[llvm] [RISCV] Remove codegen for vp_fsub (PR #191832)
Luke Lau via llvm-commits
llvm-commits at lists.llvm.org
Mon Apr 13 07:56:31 PDT 2026
https://github.com/lukel97 created https://github.com/llvm/llvm-project/pull/191832
Part of the work to remove trivial VP intrinsics from the RISC-V backend, see https://discourse.llvm.org/t/rfc-remove-codegen-support-for-trivial-vp-intrinsics-in-the-risc-v-backend/87999
This splits off the vp_fsub intrinsic from #179622.
>From 8509fc64c16d53b1574f8e8fe4aa316e1f2c412f Mon Sep 17 00:00:00 2001
From: Luke Lau <luke at igalia.com>
Date: Mon, 13 Apr 2026 15:55:06 +0100
Subject: [PATCH] [RISCV] Remove codegen for vp_fsub
Part of the work to remove trivial VP intrinsics from the RISC-V backend, see https://discourse.llvm.org/t/rfc-remove-codegen-support-for-trivial-vp-intrinsics-in-the-risc-v-backend/87999
This splits off the vp_fsub intrinsic from #179622.
---
llvm/lib/Target/RISCV/RISCVISelLowering.cpp | 5 +-
.../Target/RISCV/RISCVTargetTransformInfo.h | 1 -
.../RISCV/rvv/fixed-vectors-vfrsub-vp.ll | 72 +-
.../RISCV/rvv/fixed-vectors-vfsub-vp.ll | 278 +++---
.../RISCV/rvv/fold-vp-fsub-and-vp-fmul.ll | 15 +-
.../CodeGen/RISCV/rvv/sink-splat-operands.ll | 20 +-
llvm/test/CodeGen/RISCV/rvv/undef-vp-ops.ll | 4 +
llvm/test/CodeGen/RISCV/rvv/vfrsub-vp.ll | 90 +-
llvm/test/CodeGen/RISCV/rvv/vfsub-vp.ll | 942 +++++++-----------
9 files changed, 616 insertions(+), 811 deletions(-)
diff --git a/llvm/lib/Target/RISCV/RISCVISelLowering.cpp b/llvm/lib/Target/RISCV/RISCVISelLowering.cpp
index f4e8bebaf2814..ca84eb3507a67 100644
--- a/llvm/lib/Target/RISCV/RISCVISelLowering.cpp
+++ b/llvm/lib/Target/RISCV/RISCVISelLowering.cpp
@@ -881,7 +881,7 @@ RISCVTargetLowering::RISCVTargetLowering(const TargetMachine &TM,
ISD::VP_CTTZ_ELTS, ISD::VP_CTTZ_ELTS_ZERO_UNDEF};
static const unsigned FloatingPointVPOps[] = {
- ISD::VP_FADD, ISD::VP_FSUB, ISD::VP_FMUL,
+ ISD::VP_FADD, ISD::VP_FMUL,
ISD::VP_FNEG,
ISD::VP_FMA, ISD::VP_REDUCE_FADD, ISD::VP_REDUCE_SEQ_FADD,
ISD::VP_REDUCE_FMIN, ISD::VP_REDUCE_FMAX, ISD::VP_MERGE,
@@ -1202,7 +1202,6 @@ RISCVTargetLowering::RISCVTargetLowering(const TargetMachine &TM,
// TODO: support more vp ops.
static const unsigned ZvfhminZvfbfminPromoteVPOps[] = {
ISD::VP_FADD,
- ISD::VP_FSUB,
ISD::VP_FMUL,
ISD::VP_FMA,
ISD::VP_REDUCE_FMIN,
@@ -7547,7 +7546,6 @@ static unsigned getRISCVVLOp(SDValue Op) {
VP_CASE(UREM) // VP_UREM
VP_CASE(SHL) // VP_SHL
VP_CASE(FADD) // VP_FADD
- VP_CASE(FSUB) // VP_FSUB
VP_CASE(FMUL) // VP_FMUL
VP_CASE(FNEG) // VP_FNEG
VP_CASE(SMIN) // VP_SMIN
@@ -8949,7 +8947,6 @@ SDValue RISCVTargetLowering::LowerOperation(SDValue Op,
case ISD::VP_XOR:
return lowerLogicVPOp(Op, DAG);
case ISD::VP_FADD:
- case ISD::VP_FSUB:
case ISD::VP_FMUL:
case ISD::VP_FNEG:
case ISD::VP_SQRT:
diff --git a/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.h b/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.h
index 854157a725c9d..e03cfef8236e2 100644
--- a/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.h
+++ b/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.h
@@ -383,7 +383,6 @@ class RISCVTTIImpl final : public BasicTTIImplBase<RISCVTTIImpl> {
Intrinsic::vp_frem,
Intrinsic::vp_fshl,
Intrinsic::vp_fshr,
- Intrinsic::vp_fsub,
Intrinsic::vp_gather,
Intrinsic::vp_icmp,
Intrinsic::vp_inttoptr,
diff --git a/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vfrsub-vp.ll b/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vfrsub-vp.ll
index 30b840a2f6b4e..c118ebdcfc8ac 100644
--- a/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vfrsub-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vfrsub-vp.ll
@@ -7,8 +7,8 @@
define <2 x half> @vfrsub_vf_v2f16(<2 x half> %va, half %b, <2 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_v2f16:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
-; CHECK-NEXT: vfrsub.vf v8, v8, fa0, v0.t
+; CHECK-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
+; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <2 x half> poison, half %b, i32 0
%vb = shufflevector <2 x half> %elt.head, <2 x half> poison, <2 x i32> zeroinitializer
@@ -19,7 +19,7 @@ define <2 x half> @vfrsub_vf_v2f16(<2 x half> %va, half %b, <2 x i1> %m, i32 zer
define <2 x half> @vfrsub_vf_v2f16_unmasked(<2 x half> %va, half %b, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_v2f16_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; CHECK-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <2 x half> poison, half %b, i32 0
@@ -31,8 +31,8 @@ define <2 x half> @vfrsub_vf_v2f16_unmasked(<2 x half> %va, half %b, i32 zeroext
define <4 x half> @vfrsub_vf_v4f16(<4 x half> %va, half %b, <4 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_v4f16:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
-; CHECK-NEXT: vfrsub.vf v8, v8, fa0, v0.t
+; CHECK-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
+; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <4 x half> poison, half %b, i32 0
%vb = shufflevector <4 x half> %elt.head, <4 x half> poison, <4 x i32> zeroinitializer
@@ -43,7 +43,7 @@ define <4 x half> @vfrsub_vf_v4f16(<4 x half> %va, half %b, <4 x i1> %m, i32 zer
define <4 x half> @vfrsub_vf_v4f16_unmasked(<4 x half> %va, half %b, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_v4f16_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; CHECK-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <4 x half> poison, half %b, i32 0
@@ -55,8 +55,8 @@ define <4 x half> @vfrsub_vf_v4f16_unmasked(<4 x half> %va, half %b, i32 zeroext
define <8 x half> @vfrsub_vf_v8f16(<8 x half> %va, half %b, <8 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_v8f16:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e16, m1, ta, ma
-; CHECK-NEXT: vfrsub.vf v8, v8, fa0, v0.t
+; CHECK-NEXT: vsetivli zero, 8, e16, m1, ta, ma
+; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <8 x half> poison, half %b, i32 0
%vb = shufflevector <8 x half> %elt.head, <8 x half> poison, <8 x i32> zeroinitializer
@@ -67,7 +67,7 @@ define <8 x half> @vfrsub_vf_v8f16(<8 x half> %va, half %b, <8 x i1> %m, i32 zer
define <8 x half> @vfrsub_vf_v8f16_unmasked(<8 x half> %va, half %b, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_v8f16_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; CHECK-NEXT: vsetivli zero, 8, e16, m1, ta, ma
; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <8 x half> poison, half %b, i32 0
@@ -79,8 +79,8 @@ define <8 x half> @vfrsub_vf_v8f16_unmasked(<8 x half> %va, half %b, i32 zeroext
define <16 x half> @vfrsub_vf_v16f16(<16 x half> %va, half %b, <16 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_v16f16:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e16, m2, ta, ma
-; CHECK-NEXT: vfrsub.vf v8, v8, fa0, v0.t
+; CHECK-NEXT: vsetivli zero, 16, e16, m2, ta, ma
+; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <16 x half> poison, half %b, i32 0
%vb = shufflevector <16 x half> %elt.head, <16 x half> poison, <16 x i32> zeroinitializer
@@ -91,7 +91,7 @@ define <16 x half> @vfrsub_vf_v16f16(<16 x half> %va, half %b, <16 x i1> %m, i32
define <16 x half> @vfrsub_vf_v16f16_unmasked(<16 x half> %va, half %b, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_v16f16_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; CHECK-NEXT: vsetivli zero, 16, e16, m2, ta, ma
; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <16 x half> poison, half %b, i32 0
@@ -103,8 +103,8 @@ define <16 x half> @vfrsub_vf_v16f16_unmasked(<16 x half> %va, half %b, i32 zero
define <2 x float> @vfrsub_vf_v2f32(<2 x float> %va, float %b, <2 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_v2f32:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
-; CHECK-NEXT: vfrsub.vf v8, v8, fa0, v0.t
+; CHECK-NEXT: vsetivli zero, 2, e32, mf2, ta, ma
+; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <2 x float> poison, float %b, i32 0
%vb = shufflevector <2 x float> %elt.head, <2 x float> poison, <2 x i32> zeroinitializer
@@ -115,7 +115,7 @@ define <2 x float> @vfrsub_vf_v2f32(<2 x float> %va, float %b, <2 x i1> %m, i32
define <2 x float> @vfrsub_vf_v2f32_unmasked(<2 x float> %va, float %b, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_v2f32_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT: vsetivli zero, 2, e32, mf2, ta, ma
; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <2 x float> poison, float %b, i32 0
@@ -127,8 +127,8 @@ define <2 x float> @vfrsub_vf_v2f32_unmasked(<2 x float> %va, float %b, i32 zero
define <4 x float> @vfrsub_vf_v4f32(<4 x float> %va, float %b, <4 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_v4f32:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, m1, ta, ma
-; CHECK-NEXT: vfrsub.vf v8, v8, fa0, v0.t
+; CHECK-NEXT: vsetivli zero, 4, e32, m1, ta, ma
+; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <4 x float> poison, float %b, i32 0
%vb = shufflevector <4 x float> %elt.head, <4 x float> poison, <4 x i32> zeroinitializer
@@ -139,7 +139,7 @@ define <4 x float> @vfrsub_vf_v4f32(<4 x float> %va, float %b, <4 x i1> %m, i32
define <4 x float> @vfrsub_vf_v4f32_unmasked(<4 x float> %va, float %b, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_v4f32_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; CHECK-NEXT: vsetivli zero, 4, e32, m1, ta, ma
; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <4 x float> poison, float %b, i32 0
@@ -151,8 +151,8 @@ define <4 x float> @vfrsub_vf_v4f32_unmasked(<4 x float> %va, float %b, i32 zero
define <8 x float> @vfrsub_vf_v8f32(<8 x float> %va, float %b, <8 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_v8f32:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, m2, ta, ma
-; CHECK-NEXT: vfrsub.vf v8, v8, fa0, v0.t
+; CHECK-NEXT: vsetivli zero, 8, e32, m2, ta, ma
+; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <8 x float> poison, float %b, i32 0
%vb = shufflevector <8 x float> %elt.head, <8 x float> poison, <8 x i32> zeroinitializer
@@ -163,7 +163,7 @@ define <8 x float> @vfrsub_vf_v8f32(<8 x float> %va, float %b, <8 x i1> %m, i32
define <8 x float> @vfrsub_vf_v8f32_unmasked(<8 x float> %va, float %b, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_v8f32_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; CHECK-NEXT: vsetivli zero, 8, e32, m2, ta, ma
; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <8 x float> poison, float %b, i32 0
@@ -175,8 +175,8 @@ define <8 x float> @vfrsub_vf_v8f32_unmasked(<8 x float> %va, float %b, i32 zero
define <16 x float> @vfrsub_vf_v16f32(<16 x float> %va, float %b, <16 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_v16f32:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, m4, ta, ma
-; CHECK-NEXT: vfrsub.vf v8, v8, fa0, v0.t
+; CHECK-NEXT: vsetivli zero, 16, e32, m4, ta, ma
+; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <16 x float> poison, float %b, i32 0
%vb = shufflevector <16 x float> %elt.head, <16 x float> poison, <16 x i32> zeroinitializer
@@ -187,7 +187,7 @@ define <16 x float> @vfrsub_vf_v16f32(<16 x float> %va, float %b, <16 x i1> %m,
define <16 x float> @vfrsub_vf_v16f32_unmasked(<16 x float> %va, float %b, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_v16f32_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT: vsetivli zero, 16, e32, m4, ta, ma
; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <16 x float> poison, float %b, i32 0
@@ -199,8 +199,8 @@ define <16 x float> @vfrsub_vf_v16f32_unmasked(<16 x float> %va, float %b, i32 z
define <2 x double> @vfrsub_vf_v2f64(<2 x double> %va, double %b, <2 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_v2f64:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e64, m1, ta, ma
-; CHECK-NEXT: vfrsub.vf v8, v8, fa0, v0.t
+; CHECK-NEXT: vsetivli zero, 2, e64, m1, ta, ma
+; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <2 x double> poison, double %b, i32 0
%vb = shufflevector <2 x double> %elt.head, <2 x double> poison, <2 x i32> zeroinitializer
@@ -211,7 +211,7 @@ define <2 x double> @vfrsub_vf_v2f64(<2 x double> %va, double %b, <2 x i1> %m, i
define <2 x double> @vfrsub_vf_v2f64_unmasked(<2 x double> %va, double %b, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_v2f64_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e64, m1, ta, ma
+; CHECK-NEXT: vsetivli zero, 2, e64, m1, ta, ma
; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <2 x double> poison, double %b, i32 0
@@ -223,8 +223,8 @@ define <2 x double> @vfrsub_vf_v2f64_unmasked(<2 x double> %va, double %b, i32 z
define <4 x double> @vfrsub_vf_v4f64(<4 x double> %va, double %b, <4 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_v4f64:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e64, m2, ta, ma
-; CHECK-NEXT: vfrsub.vf v8, v8, fa0, v0.t
+; CHECK-NEXT: vsetivli zero, 4, e64, m2, ta, ma
+; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <4 x double> poison, double %b, i32 0
%vb = shufflevector <4 x double> %elt.head, <4 x double> poison, <4 x i32> zeroinitializer
@@ -235,7 +235,7 @@ define <4 x double> @vfrsub_vf_v4f64(<4 x double> %va, double %b, <4 x i1> %m, i
define <4 x double> @vfrsub_vf_v4f64_unmasked(<4 x double> %va, double %b, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_v4f64_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e64, m2, ta, ma
+; CHECK-NEXT: vsetivli zero, 4, e64, m2, ta, ma
; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <4 x double> poison, double %b, i32 0
@@ -247,8 +247,8 @@ define <4 x double> @vfrsub_vf_v4f64_unmasked(<4 x double> %va, double %b, i32 z
define <8 x double> @vfrsub_vf_v8f64(<8 x double> %va, double %b, <8 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_v8f64:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e64, m4, ta, ma
-; CHECK-NEXT: vfrsub.vf v8, v8, fa0, v0.t
+; CHECK-NEXT: vsetivli zero, 8, e64, m4, ta, ma
+; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <8 x double> poison, double %b, i32 0
%vb = shufflevector <8 x double> %elt.head, <8 x double> poison, <8 x i32> zeroinitializer
@@ -259,7 +259,7 @@ define <8 x double> @vfrsub_vf_v8f64(<8 x double> %va, double %b, <8 x i1> %m, i
define <8 x double> @vfrsub_vf_v8f64_unmasked(<8 x double> %va, double %b, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_v8f64_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e64, m4, ta, ma
+; CHECK-NEXT: vsetivli zero, 8, e64, m4, ta, ma
; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <8 x double> poison, double %b, i32 0
@@ -271,8 +271,8 @@ define <8 x double> @vfrsub_vf_v8f64_unmasked(<8 x double> %va, double %b, i32 z
define <16 x double> @vfrsub_vf_v16f64(<16 x double> %va, double %b, <16 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_v16f64:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e64, m8, ta, ma
-; CHECK-NEXT: vfrsub.vf v8, v8, fa0, v0.t
+; CHECK-NEXT: vsetivli zero, 16, e64, m8, ta, ma
+; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <16 x double> poison, double %b, i32 0
%vb = shufflevector <16 x double> %elt.head, <16 x double> poison, <16 x i32> zeroinitializer
@@ -283,7 +283,7 @@ define <16 x double> @vfrsub_vf_v16f64(<16 x double> %va, double %b, <16 x i1> %
define <16 x double> @vfrsub_vf_v16f64_unmasked(<16 x double> %va, double %b, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_v16f64_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e64, m8, ta, ma
+; CHECK-NEXT: vsetivli zero, 16, e64, m8, ta, ma
; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <16 x double> poison, double %b, i32 0
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 6299d5d86acdd..efc8e481d4cb1 100644
--- a/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vfsub-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vfsub-vp.ll
@@ -11,19 +11,19 @@
define <2 x half> @vfsub_vv_v2f16(<2 x half> %va, <2 x half> %b, <2 x i1> %m, i32 zeroext %evl) {
; ZVFH-LABEL: vfsub_vv_v2f16:
; ZVFH: # %bb.0:
-; ZVFH-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
-; ZVFH-NEXT: vfsub.vv v8, v8, v9, v0.t
+; ZVFH-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
+; ZVFH-NEXT: vfsub.vv v8, v8, v9
; ZVFH-NEXT: ret
;
; ZVFHMIN-LABEL: vfsub_vv_v2f16:
; ZVFHMIN: # %bb.0:
-; 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: 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, zero, e32, mf2, ta, ma
-; ZVFHMIN-NEXT: vfsub.vv v9, v9, v10, v0.t
+; ZVFHMIN-NEXT: vfsub.vv v9, v9, v10
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9, v0.t
+; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
; ZVFHMIN-NEXT: ret
%v = call <2 x half> @llvm.vp.fsub.v2f16(<2 x half> %va, <2 x half> %b, <2 x i1> %m, i32 %evl)
ret <2 x half> %v
@@ -32,13 +32,13 @@ define <2 x half> @vfsub_vv_v2f16(<2 x half> %va, <2 x half> %b, <2 x i1> %m, i3
define <2 x half> @vfsub_vv_v2f16_unmasked(<2 x half> %va, <2 x half> %b, i32 zeroext %evl) {
; ZVFH-LABEL: vfsub_vv_v2f16_unmasked:
; ZVFH: # %bb.0:
-; ZVFH-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFH-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
; ZVFH-NEXT: vfsub.vv v8, v8, v9
; ZVFH-NEXT: ret
;
; ZVFHMIN-LABEL: vfsub_vv_v2f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; 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, zero, e32, mf2, ta, ma
@@ -53,22 +53,21 @@ define <2 x half> @vfsub_vv_v2f16_unmasked(<2 x half> %va, <2 x half> %b, i32 ze
define <2 x half> @vfsub_vf_v2f16(<2 x half> %va, half %b, <2 x i1> %m, i32 zeroext %evl) {
; ZVFH-LABEL: vfsub_vf_v2f16:
; ZVFH: # %bb.0:
-; ZVFH-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
-; ZVFH-NEXT: vfsub.vf v8, v8, fa0, v0.t
+; ZVFH-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
+; ZVFH-NEXT: vfsub.vf v8, v8, fa0
; ZVFH-NEXT: ret
;
; ZVFHMIN-LABEL: vfsub_vf_v2f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: fmv.x.h a1, fa0
+; ZVFHMIN-NEXT: fmv.x.h a0, fa0
; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vmv.v.x v9, a1
-; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8, v0.t
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v9, v0.t
+; ZVFHMIN-NEXT: vmv.v.x v9, a0
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v9
; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
-; ZVFHMIN-NEXT: vfsub.vv v9, v10, v8, v0.t
+; ZVFHMIN-NEXT: vfsub.vv v9, v10, v8
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9, v0.t
+; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
; ZVFHMIN-NEXT: ret
%elt.head = insertelement <2 x half> poison, half %b, i32 0
%vb = shufflevector <2 x half> %elt.head, <2 x half> poison, <2 x i32> zeroinitializer
@@ -79,16 +78,15 @@ define <2 x half> @vfsub_vf_v2f16(<2 x half> %va, half %b, <2 x i1> %m, i32 zero
define <2 x half> @vfsub_vf_v2f16_unmasked(<2 x half> %va, half %b, i32 zeroext %evl) {
; ZVFH-LABEL: vfsub_vf_v2f16_unmasked:
; ZVFH: # %bb.0:
-; ZVFH-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFH-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
; ZVFH-NEXT: vfsub.vf v8, v8, fa0
; ZVFH-NEXT: ret
;
; ZVFHMIN-LABEL: vfsub_vf_v2f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: fmv.x.h a1, fa0
