[llvm] [RISCV][CostModel] Add cost for fabs/fsqrt of type bf16/f16 (PR #118608)
via llvm-commits
llvm-commits at lists.llvm.org
Thu Dec 26 18:40:55 PST 2024
https://github.com/LiqinWeng updated https://github.com/llvm/llvm-project/pull/118608
>From 48dfe67e79e7888e0456a06162be49745fabc519 Mon Sep 17 00:00:00 2001
From: LiqinWeng <liqin.weng at spacemit.com>
Date: Wed, 4 Dec 2024 17:25:09 +0800
Subject: [PATCH 1/4] [RISCV][CostModel] Add cost for fabs/fsqrt of type
bf16/f16
---
.../Target/RISCV/RISCVTargetTransformInfo.cpp | 44 +++++++++++-----
.../CostModel/RISCV/fp-min-max-abs.ll | 50 ++++++++++++-------
.../Analysis/CostModel/RISCV/fp-sqrt-pow.ll | 50 ++++++++++++-------
3 files changed, 94 insertions(+), 50 deletions(-)
diff --git a/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp b/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp
index 49192bd6380223..6298e05b9888df 100644
--- a/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp
+++ b/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp
@@ -1035,21 +1035,41 @@ RISCVTTIImpl::getIntrinsicInstrCost(const IntrinsicCostAttributes &ICA,
}
break;
}
- case Intrinsic::fabs:
+ case Intrinsic::fabs: {
+ auto LT = getTypeLegalizationCost(RetTy);
+ if (ST->hasVInstructions() && LT.second.isVector()) {
+ // lui a0, 8
+ // addi a0, a0, -1
+ // vsetvli a1, zero, e16, m1, ta, ma
+ // vand.vx v8, v8, a0
+ // f16 with zvfhmin and bf16 with zvfhbmin
+ if (LT.second.getVectorElementType() == MVT::bf16 ||
+ (LT.second.getVectorElementType() == MVT::f16 &&
+ !ST->hasVInstructionsF16()))
+ return LT.first * getRISCVInstructionCost(RISCV::VAND_VX, LT.second,
+ CostKind) +
+ 2;
+ else
+ return LT.first *
+ getRISCVInstructionCost(RISCV::VFSGNJX_VV, LT.second, CostKind);
+ }
+ break;
+ }
case Intrinsic::sqrt: {
auto LT = getTypeLegalizationCost(RetTy);
- // TODO: add f16/bf16, bf16 with zvfbfmin && f16 with zvfhmin
if (ST->hasVInstructions() && LT.second.isVector()) {
- unsigned Op;
- switch (ICA.getID()) {
- case Intrinsic::fabs:
- Op = RISCV::VFSGNJX_VV;
- break;
- case Intrinsic::sqrt:
- Op = RISCV::VFSQRT_V;
- break;
- }
- return LT.first * getRISCVInstructionCost(Op, LT.second, CostKind);
+ SmallVector<unsigned, 3> Opcodes;
+ // f16 with zvfhmin and bf16 with zvfbfmin
+ if (LT.second.getVectorElementType() == MVT::bf16 &&
+ ST->hasVInstructionsBF16Minimal())
+ Opcodes = {RISCV::VFWCVTBF16_F_F_V, RISCV::VFSQRT_V,
+ RISCV::VFNCVTBF16_F_F_W};
+ else if (LT.second.getVectorElementType() == MVT::f16 &&
+ !ST->hasVInstructionsF16())
+ Opcodes = {RISCV::VFWCVT_F_F_V, RISCV::VFSQRT_V, RISCV::VFNCVT_F_F_W};
+ else
+ Opcodes = {RISCV::VFSQRT_V};
+ return LT.first * getRISCVInstructionCost(Opcodes, LT.second, CostKind);
}
break;
}
diff --git a/llvm/test/Analysis/CostModel/RISCV/fp-min-max-abs.ll b/llvm/test/Analysis/CostModel/RISCV/fp-min-max-abs.ll
index 9eb06a07f135fa..8ff70453194381 100644
--- a/llvm/test/Analysis/CostModel/RISCV/fp-min-max-abs.ll
+++ b/llvm/test/Analysis/CostModel/RISCV/fp-min-max-abs.ll
@@ -5,14 +5,14 @@
define void @fabs() {
; CHECK-LABEL: 'fabs'
; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %1 = call bfloat @llvm.fabs.bf16(bfloat undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %2 = call <2 x bfloat> @llvm.fabs.v2bf16(<2 x bfloat> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %3 = call <4 x bfloat> @llvm.fabs.v4bf16(<4 x bfloat> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %4 = call <8 x bfloat> @llvm.fabs.v8bf16(<8 x bfloat> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %5 = call <16 x bfloat> @llvm.fabs.v16bf16(<16 x bfloat> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %6 = call <vscale x 2 x bfloat> @llvm.fabs.nxv2bf16(<vscale x 2 x bfloat> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %7 = call <vscale x 4 x bfloat> @llvm.fabs.nxv4bf16(<vscale x 4 x bfloat> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %8 = call <vscale x 8 x bfloat> @llvm.fabs.nxv8bf16(<vscale x 8 x bfloat> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %9 = call <vscale x 16 x bfloat> @llvm.fabs.nxv16bf16(<vscale x 16 x bfloat> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %2 = call <2 x bfloat> @llvm.fabs.v2bf16(<2 x bfloat> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %3 = call <4 x bfloat> @llvm.fabs.v4bf16(<4 x bfloat> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %4 = call <8 x bfloat> @llvm.fabs.v8bf16(<8 x bfloat> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %5 = call <16 x bfloat> @llvm.fabs.v16bf16(<16 x bfloat> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %6 = call <vscale x 2 x bfloat> @llvm.fabs.nxv2bf16(<vscale x 2 x bfloat> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %7 = call <vscale x 4 x bfloat> @llvm.fabs.nxv4bf16(<vscale x 4 x bfloat> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %8 = call <vscale x 8 x bfloat> @llvm.fabs.nxv8bf16(<vscale x 8 x bfloat> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %9 = call <vscale x 16 x bfloat> @llvm.fabs.nxv16bf16(<vscale x 16 x bfloat> undef)
; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %10 = call float @llvm.fabs.f32(float undef)
; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %11 = call <2 x float> @llvm.fabs.v2f32(<2 x float> undef)
; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %12 = call <4 x float> @llvm.fabs.v4f32(<4 x float> undef)
@@ -66,17 +66,29 @@ define void @fabs() {
}
define void @fabs_f16() {
-; CHECK-LABEL: 'fabs_f16'
-; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %1 = call half @llvm.fabs.f16(half undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %2 = call <2 x half> @llvm.fabs.v2f16(<2 x half> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %3 = call <4 x half> @llvm.fabs.v4f16(<4 x half> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %4 = call <8 x half> @llvm.fabs.v8f16(<8 x half> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %5 = call <16 x half> @llvm.fabs.v16f16(<16 x half> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %6 = call <vscale x 2 x half> @llvm.fabs.nxv2f16(<vscale x 2 x half> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %7 = call <vscale x 4 x half> @llvm.fabs.nxv4f16(<vscale x 4 x half> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %8 = call <vscale x 8 x half> @llvm.fabs.nxv8f16(<vscale x 8 x half> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %9 = call <vscale x 16 x half> @llvm.fabs.nxv16f16(<vscale x 16 x half> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
+; ZVFH-LABEL: 'fabs_f16'
+; ZVFH-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %1 = call half @llvm.fabs.f16(half undef)
+; ZVFH-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %2 = call <2 x half> @llvm.fabs.v2f16(<2 x half> undef)
+; ZVFH-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %3 = call <4 x half> @llvm.fabs.v4f16(<4 x half> undef)
+; ZVFH-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %4 = call <8 x half> @llvm.fabs.v8f16(<8 x half> undef)
+; ZVFH-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %5 = call <16 x half> @llvm.fabs.v16f16(<16 x half> undef)
