@ define <vscale x 1 x half> @vfsub_vv_nxv1f16_unmasked(<vscale x 1 x half> %va, <
;
; ZVFHMIN-LABEL: vfsub_vv_nxv1f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfsub.vv v9, v9, v10
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -68,9 +68,10 @@ define <vscale x 1 x half> @vfsub_vf_nxv1f16(<vscale x 1 x half> %va, half %b, <
; ZVFHMIN-NEXT: vfmv.v.f v9, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfsub.vv v9, v9, v8, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -95,9 +96,10 @@ define <vscale x 1 x half> @vfsub_vf_nxv1f16_unmasked(<vscale x 1 x half> %va, h
; ZVFHMIN-NEXT: vfmv.v.f v9, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v10, v9
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfsub.vv v9, v9, v8
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -121,10 +123,10 @@ define <vscale x 2 x half> @vfsub_vv_nxv2f16(<vscale x 2 x half> %va, <vscale x
;
; ZVFHMIN-LABEL: vfsub_vv_nxv2f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfsub.vv v9, v9, v10, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -142,10 +144,10 @@ define <vscale x 2 x half> @vfsub_vv_nxv2f16_unmasked(<vscale x 2 x half> %va, <
;
; ZVFHMIN-LABEL: vfsub_vv_nxv2f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfsub.vv v9, v9, v10
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -170,9 +172,10 @@ define <vscale x 2 x half> @vfsub_vf_nxv2f16(<vscale x 2 x half> %va, half %b, <
; ZVFHMIN-NEXT: vfmv.v.f v9, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfsub.vv v9, v9, v8, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -197,9 +200,10 @@ define <vscale x 2 x half> @vfsub_vf_nxv2f16_unmasked(<vscale x 2 x half> %va, h
; ZVFHMIN-NEXT: vfmv.v.f v9, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v10, v9
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfsub.vv v9, v9, v8
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
@@ -223,10 +227,10 @@ define <vscale x 4 x half> @vfsub_vv_nxv4f16(<vscale x 4 x half> %va, <vscale x
;
; ZVFHMIN-LABEL: vfsub_vv_nxv4f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfsub.vv v10, v12, v10, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -244,10 +248,10 @@ define <vscale x 4 x half> @vfsub_vv_nxv4f16_unmasked(<vscale x 4 x half> %va, <
;
; ZVFHMIN-LABEL: vfsub_vv_nxv4f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v9
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfsub.vv v10, v12, v10
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -272,9 +276,10 @@ define <vscale x 4 x half> @vfsub_vf_nxv4f16(<vscale x 4 x half> %va, half %b, <
; ZVFHMIN-NEXT: vfmv.v.f v10, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v9, v10
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v9
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v9, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfsub.vv v10, v10, v12, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -299,9 +304,10 @@ define <vscale x 4 x half> @vfsub_vf_nxv4f16_unmasked(<vscale x 4 x half> %va, h
; ZVFHMIN-NEXT: vfmv.v.f v10, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v9, v10
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v9
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfsub.vv v10, v10, v12
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
@@ -325,10 +331,10 @@ define <vscale x 8 x half> @vfsub_vv_nxv8f16(<vscale x 8 x half> %va, <vscale x
;
; ZVFHMIN-LABEL: vfsub_vv_nxv8f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v10
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v10, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfsub.vv v12, v16, v12, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -346,10 +352,10 @@ define <vscale x 8 x half> @vfsub_vv_nxv8f16_unmasked(<vscale x 8 x half> %va, <
;
; ZVFHMIN-LABEL: vfsub_vv_nxv8f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v10
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfsub.vv v12, v16, v12
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -374,9 +380,10 @@ define <vscale x 8 x half> @vfsub_vf_nxv8f16(<vscale x 8 x half> %va, half %b, <
; ZVFHMIN-NEXT: vfmv.v.f v12, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v10, v12