+; ZVFHMIN-NEXT: fmv.x.h a0, fa0
; ZVFHMIN-NEXT: vsetivli zero, 2, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vmv.v.x v9, a1
-; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vmv.v.x v9, a0
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v9
; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
@@ -105,19 +103,19 @@ define <2 x half> @vfsub_vf_v2f16_unmasked(<2 x half> %va, half %b, i32 zeroext
define <3 x half> @vfsub_vv_v3f16(<3 x half> %va, <3 x half> %b, <3 x i1> %m, i32 zeroext %evl) {
; ZVFH-LABEL: vfsub_vv_v3f16:
; ZVFH: # %bb.0:
-; ZVFH-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
-; ZVFH-NEXT: vfsub.vv v8, v8, v9, v0.t
+; ZVFH-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
+; ZVFH-NEXT: vfsub.vv v8, v8, v9
; ZVFH-NEXT: ret
;
; ZVFHMIN-LABEL: vfsub_vv_v3f16:
; ZVFHMIN: # %bb.0:
-; 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: 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, zero, e32, m1, ta, ma
-; ZVFHMIN-NEXT: vfsub.vv v9, v9, v10, v0.t
+; ZVFHMIN-NEXT: vfsub.vv v9, v9, v10
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9, v0.t
+; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
; ZVFHMIN-NEXT: ret
%v = call <3 x half> @llvm.vp.fsub.v3f16(<3 x half> %va, <3 x half> %b, <3 x i1> %m, i32 %evl)
ret <3 x half> %v
@@ -126,19 +124,19 @@ define <3 x half> @vfsub_vv_v3f16(<3 x half> %va, <3 x half> %b, <3 x i1> %m, i3
define <4 x half> @vfsub_vv_v4f16(<4 x half> %va, <4 x half> %b, <4 x i1> %m, i32 zeroext %evl) {
; ZVFH-LABEL: vfsub_vv_v4f16:
; ZVFH: # %bb.0:
-; ZVFH-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
-; ZVFH-NEXT: vfsub.vv v8, v8, v9, v0.t
+; ZVFH-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
+; ZVFH-NEXT: vfsub.vv v8, v8, v9
; ZVFH-NEXT: ret
;
; ZVFHMIN-LABEL: vfsub_vv_v4f16:
; ZVFHMIN: # %bb.0:
-; 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: 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, zero, e32, m1, ta, ma
-; ZVFHMIN-NEXT: vfsub.vv v9, v9, v10, v0.t
+; ZVFHMIN-NEXT: vfsub.vv v9, v9, v10
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9, v0.t
+; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
; ZVFHMIN-NEXT: ret
%v = call <4 x half> @llvm.vp.fsub.v4f16(<4 x half> %va, <4 x half> %b, <4 x i1> %m, i32 %evl)
ret <4 x half> %v
@@ -147,13 +145,13 @@ define <4 x half> @vfsub_vv_v4f16(<4 x half> %va, <4 x half> %b, <4 x i1> %m, i3
define <4 x half> @vfsub_vv_v4f16_unmasked(<4 x half> %va, <4 x half> %b, i32 zeroext %evl) {
; ZVFH-LABEL: vfsub_vv_v4f16_unmasked:
; ZVFH: # %bb.0:
-; ZVFH-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFH-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
; ZVFH-NEXT: vfsub.vv v8, v8, v9
; ZVFH-NEXT: ret
;
; ZVFHMIN-LABEL: vfsub_vv_v4f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; 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, zero, e32, m1, ta, ma
@@ -168,22 +166,21 @@ define <4 x half> @vfsub_vv_v4f16_unmasked(<4 x half> %va, <4 x half> %b, i32 ze
define <4 x half> @vfsub_vf_v4f16(<4 x half> %va, half %b, <4 x i1> %m, i32 zeroext %evl) {
; ZVFH-LABEL: vfsub_vf_v4f16:
; ZVFH: # %bb.0:
-; ZVFH-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
-; ZVFH-NEXT: vfsub.vf v8, v8, fa0, v0.t
+; ZVFH-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
+; ZVFH-NEXT: vfsub.vf v8, v8, fa0
; ZVFH-NEXT: ret
;
; ZVFHMIN-LABEL: vfsub_vf_v4f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: fmv.x.h a1, fa0
+; ZVFHMIN-NEXT: fmv.x.h a0, fa0
; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vmv.v.x v9, a1
-; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8, v0.t
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v9, v0.t
+; ZVFHMIN-NEXT: vmv.v.x v9, a0
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v9
; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
-; ZVFHMIN-NEXT: vfsub.vv v9, v10, v8, v0.t
+; ZVFHMIN-NEXT: vfsub.vv v9, v10, v8
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9, v0.t
+; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
; ZVFHMIN-NEXT: ret
%elt.head = insertelement <4 x half> poison, half %b, i32 0
%vb = shufflevector <4 x half> %elt.head, <4 x half> poison, <4 x i32> zeroinitializer
@@ -194,16 +191,15 @@ define <4 x half> @vfsub_vf_v4f16(<4 x half> %va, half %b, <4 x i1> %m, i32 zero
define <4 x half> @vfsub_vf_v4f16_unmasked(<4 x half> %va, half %b, i32 zeroext %evl) {
; ZVFH-LABEL: vfsub_vf_v4f16_unmasked:
; ZVFH: # %bb.0:
-; ZVFH-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFH-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
; ZVFH-NEXT: vfsub.vf v8, v8, fa0
; ZVFH-NEXT: ret
;
; ZVFHMIN-LABEL: vfsub_vf_v4f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: fmv.x.h a1, fa0
+; ZVFHMIN-NEXT: fmv.x.h a0, fa0
; ZVFHMIN-NEXT: vsetivli zero, 4, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vmv.v.x v9, a1
-; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vmv.v.x v9, a0
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v9
; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
@@ -220,19 +216,19 @@ define <4 x half> @vfsub_vf_v4f16_unmasked(<4 x half> %va, half %b, i32 zeroext
define <8 x half> @vfsub_vv_v8f16(<8 x half> %va, <8 x half> %b, <8 x i1> %m, i32 zeroext %evl) {
; ZVFH-LABEL: vfsub_vv_v8f16:
; ZVFH: # %bb.0:
-; ZVFH-NEXT: vsetvli zero, a0, e16, m1, ta, ma
-; ZVFH-NEXT: vfsub.vv v8, v8, v9, v0.t
+; ZVFH-NEXT: vsetivli zero, 8, e16, m1, ta, ma
+; ZVFH-NEXT: vfsub.vv v8, v8, v9
; ZVFH-NEXT: ret
;
; ZVFHMIN-LABEL: vfsub_vv_v8f16:
; ZVFHMIN: # %bb.0:
-; 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: 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, zero, e32, m2, ta, ma
-; ZVFHMIN-NEXT: vfsub.vv v10, v12, v10, v0.t
+; ZVFHMIN-NEXT: vfsub.vv v10, v12, v10
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10, v0.t
+; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
; ZVFHMIN-NEXT: ret
%v = call <8 x half> @llvm.vp.fsub.v8f16(<8 x half> %va, <8 x half> %b, <8 x i1> %m, i32 %evl)
ret <8 x half> %v
@@ -241,13 +237,13 @@ define <8 x half> @vfsub_vv_v8f16(<8 x half> %va, <8 x half> %b, <8 x i1> %m, i3
define <8 x half> @vfsub_vv_v8f16_unmasked(<8 x half> %va, <8 x half> %b, i32 zeroext %evl) {
; ZVFH-LABEL: vfsub_vv_v8f16_unmasked:
; ZVFH: # %bb.0:
-; ZVFH-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFH-NEXT: vsetivli zero, 8, e16, m1, ta, ma
; ZVFH-NEXT: vfsub.vv v8, v8, v9
; ZVFH-NEXT: ret
;
; ZVFHMIN-LABEL: vfsub_vv_v8f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; 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, zero, e32, m2, ta, ma
@@ -262,22 +258,21 @@ define <8 x half> @vfsub_vv_v8f16_unmasked(<8 x half> %va, <8 x half> %b, i32 ze
define <8 x half> @vfsub_vf_v8f16(<8 x half> %va, half %b, <8 x i1> %m, i32 zeroext %evl) {
; ZVFH-LABEL: vfsub_vf_v8f16:
; ZVFH: # %bb.0:
-; ZVFH-NEXT: vsetvli zero, a0, e16, m1, ta, ma
-; ZVFH-NEXT: vfsub.vf v8, v8, fa0, v0.t
+; ZVFH-NEXT: vsetivli zero, 8, e16, m1, ta, ma
+; ZVFH-NEXT: vfsub.vf v8, v8, fa0
; ZVFH-NEXT: ret
;
; ZVFHMIN-LABEL: vfsub_vf_v8f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: fmv.x.h a1, fa0
+; ZVFHMIN-NEXT: fmv.x.h a0, fa0
; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vmv.v.x v12, a1
-; 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 v8, v12, v0.t
+; ZVFHMIN-NEXT: vmv.v.x v12, a0
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v12
; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
-; ZVFHMIN-NEXT: vfsub.vv v10, v10, v8, v0.t
+; ZVFHMIN-NEXT: vfsub.vv v10, v10, v8
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10, v0.t
+; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
; ZVFHMIN-NEXT: ret
%elt.head = insertelement <8 x half> poison, half %b, i32 0
%vb = shufflevector <8 x half> %elt.head, <8 x half> poison, <8 x i32> zeroinitializer
@@ -288,16 +283,15 @@ define <8 x half> @vfsub_vf_v8f16(<8 x half> %va, half %b, <8 x i1> %m, i32 zero
define <8 x half> @vfsub_vf_v8f16_unmasked(<8 x half> %va, half %b, i32 zeroext %evl) {
; ZVFH-LABEL: vfsub_vf_v8f16_unmasked:
; ZVFH: # %bb.0:
-; ZVFH-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFH-NEXT: vsetivli zero, 8, e16, m1, ta, ma
; ZVFH-NEXT: vfsub.vf v8, v8, fa0
; ZVFH-NEXT: ret
;
; ZVFHMIN-LABEL: vfsub_vf_v8f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: fmv.x.h a1, fa0
+; ZVFHMIN-NEXT: fmv.x.h a0, fa0
; ZVFHMIN-NEXT: vsetivli zero, 8, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vmv.v.x v12, a1
-; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vmv.v.x v12, a0
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v12
; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
@@ -314,19 +308,19 @@ define <8 x half> @vfsub_vf_v8f16_unmasked(<8 x half> %va, half %b, i32 zeroext
define <16 x half> @vfsub_vv_v16f16(<16 x half> %va, <16 x half> %b, <16 x i1> %m, i32 zeroext %evl) {
; ZVFH-LABEL: vfsub_vv_v16f16:
; ZVFH: # %bb.0:
-; ZVFH-NEXT: vsetvli zero, a0, e16, m2, ta, ma
-; ZVFH-NEXT: vfsub.vv v8, v8, v10, v0.t
+; ZVFH-NEXT: vsetivli zero, 16, e16, m2, ta, ma
+; ZVFH-NEXT: vfsub.vv v8, v8, v10
; ZVFH-NEXT: ret
;
; ZVFHMIN-LABEL: vfsub_vv_v16f16:
; ZVFHMIN: # %bb.0:
-; 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: 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, zero, e32, m4, ta, ma
-; ZVFHMIN-NEXT: vfsub.vv v12, v16, v12, v0.t
+; ZVFHMIN-NEXT: vfsub.vv v12, v16, v12
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12, v0.t
+; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
; ZVFHMIN-NEXT: ret
%v = call <16 x half> @llvm.vp.fsub.v16f16(<16 x half> %va, <16 x half> %b, <16 x i1> %m, i32 %evl)
ret <16 x half> %v
@@ -335,13 +329,13 @@ define <16 x half> @vfsub_vv_v16f16(<16 x half> %va, <16 x half> %b, <16 x i1> %
define <16 x half> @vfsub_vv_v16f16_unmasked(<16 x half> %va, <16 x half> %b, i32 zeroext %evl) {
; ZVFH-LABEL: vfsub_vv_v16f16_unmasked:
; ZVFH: # %bb.0:
-; ZVFH-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFH-NEXT: vsetivli zero, 16, e16, m2, ta, ma
; ZVFH-NEXT: vfsub.vv v8, v8, v10
; ZVFH-NEXT: ret
;
; ZVFHMIN-LABEL: vfsub_vv_v16f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; 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, zero, e32, m4, ta, ma
@@ -356,22 +350,21 @@ define <16 x half> @vfsub_vv_v16f16_unmasked(<16 x half> %va, <16 x half> %b, i3
define <16 x half> @vfsub_vf_v16f16(<16 x half> %va, half %b, <16 x i1> %m, i32 zeroext %evl) {
; ZVFH-LABEL: vfsub_vf_v16f16:
; ZVFH: # %bb.0:
-; ZVFH-NEXT: vsetvli zero, a0, e16, m2, ta, ma
-; ZVFH-NEXT: vfsub.vf v8, v8, fa0, v0.t
+; ZVFH-NEXT: vsetivli zero, 16, e16, m2, ta, ma
+; ZVFH-NEXT: vfsub.vf v8, v8, fa0
; ZVFH-NEXT: ret
;
; ZVFHMIN-LABEL: vfsub_vf_v16f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: fmv.x.h a1, fa0
+; ZVFHMIN-NEXT: fmv.x.h a0, fa0
; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vmv.v.x v16, a1
-; 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 v8, v16, v0.t
+; ZVFHMIN-NEXT: vmv.v.x v16, a0
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v16
; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
-; ZVFHMIN-NEXT: vfsub.vv v12, v12, v8, v0.t
+; ZVFHMIN-NEXT: vfsub.vv v12, v12, v8
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12, v0.t
+; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
; ZVFHMIN-NEXT: ret
%elt.head = insertelement <16 x half> poison, half %b, i32 0
%vb = shufflevector <16 x half> %elt.head, <16 x half> poison, <16 x i32> zeroinitializer
@@ -382,16 +375,15 @@ define <16 x half> @vfsub_vf_v16f16(<16 x half> %va, half %b, <16 x i1> %m, i32
define <16 x half> @vfsub_vf_v16f16_unmasked(<16 x half> %va, half %b, i32 zeroext %evl) {
; ZVFH-LABEL: vfsub_vf_v16f16_unmasked:
; ZVFH: # %bb.0:
-; ZVFH-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFH-NEXT: vsetivli zero, 16, e16, m2, ta, ma
; ZVFH-NEXT: vfsub.vf v8, v8, fa0
; ZVFH-NEXT: ret
;
; ZVFHMIN-LABEL: vfsub_vf_v16f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: fmv.x.h a1, fa0
+; ZVFHMIN-NEXT: fmv.x.h a0, fa0
; ZVFHMIN-NEXT: vsetivli zero, 16, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vmv.v.x v16, a1
-; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vmv.v.x v16, a0
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v16
; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
@@ -408,8 +400,8 @@ define <16 x half> @vfsub_vf_v16f16_unmasked(<16 x half> %va, half %b, i32 zeroe
define <2 x float> @vfsub_vv_v2f32(<2 x float> %va, <2 x float> %b, <2 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vv_v2f32:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
-; CHECK-NEXT: vfsub.vv v8, v8, v9, v0.t
+; CHECK-NEXT: vsetivli zero, 2, e32, mf2, ta, ma
+; CHECK-NEXT: vfsub.vv v8, v8, v9
; CHECK-NEXT: ret
%v = call <2 x float> @llvm.vp.fsub.v2f32(<2 x float> %va, <2 x float> %b, <2 x i1> %m, i32 %evl)
ret <2 x float> %v
@@ -418,7 +410,7 @@ define <2 x float> @vfsub_vv_v2f32(<2 x float> %va, <2 x float> %b, <2 x i1> %m,
define <2 x float> @vfsub_vv_v2f32_unmasked(<2 x float> %va, <2 x float> %b, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vv_v2f32_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT: vsetivli zero, 2, e32, mf2, ta, ma
; CHECK-NEXT: vfsub.vv v8, v8, v9
; CHECK-NEXT: ret
%v = call <2 x float> @llvm.vp.fsub.v2f32(<2 x float> %va, <2 x float> %b, <2 x i1> splat (i1 true), i32 %evl)
@@ -428,8 +420,8 @@ define <2 x float> @vfsub_vv_v2f32_unmasked(<2 x float> %va, <2 x float> %b, i32
define <2 x float> @vfsub_vf_v2f32(<2 x float> %va, float %b, <2 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vf_v2f32:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
-; CHECK-NEXT: vfsub.vf v8, v8, fa0, v0.t
+; CHECK-NEXT: vsetivli zero, 2, e32, mf2, ta, ma
+; CHECK-NEXT: vfsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <2 x float> poison, float %b, i32 0
%vb = shufflevector <2 x float> %elt.head, <2 x float> poison, <2 x i32> zeroinitializer
@@ -440,7 +432,7 @@ define <2 x float> @vfsub_vf_v2f32(<2 x float> %va, float %b, <2 x i1> %m, i32 z
define <2 x float> @vfsub_vf_v2f32_unmasked(<2 x float> %va, float %b, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vf_v2f32_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT: vsetivli zero, 2, e32, mf2, ta, ma
; CHECK-NEXT: vfsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <2 x float> poison, float %b, i32 0
@@ -452,8 +444,8 @@ define <2 x float> @vfsub_vf_v2f32_unmasked(<2 x float> %va, float %b, i32 zeroe
define <4 x float> @vfsub_vv_v4f32(<4 x float> %va, <4 x float> %b, <4 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vv_v4f32:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, m1, ta, ma
-; CHECK-NEXT: vfsub.vv v8, v8, v9, v0.t
+; CHECK-NEXT: vsetivli zero, 4, e32, m1, ta, ma
+; CHECK-NEXT: vfsub.vv v8, v8, v9
; CHECK-NEXT: ret
%v = call <4 x float> @llvm.vp.fsub.v4f32(<4 x float> %va, <4 x float> %b, <4 x i1> %m, i32 %evl)
ret <4 x float> %v
@@ -462,7 +454,7 @@ define <4 x float> @vfsub_vv_v4f32(<4 x float> %va, <4 x float> %b, <4 x i1> %m,
define <4 x float> @vfsub_vv_v4f32_unmasked(<4 x float> %va, <4 x float> %b, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vv_v4f32_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; CHECK-NEXT: vsetivli zero, 4, e32, m1, ta, ma
; CHECK-NEXT: vfsub.vv v8, v8, v9
; CHECK-NEXT: ret
%v = call <4 x float> @llvm.vp.fsub.v4f32(<4 x float> %va, <4 x float> %b, <4 x i1> splat (i1 true), i32 %evl)
@@ -472,8 +464,8 @@ define <4 x float> @vfsub_vv_v4f32_unmasked(<4 x float> %va, <4 x float> %b, i32
define <4 x float> @vfsub_vf_v4f32(<4 x float> %va, float %b, <4 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vf_v4f32:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, m1, ta, ma
-; CHECK-NEXT: vfsub.vf v8, v8, fa0, v0.t
+; CHECK-NEXT: vsetivli zero, 4, e32, m1, ta, ma
+; CHECK-NEXT: vfsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <4 x float> poison, float %b, i32 0
%vb = shufflevector <4 x float> %elt.head, <4 x float> poison, <4 x i32> zeroinitializer
@@ -484,7 +476,7 @@ define <4 x float> @vfsub_vf_v4f32(<4 x float> %va, float %b, <4 x i1> %m, i32 z
define <4 x float> @vfsub_vf_v4f32_unmasked(<4 x float> %va, float %b, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vf_v4f32_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; CHECK-NEXT: vsetivli zero, 4, e32, m1, ta, ma
; CHECK-NEXT: vfsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <4 x float> poison, float %b, i32 0
@@ -496,8 +488,8 @@ define <4 x float> @vfsub_vf_v4f32_unmasked(<4 x float> %va, float %b, i32 zeroe
define <8 x float> @vfsub_vv_v8f32(<8 x float> %va, <8 x float> %b, <8 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vv_v8f32:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, m2, ta, ma
-; CHECK-NEXT: vfsub.vv v8, v8, v10, v0.t
+; CHECK-NEXT: vsetivli zero, 8, e32, m2, ta, ma
+; CHECK-NEXT: vfsub.vv v8, v8, v10
; CHECK-NEXT: ret
%v = call <8 x float> @llvm.vp.fsub.v8f32(<8 x float> %va, <8 x float> %b, <8 x i1> %m, i32 %evl)
ret <8 x float> %v
@@ -506,7 +498,7 @@ define <8 x float> @vfsub_vv_v8f32(<8 x float> %va, <8 x float> %b, <8 x i1> %m,
define <8 x float> @vfsub_vv_v8f32_unmasked(<8 x float> %va, <8 x float> %b, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vv_v8f32_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; CHECK-NEXT: vsetivli zero, 8, e32, m2, ta, ma
; CHECK-NEXT: vfsub.vv v8, v8, v10
; CHECK-NEXT: ret
%v = call <8 x float> @llvm.vp.fsub.v8f32(<8 x float> %va, <8 x float> %b, <8 x i1> splat (i1 true), i32 %evl)
@@ -516,8 +508,8 @@ define <8 x float> @vfsub_vv_v8f32_unmasked(<8 x float> %va, <8 x float> %b, i32
define <8 x float> @vfsub_vf_v8f32(<8 x float> %va, float %b, <8 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vf_v8f32:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, m2, ta, ma
-; CHECK-NEXT: vfsub.vf v8, v8, fa0, v0.t
+; CHECK-NEXT: vsetivli zero, 8, e32, m2, ta, ma
+; CHECK-NEXT: vfsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <8 x float> poison, float %b, i32 0
%vb = shufflevector <8 x float> %elt.head, <8 x float> poison, <8 x i32> zeroinitializer
@@ -528,7 +520,7 @@ define <8 x float> @vfsub_vf_v8f32(<8 x float> %va, float %b, <8 x i1> %m, i32 z
define <8 x float> @vfsub_vf_v8f32_unmasked(<8 x float> %va, float %b, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vf_v8f32_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; CHECK-NEXT: vsetivli zero, 8, e32, m2, ta, ma
; CHECK-NEXT: vfsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <8 x float> poison, float %b, i32 0
@@ -540,8 +532,8 @@ define <8 x float> @vfsub_vf_v8f32_unmasked(<8 x float> %va, float %b, i32 zeroe
define <16 x float> @vfsub_vv_v16f32(<16 x float> %va, <16 x float> %b, <16 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vv_v16f32:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, m4, ta, ma
-; CHECK-NEXT: vfsub.vv v8, v8, v12, v0.t
+; CHECK-NEXT: vsetivli zero, 16, e32, m4, ta, ma
+; CHECK-NEXT: vfsub.vv v8, v8, v12