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v10, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfsub.vv v12, v12, v16, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -401,9 +408,10 @@ define <vscale x 8 x half> @vfsub_vf_nxv8f16_unmasked(<vscale x 8 x half> %va, h
; ZVFHMIN-NEXT: vfmv.v.f v12, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v10, v12
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v10
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT: vfsub.vv v12, v12, v16
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
@@ -427,10 +435,10 @@ define <vscale x 16 x half> @vfsub_vv_nxv16f16(<vscale x 16 x half> %va, <vscale
;
; ZVFHMIN-LABEL: vfsub_vv_nxv16f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfsub.vv v16, v24, v16, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
@@ -448,10 +456,10 @@ define <vscale x 16 x half> @vfsub_vv_nxv16f16_unmasked(<vscale x 16 x half> %va
;
; ZVFHMIN-LABEL: vfsub_vv_nxv16f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12
; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfsub.vv v16, v24, v16
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
@@ -476,9 +484,10 @@ define <vscale x 16 x half> @vfsub_vf_nxv16f16(<vscale x 16 x half> %va, half %b
; ZVFHMIN-NEXT: vfmv.v.f v16, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v16
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v12
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v12, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfsub.vv v16, v16, v24, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
@@ -503,9 +512,10 @@ define <vscale x 16 x half> @vfsub_vf_nxv16f16_unmasked(<vscale x 16 x half> %va
; ZVFHMIN-NEXT: vfmv.v.f v16, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v16
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v12
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfsub.vv v16, v16, v24
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
@@ -532,9 +542,13 @@ define <vscale x 32 x half> @vfsub_vv_nxv32f16(<vscale x 32 x half> %va, <vscale
; ZVFHMIN-NEXT: addi sp, sp, -16
; ZVFHMIN-NEXT: .cfi_def_cfa_offset 16
; ZVFHMIN-NEXT: csrr a1, vlenb
+; ZVFHMIN-NEXT: li a2, 0
; ZVFHMIN-NEXT: slli a1, a1, 3
+; ZVFHMIN-NEXT: add a2, a2, a1
+; ZVFHMIN-NEXT: slli a1, a1, 1
+; ZVFHMIN-NEXT: add a1, a1, a2
; ZVFHMIN-NEXT: sub sp, sp, a1
-; ZVFHMIN-NEXT: .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x08, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 8 * vlenb
+; ZVFHMIN-NEXT: .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x18, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 24 * vlenb
; ZVFHMIN-NEXT: vmv1r.v v1, v0
; ZVFHMIN-NEXT: csrr a2, vlenb
; ZVFHMIN-NEXT: slli a1, a2, 1
@@ -545,30 +559,61 @@ define <vscale x 32 x half> @vfsub_vv_nxv32f16(<vscale x 32 x half> %va, <vscale
; ZVFHMIN-NEXT: srli a2, a2, 2
; ZVFHMIN-NEXT: vsetvli a4, zero, e8, mf2, ta, ma
; ZVFHMIN-NEXT: vslidedown.vx v0, v0, a2
-; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a3, e16, m4, ta, ma
+; ZVFHMIN-NEXT: csrr a2, vlenb
+; ZVFHMIN-NEXT: slli a2, a2, 3
+; ZVFHMIN-NEXT: add a2, sp, a2
+; ZVFHMIN-NEXT: addi a2, a2, 16
+; ZVFHMIN-NEXT: vs8r.v v16, (a2) # Unknown-size Folded Spill
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v20, v0.t
; ZVFHMIN-NEXT: addi a2, sp, 16
+; ZVFHMIN-NEXT: vs8r.v v24, (a2) # Unknown-size Folded Spill
+; ZVFHMIN-NEXT: vmv4r.v v16, v8
+; ZVFHMIN-NEXT: csrr a2, vlenb
+; ZVFHMIN-NEXT: slli a2, a2, 4
+; ZVFHMIN-NEXT: add a2, sp, a2
+; ZVFHMIN-NEXT: addi a2, a2, 16
; ZVFHMIN-NEXT: vs8r.v v16, (a2) # Unknown-size Folded Spill
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v20
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12
-; ZVFHMIN-NEXT: vsetvli zero, a3, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vfsub.vv v16, v16, v24, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; ZVFHMIN-NEXT: addi a2, sp, 16
+; ZVFHMIN-NEXT: vl8r.v v8, (a2) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfsub.vv v16, v16, v8, v0.t
; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v16
; ZVFHMIN-NEXT: bltu a0, a1, .LBB20_2
; ZVFHMIN-NEXT: # %bb.1:
; ZVFHMIN-NEXT: mv a0, a1
; ZVFHMIN-NEXT: .LBB20_2:
-; ZVFHMIN-NEXT: addi a1, sp, 16
-; ZVFHMIN-NEXT: vl8r.v v24, (a1) # Unknown-size Folded Reload
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v24