; CHECK-NEXT: ret
%v = call <16 x float> @llvm.vp.fsub.v16f32(<16 x float> %va, <16 x float> %b, <16 x i1> %m, i32 %evl)
ret <16 x float> %v
@@ -550,7 +542,7 @@ define <16 x float> @vfsub_vv_v16f32(<16 x float> %va, <16 x float> %b, <16 x i1
define <16 x float> @vfsub_vv_v16f32_unmasked(<16 x float> %va, <16 x float> %b, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vv_v16f32_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT: vsetivli zero, 16, e32, m4, ta, ma
; CHECK-NEXT: vfsub.vv v8, v8, v12
; CHECK-NEXT: ret
%v = call <16 x float> @llvm.vp.fsub.v16f32(<16 x float> %va, <16 x float> %b, <16 x i1> splat (i1 true), i32 %evl)
@@ -560,8 +552,8 @@ define <16 x float> @vfsub_vv_v16f32_unmasked(<16 x float> %va, <16 x float> %b,
define <16 x float> @vfsub_vf_v16f32(<16 x float> %va, float %b, <16 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vf_v16f32:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, m4, ta, ma
-; CHECK-NEXT: vfsub.vf v8, v8, fa0, v0.t
+; CHECK-NEXT: vsetivli zero, 16, e32, m4, ta, ma
+; CHECK-NEXT: vfsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <16 x float> poison, float %b, i32 0
%vb = shufflevector <16 x float> %elt.head, <16 x float> poison, <16 x i32> zeroinitializer
@@ -572,7 +564,7 @@ define <16 x float> @vfsub_vf_v16f32(<16 x float> %va, float %b, <16 x i1> %m, i
define <16 x float> @vfsub_vf_v16f32_unmasked(<16 x float> %va, float %b, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vf_v16f32_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT: vsetivli zero, 16, e32, m4, ta, ma
; CHECK-NEXT: vfsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <16 x float> poison, float %b, i32 0
@@ -584,8 +576,8 @@ define <16 x float> @vfsub_vf_v16f32_unmasked(<16 x float> %va, float %b, i32 ze
define <2 x double> @vfsub_vv_v2f64(<2 x double> %va, <2 x double> %b, <2 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vv_v2f64:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e64, m1, ta, ma
-; CHECK-NEXT: vfsub.vv v8, v8, v9, v0.t
+; CHECK-NEXT: vsetivli zero, 2, e64, m1, ta, ma
+; CHECK-NEXT: vfsub.vv v8, v8, v9
; CHECK-NEXT: ret
%v = call <2 x double> @llvm.vp.fsub.v2f64(<2 x double> %va, <2 x double> %b, <2 x i1> %m, i32 %evl)
ret <2 x double> %v
@@ -594,7 +586,7 @@ define <2 x double> @vfsub_vv_v2f64(<2 x double> %va, <2 x double> %b, <2 x i1>
define <2 x double> @vfsub_vv_v2f64_unmasked(<2 x double> %va, <2 x double> %b, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vv_v2f64_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e64, m1, ta, ma
+; CHECK-NEXT: vsetivli zero, 2, e64, m1, ta, ma
; CHECK-NEXT: vfsub.vv v8, v8, v9
; CHECK-NEXT: ret
%v = call <2 x double> @llvm.vp.fsub.v2f64(<2 x double> %va, <2 x double> %b, <2 x i1> splat (i1 true), i32 %evl)
@@ -604,8 +596,8 @@ define <2 x double> @vfsub_vv_v2f64_unmasked(<2 x double> %va, <2 x double> %b,
define <2 x double> @vfsub_vf_v2f64(<2 x double> %va, double %b, <2 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vf_v2f64:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e64, m1, ta, ma
-; CHECK-NEXT: vfsub.vf v8, v8, fa0, v0.t
+; CHECK-NEXT: vsetivli zero, 2, e64, m1, ta, ma
+; CHECK-NEXT: vfsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <2 x double> poison, double %b, i32 0
%vb = shufflevector <2 x double> %elt.head, <2 x double> poison, <2 x i32> zeroinitializer
@@ -616,7 +608,7 @@ define <2 x double> @vfsub_vf_v2f64(<2 x double> %va, double %b, <2 x i1> %m, i3
define <2 x double> @vfsub_vf_v2f64_unmasked(<2 x double> %va, double %b, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vf_v2f64_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e64, m1, ta, ma
+; CHECK-NEXT: vsetivli zero, 2, e64, m1, ta, ma
; CHECK-NEXT: vfsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <2 x double> poison, double %b, i32 0
@@ -628,8 +620,8 @@ define <2 x double> @vfsub_vf_v2f64_unmasked(<2 x double> %va, double %b, i32 ze
define <4 x double> @vfsub_vv_v4f64(<4 x double> %va, <4 x double> %b, <4 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vv_v4f64:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e64, m2, ta, ma
-; CHECK-NEXT: vfsub.vv v8, v8, v10, v0.t
+; CHECK-NEXT: vsetivli zero, 4, e64, m2, ta, ma
+; CHECK-NEXT: vfsub.vv v8, v8, v10
; CHECK-NEXT: ret
%v = call <4 x double> @llvm.vp.fsub.v4f64(<4 x double> %va, <4 x double> %b, <4 x i1> %m, i32 %evl)
ret <4 x double> %v
@@ -638,7 +630,7 @@ define <4 x double> @vfsub_vv_v4f64(<4 x double> %va, <4 x double> %b, <4 x i1>
define <4 x double> @vfsub_vv_v4f64_unmasked(<4 x double> %va, <4 x double> %b, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vv_v4f64_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e64, m2, ta, ma
+; CHECK-NEXT: vsetivli zero, 4, e64, m2, ta, ma
; CHECK-NEXT: vfsub.vv v8, v8, v10
; CHECK-NEXT: ret
%v = call <4 x double> @llvm.vp.fsub.v4f64(<4 x double> %va, <4 x double> %b, <4 x i1> splat (i1 true), i32 %evl)
@@ -648,8 +640,8 @@ define <4 x double> @vfsub_vv_v4f64_unmasked(<4 x double> %va, <4 x double> %b,
define <4 x double> @vfsub_vf_v4f64(<4 x double> %va, double %b, <4 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vf_v4f64:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e64, m2, ta, ma
-; CHECK-NEXT: vfsub.vf v8, v8, fa0, v0.t
+; CHECK-NEXT: vsetivli zero, 4, e64, m2, ta, ma
+; CHECK-NEXT: vfsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <4 x double> poison, double %b, i32 0
%vb = shufflevector <4 x double> %elt.head, <4 x double> poison, <4 x i32> zeroinitializer
@@ -660,7 +652,7 @@ define <4 x double> @vfsub_vf_v4f64(<4 x double> %va, double %b, <4 x i1> %m, i3
define <4 x double> @vfsub_vf_v4f64_unmasked(<4 x double> %va, double %b, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vf_v4f64_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e64, m2, ta, ma
+; CHECK-NEXT: vsetivli zero, 4, e64, m2, ta, ma
; CHECK-NEXT: vfsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <4 x double> poison, double %b, i32 0
@@ -672,8 +664,8 @@ define <4 x double> @vfsub_vf_v4f64_unmasked(<4 x double> %va, double %b, i32 ze
define <8 x double> @vfsub_vv_v8f64(<8 x double> %va, <8 x double> %b, <8 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vv_v8f64:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e64, m4, ta, ma
-; CHECK-NEXT: vfsub.vv v8, v8, v12, v0.t
+; CHECK-NEXT: vsetivli zero, 8, e64, m4, ta, ma
+; CHECK-NEXT: vfsub.vv v8, v8, v12
; CHECK-NEXT: ret
%v = call <8 x double> @llvm.vp.fsub.v8f64(<8 x double> %va, <8 x double> %b, <8 x i1> %m, i32 %evl)
ret <8 x double> %v
@@ -682,7 +674,7 @@ define <8 x double> @vfsub_vv_v8f64(<8 x double> %va, <8 x double> %b, <8 x i1>
define <8 x double> @vfsub_vv_v8f64_unmasked(<8 x double> %va, <8 x double> %b, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vv_v8f64_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e64, m4, ta, ma
+; CHECK-NEXT: vsetivli zero, 8, e64, m4, ta, ma
; CHECK-NEXT: vfsub.vv v8, v8, v12
; CHECK-NEXT: ret
%v = call <8 x double> @llvm.vp.fsub.v8f64(<8 x double> %va, <8 x double> %b, <8 x i1> splat (i1 true), i32 %evl)
@@ -692,8 +684,8 @@ define <8 x double> @vfsub_vv_v8f64_unmasked(<8 x double> %va, <8 x double> %b,
define <8 x double> @vfsub_vf_v8f64(<8 x double> %va, double %b, <8 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vf_v8f64:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e64, m4, ta, ma
-; CHECK-NEXT: vfsub.vf v8, v8, fa0, v0.t
+; CHECK-NEXT: vsetivli zero, 8, e64, m4, ta, ma
+; CHECK-NEXT: vfsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <8 x double> poison, double %b, i32 0
%vb = shufflevector <8 x double> %elt.head, <8 x double> poison, <8 x i32> zeroinitializer
@@ -704,7 +696,7 @@ define <8 x double> @vfsub_vf_v8f64(<8 x double> %va, double %b, <8 x i1> %m, i3
define <8 x double> @vfsub_vf_v8f64_unmasked(<8 x double> %va, double %b, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vf_v8f64_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e64, m4, ta, ma
+; CHECK-NEXT: vsetivli zero, 8, e64, m4, ta, ma
; CHECK-NEXT: vfsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <8 x double> poison, double %b, i32 0
@@ -716,8 +708,8 @@ define <8 x double> @vfsub_vf_v8f64_unmasked(<8 x double> %va, double %b, i32 ze
define <16 x double> @vfsub_vv_v16f64(<16 x double> %va, <16 x double> %b, <16 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vv_v16f64:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e64, m8, ta, ma
-; CHECK-NEXT: vfsub.vv v8, v8, v16, v0.t
+; CHECK-NEXT: vsetivli zero, 16, e64, m8, ta, ma
+; CHECK-NEXT: vfsub.vv v8, v8, v16
; CHECK-NEXT: ret
%v = call <16 x double> @llvm.vp.fsub.v16f64(<16 x double> %va, <16 x double> %b, <16 x i1> %m, i32 %evl)
ret <16 x double> %v
@@ -726,7 +718,7 @@ define <16 x double> @vfsub_vv_v16f64(<16 x double> %va, <16 x double> %b, <16 x
define <16 x double> @vfsub_vv_v16f64_unmasked(<16 x double> %va, <16 x double> %b, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vv_v16f64_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e64, m8, ta, ma
+; CHECK-NEXT: vsetivli zero, 16, e64, m8, ta, ma
; CHECK-NEXT: vfsub.vv v8, v8, v16
; CHECK-NEXT: ret
%v = call <16 x double> @llvm.vp.fsub.v16f64(<16 x double> %va, <16 x double> %b, <16 x i1> splat (i1 true), i32 %evl)
@@ -736,8 +728,8 @@ define <16 x double> @vfsub_vv_v16f64_unmasked(<16 x double> %va, <16 x double>
define <16 x double> @vfsub_vf_v16f64(<16 x double> %va, double %b, <16 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vf_v16f64:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e64, m8, ta, ma
-; CHECK-NEXT: vfsub.vf v8, v8, fa0, v0.t
+; CHECK-NEXT: vsetivli zero, 16, e64, m8, ta, ma
+; CHECK-NEXT: vfsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <16 x double> poison, double %b, i32 0
%vb = shufflevector <16 x double> %elt.head, <16 x double> poison, <16 x i32> zeroinitializer
@@ -748,7 +740,7 @@ define <16 x double> @vfsub_vf_v16f64(<16 x double> %va, double %b, <16 x i1> %m
define <16 x double> @vfsub_vf_v16f64_unmasked(<16 x double> %va, double %b, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vf_v16f64_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e64, m8, ta, ma
+; CHECK-NEXT: vsetivli zero, 16, e64, m8, ta, ma
; CHECK-NEXT: vfsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <16 x double> poison, double %b, i32 0
diff --git a/llvm/test/CodeGen/RISCV/rvv/fold-vp-fsub-and-vp-fmul.ll b/llvm/test/CodeGen/RISCV/rvv/fold-vp-fsub-and-vp-fmul.ll
index de6ada6810bd8..35cc824a6c837 100644
--- a/llvm/test/CodeGen/RISCV/rvv/fold-vp-fsub-and-vp-fmul.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/fold-vp-fsub-and-vp-fmul.ll
@@ -6,8 +6,9 @@ define <vscale x 1 x double> @test1(<vscale x 1 x double> %x, <vscale x 1 x doub
; CHECK-LABEL: test1:
; CHECK: # %bb.0:
; CHECK-NEXT: vsetvli zero, a0, e64, m1, ta, ma
-; CHECK-NEXT: vfmsub.vv v9, v8, v10, v0.t
-; CHECK-NEXT: vmv.v.v v8, v9
+; CHECK-NEXT: vfmul.vv v8, v8, v9, v0.t
+; CHECK-NEXT: vsetvli a0, zero, e64, m1, ta, ma
+; CHECK-NEXT: vfsub.vv v8, v8, v10
; CHECK-NEXT: ret
%1 = call fast <vscale x 1 x double> @llvm.vp.fmul.nxv1f64(<vscale x 1 x double> %x, <vscale x 1 x double> %y, <vscale x 1 x i1> %m, i32 %vl)
%2 = call fast <vscale x 1 x double> @llvm.vp.fsub.nxv1f64(<vscale x 1 x double> %1, <vscale x 1 x double> %z, <vscale x 1 x i1> %m, i32 %vl)
@@ -19,8 +20,9 @@ define <vscale x 1 x double> @test2(<vscale x 1 x double> %x, <vscale x 1 x doub
; CHECK-LABEL: test2:
; CHECK: # %bb.0:
; CHECK-NEXT: vsetvli zero, a0, e64, m1, ta, ma
-; CHECK-NEXT: vfnmsub.vv v9, v8, v10, v0.t
-; CHECK-NEXT: vmv.v.v v8, v9
+; CHECK-NEXT: vfmul.vv v8, v8, v9, v0.t
+; CHECK-NEXT: vsetvli a0, zero, e64, m1, ta, ma
+; CHECK-NEXT: vfsub.vv v8, v10, v8
; CHECK-NEXT: ret
%1 = call fast <vscale x 1 x double> @llvm.vp.fmul.nxv1f64(<vscale x 1 x double> %x, <vscale x 1 x double> %y, <vscale x 1 x i1> %m, i32 %vl)
%2 = call fast <vscale x 1 x double> @llvm.vp.fsub.nxv1f64(<vscale x 1 x double> %z, <vscale x 1 x double> %1, <vscale x 1 x i1> %m, i32 %vl)
@@ -32,8 +34,9 @@ define <vscale x 1 x double> @test3(<vscale x 1 x double> %x, <vscale x 1 x doub
; CHECK-LABEL: test3:
; CHECK: # %bb.0:
; CHECK-NEXT: vsetvli zero, a0, e64, m1, ta, ma
-; CHECK-NEXT: vfmsub.vv v9, v8, v10, v0.t
-; CHECK-NEXT: vmv.v.v v8, v9
+; CHECK-NEXT: vfmul.vv v8, v8, v9, v0.t
+; CHECK-NEXT: vsetvli a0, zero, e64, m1, ta, ma
+; CHECK-NEXT: vfsub.vv v8, v8, v10
; CHECK-NEXT: ret
%1 = call fast <vscale x 1 x double> @llvm.vp.fmul.nxv1f64(<vscale x 1 x double> %x, <vscale x 1 x double> %y, <vscale x 1 x i1> %m, i32 %vl)
%2 = call fast <vscale x 1 x double> @llvm.vp.fneg.nxv1f64(<vscale x 1 x double> %1, <vscale x 1 x i1> %m, i32 %vl)
diff --git a/llvm/test/CodeGen/RISCV/rvv/sink-splat-operands.ll b/llvm/test/CodeGen/RISCV/rvv/sink-splat-operands.ll
index 09cfd05202728..9e2903b98e4b9 100644
--- a/llvm/test/CodeGen/RISCV/rvv/sink-splat-operands.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/sink-splat-operands.ll
@@ -3747,18 +3747,16 @@ for.cond.cleanup: ; preds = %vector.body
define void @sink_splat_vp_fsub(ptr nocapture %a, float %x, <4 x i1> %m, i32 zeroext %vl) {
; CHECK-LABEL: sink_splat_vp_fsub:
; CHECK: # %bb.0: # %entry
-; CHECK-NEXT: lui a2, 1
-; CHECK-NEXT: add a2, a0, a2
+; CHECK-NEXT: lui a1, 1
+; CHECK-NEXT: add a1, a0, a1
; CHECK-NEXT: vsetivli zero, 4, e32, m1, ta, ma
; CHECK-NEXT: .LBB72_1: # %vector.body
; CHECK-NEXT: # =>This Inner Loop Header: Depth=1
; CHECK-NEXT: vle32.v v8, (a0)
-; CHECK-NEXT: vsetvli zero, a1, e32, m1, ta, ma
-; CHECK-NEXT: vfsub.vf v8, v8, fa0, v0.t
-; CHECK-NEXT: vsetivli zero, 4, e32, m1, ta, ma
+; CHECK-NEXT: vfsub.vf v8, v8, fa0
; CHECK-NEXT: vse32.v v8, (a0)
; CHECK-NEXT: addi a0, a0, 16
-; CHECK-NEXT: bne a0, a2, .LBB72_1
+; CHECK-NEXT: bne a0, a1, .LBB72_1
; CHECK-NEXT: # %bb.2: # %for.cond.cleanup
; CHECK-NEXT: ret
entry:
@@ -3783,18 +3781,16 @@ for.cond.cleanup: ; preds = %vector.body
define void @sink_splat_vp_frsub(ptr nocapture %a, float %x, <4 x i1> %m, i32 zeroext %vl) {
; CHECK-LABEL: sink_splat_vp_frsub:
; CHECK: # %bb.0: # %entry
-; CHECK-NEXT: lui a2, 1
-; CHECK-NEXT: add a2, a0, a2
+; CHECK-NEXT: lui a1, 1
+; CHECK-NEXT: add a1, a0, a1
; CHECK-NEXT: vsetivli zero, 4, e32, m1, ta, ma
; CHECK-NEXT: .LBB73_1: # %vector.body
; CHECK-NEXT: # =>This Inner Loop Header: Depth=1
; CHECK-NEXT: vle32.v v8, (a0)
-; CHECK-NEXT: vsetvli zero, a1, e32, m1, ta, ma
-; CHECK-NEXT: vfrsub.vf v8, v8, fa0, v0.t
-; CHECK-NEXT: vsetivli zero, 4, e32, m1, ta, ma
+; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: vse32.v v8, (a0)
; CHECK-NEXT: addi a0, a0, 16
-; CHECK-NEXT: bne a0, a2, .LBB73_1
+; CHECK-NEXT: bne a0, a1, .LBB73_1
; CHECK-NEXT: # %bb.2: # %for.cond.cleanup
; CHECK-NEXT: ret
entry:
diff --git a/llvm/test/CodeGen/RISCV/rvv/undef-vp-ops.ll b/llvm/test/CodeGen/RISCV/rvv/undef-vp-ops.ll
index 15581582de77a..8b1ccd7aff45b 100644
--- a/llvm/test/CodeGen/RISCV/rvv/undef-vp-ops.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/undef-vp-ops.ll
@@ -265,6 +265,8 @@ define <4 x float> @vfadd_v4f32_false_mask(<4 x float> %va, <4 x float> %vb, i32
define <4 x float> @vfsub_v4f32_zero_evl(<4 x float> %va, <4 x float> %vb, <4 x i1> %m) {
; CHECK-LABEL: vfsub_v4f32_zero_evl:
; CHECK: # %bb.0:
+; CHECK-NEXT: vsetivli zero, 4, e32, m1, ta, ma
+; CHECK-NEXT: vfsub.vv v8, v8, v9
; CHECK-NEXT: ret
%s = call <4 x float> @llvm.vp.fsub.v4f32(<4 x float> %va, <4 x float> %vb, <4 x i1> %m, i32 0)
ret <4 x float> %s
@@ -273,6 +275,8 @@ define <4 x float> @vfsub_v4f32_zero_evl(<4 x float> %va, <4 x float> %vb, <4 x
define <4 x float> @vfsub_v4f32_false_mask(<4 x float> %va, <4 x float> %vb, i32 %evl) {
; CHECK-LABEL: vfsub_v4f32_false_mask:
; CHECK: # %bb.0:
+; CHECK-NEXT: vsetivli zero, 4, e32, m1, ta, ma
+; CHECK-NEXT: vfsub.vv v8, v8, v9
; CHECK-NEXT: ret
%s = call <4 x float> @llvm.vp.fsub.v4f32(<4 x float> %va, <4 x float> %vb, <4 x i1> zeroinitializer, i32 %evl)
ret <4 x float> %s
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfrsub-vp.ll b/llvm/test/CodeGen/RISCV/rvv/vfrsub-vp.ll
index 2fc77437019e3..34b72b1243bb0 100644
--- a/llvm/test/CodeGen/RISCV/rvv/vfrsub-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/vfrsub-vp.ll
@@ -7,8 +7,8 @@
define <vscale x 1 x half> @vfrsub_vf_nxv1f16(<vscale x 1 x half> %va, half %b, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_nxv1f16:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
-; CHECK-NEXT: vfrsub.vf v8, v8, fa0, v0.t
+; CHECK-NEXT: vsetvli a0, zero, e16, mf4, ta, ma
+; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 1 x half> poison, half %b, i32 0
%vb = shufflevector <vscale x 1 x half> %elt.head, <vscale x 1 x half> poison, <vscale x 1 x i32> zeroinitializer
@@ -19,7 +19,7 @@ define <vscale x 1 x half> @vfrsub_vf_nxv1f16(<vscale x 1 x half> %va, half %b,
define <vscale x 1 x half> @vfrsub_vf_nxv1f16_unmasked(<vscale x 1 x half> %va, half %b, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_nxv1f16_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; CHECK-NEXT: vsetvli a0, zero, e16, mf4, ta, ma
; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 1 x half> poison, half %b, i32 0
@@ -31,8 +31,8 @@ define <vscale x 1 x half> @vfrsub_vf_nxv1f16_unmasked(<vscale x 1 x half> %va,
define <vscale x 2 x half> @vfrsub_vf_nxv2f16(<vscale x 2 x half> %va, half %b, <vscale x 2 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_nxv2f16:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
-; CHECK-NEXT: vfrsub.vf v8, v8, fa0, v0.t
+; CHECK-NEXT: vsetvli a0, zero, e16, mf2, ta, ma
+; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 2 x half> poison, half %b, i32 0
%vb = shufflevector <vscale x 2 x half> %elt.head, <vscale x 2 x half> poison, <vscale x 2 x i32> zeroinitializer
@@ -43,7 +43,7 @@ define <vscale x 2 x half> @vfrsub_vf_nxv2f16(<vscale x 2 x half> %va, half %b,
define <vscale x 2 x half> @vfrsub_vf_nxv2f16_unmasked(<vscale x 2 x half> %va, half %b, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_nxv2f16_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; CHECK-NEXT: vsetvli a0, zero, e16, mf2, ta, ma
; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 2 x half> poison, half %b, i32 0
@@ -55,8 +55,8 @@ define <vscale x 2 x half> @vfrsub_vf_nxv2f16_unmasked(<vscale x 2 x half> %va,
define <vscale x 4 x half> @vfrsub_vf_nxv4f16(<vscale x 4 x half> %va, half %b, <vscale x 4 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_nxv4f16:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e16, m1, ta, ma
-; CHECK-NEXT: vfrsub.vf v8, v8, fa0, v0.t
+; CHECK-NEXT: vsetvli a0, zero, e16, m1, ta, ma
+; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 4 x half> poison, half %b, i32 0
%vb = shufflevector <vscale x 4 x half> %elt.head, <vscale x 4 x half> poison, <vscale x 4 x i32> zeroinitializer
@@ -67,7 +67,7 @@ define <vscale x 4 x half> @vfrsub_vf_nxv4f16(<vscale x 4 x half> %va, half %b,
define <vscale x 4 x half> @vfrsub_vf_nxv4f16_unmasked(<vscale x 4 x half> %va, half %b, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_nxv4f16_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; CHECK-NEXT: vsetvli a0, zero, e16, m1, ta, ma
; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 4 x half> poison, half %b, i32 0
@@ -79,8 +79,8 @@ define <vscale x 4 x half> @vfrsub_vf_nxv4f16_unmasked(<vscale x 4 x half> %va,
define <vscale x 8 x half> @vfrsub_vf_nxv8f16(<vscale x 8 x half> %va, half %b, <vscale x 8 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_nxv8f16:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e16, m2, ta, ma
-; CHECK-NEXT: vfrsub.vf v8, v8, fa0, v0.t
+; CHECK-NEXT: vsetvli a0, zero, e16, m2, ta, ma
+; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 8 x half> poison, half %b, i32 0
%vb = shufflevector <vscale x 8 x half> %elt.head, <vscale x 8 x half> poison, <vscale x 8 x i32> zeroinitializer
@@ -91,7 +91,7 @@ define <vscale x 8 x half> @vfrsub_vf_nxv8f16(<vscale x 8 x half> %va, half %b,
define <vscale x 8 x half> @vfrsub_vf_nxv8f16_unmasked(<vscale x 8 x half> %va, half %b, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_nxv8f16_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; CHECK-NEXT: vsetvli a0, zero, e16, m2, ta, ma
; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 8 x half> poison, half %b, i32 0
@@ -103,8 +103,8 @@ define <vscale x 8 x half> @vfrsub_vf_nxv8f16_unmasked(<vscale x 8 x half> %va,
define <vscale x 16 x half> @vfrsub_vf_nxv16f16(<vscale x 16 x half> %va, half %b, <vscale x 16 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_nxv16f16:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e16, m4, ta, ma
-; CHECK-NEXT: vfrsub.vf v8, v8, fa0, v0.t
+; CHECK-NEXT: vsetvli a0, zero, e16, m4, ta, ma
+; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 16 x half> poison, half %b, i32 0
%vb = shufflevector <vscale x 16 x half> %elt.head, <vscale x 16 x half> poison, <vscale x 16 x i32> zeroinitializer
@@ -115,7 +115,7 @@ define <vscale x 16 x half> @vfrsub_vf_nxv16f16(<vscale x 16 x half> %va, half %
define <vscale x 16 x half> @vfrsub_vf_nxv16f16_unmasked(<vscale x 16 x half> %va, half %b, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_nxv16f16_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; CHECK-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 16 x half> poison, half %b, i32 0
@@ -127,8 +127,8 @@ define <vscale x 16 x half> @vfrsub_vf_nxv16f16_unmasked(<vscale x 16 x half> %v
define <vscale x 32 x half> @vfrsub_vf_nxv32f16(<vscale x 32 x half> %va, half %b, <vscale x 32 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_nxv32f16:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e16, m8, ta, ma
-; CHECK-NEXT: vfrsub.vf v8, v8, fa0, v0.t
+; CHECK-NEXT: vsetvli a0, zero, e16, m8, ta, ma
+; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-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
@@ -139,7 +139,7 @@ define <vscale x 32 x half> @vfrsub_vf_nxv32f16(<vscale x 32 x half> %va, half %
define <vscale x 32 x half> @vfrsub_vf_nxv32f16_unmasked(<vscale x 32 x half> %va, half %b, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_nxv32f16_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e16, m8, ta, ma
+; CHECK-NEXT: vsetvli a0, zero, e16, m8, ta, ma
; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 32 x half> poison, half %b, i32 0
@@ -151,8 +151,8 @@ define <vscale x 32 x half> @vfrsub_vf_nxv32f16_unmasked(<vscale x 32 x half> %v
define <vscale x 1 x float> @vfrsub_vf_nxv1f32(<vscale x 1 x float> %va, float %b, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_nxv1f32:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
-; CHECK-NEXT: vfrsub.vf v8, v8, fa0, v0.t
+; CHECK-NEXT: vsetvli a0, zero, e32, mf2, ta, ma
+; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 1 x float> poison, float %b, i32 0
%vb = shufflevector <vscale x 1 x float> %elt.head, <vscale x 1 x float> poison, <vscale x 1 x i32> zeroinitializer
@@ -163,7 +163,7 @@ define <vscale x 1 x float> @vfrsub_vf_nxv1f32(<vscale x 1 x float> %va, float %
define <vscale x 1 x float> @vfrsub_vf_nxv1f32_unmasked(<vscale x 1 x float> %va, float %b, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_nxv1f32_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT: vsetvli a0, zero, e32, mf2, ta, ma
; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 1 x float> poison, float %b, i32 0
@@ -175,8 +175,8 @@ define <vscale x 1 x float> @vfrsub_vf_nxv1f32_unmasked(<vscale x 1 x float> %va
define <vscale x 2 x float> @vfrsub_vf_nxv2f32(<vscale x 2 x float> %va, float %b, <vscale x 2 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_nxv2f32:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, m1, ta, ma
-; CHECK-NEXT: vfrsub.vf v8, v8, fa0, v0.t
+; CHECK-NEXT: vsetvli a0, zero, e32, m1, ta, ma
+; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 2 x float> poison, float %b, i32 0
%vb = shufflevector <vscale x 2 x float> %elt.head, <vscale x 2 x float> poison, <vscale x 2 x i32> zeroinitializer
@@ -187,7 +187,7 @@ define <vscale x 2 x float> @vfrsub_vf_nxv2f32(<vscale x 2 x float> %va, float %
define <vscale x 2 x float> @vfrsub_vf_nxv2f32_unmasked(<vscale x 2 x float> %va, float %b, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_nxv2f32_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; CHECK-NEXT: vsetvli a0, zero, e32, m1, ta, ma
; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 2 x float> poison, float %b, i32 0
@@ -199,8 +199,8 @@ define <vscale x 2 x float> @vfrsub_vf_nxv2f32_unmasked(<vscale x 2 x float> %va
define <vscale x 4 x float> @vfrsub_vf_nxv4f32(<vscale x 4 x float> %va, float %b, <vscale x 4 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_nxv4f32:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, m2, ta, ma
-; CHECK-NEXT: vfrsub.vf v8, v8, fa0, v0.t
+; CHECK-NEXT: vsetvli a0, zero, e32, m2, ta, ma
+; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 4 x float> poison, float %b, i32 0
%vb = shufflevector <vscale x 4 x float> %elt.head, <vscale x 4 x float> poison, <vscale x 4 x i32> zeroinitializer
@@ -211,7 +211,7 @@ define <vscale x 4 x float> @vfrsub_vf_nxv4f32(<vscale x 4 x float> %va, float %
define <vscale x 4 x float> @vfrsub_vf_nxv4f32_unmasked(<vscale x 4 x float> %va, float %b, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_nxv4f32_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; CHECK-NEXT: vsetvli a0, zero, e32, m2, ta, ma
; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 4 x float> poison, float %b, i32 0
@@ -223,8 +223,8 @@ define <vscale x 4 x float> @vfrsub_vf_nxv4f32_unmasked(<vscale x 4 x float> %va
define <vscale x 8 x float> @vfrsub_vf_nxv8f32(<vscale x 8 x float> %va, float %b, <vscale x 8 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_nxv8f32:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, m4, ta, ma
-; CHECK-NEXT: vfrsub.vf v8, v8, fa0, v0.t
+; CHECK-NEXT: vsetvli a0, zero, e32, m4, ta, ma
+; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 8 x float> poison, float %b, i32 0
%vb = shufflevector <vscale x 8 x float> %elt.head, <vscale x 8 x float> poison, <vscale x 8 x i32> zeroinitializer
@@ -235,7 +235,7 @@ define <vscale x 8 x float> @vfrsub_vf_nxv8f32(<vscale x 8 x float> %va, float %
define <vscale x 8 x float> @vfrsub_vf_nxv8f32_unmasked(<vscale x 8 x float> %va, float %b, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_nxv8f32_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT: vsetvli a0, zero, e32, m4, ta, ma
; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 8 x float> poison, float %b, i32 0
@@ -247,8 +247,8 @@ define <vscale x 8 x float> @vfrsub_vf_nxv8f32_unmasked(<vscale x 8 x float> %va
define <vscale x 16 x float> @vfrsub_vf_nxv16f32(<vscale x 16 x float> %va, float %b, <vscale x 16 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_nxv16f32:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, m8, ta, ma
-; CHECK-NEXT: vfrsub.vf v8, v8, fa0, v0.t
+; CHECK-NEXT: vsetvli a0, zero, e32, m8, ta, ma
+; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 16 x float> poison, float %b, i32 0
%vb = shufflevector <vscale x 16 x float> %elt.head, <vscale x 16 x float> poison, <vscale x 16 x i32> zeroinitializer
@@ -259,7 +259,7 @@ define <vscale x 16 x float> @vfrsub_vf_nxv16f32(<vscale x 16 x float> %va, floa
define <vscale x 16 x float> @vfrsub_vf_nxv16f32_unmasked(<vscale x 16 x float> %va, float %b, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_nxv16f32_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; CHECK-NEXT: vsetvli a0, zero, e32, m8, ta, ma
; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 16 x float> poison, float %b, i32 0
@@ -271,8 +271,8 @@ define <vscale x 16 x float> @vfrsub_vf_nxv16f32_unmasked(<vscale x 16 x float>
define <vscale x 1 x double> @vfrsub_vf_nxv1f64(<vscale x 1 x double> %va, double %b, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_nxv1f64:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e64, m1, ta, ma
-; CHECK-NEXT: vfrsub.vf v8, v8, fa0, v0.t
+; CHECK-NEXT: vsetvli a0, zero, e64, m1, ta, ma
+; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 1 x double> poison, double %b, i32 0
%vb = shufflevector <vscale x 1 x double> %elt.head, <vscale x 1 x double> poison, <vscale x 1 x i32> zeroinitializer
@@ -283,7 +283,7 @@ define <vscale x 1 x double> @vfrsub_vf_nxv1f64(<vscale x 1 x double> %va, doubl
define <vscale x 1 x double> @vfrsub_vf_nxv1f64_unmasked(<vscale x 1 x double> %va, double %b, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_nxv1f64_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e64, m1, ta, ma
+; CHECK-NEXT: vsetvli a0, zero, e64, m1, ta, ma
; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 1 x double> poison, double %b, i32 0
@@ -295,8 +295,8 @@ define <vscale x 1 x double> @vfrsub_vf_nxv1f64_unmasked(<vscale x 1 x double> %
define <vscale x 2 x double> @vfrsub_vf_nxv2f64(<vscale x 2 x double> %va, double %b, <vscale x 2 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_nxv2f64:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e64, m2, ta, ma
-; CHECK-NEXT: vfrsub.vf v8, v8, fa0, v0.t
+; CHECK-NEXT: vsetvli a0, zero, e64, m2, ta, ma
+; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 2 x double> poison, double %b, i32 0
%vb = shufflevector <vscale x 2 x double> %elt.head, <vscale x 2 x double> poison, <vscale x 2 x i32> zeroinitializer
@@ -307,7 +307,7 @@ define <vscale x 2 x double> @vfrsub_vf_nxv2f64(<vscale x 2 x double> %va, doubl
define <vscale x 2 x double> @vfrsub_vf_nxv2f64_unmasked(<vscale x 2 x double> %va, double %b, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_nxv2f64_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e64, m2, ta, ma
+; CHECK-NEXT: vsetvli a0, zero, e64, m2, ta, ma
; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 2 x double> poison, double %b, i32 0
@@ -319,8 +319,8 @@ define <vscale x 2 x double> @vfrsub_vf_nxv2f64_unmasked(<vscale x 2 x double> %
define <vscale x 4 x double> @vfrsub_vf_nxv4f64(<vscale x 4 x double> %va, double %b, <vscale x 4 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_nxv4f64:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e64, m4, ta, ma
-; CHECK-NEXT: vfrsub.vf v8, v8, fa0, v0.t
+; CHECK-NEXT: vsetvli a0, zero, e64, m4, ta, ma
+; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 4 x double> poison, double %b, i32 0
%vb = shufflevector <vscale x 4 x double> %elt.head, <vscale x 4 x double> poison, <vscale x 4 x i32> zeroinitializer
@@ -331,7 +331,7 @@ define <vscale x 4 x double> @vfrsub_vf_nxv4f64(<vscale x 4 x double> %va, doubl
define <vscale x 4 x double> @vfrsub_vf_nxv4f64_unmasked(<vscale x 4 x double> %va, double %b, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_nxv4f64_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e64, m4, ta, ma
+; CHECK-NEXT: vsetvli a0, zero, e64, m4, ta, ma
; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 4 x double> poison, double %b, i32 0
@@ -343,8 +343,8 @@ define <vscale x 4 x double> @vfrsub_vf_nxv4f64_unmasked(<vscale x 4 x double> %
define <vscale x 8 x double> @vfrsub_vf_nxv8f64(<vscale x 8 x double> %va, double %b, <vscale x 8 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_nxv8f64:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e64, m8, ta, ma
-; CHECK-NEXT: vfrsub.vf v8, v8, fa0, v0.t
+; CHECK-NEXT: vsetvli a0, zero, e64, m8, ta, ma
+; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 8 x double> poison, double %b, i32 0
%vb = shufflevector <vscale x 8 x double> %elt.head, <vscale x 8 x double> poison, <vscale x 8 x i32> zeroinitializer
@@ -355,7 +355,7 @@ define <vscale x 8 x double> @vfrsub_vf_nxv8f64(<vscale x 8 x double> %va, doubl
define <vscale x 8 x double> @vfrsub_vf_nxv8f64_unmasked(<vscale x 8 x double> %va, double %b, i32 zeroext %evl) {
; CHECK-LABEL: vfrsub_vf_nxv8f64_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e64, m8, ta, ma
+; CHECK-NEXT: vsetvli a0, zero, e64, m8, ta, ma
; CHECK-NEXT: vfrsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 8 x double> poison, double %b, i32 0
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfsub-vp.ll b/llvm/test/CodeGen/RISCV/rvv/vfsub-vp.ll
index 6637aced3cdac..20330f934aac0 100644
--- a/llvm/test/CodeGen/RISCV/rvv/vfsub-vp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/vfsub-vp.ll
@@ -15,13 +15,13 @@
define <vscale x 1 x bfloat> @vfsub_vv_nxv1bf16(<vscale x 1 x bfloat> %va, <vscale x 1 x bfloat> %b, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vv_nxv1bf16:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
-; CHECK-NEXT: vfwcvtbf16.f.f.v v10, v9, v0.t
-; CHECK-NEXT: vfwcvtbf16.f.f.v v9, v8, v0.t
+; CHECK-NEXT: vsetvli a0, zero, e16, mf4, ta, ma
+; CHECK-NEXT: vfwcvtbf16.f.f.v v10, v9
+; CHECK-NEXT: vfwcvtbf16.f.f.v v9, v8
; CHECK-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
-; CHECK-NEXT: vfsub.vv v9, v9, v10, v0.t
+; CHECK-NEXT: vfsub.vv v9, v9, v10
; CHECK-NEXT: vsetvli zero, zero, e16, mf4, ta, ma
-; CHECK-NEXT: vfncvtbf16.f.f.w v8, v9, v0.t
+; CHECK-NEXT: vfncvtbf16.f.f.w v8, v9
; CHECK-NEXT: ret
%v = call <vscale x 1 x bfloat> @llvm.vp.fsub.nxv1bf16(<vscale x 1 x bfloat> %va, <vscale x 1 x bfloat> %b, <vscale x 1 x i1> %m, i32 %evl)
ret <vscale x 1 x bfloat> %v
@@ -30,7 +30,7 @@ define <vscale x 1 x bfloat> @vfsub_vv_nxv1bf16(<vscale x 1 x bfloat> %va, <vsca
define <vscale x 1 x bfloat> @vfsub_vv_nxv1bf16_unmasked(<vscale x 1 x bfloat> %va, <vscale x 1 x bfloat> %b, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vv_nxv1bf16_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; CHECK-NEXT: vsetvli a0, zero, e16, mf4, ta, ma
; CHECK-NEXT: vfwcvtbf16.f.f.v v10, v9
; CHECK-NEXT: vfwcvtbf16.f.f.v v9, v8
; CHECK-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
@@ -45,15 +45,13 @@ define <vscale x 1 x bfloat> @vfsub_vv_nxv1bf16_unmasked(<vscale x 1 x bfloat> %
define <vscale x 1 x bfloat> @vfsub_vf_nxv1bf16(<vscale x 1 x bfloat> %va, bfloat %b, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vf_nxv1bf16:
; CHECK: # %bb.0:
-; CHECK-NEXT: fmv.x.h a1, fa0
-; CHECK-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
-; CHECK-NEXT: vmv.v.x v9, a1
-; CHECK-NEXT: vfwcvtbf16.f.f.v v10, v8, v0.t
-; CHECK-NEXT: vfwcvtbf16.f.f.v v8, v9, v0.t
+; CHECK-NEXT: fcvt.s.bf16 fa5, fa0
+; CHECK-NEXT: vsetvli a0, zero, e16, mf4, ta, ma
+; CHECK-NEXT: vfwcvtbf16.f.f.v v9, v8
; CHECK-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
-; CHECK-NEXT: vfsub.vv v9, v10, v8, v0.t
+; CHECK-NEXT: vfsub.vf v9, v9, fa5
; CHECK-NEXT: vsetvli zero, zero, e16, mf4, ta, ma
-; CHECK-NEXT: vfncvtbf16.f.f.w v8, v9, v0.t
+; CHECK-NEXT: vfncvtbf16.f.f.w v8, v9
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 1 x bfloat> poison, bfloat %b, i32 0
%vb = shufflevector <vscale x 1 x bfloat> %elt.head, <vscale x 1 x bfloat> poison, <vscale x 1 x i32> zeroinitializer
@@ -64,13 +62,11 @@ define <vscale x 1 x bfloat> @vfsub_vf_nxv1bf16(<vscale x 1 x bfloat> %va, bfloa
define <vscale x 1 x bfloat> @vfsub_vf_nxv1bf16_unmasked(<vscale x 1 x bfloat> %va, bfloat %b, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vf_nxv1bf16_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: fmv.x.h a1, fa0
-; CHECK-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
-; CHECK-NEXT: vmv.v.x v9, a1
-; CHECK-NEXT: vfwcvtbf16.f.f.v v10, v8
-; CHECK-NEXT: vfwcvtbf16.f.f.v v8, v9
+; CHECK-NEXT: fcvt.s.bf16 fa5, fa0
+; CHECK-NEXT: vsetvli a0, zero, e16, mf4, ta, ma
+; CHECK-NEXT: vfwcvtbf16.f.f.v v9, v8
; CHECK-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
-; CHECK-NEXT: vfsub.vv v9, v10, v8
+; CHECK-NEXT: vfsub.vf v9, v9, fa5
; CHECK-NEXT: vsetvli zero, zero, e16, mf4, ta, ma
; CHECK-NEXT: vfncvtbf16.f.f.w v8, v9
; CHECK-NEXT: ret
@@ -83,13 +79,13 @@ define <vscale x 1 x bfloat> @vfsub_vf_nxv1bf16_unmasked(<vscale x 1 x bfloat> %
define <vscale x 2 x bfloat> @vfsub_vv_nxv2bf16(<vscale x 2 x bfloat> %va, <vscale x 2 x bfloat> %b, <vscale x 2 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vv_nxv2bf16:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
-; CHECK-NEXT: vfwcvtbf16.f.f.v v10, v9, v0.t
-; CHECK-NEXT: vfwcvtbf16.f.f.v v9, v8, v0.t
+; CHECK-NEXT: vsetvli a0, zero, e16, mf2, ta, ma
+; CHECK-NEXT: vfwcvtbf16.f.f.v v10, v9
+; CHECK-NEXT: vfwcvtbf16.f.f.v v9, v8
; CHECK-NEXT: vsetvli zero, zero, e32, m1, ta, ma
-; CHECK-NEXT: vfsub.vv v9, v9, v10, v0.t
+; CHECK-NEXT: vfsub.vv v9, v9, v10
; CHECK-NEXT: vsetvli zero, zero, e16, mf2, ta, ma
-; CHECK-NEXT: vfncvtbf16.f.f.w v8, v9, v0.t
+; CHECK-NEXT: vfncvtbf16.f.f.w v8, v9
; CHECK-NEXT: ret
%v = call <vscale x 2 x bfloat> @llvm.vp.fsub.nxv2bf16(<vscale x 2 x bfloat> %va, <vscale x 2 x bfloat> %b, <vscale x 2 x i1> %m, i32 %evl)
ret <vscale x 2 x bfloat> %v
@@ -98,7 +94,7 @@ define <vscale x 2 x bfloat> @vfsub_vv_nxv2bf16(<vscale x 2 x bfloat> %va, <vsca
define <vscale x 2 x bfloat> @vfsub_vv_nxv2bf16_unmasked(<vscale x 2 x bfloat> %va, <vscale x 2 x bfloat> %b, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vv_nxv2bf16_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; CHECK-NEXT: vsetvli a0, zero, e16, mf2, ta, ma
; CHECK-NEXT: vfwcvtbf16.f.f.v v10, v9
; CHECK-NEXT: vfwcvtbf16.f.f.v v9, v8
; CHECK-NEXT: vsetvli zero, zero, e32, m1, ta, ma
@@ -113,15 +109,13 @@ define <vscale x 2 x bfloat> @vfsub_vv_nxv2bf16_unmasked(<vscale x 2 x bfloat> %
define <vscale x 2 x bfloat> @vfsub_vf_nxv2bf16(<vscale x 2 x bfloat> %va, bfloat %b, <vscale x 2 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vf_nxv2bf16:
; CHECK: # %bb.0:
-; CHECK-NEXT: fmv.x.h a1, fa0
-; CHECK-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
-; CHECK-NEXT: vmv.v.x v9, a1
-; CHECK-NEXT: vfwcvtbf16.f.f.v v10, v8, v0.t
-; CHECK-NEXT: vfwcvtbf16.f.f.v v8, v9, v0.t
+; CHECK-NEXT: fcvt.s.bf16 fa5, fa0
+; CHECK-NEXT: vsetvli a0, zero, e16, mf2, ta, ma