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
; ZVFHMIN-NEXT: vmv1r.v v0, v1
-; ZVFHMIN-NEXT: vfsub.vv v16, v24, v16, v0.t
+; ZVFHMIN-NEXT: csrr a0, vlenb
+; ZVFHMIN-NEXT: slli a0, a0, 3
+; ZVFHMIN-NEXT: add a0, sp, a0
+; ZVFHMIN-NEXT: addi a0, a0, 16
+; ZVFHMIN-NEXT: vl8r.v v16, (a0) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v16, v0.t
+; ZVFHMIN-NEXT: addi a0, sp, 16
+; ZVFHMIN-NEXT: vs8r.v v24, (a0) # Unknown-size Folded Spill
+; ZVFHMIN-NEXT: csrr a0, vlenb
+; ZVFHMIN-NEXT: slli a0, a0, 4
+; ZVFHMIN-NEXT: add a0, sp, a0
+; ZVFHMIN-NEXT: addi a0, a0, 16
+; ZVFHMIN-NEXT: vl8r.v v16, (a0) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v16, v0.t
+; ZVFHMIN-NEXT: vmv8r.v v16, v24
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; ZVFHMIN-NEXT: addi a0, sp, 16
+; ZVFHMIN-NEXT: vl8r.v v24, (a0) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfsub.vv v16, v16, v24, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
; ZVFHMIN-NEXT: csrr a0, vlenb
+; ZVFHMIN-NEXT: li a1, 0
; ZVFHMIN-NEXT: slli a0, a0, 3
+; ZVFHMIN-NEXT: add a1, a1, a0
+; ZVFHMIN-NEXT: slli a0, a0, 1
+; ZVFHMIN-NEXT: add a0, a0, a1
; ZVFHMIN-NEXT: add sp, sp, a0
; ZVFHMIN-NEXT: addi sp, sp, 16
; ZVFHMIN-NEXT: ret
@@ -602,12 +647,12 @@ define <vscale x 32 x half> @vfsub_vv_nxv32f16_unmasked(<vscale x 32 x half> %va
; ZVFHMIN-NEXT: srli a2, a2, 2
; ZVFHMIN-NEXT: vsetvli a4, zero, e8, mf2, ta, ma
; ZVFHMIN-NEXT: vslidedown.vx v0, v24, a2
-; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a3, e16, m4, ta, ma
; ZVFHMIN-NEXT: addi a2, sp, 16
; ZVFHMIN-NEXT: vs8r.v v16, (a2) # Unknown-size Folded Spill
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v20
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12
-; ZVFHMIN-NEXT: vsetvli zero, a3, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v20, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfsub.vv v16, v16, v24, v0.t
; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v16
@@ -615,11 +660,12 @@ define <vscale x 32 x half> @vfsub_vv_nxv32f16_unmasked(<vscale x 32 x half> %va
; ZVFHMIN-NEXT: # %bb.1:
; ZVFHMIN-NEXT: mv a0, a1
; ZVFHMIN-NEXT: .LBB21_2:
-; ZVFHMIN-NEXT: addi a1, sp, 16
-; ZVFHMIN-NEXT: vl8r.v v24, (a1) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; ZVFHMIN-NEXT: addi a0, sp, 16
+; ZVFHMIN-NEXT: vl8r.v v24, (a0) # Unknown-size Folded Reload
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v24
; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfsub.vv v16, v24, v16
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
@@ -643,7 +689,14 @@ define <vscale x 32 x half> @vfsub_vf_nxv32f16(<vscale x 32 x half> %va, half %b
;
; ZVFHMIN-LABEL: vfsub_vf_nxv32f16:
; ZVFHMIN: # %bb.0:
+; ZVFHMIN-NEXT: addi sp, sp, -16
+; ZVFHMIN-NEXT: .cfi_def_cfa_offset 16
+; ZVFHMIN-NEXT: csrr a1, vlenb
+; ZVFHMIN-NEXT: slli a1, a1, 3
+; ZVFHMIN-NEXT: sub sp, sp, a1
+; ZVFHMIN-NEXT: .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x08, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 8 * vlenb
; ZVFHMIN-NEXT: vmv1r.v v1, v0
+; ZVFHMIN-NEXT: vmv8r.v v16, v8
; ZVFHMIN-NEXT: fcvt.s.h fa5, fa0
; ZVFHMIN-NEXT: vsetvli a1, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfmv.v.f v24, fa5
@@ -658,23 +711,33 @@ define <vscale x 32 x half> @vfsub_vf_nxv32f16(<vscale x 32 x half> %va, half %b
; ZVFHMIN-NEXT: srli a2, a2, 2
; ZVFHMIN-NEXT: vsetvli a4, zero, e8, mf2, ta, ma
; ZVFHMIN-NEXT: vslidedown.vx v0, v0, a2
+; ZVFHMIN-NEXT: vsetvli zero, a3, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v4, v0.t
+; ZVFHMIN-NEXT: addi a2, sp, 16
+; ZVFHMIN-NEXT: vs8r.v v8, (a2) # Unknown-size Folded Spill
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v20, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vfsub.vv v24, v8, v24, v0.t
; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v4
-; ZVFHMIN-NEXT: vsetvli zero, a3, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vfsub.vv v16, v16, v24, v0.t
-; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v16
+; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v24
; ZVFHMIN-NEXT: bltu a0, a1, .LBB22_2
; ZVFHMIN-NEXT: # %bb.1:
; ZVFHMIN-NEXT: mv a0, a1
; ZVFHMIN-NEXT: .LBB22_2:
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
; ZVFHMIN-NEXT: vmv1r.v v0, v1
+; ZVFHMIN-NEXT: addi a0, sp, 16
+; ZVFHMIN-NEXT: vl8r.v v24, (a0) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v24, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v4, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfsub.vv v16, v16, v24, v0.t
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
+; ZVFHMIN-NEXT: csrr a0, vlenb
+; ZVFHMIN-NEXT: slli a0, a0, 3
+; ZVFHMIN-NEXT: add sp, sp, a0
+; ZVFHMIN-NEXT: addi sp, sp, 16
; ZVFHMIN-NEXT: ret
%elt.head = insertelement <vscale x 32 x half> poison, half %b, i32 0
%vb = shufflevector <vscale x 32 x half> %elt.head, <vscale x 32 x half> poison, <vscale x 32 x i32> zeroinitializer
@@ -693,11 +756,11 @@ define <vscale x 32 x half> @vfsub_vf_nxv32f16_unmasked(<vscale x 32 x half> %va
; ZVFHMIN: # %bb.0:
; ZVFHMIN-NEXT: fcvt.s.h fa5, fa0
; ZVFHMIN-NEXT: vsetvli a1, zero, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vfmv.v.f v16, fa5
+; ZVFHMIN-NEXT: vfmv.v.f v24, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfncvt.f.f.w v4, v16
+; ZVFHMIN-NEXT: vfncvt.f.f.w v4, v24
; ZVFHMIN-NEXT: vsetvli a1, zero, e8, m4, ta, ma
-; ZVFHMIN-NEXT: vmset.m v16
+; ZVFHMIN-NEXT: vmset.m v20
; ZVFHMIN-NEXT: csrr a2, vlenb
; ZVFHMIN-NEXT: slli a1, a2, 1
; ZVFHMIN-NEXT: sub a3, a0, a1
@@ -706,21 +769,23 @@ define <vscale x 32 x half> @vfsub_vf_nxv32f16_unmasked(<vscale x 32 x half> %va
; ZVFHMIN-NEXT: and a3, a4, a3
; ZVFHMIN-NEXT: srli a2, a2, 2
; ZVFHMIN-NEXT: vsetvli a4, zero, e8, mf2, ta, ma
-; ZVFHMIN-NEXT: vslidedown.vx v0, v16, a2
-; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v12
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v4
-; ZVFHMIN-NEXT: vsetvli zero, a3, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vfsub.vv v24, v24, v16, v0.t
+; ZVFHMIN-NEXT: vslidedown.vx v0, v20, a2
+; ZVFHMIN-NEXT: vsetvli zero, a3, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v4, v0.t
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vfsub.vv v16, v16, v24, v0.t
; ZVFHMIN-NEXT: vsetvli a2, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v24
+; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v16
; ZVFHMIN-NEXT: bltu a0, a1, .LBB23_2
; ZVFHMIN-NEXT: # %bb.1:
; ZVFHMIN-NEXT: mv a0, a1
; ZVFHMIN-NEXT: .LBB23_2:
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vfsub.vv v16, v24, v16
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v4
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vfsub.vv v16, v16, v24
; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
; ZVFHMIN-NEXT: ret
diff --git a/llvm/test/CodeGen/RISCV/rvv/vreductions-fp-vp.ll b/llvm/test/CodeGen/RISCV/rvv/vreductions-fp-vp.ll
index bd510d26279c48..a4b77904adf7fe 100644
--- a/llvm/test/CodeGen/RISCV/rvv/vreductions-fp-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/vreductions-fp-vp.ll
@@ -22,10 +22,10 @@ define half @vpreduce_fadd_nxv1f16(half %s, <vscale x 1 x half> %v, <vscale x 1
;
; ZVFHMIN-LABEL: vpreduce_fadd_nxv1f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
; ZVFHMIN-NEXT: fcvt.s.h fa5, fa0
-; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetivli zero, 1, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfmv.s.f v8, fa5
; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfredusum.vs v8, v9, v8, v0.t
@@ -48,10 +48,10 @@ define half @vpreduce_ord_fadd_nxv1f16(half %s, <vscale x 1 x half> %v, <vscale
;
; ZVFHMIN-LABEL: vpreduce_ord_fadd_nxv1f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
; ZVFHMIN-NEXT: fcvt.s.h fa5, fa0
-; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetivli zero, 1, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfmv.s.f v8, fa5
; ZVFHMIN-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
; ZVFHMIN-NEXT: vfredosum.vs v8, v9, v8, v0.t
@@ -76,10 +76,10 @@ define half @vpreduce_fadd_nxv2f16(half %s, <vscale x 2 x half> %v, <vscale x 2
;
; ZVFHMIN-LABEL: vpreduce_fadd_nxv2f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
; ZVFHMIN-NEXT: fcvt.s.h fa5, fa0
-; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetivli zero, 1, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfmv.s.f v8, fa5
; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfredusum.vs v8, v9, v8, v0.t