+; CHECK-NEXT: vfwcvtbf16.f.f.v v9, v8
; CHECK-NEXT: vsetvli zero, zero, e32, m1, ta, ma
-; CHECK-NEXT: vfsub.vv v9, v10, v8, v0.t
+; CHECK-NEXT: vfsub.vf v9, v9, fa5
; CHECK-NEXT: vsetvli zero, zero, e16, mf2, ta, ma
-; CHECK-NEXT: vfncvtbf16.f.f.w v8, v9, v0.t
+; CHECK-NEXT: vfncvtbf16.f.f.w v8, v9
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 2 x bfloat> poison, bfloat %b, i32 0
%vb = shufflevector <vscale x 2 x bfloat> %elt.head, <vscale x 2 x bfloat> poison, <vscale x 2 x i32> zeroinitializer
@@ -132,13 +126,11 @@ define <vscale x 2 x bfloat> @vfsub_vf_nxv2bf16(<vscale x 2 x bfloat> %va, bfloa
define <vscale x 2 x bfloat> @vfsub_vf_nxv2bf16_unmasked(<vscale x 2 x bfloat> %va, bfloat %b, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vf_nxv2bf16_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: fmv.x.h a1, fa0
-; CHECK-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
-; CHECK-NEXT: vmv.v.x v9, a1
-; CHECK-NEXT: vfwcvtbf16.f.f.v v10, v8
-; CHECK-NEXT: vfwcvtbf16.f.f.v v8, v9
+; CHECK-NEXT: fcvt.s.bf16 fa5, fa0
+; CHECK-NEXT: vsetvli a0, zero, e16, mf2, ta, ma
+; CHECK-NEXT: vfwcvtbf16.f.f.v v9, v8
; CHECK-NEXT: vsetvli zero, zero, e32, m1, ta, ma
-; CHECK-NEXT: vfsub.vv v9, v10, v8
+; CHECK-NEXT: vfsub.vf v9, v9, fa5
; CHECK-NEXT: vsetvli zero, zero, e16, mf2, ta, ma
; CHECK-NEXT: vfncvtbf16.f.f.w v8, v9
; CHECK-NEXT: ret
@@ -151,13 +143,13 @@ define <vscale x 2 x bfloat> @vfsub_vf_nxv2bf16_unmasked(<vscale x 2 x bfloat> %
define <vscale x 4 x bfloat> @vfsub_vv_nxv4bf16(<vscale x 4 x bfloat> %va, <vscale x 4 x bfloat> %b, <vscale x 4 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vv_nxv4bf16:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e16, m1, ta, ma
-; CHECK-NEXT: vfwcvtbf16.f.f.v v10, v9, v0.t
-; CHECK-NEXT: vfwcvtbf16.f.f.v v12, v8, v0.t
+; CHECK-NEXT: vsetvli a0, zero, e16, m1, ta, ma
+; CHECK-NEXT: vfwcvtbf16.f.f.v v10, v9
+; CHECK-NEXT: vfwcvtbf16.f.f.v v12, v8
; CHECK-NEXT: vsetvli zero, zero, e32, m2, ta, ma
-; CHECK-NEXT: vfsub.vv v10, v12, v10, v0.t
+; CHECK-NEXT: vfsub.vv v10, v12, v10
; CHECK-NEXT: vsetvli zero, zero, e16, m1, ta, ma
-; CHECK-NEXT: vfncvtbf16.f.f.w v8, v10, v0.t
+; CHECK-NEXT: vfncvtbf16.f.f.w v8, v10
; CHECK-NEXT: ret
%v = call <vscale x 4 x bfloat> @llvm.vp.fsub.nxv4bf16(<vscale x 4 x bfloat> %va, <vscale x 4 x bfloat> %b, <vscale x 4 x i1> %m, i32 %evl)
ret <vscale x 4 x bfloat> %v
@@ -166,7 +158,7 @@ define <vscale x 4 x bfloat> @vfsub_vv_nxv4bf16(<vscale x 4 x bfloat> %va, <vsca
define <vscale x 4 x bfloat> @vfsub_vv_nxv4bf16_unmasked(<vscale x 4 x bfloat> %va, <vscale x 4 x bfloat> %b, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vv_nxv4bf16_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; CHECK-NEXT: vsetvli a0, zero, e16, m1, ta, ma
; CHECK-NEXT: vfwcvtbf16.f.f.v v10, v9
; CHECK-NEXT: vfwcvtbf16.f.f.v v12, v8
; CHECK-NEXT: vsetvli zero, zero, e32, m2, ta, ma
@@ -181,15 +173,13 @@ define <vscale x 4 x bfloat> @vfsub_vv_nxv4bf16_unmasked(<vscale x 4 x bfloat> %
define <vscale x 4 x bfloat> @vfsub_vf_nxv4bf16(<vscale x 4 x bfloat> %va, bfloat %b, <vscale x 4 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vf_nxv4bf16:
; CHECK: # %bb.0:
-; CHECK-NEXT: fmv.x.h a1, fa0
-; CHECK-NEXT: vsetvli zero, a0, e16, m1, ta, ma
-; CHECK-NEXT: vmv.v.x v12, a1
-; CHECK-NEXT: vfwcvtbf16.f.f.v v10, v8, v0.t
-; CHECK-NEXT: vfwcvtbf16.f.f.v v8, v12, v0.t
+; CHECK-NEXT: fcvt.s.bf16 fa5, fa0
+; CHECK-NEXT: vsetvli a0, zero, e16, m1, ta, ma
+; CHECK-NEXT: vfwcvtbf16.f.f.v v10, v8
; CHECK-NEXT: vsetvli zero, zero, e32, m2, ta, ma
-; CHECK-NEXT: vfsub.vv v10, v10, v8, v0.t
+; CHECK-NEXT: vfsub.vf v10, v10, fa5
; CHECK-NEXT: vsetvli zero, zero, e16, m1, ta, ma
-; CHECK-NEXT: vfncvtbf16.f.f.w v8, v10, v0.t
+; CHECK-NEXT: vfncvtbf16.f.f.w v8, v10
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 4 x bfloat> poison, bfloat %b, i32 0
%vb = shufflevector <vscale x 4 x bfloat> %elt.head, <vscale x 4 x bfloat> poison, <vscale x 4 x i32> zeroinitializer
@@ -200,13 +190,11 @@ define <vscale x 4 x bfloat> @vfsub_vf_nxv4bf16(<vscale x 4 x bfloat> %va, bfloa
define <vscale x 4 x bfloat> @vfsub_vf_nxv4bf16_unmasked(<vscale x 4 x bfloat> %va, bfloat %b, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vf_nxv4bf16_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: fmv.x.h a1, fa0
-; CHECK-NEXT: vsetvli zero, a0, e16, m1, ta, ma
-; CHECK-NEXT: vmv.v.x v12, a1
+; CHECK-NEXT: fcvt.s.bf16 fa5, fa0
+; CHECK-NEXT: vsetvli a0, zero, e16, m1, ta, ma
; CHECK-NEXT: vfwcvtbf16.f.f.v v10, v8
-; CHECK-NEXT: vfwcvtbf16.f.f.v v8, v12
; CHECK-NEXT: vsetvli zero, zero, e32, m2, ta, ma
-; CHECK-NEXT: vfsub.vv v10, v10, v8
+; CHECK-NEXT: vfsub.vf v10, v10, fa5
; CHECK-NEXT: vsetvli zero, zero, e16, m1, ta, ma
; CHECK-NEXT: vfncvtbf16.f.f.w v8, v10
; CHECK-NEXT: ret
@@ -219,13 +207,13 @@ define <vscale x 4 x bfloat> @vfsub_vf_nxv4bf16_unmasked(<vscale x 4 x bfloat> %
define <vscale x 8 x bfloat> @vfsub_vv_nxv8bf16(<vscale x 8 x bfloat> %va, <vscale x 8 x bfloat> %b, <vscale x 8 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vv_nxv8bf16:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e16, m2, ta, ma
-; CHECK-NEXT: vfwcvtbf16.f.f.v v12, v10, v0.t
-; CHECK-NEXT: vfwcvtbf16.f.f.v v16, v8, v0.t
+; CHECK-NEXT: vsetvli a0, zero, e16, m2, ta, ma
+; CHECK-NEXT: vfwcvtbf16.f.f.v v12, v10
+; CHECK-NEXT: vfwcvtbf16.f.f.v v16, v8
; CHECK-NEXT: vsetvli zero, zero, e32, m4, ta, ma
-; CHECK-NEXT: vfsub.vv v12, v16, v12, v0.t
+; CHECK-NEXT: vfsub.vv v12, v16, v12
; CHECK-NEXT: vsetvli zero, zero, e16, m2, ta, ma
-; CHECK-NEXT: vfncvtbf16.f.f.w v8, v12, v0.t
+; CHECK-NEXT: vfncvtbf16.f.f.w v8, v12
; CHECK-NEXT: ret
%v = call <vscale x 8 x bfloat> @llvm.vp.fsub.nxv8bf16(<vscale x 8 x bfloat> %va, <vscale x 8 x bfloat> %b, <vscale x 8 x i1> %m, i32 %evl)
ret <vscale x 8 x bfloat> %v
@@ -234,7 +222,7 @@ define <vscale x 8 x bfloat> @vfsub_vv_nxv8bf16(<vscale x 8 x bfloat> %va, <vsca
define <vscale x 8 x bfloat> @vfsub_vv_nxv8bf16_unmasked(<vscale x 8 x bfloat> %va, <vscale x 8 x bfloat> %b, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vv_nxv8bf16_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; CHECK-NEXT: vsetvli a0, zero, e16, m2, ta, ma
; CHECK-NEXT: vfwcvtbf16.f.f.v v12, v10
; CHECK-NEXT: vfwcvtbf16.f.f.v v16, v8
; CHECK-NEXT: vsetvli zero, zero, e32, m4, ta, ma
@@ -249,15 +237,13 @@ define <vscale x 8 x bfloat> @vfsub_vv_nxv8bf16_unmasked(<vscale x 8 x bfloat> %
define <vscale x 8 x bfloat> @vfsub_vf_nxv8bf16(<vscale x 8 x bfloat> %va, bfloat %b, <vscale x 8 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vf_nxv8bf16:
; CHECK: # %bb.0:
-; CHECK-NEXT: fmv.x.h a1, fa0
-; CHECK-NEXT: vsetvli zero, a0, e16, m2, ta, ma
-; CHECK-NEXT: vmv.v.x v16, a1
-; CHECK-NEXT: vfwcvtbf16.f.f.v v12, v8, v0.t
-; CHECK-NEXT: vfwcvtbf16.f.f.v v8, v16, v0.t
+; CHECK-NEXT: fcvt.s.bf16 fa5, fa0
+; CHECK-NEXT: vsetvli a0, zero, e16, m2, ta, ma
+; CHECK-NEXT: vfwcvtbf16.f.f.v v12, v8
; CHECK-NEXT: vsetvli zero, zero, e32, m4, ta, ma
-; CHECK-NEXT: vfsub.vv v12, v12, v8, v0.t
+; CHECK-NEXT: vfsub.vf v12, v12, fa5
; CHECK-NEXT: vsetvli zero, zero, e16, m2, ta, ma
-; CHECK-NEXT: vfncvtbf16.f.f.w v8, v12, v0.t
+; CHECK-NEXT: vfncvtbf16.f.f.w v8, v12
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 8 x bfloat> poison, bfloat %b, i32 0
%vb = shufflevector <vscale x 8 x bfloat> %elt.head, <vscale x 8 x bfloat> poison, <vscale x 8 x i32> zeroinitializer
@@ -268,13 +254,11 @@ define <vscale x 8 x bfloat> @vfsub_vf_nxv8bf16(<vscale x 8 x bfloat> %va, bfloa
define <vscale x 8 x bfloat> @vfsub_vf_nxv8bf16_unmasked(<vscale x 8 x bfloat> %va, bfloat %b, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vf_nxv8bf16_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: fmv.x.h a1, fa0
-; CHECK-NEXT: vsetvli zero, a0, e16, m2, ta, ma
-; CHECK-NEXT: vmv.v.x v16, a1
+; CHECK-NEXT: fcvt.s.bf16 fa5, fa0
+; CHECK-NEXT: vsetvli a0, zero, e16, m2, ta, ma
; CHECK-NEXT: vfwcvtbf16.f.f.v v12, v8
-; CHECK-NEXT: vfwcvtbf16.f.f.v v8, v16
; CHECK-NEXT: vsetvli zero, zero, e32, m4, ta, ma
-; CHECK-NEXT: vfsub.vv v12, v12, v8
+; CHECK-NEXT: vfsub.vf v12, v12, fa5
; CHECK-NEXT: vsetvli zero, zero, e16, m2, ta, ma
; CHECK-NEXT: vfncvtbf16.f.f.w v8, v12
; CHECK-NEXT: ret
@@ -287,13 +271,13 @@ define <vscale x 8 x bfloat> @vfsub_vf_nxv8bf16_unmasked(<vscale x 8 x bfloat> %
define <vscale x 16 x bfloat> @vfsub_vv_nxv16bf16(<vscale x 16 x bfloat> %va, <vscale x 16 x bfloat> %b, <vscale x 16 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vv_nxv16bf16:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e16, m4, ta, ma
-; CHECK-NEXT: vfwcvtbf16.f.f.v v16, v12, v0.t
-; CHECK-NEXT: vfwcvtbf16.f.f.v v24, v8, v0.t
+; CHECK-NEXT: vsetvli a0, zero, e16, m4, ta, ma
+; CHECK-NEXT: vfwcvtbf16.f.f.v v16, v12
+; CHECK-NEXT: vfwcvtbf16.f.f.v v24, v8
; CHECK-NEXT: vsetvli zero, zero, e32, m8, ta, ma
-; CHECK-NEXT: vfsub.vv v16, v24, v16, v0.t
+; CHECK-NEXT: vfsub.vv v16, v24, v16
; CHECK-NEXT: vsetvli zero, zero, e16, m4, ta, ma
-; CHECK-NEXT: vfncvtbf16.f.f.w v8, v16, v0.t
+; CHECK-NEXT: vfncvtbf16.f.f.w v8, v16
; CHECK-NEXT: ret
%v = call <vscale x 16 x bfloat> @llvm.vp.fsub.nxv16bf16(<vscale x 16 x bfloat> %va, <vscale x 16 x bfloat> %b, <vscale x 16 x i1> %m, i32 %evl)
ret <vscale x 16 x bfloat> %v
@@ -302,7 +286,7 @@ define <vscale x 16 x bfloat> @vfsub_vv_nxv16bf16(<vscale x 16 x bfloat> %va, <v
define <vscale x 16 x bfloat> @vfsub_vv_nxv16bf16_unmasked(<vscale x 16 x bfloat> %va, <vscale x 16 x bfloat> %b, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vv_nxv16bf16_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; CHECK-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; CHECK-NEXT: vfwcvtbf16.f.f.v v16, v12
; CHECK-NEXT: vfwcvtbf16.f.f.v v24, v8
; CHECK-NEXT: vsetvli zero, zero, e32, m8, ta, ma
@@ -317,15 +301,13 @@ define <vscale x 16 x bfloat> @vfsub_vv_nxv16bf16_unmasked(<vscale x 16 x bfloat
define <vscale x 16 x bfloat> @vfsub_vf_nxv16bf16(<vscale x 16 x bfloat> %va, bfloat %b, <vscale x 16 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vf_nxv16bf16:
; CHECK: # %bb.0:
-; CHECK-NEXT: fmv.x.h a1, fa0
-; CHECK-NEXT: vsetvli zero, a0, e16, m4, ta, ma
-; CHECK-NEXT: vmv.v.x v24, a1
-; CHECK-NEXT: vfwcvtbf16.f.f.v v16, v8, v0.t
-; CHECK-NEXT: vfwcvtbf16.f.f.v v8, v24, v0.t
+; CHECK-NEXT: fcvt.s.bf16 fa5, fa0
+; CHECK-NEXT: vsetvli a0, zero, e16, m4, ta, ma
+; CHECK-NEXT: vfwcvtbf16.f.f.v v16, v8
; CHECK-NEXT: vsetvli zero, zero, e32, m8, ta, ma
-; CHECK-NEXT: vfsub.vv v16, v16, v8, v0.t
+; CHECK-NEXT: vfsub.vf v16, v16, fa5
; CHECK-NEXT: vsetvli zero, zero, e16, m4, ta, ma
-; CHECK-NEXT: vfncvtbf16.f.f.w v8, v16, v0.t
+; CHECK-NEXT: vfncvtbf16.f.f.w v8, v16
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 16 x bfloat> poison, bfloat %b, i32 0
%vb = shufflevector <vscale x 16 x bfloat> %elt.head, <vscale x 16 x bfloat> poison, <vscale x 16 x i32> zeroinitializer
@@ -336,13 +318,11 @@ define <vscale x 16 x bfloat> @vfsub_vf_nxv16bf16(<vscale x 16 x bfloat> %va, bf
define <vscale x 16 x bfloat> @vfsub_vf_nxv16bf16_unmasked(<vscale x 16 x bfloat> %va, bfloat %b, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vf_nxv16bf16_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: fmv.x.h a1, fa0
-; CHECK-NEXT: vsetvli zero, a0, e16, m4, ta, ma
-; CHECK-NEXT: vmv.v.x v24, a1
+; CHECK-NEXT: fcvt.s.bf16 fa5, fa0
+; CHECK-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; CHECK-NEXT: vfwcvtbf16.f.f.v v16, v8
-; CHECK-NEXT: vfwcvtbf16.f.f.v v8, v24
; CHECK-NEXT: vsetvli zero, zero, e32, m8, ta, ma
-; CHECK-NEXT: vfsub.vv v16, v16, v8
+; CHECK-NEXT: vfsub.vf v16, v16, fa5
; CHECK-NEXT: vsetvli zero, zero, e16, m4, ta, ma
; CHECK-NEXT: vfncvtbf16.f.f.w v8, v16
; CHECK-NEXT: ret
@@ -357,43 +337,26 @@ define <vscale x 32 x bfloat> @vfsub_vv_nxv32bf16(<vscale x 32 x bfloat> %va, <v
; CHECK: # %bb.0:
; CHECK-NEXT: addi sp, sp, -16
; CHECK-NEXT: .cfi_def_cfa_offset 16
-; CHECK-NEXT: csrr a1, vlenb
-; CHECK-NEXT: slli a1, a1, 3
-; CHECK-NEXT: sub sp, sp, a1
+; CHECK-NEXT: csrr a0, vlenb
+; CHECK-NEXT: slli a0, a0, 3
+; CHECK-NEXT: sub sp, sp, a0
; CHECK-NEXT: .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x08, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 8 * vlenb
-; CHECK-NEXT: vsetvli a1, zero, e8, mf2, ta, ma
-; CHECK-NEXT: vmv1r.v v7, v0
-; CHECK-NEXT: csrr a2, vlenb
-; CHECK-NEXT: slli a1, a2, 1
-; CHECK-NEXT: srli a2, a2, 2
-; CHECK-NEXT: sub a3, a0, a1
-; CHECK-NEXT: vslidedown.vx v0, v0, a2
-; CHECK-NEXT: sltu a2, a0, a3
-; CHECK-NEXT: addi a2, a2, -1
-; CHECK-NEXT: and a2, a2, a3
-; CHECK-NEXT: addi a3, sp, 16
-; CHECK-NEXT: vs8r.v v16, (a3) # vscale x 64-byte Folded Spill
-; CHECK-NEXT: vsetvli zero, a2, e16, m4, ta, ma
-; CHECK-NEXT: vfwcvtbf16.f.f.v v24, v20, v0.t
-; CHECK-NEXT: vfwcvtbf16.f.f.v v16, v12, v0.t
+; CHECK-NEXT: vsetvli a0, zero, e16, m4, ta, ma
+; CHECK-NEXT: vfwcvtbf16.f.f.v v24, v16
+; CHECK-NEXT: addi a0, sp, 16
+; CHECK-NEXT: vs8r.v v24, (a0) # vscale x 64-byte Folded Spill
+; CHECK-NEXT: vfwcvtbf16.f.f.v v0, v8
+; CHECK-NEXT: vfwcvtbf16.f.f.v v24, v20
+; CHECK-NEXT: vfwcvtbf16.f.f.v v16, v12
+; CHECK-NEXT: vl8r.v v8, (a0) # vscale x 64-byte Folded Reload
; CHECK-NEXT: vsetvli zero, zero, e32, m8, ta, ma
-; CHECK-NEXT: vfsub.vv v16, v16, v24, v0.t
+; CHECK-NEXT: vfsub.vv v0, v0, v8
; CHECK-NEXT: vsetvli zero, zero, e16, m4, ta, ma
-; CHECK-NEXT: vfncvtbf16.f.f.w v12, v16, v0.t
-; CHECK-NEXT: bltu a0, a1, .LBB20_2
-; CHECK-NEXT: # %bb.1:
-; CHECK-NEXT: mv a0, a1
-; CHECK-NEXT: .LBB20_2:
-; CHECK-NEXT: vmv1r.v v0, v7
-; CHECK-NEXT: addi a1, sp, 16
-; CHECK-NEXT: vl8r.v v24, (a1) # vscale x 64-byte Folded Reload
-; CHECK-NEXT: vsetvli zero, a0, e16, m4, ta, ma
-; CHECK-NEXT: vfwcvtbf16.f.f.v v16, v24, v0.t
-; CHECK-NEXT: vfwcvtbf16.f.f.v v24, v8, v0.t
+; CHECK-NEXT: vfncvtbf16.f.f.w v8, v0
; CHECK-NEXT: vsetvli zero, zero, e32, m8, ta, ma
-; CHECK-NEXT: vfsub.vv v16, v24, v16, v0.t
+; CHECK-NEXT: vfsub.vv v16, v16, v24
; CHECK-NEXT: vsetvli zero, zero, e16, m4, ta, ma
-; CHECK-NEXT: vfncvtbf16.f.f.w v8, v16, v0.t
+; CHECK-NEXT: vfncvtbf16.f.f.w v12, v16
; CHECK-NEXT: csrr a0, vlenb
; CHECK-NEXT: slli a0, a0, 3
; CHECK-NEXT: add sp, sp, a0
@@ -410,43 +373,26 @@ define <vscale x 32 x bfloat> @vfsub_vv_nxv32bf16_unmasked(<vscale x 32 x bfloat
; CHECK: # %bb.0:
; CHECK-NEXT: addi sp, sp, -16
; CHECK-NEXT: .cfi_def_cfa_offset 16
-; CHECK-NEXT: csrr a1, vlenb
-; CHECK-NEXT: slli a1, a1, 3
-; CHECK-NEXT: sub sp, sp, a1
+; CHECK-NEXT: csrr a0, vlenb
+; CHECK-NEXT: slli a0, a0, 3
+; CHECK-NEXT: sub sp, sp, a0
; CHECK-NEXT: .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x08, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 8 * vlenb
-; CHECK-NEXT: csrr a2, vlenb
-; CHECK-NEXT: vsetvli a1, zero, e8, m4, ta, ma
-; CHECK-NEXT: vmset.m v24
-; CHECK-NEXT: slli a1, a2, 1
-; CHECK-NEXT: srli a2, a2, 2
-; CHECK-NEXT: sub a3, a0, a1
-; CHECK-NEXT: vsetvli a4, zero, e8, mf2, ta, ma
-; CHECK-NEXT: vslidedown.vx v0, v24, a2
-; CHECK-NEXT: sltu a2, a0, a3
-; CHECK-NEXT: addi a2, a2, -1
-; CHECK-NEXT: and a2, a2, a3
-; CHECK-NEXT: addi a3, sp, 16
-; CHECK-NEXT: vs8r.v v16, (a3) # vscale x 64-byte Folded Spill
-; CHECK-NEXT: vsetvli zero, a2, e16, m4, ta, ma
-; CHECK-NEXT: vfwcvtbf16.f.f.v v24, v20, v0.t
-; CHECK-NEXT: vfwcvtbf16.f.f.v v16, v12, v0.t
+; CHECK-NEXT: vsetvli a0, zero, e16, m4, ta, ma
+; CHECK-NEXT: vfwcvtbf16.f.f.v v24, v16
+; CHECK-NEXT: addi a0, sp, 16
+; CHECK-NEXT: vs8r.v v24, (a0) # vscale x 64-byte Folded Spill
+; CHECK-NEXT: vfwcvtbf16.f.f.v v0, v8
+; CHECK-NEXT: vfwcvtbf16.f.f.v v24, v20
+; CHECK-NEXT: vfwcvtbf16.f.f.v v16, v12
+; CHECK-NEXT: vl8r.v v8, (a0) # vscale x 64-byte Folded Reload
; CHECK-NEXT: vsetvli zero, zero, e32, m8, ta, ma
-; CHECK-NEXT: vfsub.vv v16, v16, v24, v0.t
+; CHECK-NEXT: vfsub.vv v0, v0, v8
; CHECK-NEXT: vsetvli zero, zero, e16, m4, ta, ma
-; CHECK-NEXT: vfncvtbf16.f.f.w v12, v16, v0.t
-; CHECK-NEXT: bltu a0, a1, .LBB21_2
-; CHECK-NEXT: # %bb.1:
-; CHECK-NEXT: mv a0, a1
-; CHECK-NEXT: .LBB21_2:
-; CHECK-NEXT: addi a1, sp, 16
-; CHECK-NEXT: vl8r.v v24, (a1) # vscale x 64-byte Folded Reload
-; CHECK-NEXT: vsetvli zero, a0, e16, m4, ta, ma
-; CHECK-NEXT: vfwcvtbf16.f.f.v v16, v24
-; CHECK-NEXT: vfwcvtbf16.f.f.v v24, v8
+; CHECK-NEXT: vfncvtbf16.f.f.w v8, v0
; CHECK-NEXT: vsetvli zero, zero, e32, m8, ta, ma
-; CHECK-NEXT: vfsub.vv v16, v24, v16
+; CHECK-NEXT: vfsub.vv v16, v16, v24
; CHECK-NEXT: vsetvli zero, zero, e16, m4, ta, ma
-; CHECK-NEXT: vfncvtbf16.f.f.w v8, v16
+; CHECK-NEXT: vfncvtbf16.f.f.w v12, v16
; CHECK-NEXT: csrr a0, vlenb
; CHECK-NEXT: slli a0, a0, 3
; CHECK-NEXT: add sp, sp, a0
@@ -463,58 +409,33 @@ define <vscale x 32 x bfloat> @vfsub_vf_nxv32bf16(<vscale x 32 x bfloat> %va, bf
; CHECK: # %bb.0:
; CHECK-NEXT: addi sp, sp, -16
; CHECK-NEXT: .cfi_def_cfa_offset 16
-; CHECK-NEXT: csrr a1, vlenb
-; CHECK-NEXT: slli a1, a1, 4
-; CHECK-NEXT: sub sp, sp, a1
-; CHECK-NEXT: .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x10, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 16 * vlenb
-; CHECK-NEXT: vsetvli a1, zero, e16, m8, ta, ma
-; CHECK-NEXT: vmv1r.v v7, v0
-; CHECK-NEXT: fmv.x.h a1, fa0
-; CHECK-NEXT: csrr a2, vlenb
-; CHECK-NEXT: vmv.v.x v24, a1
-; CHECK-NEXT: slli a1, a2, 1
-; CHECK-NEXT: srli a2, a2, 2
-; CHECK-NEXT: sub a3, a0, a1
-; CHECK-NEXT: vsetvli a4, zero, e8, mf2, ta, ma
-; CHECK-NEXT: vslidedown.vx v0, v0, a2
-; CHECK-NEXT: sltu a2, a0, a3
-; CHECK-NEXT: addi a2, a2, -1
-; CHECK-NEXT: and a2, a2, a3
-; CHECK-NEXT: csrr a3, vlenb
-; CHECK-NEXT: slli a3, a3, 3
-; CHECK-NEXT: add a3, sp, a3
-; CHECK-NEXT: addi a3, a3, 16
-; CHECK-NEXT: vs8r.v v24, (a3) # vscale x 64-byte Folded Spill
-; CHECK-NEXT: vsetvli zero, a2, e16, m4, ta, ma
-; CHECK-NEXT: vfwcvtbf16.f.f.v v16, v28, v0.t
-; CHECK-NEXT: vfwcvtbf16.f.f.v v24, v12, v0.t
-; CHECK-NEXT: vsetvli zero, zero, e32, m8, ta, ma
-; CHECK-NEXT: vfsub.vv v16, v24, v16, v0.t
-; CHECK-NEXT: vsetvli zero, zero, e16, m4, ta, ma
-; CHECK-NEXT: vfncvtbf16.f.f.w v12, v16, v0.t
-; CHECK-NEXT: bltu a0, a1, .LBB22_2
-; CHECK-NEXT: # %bb.1:
-; CHECK-NEXT: mv a0, a1
-; CHECK-NEXT: .LBB22_2:
-; CHECK-NEXT: vmv1r.v v0, v7
-; CHECK-NEXT: vsetvli zero, a0, e16, m4, ta, ma
-; CHECK-NEXT: vfwcvtbf16.f.f.v v16, v8, v0.t
-; CHECK-NEXT: addi a0, sp, 16
-; CHECK-NEXT: vs8r.v v16, (a0) # vscale x 64-byte Folded Spill
; CHECK-NEXT: csrr a0, vlenb
; CHECK-NEXT: slli a0, a0, 3
-; CHECK-NEXT: add a0, sp, a0
-; CHECK-NEXT: addi a0, a0, 16
-; CHECK-NEXT: vl8r.v v16, (a0) # vscale x 64-byte Folded Reload
-; CHECK-NEXT: vfwcvtbf16.f.f.v v24, v16, v0.t
+; CHECK-NEXT: sub sp, sp, a0
+; CHECK-NEXT: .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x08, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 8 * vlenb
+; CHECK-NEXT: fmv.x.h a0, fa0
+; CHECK-NEXT: vsetvli a1, zero, e16, m4, ta, ma
+; CHECK-NEXT: vfwcvtbf16.f.f.v v16, v8
+; CHECK-NEXT: addi a1, sp, 16