@@ -102,10 +102,10 @@ define half @vpreduce_ord_fadd_nxv2f16(half %s, <vscale x 2 x half> %v, <vscale
;
; ZVFHMIN-LABEL: vpreduce_ord_fadd_nxv2f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8, v0.t
; ZVFHMIN-NEXT: fcvt.s.h fa5, fa0
-; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
+; ZVFHMIN-NEXT: vsetivli zero, 1, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfmv.s.f v8, fa5
; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m1, ta, ma
; ZVFHMIN-NEXT: vfredosum.vs v8, v9, v8, v0.t
@@ -130,10 +130,10 @@ define half @vpreduce_fadd_nxv4f16(half %s, <vscale x 4 x half> %v, <vscale x 4
;
; ZVFHMIN-LABEL: vpreduce_fadd_nxv4f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8, v0.t
; ZVFHMIN-NEXT: fcvt.s.h fa5, fa0
-; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetivli zero, 1, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfmv.s.f v8, fa5
; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfredusum.vs v8, v10, v8, v0.t
@@ -156,10 +156,10 @@ define half @vpreduce_ord_fadd_nxv4f16(half %s, <vscale x 4 x half> %v, <vscale
;
; ZVFHMIN-LABEL: vpreduce_ord_fadd_nxv4f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8, v0.t
; ZVFHMIN-NEXT: fcvt.s.h fa5, fa0
-; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
+; ZVFHMIN-NEXT: vsetivli zero, 1, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfmv.s.f v8, fa5
; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m2, ta, ma
; ZVFHMIN-NEXT: vfredosum.vs v8, v10, v8, v0.t
@@ -200,79 +200,108 @@ define half @vpreduce_fadd_nxv64f16(half %s, <vscale x 64 x half> %v, <vscale x
;
; ZVFHMIN-LABEL: vpreduce_fadd_nxv64f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: csrr a3, vlenb
-; ZVFHMIN-NEXT: srli a1, a3, 1
-; ZVFHMIN-NEXT: vsetvli a2, zero, e8, m1, ta, ma
+; ZVFHMIN-NEXT: addi sp, sp, -16
+; ZVFHMIN-NEXT: .cfi_def_cfa_offset 16
+; ZVFHMIN-NEXT: csrr a1, vlenb
+; ZVFHMIN-NEXT: slli a1, a1, 4
+; ZVFHMIN-NEXT: sub sp, sp, a1
+; ZVFHMIN-NEXT: .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x10, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 16 * vlenb
+; ZVFHMIN-NEXT: vmv1r.v v24, v0
+; ZVFHMIN-NEXT: csrr a1, vlenb
+; ZVFHMIN-NEXT: slli a1, a1, 3
+; ZVFHMIN-NEXT: add a1, sp, a1
+; ZVFHMIN-NEXT: addi a1, a1, 16
+; ZVFHMIN-NEXT: vs8r.v v16, (a1) # Unknown-size Folded Spill
+; ZVFHMIN-NEXT: csrr a2, vlenb
+; ZVFHMIN-NEXT: srli a1, a2, 1
+; ZVFHMIN-NEXT: vsetvli a3, zero, e8, m1, ta, ma
+; ZVFHMIN-NEXT: slli a3, a2, 2
; ZVFHMIN-NEXT: vslidedown.vx v1, v0, a1
-; ZVFHMIN-NEXT: slli a5, a3, 2
-; ZVFHMIN-NEXT: sub a1, a0, a5
-; ZVFHMIN-NEXT: sltu a2, a0, a1
-; ZVFHMIN-NEXT: addi a2, a2, -1
-; ZVFHMIN-NEXT: and a1, a2, a1
-; ZVFHMIN-NEXT: slli a4, a3, 1
-; ZVFHMIN-NEXT: sub a2, a1, a4
-; ZVFHMIN-NEXT: sltu a6, a1, a2
-; ZVFHMIN-NEXT: bltu a1, a4, .LBB6_2
+; ZVFHMIN-NEXT: mv a4, a0
+; ZVFHMIN-NEXT: bltu a0, a3, .LBB6_2
; ZVFHMIN-NEXT: # %bb.1:
-; ZVFHMIN-NEXT: mv a1, a4
+; ZVFHMIN-NEXT: mv a4, a3
; ZVFHMIN-NEXT: .LBB6_2:
-; ZVFHMIN-NEXT: addi a6, a6, -1
-; ZVFHMIN-NEXT: bltu a0, a5, .LBB6_4
-; ZVFHMIN-NEXT: # %bb.3:
-; ZVFHMIN-NEXT: mv a0, a5
-; ZVFHMIN-NEXT: .LBB6_4:
-; ZVFHMIN-NEXT: and a2, a6, a2
-; ZVFHMIN-NEXT: sub a5, a0, a4
-; ZVFHMIN-NEXT: sltu a6, a0, a5
+; ZVFHMIN-NEXT: slli a1, a2, 1
+; ZVFHMIN-NEXT: sub a5, a4, a1
+; ZVFHMIN-NEXT: sltu a6, a4, a5
; ZVFHMIN-NEXT: addi a6, a6, -1
; ZVFHMIN-NEXT: and a5, a6, a5
-; ZVFHMIN-NEXT: srli a3, a3, 2
+; ZVFHMIN-NEXT: srli a2, a2, 2
; ZVFHMIN-NEXT: vsetvli a6, zero, e8, mf2, ta, ma
-; ZVFHMIN-NEXT: vslidedown.vx v2, v0, a3
-; ZVFHMIN-NEXT: bltu a0, a4, .LBB6_6
-; ZVFHMIN-NEXT: # %bb.5:
-; ZVFHMIN-NEXT: mv a0, a4
-; ZVFHMIN-NEXT: .LBB6_6:
-; ZVFHMIN-NEXT: vsetvli a4, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8
+; ZVFHMIN-NEXT: vslidedown.vx v26, v24, a2
+; ZVFHMIN-NEXT: vsetvli zero, a5, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vmv1r.v v0, v26
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12, v0.t
+; ZVFHMIN-NEXT: addi a6, sp, 16
+; ZVFHMIN-NEXT: vs8r.v v16, (a6) # Unknown-size Folded Spill
+; ZVFHMIN-NEXT: bltu a4, a1, .LBB6_4
+; ZVFHMIN-NEXT: # %bb.3:
+; ZVFHMIN-NEXT: mv a4, a1
+; ZVFHMIN-NEXT: .LBB6_4:
+; ZVFHMIN-NEXT: vsetvli zero, a4, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vmv1r.v v0, v24
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8, v0.t
; ZVFHMIN-NEXT: fcvt.s.h fa5, fa0
-; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetivli zero, 1, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfmv.s.f v8, fa5
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vfredusum.vs v8, v24, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, a4, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vfredusum.vs v8, v16, v8, v0.t
; ZVFHMIN-NEXT: vfmv.f.s fa5, v8
; ZVFHMIN-NEXT: fcvt.h.s fa5, fa5
; ZVFHMIN-NEXT: fcvt.s.h fa5, fa5
; ZVFHMIN-NEXT: vsetivli zero, 1, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfmv.s.f v8, fa5
-; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v12
; ZVFHMIN-NEXT: vsetvli zero, a5, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vmv1r.v v0, v2
-; ZVFHMIN-NEXT: vfredusum.vs v8, v24, v8, v0.t
+; ZVFHMIN-NEXT: vmv1r.v v0, v26
+; ZVFHMIN-NEXT: addi a4, sp, 16
+; ZVFHMIN-NEXT: vl8r.v v16, (a4) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfredusum.vs v8, v16, v8, v0.t
; ZVFHMIN-NEXT: vfmv.f.s fa5, v8
; ZVFHMIN-NEXT: fcvt.h.s fa5, fa5
; ZVFHMIN-NEXT: fcvt.s.h fa5, fa5
; ZVFHMIN-NEXT: vsetivli zero, 1, e32, m8, ta, ma
+; ZVFHMIN-NEXT: sub a3, a0, a3
+; ZVFHMIN-NEXT: sltu a0, a0, a3
+; ZVFHMIN-NEXT: addi a0, a0, -1
+; ZVFHMIN-NEXT: and a0, a0, a3
; ZVFHMIN-NEXT: vfmv.s.f v8, fa5
-; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v16
-; ZVFHMIN-NEXT: vsetvli zero, a1, e32, m8, ta, ma
+; ZVFHMIN-NEXT: mv a3, a0
+; ZVFHMIN-NEXT: bltu a0, a1, .LBB6_6
+; ZVFHMIN-NEXT: # %bb.5:
+; ZVFHMIN-NEXT: mv a3, a1
+; ZVFHMIN-NEXT: .LBB6_6:
+; ZVFHMIN-NEXT: vsetvli zero, a3, e16, m4, ta, ma
; ZVFHMIN-NEXT: vmv1r.v v0, v1
-; ZVFHMIN-NEXT: vfredusum.vs v8, v24, v8, v0.t
+; ZVFHMIN-NEXT: csrr a3, vlenb
+; ZVFHMIN-NEXT: slli a3, a3, 3
+; ZVFHMIN-NEXT: add a3, sp, a3
+; ZVFHMIN-NEXT: addi a3, a3, 16
+; ZVFHMIN-NEXT: vl8r.v v24, (a3) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v24, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vfredusum.vs v8, v16, v8, v0.t
; ZVFHMIN-NEXT: vfmv.f.s fa5, v8
; ZVFHMIN-NEXT: fcvt.h.s fa5, fa5
; ZVFHMIN-NEXT: fcvt.s.h fa5, fa5
; ZVFHMIN-NEXT: vsetivli zero, 1, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfmv.s.f v8, fa5
-; ZVFHMIN-NEXT: vsetvli a0, zero, e8, mf2, ta, ma
-; ZVFHMIN-NEXT: vslidedown.vx v0, v1, a3
-; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v20
-; ZVFHMIN-NEXT: vsetvli zero, a2, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vfredusum.vs v8, v24, v8, v0.t
+; ZVFHMIN-NEXT: sub a1, a0, a1
+; ZVFHMIN-NEXT: sltu a0, a0, a1
+; ZVFHMIN-NEXT: addi a0, a0, -1
+; ZVFHMIN-NEXT: and a0, a0, a1
+; ZVFHMIN-NEXT: vsetvli a1, zero, e8, mf2, ta, ma
+; ZVFHMIN-NEXT: vslidedown.vx v0, v1, a2
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v28, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vfredusum.vs v8, v16, v8, v0.t
; ZVFHMIN-NEXT: vfmv.f.s fa5, v8
; ZVFHMIN-NEXT: fcvt.h.s fa0, fa5
+; ZVFHMIN-NEXT: csrr a0, vlenb
+; ZVFHMIN-NEXT: slli a0, a0, 4
+; ZVFHMIN-NEXT: add sp, sp, a0
+; ZVFHMIN-NEXT: addi sp, sp, 16
; ZVFHMIN-NEXT: ret
%r = call reassoc half @llvm.vp.reduce.fadd.nxv64f16(half %s, <vscale x 64 x half> %v, <vscale x 64 x i1> %m, i32 %evl)
ret half %r
@@ -306,79 +335,108 @@ define half @vpreduce_ord_fadd_nxv64f16(half %s, <vscale x 64 x half> %v, <vscal
;
; ZVFHMIN-LABEL: vpreduce_ord_fadd_nxv64f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: csrr a3, vlenb
-; ZVFHMIN-NEXT: srli a1, a3, 1
-; ZVFHMIN-NEXT: vsetvli a2, zero, e8, m1, ta, ma
+; ZVFHMIN-NEXT: addi sp, sp, -16
+; ZVFHMIN-NEXT: .cfi_def_cfa_offset 16
+; ZVFHMIN-NEXT: csrr a1, vlenb
+; ZVFHMIN-NEXT: slli a1, a1, 4
+; ZVFHMIN-NEXT: sub sp, sp, a1
+; ZVFHMIN-NEXT: .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x10, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 16 * vlenb
+; ZVFHMIN-NEXT: vmv1r.v v24, v0
+; ZVFHMIN-NEXT: csrr a1, vlenb
+; ZVFHMIN-NEXT: slli a1, a1, 3
+; ZVFHMIN-NEXT: add a1, sp, a1
+; ZVFHMIN-NEXT: addi a1, a1, 16
+; ZVFHMIN-NEXT: vs8r.v v16, (a1) # Unknown-size Folded Spill
+; ZVFHMIN-NEXT: csrr a2, vlenb
+; ZVFHMIN-NEXT: srli a1, a2, 1
+; ZVFHMIN-NEXT: vsetvli a3, zero, e8, m1, ta, ma
+; ZVFHMIN-NEXT: slli a3, a2, 2
; ZVFHMIN-NEXT: vslidedown.vx v1, v0, a1