+; CHECK-NEXT: vs8r.v v16, (a1) # vscale x 64-byte Folded Spill
+; CHECK-NEXT: vfwcvtbf16.f.f.v v24, v12
+; CHECK-NEXT: vsetvli a1, zero, e16, m8, ta, ma
+; CHECK-NEXT: vmv.v.x v8, a0
+; CHECK-NEXT: vsetvli a0, zero, e16, m4, ta, ma
+; CHECK-NEXT: vfwcvtbf16.f.f.v v0, v8
+; CHECK-NEXT: vfwcvtbf16.f.f.v v16, v12
; CHECK-NEXT: addi a0, sp, 16
-; CHECK-NEXT: vl8r.v v16, (a0) # vscale x 64-byte Folded Reload
+; CHECK-NEXT: vl8r.v v8, (a0) # vscale x 64-byte Folded Reload
+; CHECK-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; CHECK-NEXT: vfsub.vv v0, v8, v0
+; CHECK-NEXT: vsetvli zero, zero, e16, m4, ta, ma
+; CHECK-NEXT: vfncvtbf16.f.f.w v8, v0
; CHECK-NEXT: vsetvli zero, zero, e32, m8, ta, ma
-; CHECK-NEXT: vfsub.vv v16, v16, v24, v0.t
+; CHECK-NEXT: vfsub.vv v16, v24, v16
; CHECK-NEXT: vsetvli zero, zero, e16, m4, ta, ma
-; CHECK-NEXT: vfncvtbf16.f.f.w v8, v16, v0.t
+; CHECK-NEXT: vfncvtbf16.f.f.w v12, v16
; CHECK-NEXT: csrr a0, vlenb
-; CHECK-NEXT: slli a0, a0, 4
+; CHECK-NEXT: slli a0, a0, 3
; CHECK-NEXT: add sp, sp, a0
; CHECK-NEXT: .cfi_def_cfa sp, 16
; CHECK-NEXT: addi sp, sp, 16
@@ -531,45 +452,31 @@ define <vscale x 32 x bfloat> @vfsub_vf_nxv32bf16_unmasked(<vscale x 32 x bfloat
; CHECK: # %bb.0:
; CHECK-NEXT: addi sp, sp, -16
; CHECK-NEXT: .cfi_def_cfa_offset 16
-; CHECK-NEXT: csrr a1, vlenb
-; CHECK-NEXT: slli a1, a1, 3
-; CHECK-NEXT: sub sp, sp, a1
+; CHECK-NEXT: csrr a0, vlenb
+; CHECK-NEXT: slli a0, a0, 3
+; CHECK-NEXT: sub sp, sp, a0
; CHECK-NEXT: .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x10, 0x22, 0x11, 0x08, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 16 + 8 * vlenb
-; CHECK-NEXT: fmv.x.h a1, fa0
-; CHECK-NEXT: csrr a2, vlenb
-; CHECK-NEXT: vsetvli a3, zero, e16, m8, ta, ma
-; CHECK-NEXT: vmset.m v24
-; CHECK-NEXT: vmv.v.x v16, a1
-; CHECK-NEXT: slli a1, a2, 1
-; CHECK-NEXT: srli a2, a2, 2
-; CHECK-NEXT: sub a3, a0, a1
-; CHECK-NEXT: vsetvli a4, zero, e8, mf2, ta, ma
-; CHECK-NEXT: vslidedown.vx v0, v24, a2
-; CHECK-NEXT: sltu a2, a0, a3
-; CHECK-NEXT: addi a2, a2, -1
-; CHECK-NEXT: and a2, a2, a3
-; CHECK-NEXT: addi a3, sp, 16
-; CHECK-NEXT: vs8r.v v16, (a3) # vscale x 64-byte Folded Spill
-; CHECK-NEXT: vsetvli zero, a2, e16, m4, ta, ma
-; CHECK-NEXT: vfwcvtbf16.f.f.v v24, v20, v0.t
-; CHECK-NEXT: vfwcvtbf16.f.f.v v16, v12, v0.t
-; CHECK-NEXT: vsetvli zero, zero, e32, m8, ta, ma
-; CHECK-NEXT: vfsub.vv v16, v16, v24, v0.t
-; CHECK-NEXT: vsetvli zero, zero, e16, m4, ta, ma
-; CHECK-NEXT: vfncvtbf16.f.f.w v12, v16, v0.t
-; CHECK-NEXT: bltu a0, a1, .LBB23_2
-; CHECK-NEXT: # %bb.1:
-; CHECK-NEXT: mv a0, a1
-; CHECK-NEXT: .LBB23_2:
-; CHECK-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; CHECK-NEXT: fmv.x.h a0, fa0
+; CHECK-NEXT: vsetvli a1, zero, e16, m4, ta, ma
; CHECK-NEXT: vfwcvtbf16.f.f.v v16, v8
+; CHECK-NEXT: addi a1, sp, 16
+; CHECK-NEXT: vs8r.v v16, (a1) # vscale x 64-byte Folded Spill
+; CHECK-NEXT: vfwcvtbf16.f.f.v v24, v12
+; CHECK-NEXT: vsetvli a1, zero, e16, m8, ta, ma
+; CHECK-NEXT: vmv.v.x v8, a0
+; CHECK-NEXT: vsetvli a0, zero, e16, m4, ta, ma
+; CHECK-NEXT: vfwcvtbf16.f.f.v v0, v8
+; CHECK-NEXT: vfwcvtbf16.f.f.v v16, v12
; CHECK-NEXT: addi a0, sp, 16
-; CHECK-NEXT: vl8r.v v0, (a0) # vscale x 64-byte Folded Reload
-; CHECK-NEXT: vfwcvtbf16.f.f.v v24, v0
+; CHECK-NEXT: vl8r.v v8, (a0) # vscale x 64-byte Folded Reload
; CHECK-NEXT: vsetvli zero, zero, e32, m8, ta, ma
-; CHECK-NEXT: vfsub.vv v16, v16, v24
+; CHECK-NEXT: vfsub.vv v0, v8, v0
; CHECK-NEXT: vsetvli zero, zero, e16, m4, ta, ma
-; CHECK-NEXT: vfncvtbf16.f.f.w v8, v16
+; CHECK-NEXT: vfncvtbf16.f.f.w v8, v0
+; CHECK-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; CHECK-NEXT: vfsub.vv v16, v24, v16
+; CHECK-NEXT: vsetvli zero, zero, e16, m4, ta, ma
+; CHECK-NEXT: vfncvtbf16.f.f.w v12, v16
; CHECK-NEXT: csrr a0, vlenb
; CHECK-NEXT: slli a0, a0, 3
; CHECK-NEXT: add sp, sp, a0
@@ -586,19 +493,19 @@ define <vscale x 32 x bfloat> @vfsub_vf_nxv32bf16_unmasked(<vscale x 32 x bfloat
define <vscale x 1 x half> @vfsub_vv_nxv1f16(<vscale x 1 x half> %va, <vscale x 1 x half> %b, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; ZVFH-LABEL: vfsub_vv_nxv1f16:
; ZVFH: # %bb.0:
-; ZVFH-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
-; ZVFH-NEXT: vfsub.vv v8, v8, v9, v0.t
+; ZVFH-NEXT: vsetvli a0, zero, e16, mf4, ta, ma
+; ZVFH-NEXT: vfsub.vv v8, v8, v9
; ZVFH-NEXT: ret
;
; ZVFHMIN-LABEL: vfsub_vv_nxv1f16:
; ZVFHMIN: # %bb.0:
-; 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 a0, 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, zero, e32, mf2, ta, ma
-; ZVFHMIN-NEXT: vfsub.vv v9, v9, v10, v0.t
+; ZVFHMIN-NEXT: vfsub.vv v9, v9, v10
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9, v0.t
+; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
; ZVFHMIN-NEXT: ret
%v = call <vscale x 1 x half> @llvm.vp.fsub.nxv1f16(<vscale x 1 x half> %va, <vscale x 1 x half> %b, <vscale x 1 x i1> %m, i32 %evl)
ret <vscale x 1 x half> %v
@@ -607,13 +514,13 @@ define <vscale x 1 x half> @vfsub_vv_nxv1f16(<vscale x 1 x half> %va, <vscale x
define <vscale x 1 x half> @vfsub_vv_nxv1f16_unmasked(<vscale x 1 x half> %va, <vscale x 1 x half> %b, i32 zeroext %evl) {
; ZVFH-LABEL: vfsub_vv_nxv1f16_unmasked:
; ZVFH: # %bb.0:
-; ZVFH-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFH-NEXT: vsetvli a0, zero, e16, mf4, ta, ma
; ZVFH-NEXT: vfsub.vv v8, v8, v9
; ZVFH-NEXT: ret
;
; ZVFHMIN-LABEL: vfsub_vv_nxv1f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vsetvli a0, 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, zero, e32, mf2, ta, ma
@@ -628,21 +535,19 @@ define <vscale x 1 x half> @vfsub_vv_nxv1f16_unmasked(<vscale x 1 x half> %va, <
define <vscale x 1 x half> @vfsub_vf_nxv1f16(<vscale x 1 x half> %va, half %b, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; ZVFH-LABEL: vfsub_vf_nxv1f16:
; ZVFH: # %bb.0:
-; ZVFH-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
-; ZVFH-NEXT: vfsub.vf v8, v8, fa0, v0.t
+; ZVFH-NEXT: vsetvli a0, zero, e16, mf4, ta, ma
+; ZVFH-NEXT: vfsub.vf v8, v8, fa0
; ZVFH-NEXT: ret
;
; ZVFHMIN-LABEL: vfsub_vf_nxv1f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: fmv.x.h a1, fa0
-; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vmv.v.x v9, a1
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8, v0.t
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v9, v0.t
+; ZVFHMIN-NEXT: fcvt.s.h fa5, fa0
+; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
-; ZVFHMIN-NEXT: vfsub.vv v9, v10, v8, v0.t
+; ZVFHMIN-NEXT: vfsub.vf v9, v9, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9, v0.t
+; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
; ZVFHMIN-NEXT: ret
%elt.head = insertelement <vscale x 1 x half> poison, half %b, i32 0
%vb = shufflevector <vscale x 1 x half> %elt.head, <vscale x 1 x half> poison, <vscale x 1 x i32> zeroinitializer
@@ -653,19 +558,17 @@ define <vscale x 1 x half> @vfsub_vf_nxv1f16(<vscale x 1 x half> %va, half %b, <
define <vscale x 1 x half> @vfsub_vf_nxv1f16_unmasked(<vscale x 1 x half> %va, half %b, i32 zeroext %evl) {
; ZVFH-LABEL: vfsub_vf_nxv1f16_unmasked:
; ZVFH: # %bb.0:
-; ZVFH-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
+; ZVFH-NEXT: vsetvli a0, zero, e16, mf4, ta, ma
; ZVFH-NEXT: vfsub.vf v8, v8, fa0
; ZVFH-NEXT: ret
;
; ZVFHMIN-LABEL: vfsub_vf_nxv1f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: fmv.x.h a1, fa0
-; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf4, ta, ma
-; ZVFHMIN-NEXT: vmv.v.x v9, a1
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v9
+; ZVFHMIN-NEXT: fcvt.s.h fa5, fa0
+; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
; ZVFHMIN-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
-; ZVFHMIN-NEXT: vfsub.vv v9, v10, v8
+; ZVFHMIN-NEXT: vfsub.vf v9, v9, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
; ZVFHMIN-NEXT: ret
@@ -678,19 +581,19 @@ define <vscale x 1 x half> @vfsub_vf_nxv1f16_unmasked(<vscale x 1 x half> %va, h
define <vscale x 2 x half> @vfsub_vv_nxv2f16(<vscale x 2 x half> %va, <vscale x 2 x half> %b, <vscale x 2 x i1> %m, i32 zeroext %evl) {
; ZVFH-LABEL: vfsub_vv_nxv2f16:
; ZVFH: # %bb.0:
-; ZVFH-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
-; ZVFH-NEXT: vfsub.vv v8, v8, v9, v0.t
+; ZVFH-NEXT: vsetvli a0, zero, e16, mf2, ta, ma
+; ZVFH-NEXT: vfsub.vv v8, v8, v9
; ZVFH-NEXT: ret
;
; ZVFHMIN-LABEL: vfsub_vv_nxv2f16:
; ZVFHMIN: # %bb.0:
-; 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 a0, 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, zero, e32, m1, ta, ma
-; ZVFHMIN-NEXT: vfsub.vv v9, v9, v10, v0.t
+; ZVFHMIN-NEXT: vfsub.vv v9, v9, v10
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9, v0.t
+; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
; ZVFHMIN-NEXT: ret
%v = call <vscale x 2 x half> @llvm.vp.fsub.nxv2f16(<vscale x 2 x half> %va, <vscale x 2 x half> %b, <vscale x 2 x i1> %m, i32 %evl)
ret <vscale x 2 x half> %v
@@ -699,13 +602,13 @@ define <vscale x 2 x half> @vfsub_vv_nxv2f16(<vscale x 2 x half> %va, <vscale x
define <vscale x 2 x half> @vfsub_vv_nxv2f16_unmasked(<vscale x 2 x half> %va, <vscale x 2 x half> %b, i32 zeroext %evl) {
; ZVFH-LABEL: vfsub_vv_nxv2f16_unmasked:
; ZVFH: # %bb.0:
-; ZVFH-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFH-NEXT: vsetvli a0, zero, e16, mf2, ta, ma
; ZVFH-NEXT: vfsub.vv v8, v8, v9
; ZVFH-NEXT: ret
;
; ZVFHMIN-LABEL: vfsub_vv_nxv2f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vsetvli a0, 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, zero, e32, m1, ta, ma
@@ -720,21 +623,19 @@ define <vscale x 2 x half> @vfsub_vv_nxv2f16_unmasked(<vscale x 2 x half> %va, <
define <vscale x 2 x half> @vfsub_vf_nxv2f16(<vscale x 2 x half> %va, half %b, <vscale x 2 x i1> %m, i32 zeroext %evl) {
; ZVFH-LABEL: vfsub_vf_nxv2f16:
; ZVFH: # %bb.0:
-; ZVFH-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
-; ZVFH-NEXT: vfsub.vf v8, v8, fa0, v0.t
+; ZVFH-NEXT: vsetvli a0, zero, e16, mf2, ta, ma
+; ZVFH-NEXT: vfsub.vf v8, v8, fa0
; ZVFH-NEXT: ret
;
; ZVFHMIN-LABEL: vfsub_vf_nxv2f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: fmv.x.h a1, fa0
-; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vmv.v.x v9, a1
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8, v0.t
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v9, v0.t
+; ZVFHMIN-NEXT: fcvt.s.h fa5, fa0
+; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
-; ZVFHMIN-NEXT: vfsub.vv v9, v10, v8, v0.t
+; ZVFHMIN-NEXT: vfsub.vf v9, v9, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9, v0.t
+; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
; ZVFHMIN-NEXT: ret
%elt.head = insertelement <vscale x 2 x half> poison, half %b, i32 0
%vb = shufflevector <vscale x 2 x half> %elt.head, <vscale x 2 x half> poison, <vscale x 2 x i32> zeroinitializer
@@ -745,19 +646,17 @@ define <vscale x 2 x half> @vfsub_vf_nxv2f16(<vscale x 2 x half> %va, half %b, <
define <vscale x 2 x half> @vfsub_vf_nxv2f16_unmasked(<vscale x 2 x half> %va, half %b, i32 zeroext %evl) {
; ZVFH-LABEL: vfsub_vf_nxv2f16_unmasked:
; ZVFH: # %bb.0:
-; ZVFH-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
+; ZVFH-NEXT: vsetvli a0, zero, e16, mf2, ta, ma
; ZVFH-NEXT: vfsub.vf v8, v8, fa0
; ZVFH-NEXT: ret
;
; ZVFHMIN-LABEL: vfsub_vf_nxv2f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: fmv.x.h a1, fa0
-; ZVFHMIN-NEXT: vsetvli zero, a0, e16, mf2, ta, ma
-; ZVFHMIN-NEXT: vmv.v.x v9, a1
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v9
+; ZVFHMIN-NEXT: fcvt.s.h fa5, fa0
+; ZVFHMIN-NEXT: vsetvli a0, zero, e16, mf2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v9, v8
; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m1, ta, ma
-; ZVFHMIN-NEXT: vfsub.vv v9, v10, v8
+; ZVFHMIN-NEXT: vfsub.vf v9, v9, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v9
; ZVFHMIN-NEXT: ret
@@ -770,19 +669,19 @@ define <vscale x 2 x half> @vfsub_vf_nxv2f16_unmasked(<vscale x 2 x half> %va, h
define <vscale x 4 x half> @vfsub_vv_nxv4f16(<vscale x 4 x half> %va, <vscale x 4 x half> %b, <vscale x 4 x i1> %m, i32 zeroext %evl) {
; ZVFH-LABEL: vfsub_vv_nxv4f16:
; ZVFH: # %bb.0:
-; ZVFH-NEXT: vsetvli zero, a0, e16, m1, ta, ma
-; ZVFH-NEXT: vfsub.vv v8, v8, v9, v0.t
+; ZVFH-NEXT: vsetvli a0, zero, e16, m1, ta, ma
+; ZVFH-NEXT: vfsub.vv v8, v8, v9
; ZVFH-NEXT: ret
;
; ZVFHMIN-LABEL: vfsub_vv_nxv4f16:
; ZVFHMIN: # %bb.0:
-; 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 a0, 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, zero, e32, m2, ta, ma
-; ZVFHMIN-NEXT: vfsub.vv v10, v12, v10, v0.t
+; ZVFHMIN-NEXT: vfsub.vv v10, v12, v10
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10, v0.t
+; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
; ZVFHMIN-NEXT: ret
%v = call <vscale x 4 x half> @llvm.vp.fsub.nxv4f16(<vscale x 4 x half> %va, <vscale x 4 x half> %b, <vscale x 4 x i1> %m, i32 %evl)
ret <vscale x 4 x half> %v
@@ -791,13 +690,13 @@ define <vscale x 4 x half> @vfsub_vv_nxv4f16(<vscale x 4 x half> %va, <vscale x
define <vscale x 4 x half> @vfsub_vv_nxv4f16_unmasked(<vscale x 4 x half> %va, <vscale x 4 x half> %b, i32 zeroext %evl) {
; ZVFH-LABEL: vfsub_vv_nxv4f16_unmasked:
; ZVFH: # %bb.0:
-; ZVFH-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFH-NEXT: vsetvli a0, zero, e16, m1, ta, ma
; ZVFH-NEXT: vfsub.vv v8, v8, v9
; ZVFH-NEXT: ret
;
; ZVFHMIN-LABEL: vfsub_vv_nxv4f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vsetvli a0, 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, zero, e32, m2, ta, ma
@@ -812,21 +711,19 @@ define <vscale x 4 x half> @vfsub_vv_nxv4f16_unmasked(<vscale x 4 x half> %va, <
define <vscale x 4 x half> @vfsub_vf_nxv4f16(<vscale x 4 x half> %va, half %b, <vscale x 4 x i1> %m, i32 zeroext %evl) {
; ZVFH-LABEL: vfsub_vf_nxv4f16:
; ZVFH: # %bb.0:
-; ZVFH-NEXT: vsetvli zero, a0, e16, m1, ta, ma
-; ZVFH-NEXT: vfsub.vf v8, v8, fa0, v0.t
+; ZVFH-NEXT: vsetvli a0, zero, e16, m1, ta, ma
+; ZVFH-NEXT: vfsub.vf v8, v8, fa0
; ZVFH-NEXT: ret
;
; ZVFHMIN-LABEL: vfsub_vf_nxv4f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: fmv.x.h a1, fa0
-; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vmv.v.x v12, a1
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8, v0.t
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v12, v0.t
+; ZVFHMIN-NEXT: fcvt.s.h fa5, fa0
+; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m1, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
-; ZVFHMIN-NEXT: vfsub.vv v10, v10, v8, v0.t
+; ZVFHMIN-NEXT: vfsub.vf v10, v10, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10, v0.t
+; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
; ZVFHMIN-NEXT: ret
%elt.head = insertelement <vscale x 4 x half> poison, half %b, i32 0
%vb = shufflevector <vscale x 4 x half> %elt.head, <vscale x 4 x half> poison, <vscale x 4 x i32> zeroinitializer
@@ -837,19 +734,17 @@ define <vscale x 4 x half> @vfsub_vf_nxv4f16(<vscale x 4 x half> %va, half %b, <
define <vscale x 4 x half> @vfsub_vf_nxv4f16_unmasked(<vscale x 4 x half> %va, half %b, i32 zeroext %evl) {
; ZVFH-LABEL: vfsub_vf_nxv4f16_unmasked:
; ZVFH: # %bb.0:
-; ZVFH-NEXT: vsetvli zero, a0, e16, m1, ta, ma
+; ZVFH-NEXT: vsetvli a0, zero, e16, m1, ta, ma
; ZVFH-NEXT: vfsub.vf v8, v8, fa0
; ZVFH-NEXT: ret
;
; ZVFHMIN-LABEL: vfsub_vf_nxv4f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: fmv.x.h a1, fa0
-; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m1, ta, ma
-; ZVFHMIN-NEXT: vmv.v.x v12, a1
+; ZVFHMIN-NEXT: fcvt.s.h fa5, fa0
+; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v10, v8
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v12
; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m2, ta, ma
-; ZVFHMIN-NEXT: vfsub.vv v10, v10, v8
+; ZVFHMIN-NEXT: vfsub.vf v10, v10, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v10
; ZVFHMIN-NEXT: ret
@@ -862,19 +757,19 @@ define <vscale x 4 x half> @vfsub_vf_nxv4f16_unmasked(<vscale x 4 x half> %va, h
define <vscale x 8 x half> @vfsub_vv_nxv8f16(<vscale x 8 x half> %va, <vscale x 8 x half> %b, <vscale x 8 x i1> %m, i32 zeroext %evl) {
; ZVFH-LABEL: vfsub_vv_nxv8f16:
; ZVFH: # %bb.0:
-; ZVFH-NEXT: vsetvli zero, a0, e16, m2, ta, ma
-; ZVFH-NEXT: vfsub.vv v8, v8, v10, v0.t
+; ZVFH-NEXT: vsetvli a0, zero, e16, m2, ta, ma
+; ZVFH-NEXT: vfsub.vv v8, v8, v10
; ZVFH-NEXT: ret
;
; ZVFHMIN-LABEL: vfsub_vv_nxv8f16:
; ZVFHMIN: # %bb.0:
-; 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 a0, 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, zero, e32, m4, ta, ma
-; ZVFHMIN-NEXT: vfsub.vv v12, v16, v12, v0.t
+; ZVFHMIN-NEXT: vfsub.vv v12, v16, v12
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12, v0.t
+; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
; ZVFHMIN-NEXT: ret
%v = call <vscale x 8 x half> @llvm.vp.fsub.nxv8f16(<vscale x 8 x half> %va, <vscale x 8 x half> %b, <vscale x 8 x i1> %m, i32 %evl)
ret <vscale x 8 x half> %v
@@ -883,13 +778,13 @@ define <vscale x 8 x half> @vfsub_vv_nxv8f16(<vscale x 8 x half> %va, <vscale x
define <vscale x 8 x half> @vfsub_vv_nxv8f16_unmasked(<vscale x 8 x half> %va, <vscale x 8 x half> %b, i32 zeroext %evl) {
; ZVFH-LABEL: vfsub_vv_nxv8f16_unmasked:
; ZVFH: # %bb.0:
-; ZVFH-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFH-NEXT: vsetvli a0, zero, e16, m2, ta, ma
; ZVFH-NEXT: vfsub.vv v8, v8, v10
; ZVFH-NEXT: ret
;
; ZVFHMIN-LABEL: vfsub_vv_nxv8f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vsetvli a0, 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, zero, e32, m4, ta, ma
@@ -904,21 +799,19 @@ define <vscale x 8 x half> @vfsub_vv_nxv8f16_unmasked(<vscale x 8 x half> %va, <
define <vscale x 8 x half> @vfsub_vf_nxv8f16(<vscale x 8 x half> %va, half %b, <vscale x 8 x i1> %m, i32 zeroext %evl) {
; ZVFH-LABEL: vfsub_vf_nxv8f16:
; ZVFH: # %bb.0:
-; ZVFH-NEXT: vsetvli zero, a0, e16, m2, ta, ma
-; ZVFH-NEXT: vfsub.vf v8, v8, fa0, v0.t
+; ZVFH-NEXT: vsetvli a0, zero, e16, m2, ta, ma
+; ZVFH-NEXT: vfsub.vf v8, v8, fa0
; ZVFH-NEXT: ret
;
; ZVFHMIN-LABEL: vfsub_vf_nxv8f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: fmv.x.h a1, fa0
-; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vmv.v.x v16, a1
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8, v0.t
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v16, v0.t
+; ZVFHMIN-NEXT: fcvt.s.h fa5, fa0
+; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m2, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
-; ZVFHMIN-NEXT: vfsub.vv v12, v12, v8, v0.t
+; ZVFHMIN-NEXT: vfsub.vf v12, v12, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12, v0.t
+; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
; ZVFHMIN-NEXT: ret
%elt.head = insertelement <vscale x 8 x half> poison, half %b, i32 0
%vb = shufflevector <vscale x 8 x half> %elt.head, <vscale x 8 x half> poison, <vscale x 8 x i32> zeroinitializer
@@ -929,19 +822,17 @@ define <vscale x 8 x half> @vfsub_vf_nxv8f16(<vscale x 8 x half> %va, half %b, <
define <vscale x 8 x half> @vfsub_vf_nxv8f16_unmasked(<vscale x 8 x half> %va, half %b, i32 zeroext %evl) {