-; ZVFHMIN-NEXT: slli a5, a3, 2
-; ZVFHMIN-NEXT: sub a1, a0, a5
-; ZVFHMIN-NEXT: sltu a2, a0, a1
-; ZVFHMIN-NEXT: addi a2, a2, -1
-; ZVFHMIN-NEXT: and a1, a2, a1
-; ZVFHMIN-NEXT: slli a4, a3, 1
-; ZVFHMIN-NEXT: sub a2, a1, a4
-; ZVFHMIN-NEXT: sltu a6, a1, a2
-; ZVFHMIN-NEXT: bltu a1, a4, .LBB7_2
+; ZVFHMIN-NEXT: mv a4, a0
+; ZVFHMIN-NEXT: bltu a0, a3, .LBB7_2
; ZVFHMIN-NEXT: # %bb.1:
-; ZVFHMIN-NEXT: mv a1, a4
+; ZVFHMIN-NEXT: mv a4, a3
; ZVFHMIN-NEXT: .LBB7_2:
-; ZVFHMIN-NEXT: addi a6, a6, -1
-; ZVFHMIN-NEXT: bltu a0, a5, .LBB7_4
-; ZVFHMIN-NEXT: # %bb.3:
-; ZVFHMIN-NEXT: mv a0, a5
-; ZVFHMIN-NEXT: .LBB7_4:
-; ZVFHMIN-NEXT: and a2, a6, a2
-; ZVFHMIN-NEXT: sub a5, a0, a4
-; ZVFHMIN-NEXT: sltu a6, a0, a5
+; ZVFHMIN-NEXT: slli a1, a2, 1
+; ZVFHMIN-NEXT: sub a5, a4, a1
+; ZVFHMIN-NEXT: sltu a6, a4, a5
; ZVFHMIN-NEXT: addi a6, a6, -1
; ZVFHMIN-NEXT: and a5, a6, a5
-; ZVFHMIN-NEXT: srli a3, a3, 2
+; ZVFHMIN-NEXT: srli a2, a2, 2
; ZVFHMIN-NEXT: vsetvli a6, zero, e8, mf2, ta, ma
-; ZVFHMIN-NEXT: vslidedown.vx v2, v0, a3
-; ZVFHMIN-NEXT: bltu a0, a4, .LBB7_6
-; ZVFHMIN-NEXT: # %bb.5:
-; ZVFHMIN-NEXT: mv a0, a4
-; ZVFHMIN-NEXT: .LBB7_6:
-; ZVFHMIN-NEXT: vsetvli a4, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8
+; ZVFHMIN-NEXT: vslidedown.vx v26, v24, a2
+; ZVFHMIN-NEXT: vsetvli zero, a5, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vmv1r.v v0, v26
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12, v0.t
+; ZVFHMIN-NEXT: addi a6, sp, 16
+; ZVFHMIN-NEXT: vs8r.v v16, (a6) # Unknown-size Folded Spill
+; ZVFHMIN-NEXT: bltu a4, a1, .LBB7_4
+; ZVFHMIN-NEXT: # %bb.3:
+; ZVFHMIN-NEXT: mv a4, a1
+; ZVFHMIN-NEXT: .LBB7_4:
+; ZVFHMIN-NEXT: vsetvli zero, a4, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vmv1r.v v0, v24
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8, v0.t
; ZVFHMIN-NEXT: fcvt.s.h fa5, fa0
-; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vsetivli zero, 1, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfmv.s.f v8, fa5
-; ZVFHMIN-NEXT: vsetvli zero, a0, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vfredosum.vs v8, v24, v8, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, a4, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vfredosum.vs v8, v16, v8, v0.t
; ZVFHMIN-NEXT: vfmv.f.s fa5, v8
; ZVFHMIN-NEXT: fcvt.h.s fa5, fa5
; ZVFHMIN-NEXT: fcvt.s.h fa5, fa5
; ZVFHMIN-NEXT: vsetivli zero, 1, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfmv.s.f v8, fa5
-; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v12
; ZVFHMIN-NEXT: vsetvli zero, a5, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vmv1r.v v0, v2
-; ZVFHMIN-NEXT: vfredosum.vs v8, v24, v8, v0.t
+; ZVFHMIN-NEXT: vmv1r.v v0, v26
+; ZVFHMIN-NEXT: addi a4, sp, 16
+; ZVFHMIN-NEXT: vl8r.v v16, (a4) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfredosum.vs v8, v16, v8, v0.t
; ZVFHMIN-NEXT: vfmv.f.s fa5, v8
; ZVFHMIN-NEXT: fcvt.h.s fa5, fa5
; ZVFHMIN-NEXT: fcvt.s.h fa5, fa5
; ZVFHMIN-NEXT: vsetivli zero, 1, e32, m8, ta, ma
+; ZVFHMIN-NEXT: sub a3, a0, a3
+; ZVFHMIN-NEXT: sltu a0, a0, a3
+; ZVFHMIN-NEXT: addi a0, a0, -1
+; ZVFHMIN-NEXT: and a0, a0, a3
; ZVFHMIN-NEXT: vfmv.s.f v8, fa5
-; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v16
-; ZVFHMIN-NEXT: vsetvli zero, a1, e32, m8, ta, ma
+; ZVFHMIN-NEXT: mv a3, a0
+; ZVFHMIN-NEXT: bltu a0, a1, .LBB7_6
+; ZVFHMIN-NEXT: # %bb.5:
+; ZVFHMIN-NEXT: mv a3, a1
+; ZVFHMIN-NEXT: .LBB7_6:
+; ZVFHMIN-NEXT: vsetvli zero, a3, e16, m4, ta, ma
; ZVFHMIN-NEXT: vmv1r.v v0, v1
-; ZVFHMIN-NEXT: vfredosum.vs v8, v24, v8, v0.t
+; ZVFHMIN-NEXT: csrr a3, vlenb
+; ZVFHMIN-NEXT: slli a3, a3, 3
+; ZVFHMIN-NEXT: add a3, sp, a3
+; ZVFHMIN-NEXT: addi a3, a3, 16
+; ZVFHMIN-NEXT: vl8r.v v24, (a3) # Unknown-size Folded Reload
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v24, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vfredosum.vs v8, v16, v8, v0.t
; ZVFHMIN-NEXT: vfmv.f.s fa5, v8
; ZVFHMIN-NEXT: fcvt.h.s fa5, fa5
; ZVFHMIN-NEXT: fcvt.s.h fa5, fa5
; ZVFHMIN-NEXT: vsetivli zero, 1, e32, m8, ta, ma
; ZVFHMIN-NEXT: vfmv.s.f v8, fa5
-; ZVFHMIN-NEXT: vsetvli a0, zero, e8, mf2, ta, ma
-; ZVFHMIN-NEXT: vslidedown.vx v0, v1, a3
-; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v20