; ZVFH-LABEL: vfsub_vf_nxv8f16_unmasked:
; ZVFH: # %bb.0:
-; ZVFH-NEXT: vsetvli zero, a0, e16, m2, ta, ma
+; ZVFH-NEXT: vsetvli a0, zero, e16, m2, ta, ma
; ZVFH-NEXT: vfsub.vf v8, v8, fa0
; ZVFH-NEXT: ret
;
; ZVFHMIN-LABEL: vfsub_vf_nxv8f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: fmv.x.h a1, fa0
-; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m2, ta, ma
-; ZVFHMIN-NEXT: vmv.v.x v16, a1
+; ZVFHMIN-NEXT: fcvt.s.h fa5, fa0
+; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v12, v8
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v16
; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m4, ta, ma
-; ZVFHMIN-NEXT: vfsub.vv v12, v12, v8
+; ZVFHMIN-NEXT: vfsub.vf v12, v12, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v12
; ZVFHMIN-NEXT: ret
@@ -954,19 +845,19 @@ define <vscale x 8 x half> @vfsub_vf_nxv8f16_unmasked(<vscale x 8 x half> %va, h
define <vscale x 16 x half> @vfsub_vv_nxv16f16(<vscale x 16 x half> %va, <vscale x 16 x half> %b, <vscale x 16 x i1> %m, i32 zeroext %evl) {
; ZVFH-LABEL: vfsub_vv_nxv16f16:
; ZVFH: # %bb.0:
-; ZVFH-NEXT: vsetvli zero, a0, e16, m4, ta, ma
-; ZVFH-NEXT: vfsub.vv v8, v8, v12, v0.t
+; ZVFH-NEXT: vsetvli a0, zero, e16, m4, ta, ma
+; ZVFH-NEXT: vfsub.vv v8, v8, v12
; ZVFH-NEXT: ret
;
; ZVFHMIN-LABEL: vfsub_vv_nxv16f16:
; ZVFHMIN: # %bb.0:
-; 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 a0, 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, zero, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vfsub.vv v16, v24, v16, v0.t
+; ZVFHMIN-NEXT: vfsub.vv v16, v24, v16
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16, v0.t
+; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
; ZVFHMIN-NEXT: ret
%v = call <vscale x 16 x half> @llvm.vp.fsub.nxv16f16(<vscale x 16 x half> %va, <vscale x 16 x half> %b, <vscale x 16 x i1> %m, i32 %evl)
ret <vscale x 16 x half> %v
@@ -975,13 +866,13 @@ define <vscale x 16 x half> @vfsub_vv_nxv16f16(<vscale x 16 x half> %va, <vscale
define <vscale x 16 x half> @vfsub_vv_nxv16f16_unmasked(<vscale x 16 x half> %va, <vscale x 16 x half> %b, i32 zeroext %evl) {
; ZVFH-LABEL: vfsub_vv_nxv16f16_unmasked:
; ZVFH: # %bb.0:
-; ZVFH-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; ZVFH-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFH-NEXT: vfsub.vv v8, v8, v12
; ZVFH-NEXT: ret
;
; ZVFHMIN-LABEL: vfsub_vv_nxv16f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vsetvli a0, 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, zero, e32, m8, ta, ma
@@ -996,21 +887,19 @@ define <vscale x 16 x half> @vfsub_vv_nxv16f16_unmasked(<vscale x 16 x half> %va
define <vscale x 16 x half> @vfsub_vf_nxv16f16(<vscale x 16 x half> %va, half %b, <vscale x 16 x i1> %m, i32 zeroext %evl) {
; ZVFH-LABEL: vfsub_vf_nxv16f16:
; ZVFH: # %bb.0:
-; ZVFH-NEXT: vsetvli zero, a0, e16, m4, ta, ma
-; ZVFH-NEXT: vfsub.vf v8, v8, fa0, v0.t
+; ZVFH-NEXT: vsetvli a0, zero, e16, m4, ta, ma
+; ZVFH-NEXT: vfsub.vf v8, v8, fa0
; ZVFH-NEXT: ret
;
; ZVFHMIN-LABEL: vfsub_vf_nxv16f16:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: fmv.x.h a1, fa0
-; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vmv.v.x v24, a1
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8, v0.t
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v24, v0.t
+; ZVFHMIN-NEXT: fcvt.s.h fa5, fa0
+; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vfsub.vv v16, v16, v8, v0.t
+; ZVFHMIN-NEXT: vfsub.vf v16, v16, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16, v0.t
+; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
; ZVFHMIN-NEXT: ret
%elt.head = insertelement <vscale x 16 x half> poison, half %b, i32 0
%vb = shufflevector <vscale x 16 x half> %elt.head, <vscale x 16 x half> poison, <vscale x 16 x i32> zeroinitializer
@@ -1021,19 +910,17 @@ define <vscale x 16 x half> @vfsub_vf_nxv16f16(<vscale x 16 x half> %va, half %b
define <vscale x 16 x half> @vfsub_vf_nxv16f16_unmasked(<vscale x 16 x half> %va, half %b, i32 zeroext %evl) {
; ZVFH-LABEL: vfsub_vf_nxv16f16_unmasked:
; ZVFH: # %bb.0:
-; ZVFH-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; ZVFH-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFH-NEXT: vfsub.vf v8, v8, fa0
; ZVFH-NEXT: ret
;
; ZVFHMIN-LABEL: vfsub_vf_nxv16f16_unmasked:
; ZVFHMIN: # %bb.0:
-; ZVFHMIN-NEXT: fmv.x.h a1, fa0
-; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vmv.v.x v24, a1
+; ZVFHMIN-NEXT: fcvt.s.h fa5, fa0
+; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v8, v24
; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vfsub.vv v16, v16, v8
+; ZVFHMIN-NEXT: vfsub.vf v16, v16, fa5
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
; ZVFHMIN-NEXT: ret
@@ -1046,51 +933,34 @@ define <vscale x 16 x half> @vfsub_vf_nxv16f16_unmasked(<vscale x 16 x half> %va
define <vscale x 32 x half> @vfsub_vv_nxv32f16(<vscale x 32 x half> %va, <vscale x 32 x half> %b, <vscale x 32 x i1> %m, i32 zeroext %evl) {
; ZVFH-LABEL: vfsub_vv_nxv32f16:
; ZVFH: # %bb.0:
-; ZVFH-NEXT: vsetvli zero, a0, e16, m8, ta, ma
-; ZVFH-NEXT: vfsub.vv v8, v8, v16, v0.t
+; ZVFH-NEXT: vsetvli a0, zero, e16, m8, ta, ma
+; ZVFH-NEXT: vfsub.vv v8, v8, v16
; ZVFH-NEXT: ret
;
; ZVFHMIN-LABEL: vfsub_vv_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: csrr a0, vlenb
+; ZVFHMIN-NEXT: slli a0, a0, 3
+; ZVFHMIN-NEXT: sub sp, sp, a0
; 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, e8, mf2, ta, ma
-; ZVFHMIN-NEXT: vmv1r.v v7, v0
-; ZVFHMIN-NEXT: csrr a2, vlenb
-; ZVFHMIN-NEXT: slli a1, a2, 1
-; ZVFHMIN-NEXT: srli a2, a2, 2
-; ZVFHMIN-NEXT: sub a3, a0, a1
-; ZVFHMIN-NEXT: vslidedown.vx v0, v0, a2
-; ZVFHMIN-NEXT: sltu a2, a0, a3
-; ZVFHMIN-NEXT: addi a2, a2, -1
-; ZVFHMIN-NEXT: and a2, a2, a3
-; ZVFHMIN-NEXT: addi a3, sp, 16
-; ZVFHMIN-NEXT: vs8r.v v16, (a3) # vscale x 64-byte Folded Spill
-; ZVFHMIN-NEXT: vsetvli zero, a2, e16, m4, 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 a0, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v16
+; ZVFHMIN-NEXT: addi a0, sp, 16
+; ZVFHMIN-NEXT: vs8r.v v24, (a0) # vscale x 64-byte Folded Spill
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v0, v8
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v20
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12
+; ZVFHMIN-NEXT: vl8r.v v8, (a0) # vscale x 64-byte Folded Reload
; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vfsub.vv v16, v16, v24, v0.t
+; ZVFHMIN-NEXT: vfsub.vv v0, v0, v8
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v16, v0.t
-; ZVFHMIN-NEXT: bltu a0, a1, .LBB44_2
-; ZVFHMIN-NEXT: # %bb.1:
-; ZVFHMIN-NEXT: mv a0, a1
-; ZVFHMIN-NEXT: .LBB44_2:
-; ZVFHMIN-NEXT: vmv1r.v v0, v7
-; ZVFHMIN-NEXT: addi a1, sp, 16
-; ZVFHMIN-NEXT: vl8r.v v24, (a1) # vscale x 64-byte Folded Reload
-; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v24, v0.t
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8, v0.t
+; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v0
; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vfsub.vv v16, v24, v16, v0.t
+; ZVFHMIN-NEXT: vfsub.vv v16, v16, v24
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16, v0.t
+; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v16
; ZVFHMIN-NEXT: csrr a0, vlenb
; ZVFHMIN-NEXT: slli a0, a0, 3
; ZVFHMIN-NEXT: add sp, sp, a0
@@ -1105,7 +975,7 @@ define <vscale x 32 x half> @vfsub_vv_nxv32f16(<vscale x 32 x half> %va, <vscale
define <vscale x 32 x half> @vfsub_vv_nxv32f16_unmasked(<vscale x 32 x half> %va, <vscale x 32 x half> %b, i32 zeroext %evl) {
; ZVFH-LABEL: vfsub_vv_nxv32f16_unmasked:
; ZVFH: # %bb.0:
-; ZVFH-NEXT: vsetvli zero, a0, e16, m8, ta, ma
+; ZVFH-NEXT: vsetvli a0, zero, e16, m8, ta, ma
; ZVFH-NEXT: vfsub.vv v8, v8, v16
; ZVFH-NEXT: ret
;
@@ -1113,43 +983,26 @@ define <vscale x 32 x half> @vfsub_vv_nxv32f16_unmasked(<vscale x 32 x half> %va
; 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: csrr a0, vlenb
+; ZVFHMIN-NEXT: slli a0, a0, 3
+; ZVFHMIN-NEXT: sub sp, sp, a0
; 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: csrr a2, vlenb
-; ZVFHMIN-NEXT: vsetvli a1, zero, e8, m4, ta, ma
-; ZVFHMIN-NEXT: vmset.m v24
-; ZVFHMIN-NEXT: slli a1, a2, 1
-; ZVFHMIN-NEXT: srli a2, a2, 2
-; ZVFHMIN-NEXT: sub a3, a0, a1
-; ZVFHMIN-NEXT: vsetvli a4, zero, e8, mf2, ta, ma
-; ZVFHMIN-NEXT: vslidedown.vx v0, v24, a2
-; ZVFHMIN-NEXT: sltu a2, a0, a3
-; ZVFHMIN-NEXT: addi a2, a2, -1
-; ZVFHMIN-NEXT: and a2, a2, a3
-; ZVFHMIN-NEXT: addi a3, sp, 16
-; ZVFHMIN-NEXT: vs8r.v v16, (a3) # vscale x 64-byte Folded Spill
-; ZVFHMIN-NEXT: vsetvli zero, a2, e16, m4, 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 a0, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v16
+; ZVFHMIN-NEXT: addi a0, sp, 16
+; ZVFHMIN-NEXT: vs8r.v v24, (a0) # vscale x 64-byte Folded Spill
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v0, v8
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v20
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12
+; ZVFHMIN-NEXT: vl8r.v v8, (a0) # vscale x 64-byte Folded Reload
; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vfsub.vv v16, v16, v24, v0.t
+; ZVFHMIN-NEXT: vfsub.vv v0, v0, v8
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v16, v0.t
-; ZVFHMIN-NEXT: bltu a0, a1, .LBB45_2
-; ZVFHMIN-NEXT: # %bb.1:
-; ZVFHMIN-NEXT: mv a0, a1
-; ZVFHMIN-NEXT: .LBB45_2:
-; ZVFHMIN-NEXT: addi a1, sp, 16
-; ZVFHMIN-NEXT: vl8r.v v24, (a1) # vscale x 64-byte Folded Reload
-; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v24
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v8
+; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v0
; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vfsub.vv v16, v24, v16
+; ZVFHMIN-NEXT: vfsub.vv v16, v16, v24
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
+; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v16
; ZVFHMIN-NEXT: csrr a0, vlenb
; ZVFHMIN-NEXT: slli a0, a0, 3
; ZVFHMIN-NEXT: add sp, sp, a0
@@ -1164,66 +1017,41 @@ define <vscale x 32 x half> @vfsub_vv_nxv32f16_unmasked(<vscale x 32 x half> %va
define <vscale x 32 x half> @vfsub_vf_nxv32f16(<vscale x 32 x half> %va, half %b, <vscale x 32 x i1> %m, i32 zeroext %evl) {
; ZVFH-LABEL: vfsub_vf_nxv32f16:
; ZVFH: # %bb.0:
-; ZVFH-NEXT: vsetvli zero, a0, e16, m8, ta, ma
-; ZVFH-NEXT: vfsub.vf v8, v8, fa0, v0.t
+; ZVFH-NEXT: vsetvli a0, zero, e16, m8, ta, ma
+; ZVFH-NEXT: vfsub.vf v8, v8, fa0
; ZVFH-NEXT: ret
;
; 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, 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: vsetvli a1, zero, e16, m8, ta, ma
-; ZVFHMIN-NEXT: vmv1r.v v7, v0
-; ZVFHMIN-NEXT: fmv.x.h a1, fa0
-; ZVFHMIN-NEXT: csrr a2, vlenb
-; ZVFHMIN-NEXT: vmv.v.x v24, a1
-; ZVFHMIN-NEXT: slli a1, a2, 1
-; ZVFHMIN-NEXT: srli a2, a2, 2
-; ZVFHMIN-NEXT: sub a3, a0, a1
-; ZVFHMIN-NEXT: vsetvli a4, zero, e8, mf2, ta, ma
-; ZVFHMIN-NEXT: vslidedown.vx v0, v0, a2
-; ZVFHMIN-NEXT: sltu a2, a0, a3
-; ZVFHMIN-NEXT: addi a2, a2, -1
-; ZVFHMIN-NEXT: and a2, a2, a3
-; 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: vs8r.v v24, (a3) # vscale x 64-byte Folded Spill
-; ZVFHMIN-NEXT: vsetvli zero, a2, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v28, 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, v24, v16, v0.t
-; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v16, v0.t
-; ZVFHMIN-NEXT: bltu a0, a1, .LBB46_2
-; ZVFHMIN-NEXT: # %bb.1:
-; ZVFHMIN-NEXT: mv a0, a1
-; ZVFHMIN-NEXT: .LBB46_2:
-; ZVFHMIN-NEXT: vmv1r.v v0, v7
-; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8, v0.t
-; ZVFHMIN-NEXT: addi a0, sp, 16
-; ZVFHMIN-NEXT: vs8r.v v16, (a0) # vscale x 64-byte Folded Spill
; 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) # vscale x 64-byte Folded Reload
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v16, v0.t
+; ZVFHMIN-NEXT: sub sp, sp, a0
+; 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: fmv.x.h a0, fa0
+; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
+; ZVFHMIN-NEXT: addi a1, sp, 16
+; ZVFHMIN-NEXT: vs8r.v v16, (a1) # vscale x 64-byte Folded Spill
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v12
+; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m8, ta, ma
+; ZVFHMIN-NEXT: vmv.v.x v8, a0
+; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v0, v8
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12
; ZVFHMIN-NEXT: addi a0, sp, 16
-; ZVFHMIN-NEXT: vl8r.v v16, (a0) # vscale x 64-byte Folded Reload
+; ZVFHMIN-NEXT: vl8r.v v8, (a0) # vscale x 64-byte Folded Reload
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vfsub.vv v0, v8, v0
+; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v0
; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vfsub.vv v16, v16, v24, v0.t
+; ZVFHMIN-NEXT: vfsub.vv v16, v24, v16
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16, v0.t
+; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v16
; ZVFHMIN-NEXT: csrr a0, vlenb
-; ZVFHMIN-NEXT: slli a0, a0, 4
+; ZVFHMIN-NEXT: slli a0, a0, 3
; ZVFHMIN-NEXT: add sp, sp, a0
; ZVFHMIN-NEXT: .cfi_def_cfa sp, 16
; ZVFHMIN-NEXT: addi sp, sp, 16
@@ -1238,7 +1066,7 @@ define <vscale x 32 x half> @vfsub_vf_nxv32f16(<vscale x 32 x half> %va, half %b
define <vscale x 32 x half> @vfsub_vf_nxv32f16_unmasked(<vscale x 32 x half> %va, half %b, i32 zeroext %evl) {
; ZVFH-LABEL: vfsub_vf_nxv32f16_unmasked:
; ZVFH: # %bb.0:
-; ZVFH-NEXT: vsetvli zero, a0, e16, m8, ta, ma
+; ZVFH-NEXT: vsetvli a0, zero, e16, m8, ta, ma
; ZVFH-NEXT: vfsub.vf v8, v8, fa0
; ZVFH-NEXT: ret
;
@@ -1246,45 +1074,31 @@ define <vscale x 32 x half> @vfsub_vf_nxv32f16_unmasked(<vscale x 32 x half> %va
; 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: csrr a0, vlenb
+; ZVFHMIN-NEXT: slli a0, a0, 3
+; ZVFHMIN-NEXT: sub sp, sp, a0
; 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: fmv.x.h a1, fa0
-; ZVFHMIN-NEXT: csrr a2, vlenb
-; ZVFHMIN-NEXT: vsetvli a3, zero, e16, m8, ta, ma
-; ZVFHMIN-NEXT: vmset.m v24
-; ZVFHMIN-NEXT: vmv.v.x v16, a1
-; ZVFHMIN-NEXT: slli a1, a2, 1
-; ZVFHMIN-NEXT: srli a2, a2, 2
-; ZVFHMIN-NEXT: sub a3, a0, a1
-; ZVFHMIN-NEXT: vsetvli a4, zero, e8, mf2, ta, ma
-; ZVFHMIN-NEXT: vslidedown.vx v0, v24, a2
-; ZVFHMIN-NEXT: sltu a2, a0, a3
-; ZVFHMIN-NEXT: addi a2, a2, -1
-; ZVFHMIN-NEXT: and a2, a2, a3
-; ZVFHMIN-NEXT: addi a3, sp, 16
-; ZVFHMIN-NEXT: vs8r.v v16, (a3) # vscale x 64-byte Folded Spill
-; ZVFHMIN-NEXT: vsetvli zero, a2, e16, m4, 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 zero, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v16, v0.t
-; ZVFHMIN-NEXT: bltu a0, a1, .LBB47_2
-; ZVFHMIN-NEXT: # %bb.1:
-; ZVFHMIN-NEXT: mv a0, a1
-; ZVFHMIN-NEXT: .LBB47_2:
-; ZVFHMIN-NEXT: vsetvli zero, a0, e16, m4, ta, ma
+; ZVFHMIN-NEXT: fmv.x.h a0, fa0
+; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v8
+; ZVFHMIN-NEXT: addi a1, sp, 16
+; ZVFHMIN-NEXT: vs8r.v v16, (a1) # vscale x 64-byte Folded Spill
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v12
+; ZVFHMIN-NEXT: vsetvli a1, zero, e16, m8, ta, ma
+; ZVFHMIN-NEXT: vmv.v.x v8, a0
+; ZVFHMIN-NEXT: vsetvli a0, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v0, v8
+; ZVFHMIN-NEXT: vfwcvt.f.f.v v16, v12
; ZVFHMIN-NEXT: addi a0, sp, 16
-; ZVFHMIN-NEXT: vl8r.v v0, (a0) # vscale x 64-byte Folded Reload
-; ZVFHMIN-NEXT: vfwcvt.f.f.v v24, v0
+; ZVFHMIN-NEXT: vl8r.v v8, (a0) # vscale x 64-byte Folded Reload
; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
-; ZVFHMIN-NEXT: vfsub.vv v16, v16, v24
+; ZVFHMIN-NEXT: vfsub.vv v0, v8, v0
; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m4, ta, ma
-; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v16
+; ZVFHMIN-NEXT: vfncvt.f.f.w v8, v0
+; ZVFHMIN-NEXT: vsetvli zero, zero, e32, m8, ta, ma
+; ZVFHMIN-NEXT: vfsub.vv v16, v24, v16
+; ZVFHMIN-NEXT: vsetvli zero, zero, e16, m4, ta, ma
+; ZVFHMIN-NEXT: vfncvt.f.f.w v12, v16
; ZVFHMIN-NEXT: csrr a0, vlenb
; ZVFHMIN-NEXT: slli a0, a0, 3
; ZVFHMIN-NEXT: add sp, sp, a0
@@ -1301,8 +1115,8 @@ define <vscale x 32 x half> @vfsub_vf_nxv32f16_unmasked(<vscale x 32 x half> %va
define <vscale x 1 x float> @vfsub_vv_nxv1f32(<vscale x 1 x float> %va, <vscale x 1 x float> %b, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vv_nxv1f32:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
-; CHECK-NEXT: vfsub.vv v8, v8, v9, v0.t
+; CHECK-NEXT: vsetvli a0, zero, e32, mf2, ta, ma
+; CHECK-NEXT: vfsub.vv v8, v8, v9
; CHECK-NEXT: ret
%v = call <vscale x 1 x float> @llvm.vp.fsub.nxv1f32(<vscale x 1 x float> %va, <vscale x 1 x float> %b, <vscale x 1 x i1> %m, i32 %evl)
ret <vscale x 1 x float> %v
@@ -1311,7 +1125,7 @@ define <vscale x 1 x float> @vfsub_vv_nxv1f32(<vscale x 1 x float> %va, <vscale
define <vscale x 1 x float> @vfsub_vv_nxv1f32_unmasked(<vscale x 1 x float> %va, <vscale x 1 x float> %b, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vv_nxv1f32_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT: vsetvli a0, zero, e32, mf2, ta, ma
; CHECK-NEXT: vfsub.vv v8, v8, v9
; CHECK-NEXT: ret
%v = call <vscale x 1 x float> @llvm.vp.fsub.nxv1f32(<vscale x 1 x float> %va, <vscale x 1 x float> %b, <vscale x 1 x i1> splat (i1 true), i32 %evl)
@@ -1321,8 +1135,8 @@ define <vscale x 1 x float> @vfsub_vv_nxv1f32_unmasked(<vscale x 1 x float> %va,
define <vscale x 1 x float> @vfsub_vf_nxv1f32(<vscale x 1 x float> %va, float %b, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vf_nxv1f32:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
-; CHECK-NEXT: vfsub.vf v8, v8, fa0, v0.t
+; CHECK-NEXT: vsetvli a0, zero, e32, mf2, ta, ma
+; CHECK-NEXT: vfsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 1 x float> poison, float %b, i32 0
%vb = shufflevector <vscale x 1 x float> %elt.head, <vscale x 1 x float> poison, <vscale x 1 x i32> zeroinitializer
@@ -1333,7 +1147,7 @@ define <vscale x 1 x float> @vfsub_vf_nxv1f32(<vscale x 1 x float> %va, float %b
define <vscale x 1 x float> @vfsub_vf_nxv1f32_unmasked(<vscale x 1 x float> %va, float %b, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vf_nxv1f32_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT: vsetvli a0, zero, e32, mf2, ta, ma
; CHECK-NEXT: vfsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 1 x float> poison, float %b, i32 0
@@ -1345,8 +1159,8 @@ define <vscale x 1 x float> @vfsub_vf_nxv1f32_unmasked(<vscale x 1 x float> %va,
define <vscale x 2 x float> @vfsub_vv_nxv2f32(<vscale x 2 x float> %va, <vscale x 2 x float> %b, <vscale x 2 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vv_nxv2f32:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, m1, ta, ma
-; CHECK-NEXT: vfsub.vv v8, v8, v9, v0.t
+; CHECK-NEXT: vsetvli a0, zero, e32, m1, ta, ma
+; CHECK-NEXT: vfsub.vv v8, v8, v9
; CHECK-NEXT: ret
%v = call <vscale x 2 x float> @llvm.vp.fsub.nxv2f32(<vscale x 2 x float> %va, <vscale x 2 x float> %b, <vscale x 2 x i1> %m, i32 %evl)
ret <vscale x 2 x float> %v
@@ -1355,7 +1169,7 @@ define <vscale x 2 x float> @vfsub_vv_nxv2f32(<vscale x 2 x float> %va, <vscale
define <vscale x 2 x float> @vfsub_vv_nxv2f32_unmasked(<vscale x 2 x float> %va, <vscale x 2 x float> %b, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vv_nxv2f32_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; CHECK-NEXT: vsetvli a0, zero, e32, m1, ta, ma
; CHECK-NEXT: vfsub.vv v8, v8, v9
; CHECK-NEXT: ret
%v = call <vscale x 2 x float> @llvm.vp.fsub.nxv2f32(<vscale x 2 x float> %va, <vscale x 2 x float> %b, <vscale x 2 x i1> splat (i1 true), i32 %evl)
@@ -1365,8 +1179,8 @@ define <vscale x 2 x float> @vfsub_vv_nxv2f32_unmasked(<vscale x 2 x float> %va,
define <vscale x 2 x float> @vfsub_vf_nxv2f32(<vscale x 2 x float> %va, float %b, <vscale x 2 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vf_nxv2f32:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, m1, ta, ma
-; CHECK-NEXT: vfsub.vf v8, v8, fa0, v0.t
+; CHECK-NEXT: vsetvli a0, zero, e32, m1, ta, ma
+; CHECK-NEXT: vfsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 2 x float> poison, float %b, i32 0
%vb = shufflevector <vscale x 2 x float> %elt.head, <vscale x 2 x float> poison, <vscale x 2 x i32> zeroinitializer
@@ -1377,7 +1191,7 @@ define <vscale x 2 x float> @vfsub_vf_nxv2f32(<vscale x 2 x float> %va, float %b
define <vscale x 2 x float> @vfsub_vf_nxv2f32_unmasked(<vscale x 2 x float> %va, float %b, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vf_nxv2f32_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, m1, ta, ma
+; CHECK-NEXT: vsetvli a0, zero, e32, m1, ta, ma
; CHECK-NEXT: vfsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 2 x float> poison, float %b, i32 0
@@ -1389,8 +1203,8 @@ define <vscale x 2 x float> @vfsub_vf_nxv2f32_unmasked(<vscale x 2 x float> %va,
define <vscale x 4 x float> @vfsub_vv_nxv4f32(<vscale x 4 x float> %va, <vscale x 4 x float> %b, <vscale x 4 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vv_nxv4f32:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, m2, ta, ma
-; CHECK-NEXT: vfsub.vv v8, v8, v10, v0.t
+; CHECK-NEXT: vsetvli a0, zero, e32, m2, ta, ma
+; CHECK-NEXT: vfsub.vv v8, v8, v10
; CHECK-NEXT: ret
%v = call <vscale x 4 x float> @llvm.vp.fsub.nxv4f32(<vscale x 4 x float> %va, <vscale x 4 x float> %b, <vscale x 4 x i1> %m, i32 %evl)
ret <vscale x 4 x float> %v
@@ -1399,7 +1213,7 @@ define <vscale x 4 x float> @vfsub_vv_nxv4f32(<vscale x 4 x float> %va, <vscale
define <vscale x 4 x float> @vfsub_vv_nxv4f32_unmasked(<vscale x 4 x float> %va, <vscale x 4 x float> %b, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vv_nxv4f32_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; CHECK-NEXT: vsetvli a0, zero, e32, m2, ta, ma
; CHECK-NEXT: vfsub.vv v8, v8, v10
; CHECK-NEXT: ret
%v = call <vscale x 4 x float> @llvm.vp.fsub.nxv4f32(<vscale x 4 x float> %va, <vscale x 4 x float> %b, <vscale x 4 x i1> splat (i1 true), i32 %evl)
@@ -1409,8 +1223,8 @@ define <vscale x 4 x float> @vfsub_vv_nxv4f32_unmasked(<vscale x 4 x float> %va,
define <vscale x 4 x float> @vfsub_vf_nxv4f32(<vscale x 4 x float> %va, float %b, <vscale x 4 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vf_nxv4f32:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, m2, ta, ma
-; CHECK-NEXT: vfsub.vf v8, v8, fa0, v0.t
+; CHECK-NEXT: vsetvli a0, zero, e32, m2, ta, ma
+; CHECK-NEXT: vfsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 4 x float> poison, float %b, i32 0
%vb = shufflevector <vscale x 4 x float> %elt.head, <vscale x 4 x float> poison, <vscale x 4 x i32> zeroinitializer
@@ -1421,7 +1235,7 @@ define <vscale x 4 x float> @vfsub_vf_nxv4f32(<vscale x 4 x float> %va, float %b
define <vscale x 4 x float> @vfsub_vf_nxv4f32_unmasked(<vscale x 4 x float> %va, float %b, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vf_nxv4f32_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, m2, ta, ma
+; CHECK-NEXT: vsetvli a0, zero, e32, m2, ta, ma
; CHECK-NEXT: vfsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 4 x float> poison, float %b, i32 0
@@ -1433,8 +1247,8 @@ define <vscale x 4 x float> @vfsub_vf_nxv4f32_unmasked(<vscale x 4 x float> %va,
define <vscale x 8 x float> @vfsub_vv_nxv8f32(<vscale x 8 x float> %va, <vscale x 8 x float> %b, <vscale x 8 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vv_nxv8f32:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, m4, ta, ma
-; CHECK-NEXT: vfsub.vv v8, v8, v12, v0.t
+; CHECK-NEXT: vsetvli a0, zero, e32, m4, ta, ma
+; CHECK-NEXT: vfsub.vv v8, v8, v12
; CHECK-NEXT: ret
%v = call <vscale x 8 x float> @llvm.vp.fsub.nxv8f32(<vscale x 8 x float> %va, <vscale x 8 x float> %b, <vscale x 8 x i1> %m, i32 %evl)
ret <vscale x 8 x float> %v
@@ -1443,7 +1257,7 @@ define <vscale x 8 x float> @vfsub_vv_nxv8f32(<vscale x 8 x float> %va, <vscale
define <vscale x 8 x float> @vfsub_vv_nxv8f32_unmasked(<vscale x 8 x float> %va, <vscale x 8 x float> %b, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vv_nxv8f32_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT: vsetvli a0, zero, e32, m4, ta, ma
; CHECK-NEXT: vfsub.vv v8, v8, v12
; CHECK-NEXT: ret
%v = call <vscale x 8 x float> @llvm.vp.fsub.nxv8f32(<vscale x 8 x float> %va, <vscale x 8 x float> %b, <vscale x 8 x i1> splat (i1 true), i32 %evl)
@@ -1453,8 +1267,8 @@ define <vscale x 8 x float> @vfsub_vv_nxv8f32_unmasked(<vscale x 8 x float> %va,
define <vscale x 8 x float> @vfsub_vf_nxv8f32(<vscale x 8 x float> %va, float %b, <vscale x 8 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vf_nxv8f32:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, m4, ta, ma
-; CHECK-NEXT: vfsub.vf v8, v8, fa0, v0.t
+; CHECK-NEXT: vsetvli a0, zero, e32, m4, ta, ma
+; CHECK-NEXT: vfsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 8 x float> poison, float %b, i32 0
%vb = shufflevector <vscale x 8 x float> %elt.head, <vscale x 8 x float> poison, <vscale x 8 x i32> zeroinitializer
@@ -1465,7 +1279,7 @@ define <vscale x 8 x float> @vfsub_vf_nxv8f32(<vscale x 8 x float> %va, float %b
define <vscale x 8 x float> @vfsub_vf_nxv8f32_unmasked(<vscale x 8 x float> %va, float %b, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vf_nxv8f32_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, m4, ta, ma
+; CHECK-NEXT: vsetvli a0, zero, e32, m4, ta, ma
; CHECK-NEXT: vfsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 8 x float> poison, float %b, i32 0
@@ -1477,8 +1291,8 @@ define <vscale x 8 x float> @vfsub_vf_nxv8f32_unmasked(<vscale x 8 x float> %va,
define <vscale x 16 x float> @vfsub_vv_nxv16f32(<vscale x 16 x float> %va, <vscale x 16 x float> %b, <vscale x 16 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vv_nxv16f32:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, m8, ta, ma
-; CHECK-NEXT: vfsub.vv v8, v8, v16, v0.t
+; CHECK-NEXT: vsetvli a0, zero, e32, m8, ta, ma
+; CHECK-NEXT: vfsub.vv v8, v8, v16
; CHECK-NEXT: ret
%v = call <vscale x 16 x float> @llvm.vp.fsub.nxv16f32(<vscale x 16 x float> %va, <vscale x 16 x float> %b, <vscale x 16 x i1> %m, i32 %evl)
ret <vscale x 16 x float> %v
@@ -1487,7 +1301,7 @@ define <vscale x 16 x float> @vfsub_vv_nxv16f32(<vscale x 16 x float> %va, <vsca
define <vscale x 16 x float> @vfsub_vv_nxv16f32_unmasked(<vscale x 16 x float> %va, <vscale x 16 x float> %b, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vv_nxv16f32_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; CHECK-NEXT: vsetvli a0, zero, e32, m8, ta, ma
; CHECK-NEXT: vfsub.vv v8, v8, v16
; CHECK-NEXT: ret
%v = call <vscale x 16 x float> @llvm.vp.fsub.nxv16f32(<vscale x 16 x float> %va, <vscale x 16 x float> %b, <vscale x 16 x i1> splat (i1 true), i32 %evl)
@@ -1497,8 +1311,8 @@ define <vscale x 16 x float> @vfsub_vv_nxv16f32_unmasked(<vscale x 16 x float> %
define <vscale x 16 x float> @vfsub_vf_nxv16f32(<vscale x 16 x float> %va, float %b, <vscale x 16 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vf_nxv16f32:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, m8, ta, ma
-; CHECK-NEXT: vfsub.vf v8, v8, fa0, v0.t
+; CHECK-NEXT: vsetvli a0, zero, e32, m8, ta, ma
+; CHECK-NEXT: vfsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 16 x float> poison, float %b, i32 0
%vb = shufflevector <vscale x 16 x float> %elt.head, <vscale x 16 x float> poison, <vscale x 16 x i32> zeroinitializer
@@ -1509,7 +1323,7 @@ define <vscale x 16 x float> @vfsub_vf_nxv16f32(<vscale x 16 x float> %va, float
define <vscale x 16 x float> @vfsub_vf_nxv16f32_unmasked(<vscale x 16 x float> %va, float %b, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vf_nxv16f32_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e32, m8, ta, ma
+; CHECK-NEXT: vsetvli a0, zero, e32, m8, ta, ma
; CHECK-NEXT: vfsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 16 x float> poison, float %b, i32 0
@@ -1521,8 +1335,8 @@ define <vscale x 16 x float> @vfsub_vf_nxv16f32_unmasked(<vscale x 16 x float> %
define <vscale x 1 x double> @vfsub_vv_nxv1f64(<vscale x 1 x double> %va, <vscale x 1 x double> %b, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vv_nxv1f64:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e64, m1, ta, ma
-; CHECK-NEXT: vfsub.vv v8, v8, v9, v0.t
+; CHECK-NEXT: vsetvli a0, zero, e64, m1, ta, ma
+; CHECK-NEXT: vfsub.vv v8, v8, v9
; CHECK-NEXT: ret
%v = call <vscale x 1 x double> @llvm.vp.fsub.nxv1f64(<vscale x 1 x double> %va, <vscale x 1 x double> %b, <vscale x 1 x i1> %m, i32 %evl)
ret <vscale x 1 x double> %v
@@ -1531,7 +1345,7 @@ define <vscale x 1 x double> @vfsub_vv_nxv1f64(<vscale x 1 x double> %va, <vscal
define <vscale x 1 x double> @vfsub_vv_nxv1f64_unmasked(<vscale x 1 x double> %va, <vscale x 1 x double> %b, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vv_nxv1f64_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e64, m1, ta, ma
+; CHECK-NEXT: vsetvli a0, zero, e64, m1, ta, ma
; CHECK-NEXT: vfsub.vv v8, v8, v9
; CHECK-NEXT: ret
%v = call <vscale x 1 x double> @llvm.vp.fsub.nxv1f64(<vscale x 1 x double> %va, <vscale x 1 x double> %b, <vscale x 1 x i1> splat (i1 true), i32 %evl)
@@ -1541,8 +1355,8 @@ define <vscale x 1 x double> @vfsub_vv_nxv1f64_unmasked(<vscale x 1 x double> %v
define <vscale x 1 x double> @vfsub_vf_nxv1f64(<vscale x 1 x double> %va, double %b, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vf_nxv1f64:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e64, m1, ta, ma
-; CHECK-NEXT: vfsub.vf v8, v8, fa0, v0.t
+; CHECK-NEXT: vsetvli a0, zero, e64, m1, ta, ma
+; CHECK-NEXT: vfsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 1 x double> poison, double %b, i32 0
%vb = shufflevector <vscale x 1 x double> %elt.head, <vscale x 1 x double> poison, <vscale x 1 x i32> zeroinitializer
@@ -1553,7 +1367,7 @@ define <vscale x 1 x double> @vfsub_vf_nxv1f64(<vscale x 1 x double> %va, double
define <vscale x 1 x double> @vfsub_vf_nxv1f64_unmasked(<vscale x 1 x double> %va, double %b, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vf_nxv1f64_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e64, m1, ta, ma
+; CHECK-NEXT: vsetvli a0, zero, e64, m1, ta, ma
; CHECK-NEXT: vfsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 1 x double> poison, double %b, i32 0
@@ -1565,8 +1379,8 @@ define <vscale x 1 x double> @vfsub_vf_nxv1f64_unmasked(<vscale x 1 x double> %v
define <vscale x 2 x double> @vfsub_vv_nxv2f64(<vscale x 2 x double> %va, <vscale x 2 x double> %b, <vscale x 2 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vv_nxv2f64:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e64, m2, ta, ma
-; CHECK-NEXT: vfsub.vv v8, v8, v10, v0.t
+; CHECK-NEXT: vsetvli a0, zero, e64, m2, ta, ma
+; CHECK-NEXT: vfsub.vv v8, v8, v10
; CHECK-NEXT: ret
%v = call <vscale x 2 x double> @llvm.vp.fsub.nxv2f64(<vscale x 2 x double> %va, <vscale x 2 x double> %b, <vscale x 2 x i1> %m, i32 %evl)
ret <vscale x 2 x double> %v
@@ -1575,7 +1389,7 @@ define <vscale x 2 x double> @vfsub_vv_nxv2f64(<vscale x 2 x double> %va, <vscal
define <vscale x 2 x double> @vfsub_vv_nxv2f64_unmasked(<vscale x 2 x double> %va, <vscale x 2 x double> %b, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vv_nxv2f64_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e64, m2, ta, ma
+; CHECK-NEXT: vsetvli a0, zero, e64, m2, ta, ma
; CHECK-NEXT: vfsub.vv v8, v8, v10
; CHECK-NEXT: ret
%v = call <vscale x 2 x double> @llvm.vp.fsub.nxv2f64(<vscale x 2 x double> %va, <vscale x 2 x double> %b, <vscale x 2 x i1> splat (i1 true), i32 %evl)
@@ -1585,8 +1399,8 @@ define <vscale x 2 x double> @vfsub_vv_nxv2f64_unmasked(<vscale x 2 x double> %v
define <vscale x 2 x double> @vfsub_vf_nxv2f64(<vscale x 2 x double> %va, double %b, <vscale x 2 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vf_nxv2f64:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e64, m2, ta, ma
-; CHECK-NEXT: vfsub.vf v8, v8, fa0, v0.t
+; CHECK-NEXT: vsetvli a0, zero, e64, m2, ta, ma
+; CHECK-NEXT: vfsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 2 x double> poison, double %b, i32 0
%vb = shufflevector <vscale x 2 x double> %elt.head, <vscale x 2 x double> poison, <vscale x 2 x i32> zeroinitializer
@@ -1597,7 +1411,7 @@ define <vscale x 2 x double> @vfsub_vf_nxv2f64(<vscale x 2 x double> %va, double
define <vscale x 2 x double> @vfsub_vf_nxv2f64_unmasked(<vscale x 2 x double> %va, double %b, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vf_nxv2f64_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e64, m2, ta, ma
+; CHECK-NEXT: vsetvli a0, zero, e64, m2, ta, ma
; CHECK-NEXT: vfsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 2 x double> poison, double %b, i32 0
@@ -1609,8 +1423,8 @@ define <vscale x 2 x double> @vfsub_vf_nxv2f64_unmasked(<vscale x 2 x double> %v
define <vscale x 4 x double> @vfsub_vv_nxv4f64(<vscale x 4 x double> %va, <vscale x 4 x double> %b, <vscale x 4 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vv_nxv4f64:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e64, m4, ta, ma
-; CHECK-NEXT: vfsub.vv v8, v8, v12, v0.t
+; CHECK-NEXT: vsetvli a0, zero, e64, m4, ta, ma
+; CHECK-NEXT: vfsub.vv v8, v8, v12
; CHECK-NEXT: ret
%v = call <vscale x 4 x double> @llvm.vp.fsub.nxv4f64(<vscale x 4 x double> %va, <vscale x 4 x double> %b, <vscale x 4 x i1> %m, i32 %evl)
ret <vscale x 4 x double> %v
@@ -1619,7 +1433,7 @@ define <vscale x 4 x double> @vfsub_vv_nxv4f64(<vscale x 4 x double> %va, <vscal
define <vscale x 4 x double> @vfsub_vv_nxv4f64_unmasked(<vscale x 4 x double> %va, <vscale x 4 x double> %b, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vv_nxv4f64_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e64, m4, ta, ma
+; CHECK-NEXT: vsetvli a0, zero, e64, m4, ta, ma
; CHECK-NEXT: vfsub.vv v8, v8, v12
; CHECK-NEXT: ret
%v = call <vscale x 4 x double> @llvm.vp.fsub.nxv4f64(<vscale x 4 x double> %va, <vscale x 4 x double> %b, <vscale x 4 x i1> splat (i1 true), i32 %evl)
@@ -1629,8 +1443,8 @@ define <vscale x 4 x double> @vfsub_vv_nxv4f64_unmasked(<vscale x 4 x double> %v
define <vscale x 4 x double> @vfsub_vf_nxv4f64(<vscale x 4 x double> %va, double %b, <vscale x 4 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vf_nxv4f64:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e64, m4, ta, ma
-; CHECK-NEXT: vfsub.vf v8, v8, fa0, v0.t
+; CHECK-NEXT: vsetvli a0, zero, e64, m4, ta, ma
+; CHECK-NEXT: vfsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 4 x double> poison, double %b, i32 0
%vb = shufflevector <vscale x 4 x double> %elt.head, <vscale x 4 x double> poison, <vscale x 4 x i32> zeroinitializer
@@ -1641,7 +1455,7 @@ define <vscale x 4 x double> @vfsub_vf_nxv4f64(<vscale x 4 x double> %va, double
define <vscale x 4 x double> @vfsub_vf_nxv4f64_unmasked(<vscale x 4 x double> %va, double %b, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vf_nxv4f64_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e64, m4, ta, ma
+; CHECK-NEXT: vsetvli a0, zero, e64, m4, ta, ma
; CHECK-NEXT: vfsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 4 x double> poison, double %b, i32 0
@@ -1653,8 +1467,8 @@ define <vscale x 4 x double> @vfsub_vf_nxv4f64_unmasked(<vscale x 4 x double> %v
define <vscale x 7 x double> @vfsub_vv_nxv7f64(<vscale x 7 x double> %va, <vscale x 7 x double> %b, <vscale x 7 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vv_nxv7f64:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e64, m8, ta, ma
-; CHECK-NEXT: vfsub.vv v8, v8, v16, v0.t
+; CHECK-NEXT: vsetvli a0, zero, e64, m8, ta, ma
+; CHECK-NEXT: vfsub.vv v8, v8, v16
; CHECK-NEXT: ret
%v = call <vscale x 7 x double> @llvm.vp.fsub.nxv7f64(<vscale x 7 x double> %va, <vscale x 7 x double> %b, <vscale x 7 x i1> %m, i32 %evl)
ret <vscale x 7 x double> %v
@@ -1663,8 +1477,8 @@ define <vscale x 7 x double> @vfsub_vv_nxv7f64(<vscale x 7 x double> %va, <vscal
define <vscale x 8 x double> @vfsub_vv_nxv8f64(<vscale x 8 x double> %va, <vscale x 8 x double> %b, <vscale x 8 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vv_nxv8f64:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e64, m8, ta, ma
-; CHECK-NEXT: vfsub.vv v8, v8, v16, v0.t
+; CHECK-NEXT: vsetvli a0, zero, e64, m8, ta, ma
+; CHECK-NEXT: vfsub.vv v8, v8, v16
; CHECK-NEXT: ret
%v = call <vscale x 8 x double> @llvm.vp.fsub.nxv8f64(<vscale x 8 x double> %va, <vscale x 8 x double> %b, <vscale x 8 x i1> %m, i32 %evl)
ret <vscale x 8 x double> %v
@@ -1673,7 +1487,7 @@ define <vscale x 8 x double> @vfsub_vv_nxv8f64(<vscale x 8 x double> %va, <vscal
define <vscale x 8 x double> @vfsub_vv_nxv8f64_unmasked(<vscale x 8 x double> %va, <vscale x 8 x double> %b, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vv_nxv8f64_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e64, m8, ta, ma
+; CHECK-NEXT: vsetvli a0, zero, e64, m8, ta, ma
; CHECK-NEXT: vfsub.vv v8, v8, v16
; CHECK-NEXT: ret
%v = call <vscale x 8 x double> @llvm.vp.fsub.nxv8f64(<vscale x 8 x double> %va, <vscale x 8 x double> %b, <vscale x 8 x i1> splat (i1 true), i32 %evl)
@@ -1683,8 +1497,8 @@ define <vscale x 8 x double> @vfsub_vv_nxv8f64_unmasked(<vscale x 8 x double> %v
define <vscale x 8 x double> @vfsub_vf_nxv8f64(<vscale x 8 x double> %va, double %b, <vscale x 8 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vf_nxv8f64:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e64, m8, ta, ma
-; CHECK-NEXT: vfsub.vf v8, v8, fa0, v0.t
+; CHECK-NEXT: vsetvli a0, zero, e64, m8, ta, ma
+; CHECK-NEXT: vfsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 8 x double> poison, double %b, i32 0
%vb = shufflevector <vscale x 8 x double> %elt.head, <vscale x 8 x double> poison, <vscale x 8 x i32> zeroinitializer
@@ -1695,7 +1509,7 @@ define <vscale x 8 x double> @vfsub_vf_nxv8f64(<vscale x 8 x double> %va, double
define <vscale x 8 x double> @vfsub_vf_nxv8f64_unmasked(<vscale x 8 x double> %va, double %b, i32 zeroext %evl) {
; CHECK-LABEL: vfsub_vf_nxv8f64_unmasked:
; CHECK: # %bb.0:
-; CHECK-NEXT: vsetvli zero, a0, e64, m8, ta, ma
+; CHECK-NEXT: vsetvli a0, zero, e64, m8, ta, ma
; CHECK-NEXT: vfsub.vf v8, v8, fa0
; CHECK-NEXT: ret
%elt.head = insertelement <vscale x 8 x double> poison, double %b, i32 0
More information about the llvm-commits
mailing list