-; ZVFHMIN-NEXT: vsetvli zero, a2, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vfredosum.vs v8, v24, v8, v0.t
+; ZVFHMIN-NEXT: sub a1, a0, a1
+; ZVFHMIN-NEXT: sltu a0, a0, a1
+; ZVFHMIN-NEXT: addi a0, a0, -1
+; ZVFHMIN-NEXT: and a0, a0, a1
+; ZVFHMIN-NEXT: vsetvli a1, zero, e8, mf2, ta, ma
+; ZVFHMIN-NEXT: vslidedown.vx v0, v1, a2
+; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v28, v0.t
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vfredosum.vs v8, v16, v8, v0.t
; ZVFHMIN-NEXT: vfmv.f.s fa5, v8
; ZVFHMIN-NEXT: fcvt.h.s fa0, fa5
+; ZVFHMIN-NEXT: csrr a0, vlenb
+; ZVFHMIN-NEXT: slli a0, a0, 4
+; ZVFHMIN-NEXT: add sp, sp, a0
+; ZVFHMIN-NEXT: addi sp, sp, 16
; ZVFHMIN-NEXT: ret
%r = call half @llvm.vp.reduce.fadd.nxv64f16(half %s, <vscale x 64 x half> %v, <vscale x 64 x i1> %m, i32 %evl)
ret half %r
>From aa93433c8d92436ec15cecb9274937828bd32a50 Mon Sep 17 00:00:00 2001
From: Simeon Krastnikov <simeon.krastnikov at imgtec.com>
Date: Fri, 23 Feb 2024 16:57:49 +0000
Subject: [PATCH 2/2] fixup! [RISCV] Use predicated FP extension when promoting
VP operands
---
.../SelectionDAG/LegalizeVectorOps.cpp | 38 +++++++++----------
1 file changed, 18 insertions(+), 20 deletions(-)
diff --git a/llvm/lib/CodeGen/SelectionDAG/LegalizeVectorOps.cpp b/llvm/lib/CodeGen/SelectionDAG/LegalizeVectorOps.cpp
index 937a6a0df65586..85ee41a5b786cb 100644
--- a/llvm/lib/CodeGen/SelectionDAG/LegalizeVectorOps.cpp
+++ b/llvm/lib/CodeGen/SelectionDAG/LegalizeVectorOps.cpp
@@ -166,6 +166,10 @@ class VectorLegalizer {
/// truncated back to the original type.
void PromoteFP_TO_INT(SDNode *Node, SmallVectorImpl<SDValue> &Results);
+ // Promotes the floating-point operand at the given index of the node into
+ // an FP_EXTEND/VP_FP_EXTEND node of type NewVT
+ SDValue PromoteFPOperand(SDNode *Node, MVT NewVT, unsigned Index);
+
/// Implements vector reduce operation promotion.
///
/// All vector operands are promoted to a vector type with larger element
@@ -569,6 +573,18 @@ bool VectorLegalizer::LowerOperationWrapper(SDNode *Node,
return true;
}
+SDValue VectorLegalizer::PromoteFPOperand(SDNode *Node, MVT NewVT,
+ unsigned Index) {
+ SDLoc DL(Node);
+ if (ISD::isVPOpcode(Node->getOpcode())) {
+ auto MaskIdx = *ISD::getVPMaskIdx(Node->getOpcode());
+ auto EVLIdx = *ISD::getVPExplicitVectorLengthIdx(Node->getOpcode());
+ return DAG.getNode(ISD::VP_FP_EXTEND, DL, NewVT, Node->getOperand(Index),
+ Node->getOperand(MaskIdx), Node->getOperand(EVLIdx));
+ }
+ return DAG.getNode(ISD::FP_EXTEND, DL, NewVT, Node->getOperand(Index));
+}
+
void VectorLegalizer::PromoteReduction(SDNode *Node,
SmallVectorImpl<SDValue> &Results) {
MVT VecVT = Node->getOperand(1).getSimpleValueType();
@@ -593,16 +609,7 @@ void VectorLegalizer::PromoteReduction(SDNode *Node,
ISD::getVPMaskIdx(Node->getOpcode()) == j)) // Skip mask operand.
// promote the vector operand.
if (Node->getOperand(j).getValueType().isFloatingPoint())
- if (!ISD::isVPOpcode(Node->getOpcode()))
- Operands[j] =
- DAG.getNode(ISD::FP_EXTEND, DL, NewVecVT, Node->getOperand(j));
- else {
- auto MaskIdx = *ISD::getVPMaskIdx(Node->getOpcode());
- auto EVLIdx = *ISD::getVPExplicitVectorLengthIdx(Node->getOpcode());
- Operands[j] =
- DAG.getNode(ISD::VP_FP_EXTEND, DL, NewVecVT, Node->getOperand(j),
- Node->getOperand(MaskIdx), Node->getOperand(EVLIdx));
- }
+ Operands[j] = PromoteFPOperand(Node, NewVecVT, j);
else
Operands[j] =
DAG.getNode(ISD::ANY_EXTEND, DL, NewVecVT, Node->getOperand(j));
@@ -764,16 +771,7 @@ void VectorLegalizer::Promote(SDNode *Node, SmallVectorImpl<SDValue> &Results) {
.getVectorElementType()
.isFloatingPoint() &&
NVT.isVector() && NVT.getVectorElementType().isFloatingPoint())
- if (!ISD::isVPOpcode(Node->getOpcode()))
- Operands[j] =
- DAG.getNode(ISD::FP_EXTEND, dl, NVT, Node->getOperand(j));
- else {
- auto MaskIdx = *ISD::getVPMaskIdx(Node->getOpcode());
- auto EVLIdx = *ISD::getVPExplicitVectorLengthIdx(Node->getOpcode());
- Operands[j] =
- DAG.getNode(ISD::VP_FP_EXTEND, dl, NVT, Node->getOperand(j),
- Node->getOperand(MaskIdx), Node->getOperand(EVLIdx));
- }
+ Operands[j] = PromoteFPOperand(Node, NVT, j);
else
Operands[j] = DAG.getNode(ISD::BITCAST, dl, NVT, Node->getOperand(j));
else
More information about the llvm-commits
mailing list