+; ZVFH-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %6 = call <vscale x 2 x half> @llvm.fabs.nxv2f16(<vscale x 2 x half> undef)
+; ZVFH-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %7 = call <vscale x 4 x half> @llvm.fabs.nxv4f16(<vscale x 4 x half> undef)
+; ZVFH-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %8 = call <vscale x 8 x half> @llvm.fabs.nxv8f16(<vscale x 8 x half> undef)
+; ZVFH-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %9 = call <vscale x 16 x half> @llvm.fabs.nxv16f16(<vscale x 16 x half> undef)
+; ZVFH-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
+;
+; ZVFHMIN-LABEL: 'fabs_f16'
+; ZVFHMIN-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %1 = call half @llvm.fabs.f16(half undef)
+; ZVFHMIN-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %2 = call <2 x half> @llvm.fabs.v2f16(<2 x half> undef)
+; ZVFHMIN-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %3 = call <4 x half> @llvm.fabs.v4f16(<4 x half> undef)
+; ZVFHMIN-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %4 = call <8 x half> @llvm.fabs.v8f16(<8 x half> undef)
+; ZVFHMIN-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %5 = call <16 x half> @llvm.fabs.v16f16(<16 x half> undef)
+; ZVFHMIN-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %6 = call <vscale x 2 x half> @llvm.fabs.nxv2f16(<vscale x 2 x half> undef)
+; ZVFHMIN-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %7 = call <vscale x 4 x half> @llvm.fabs.nxv4f16(<vscale x 4 x half> undef)
+; ZVFHMIN-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %8 = call <vscale x 8 x half> @llvm.fabs.nxv8f16(<vscale x 8 x half> undef)
+; ZVFHMIN-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %9 = call <vscale x 16 x half> @llvm.fabs.nxv16f16(<vscale x 16 x half> undef)
+; ZVFHMIN-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
call half @llvm.fabs.f16(half undef)
call <2 x half> @llvm.fabs.v2f16(<2 x half> undef)
diff --git a/llvm/test/Analysis/CostModel/RISCV/fp-sqrt-pow.ll b/llvm/test/Analysis/CostModel/RISCV/fp-sqrt-pow.ll
index 446627f6bf3c0e..d3ceb162c13364 100644
--- a/llvm/test/Analysis/CostModel/RISCV/fp-sqrt-pow.ll
+++ b/llvm/test/Analysis/CostModel/RISCV/fp-sqrt-pow.ll
@@ -5,14 +5,14 @@
define void @sqrt() {
; CHECK-LABEL: 'sqrt'
; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %1 = call bfloat @llvm.sqrt.bf16(bfloat undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %2 = call <2 x bfloat> @llvm.sqrt.v2bf16(<2 x bfloat> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %3 = call <4 x bfloat> @llvm.sqrt.v4bf16(<4 x bfloat> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %4 = call <8 x bfloat> @llvm.sqrt.v8bf16(<8 x bfloat> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %5 = call <16 x bfloat> @llvm.sqrt.v16bf16(<16 x bfloat> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %6 = call <vscale x 2 x bfloat> @llvm.sqrt.nxv2bf16(<vscale x 2 x bfloat> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %7 = call <vscale x 4 x bfloat> @llvm.sqrt.nxv4bf16(<vscale x 4 x bfloat> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %8 = call <vscale x 8 x bfloat> @llvm.sqrt.nxv8bf16(<vscale x 8 x bfloat> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %9 = call <vscale x 16 x bfloat> @llvm.sqrt.nxv16bf16(<vscale x 16 x bfloat> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %2 = call <2 x bfloat> @llvm.sqrt.v2bf16(<2 x bfloat> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %3 = call <4 x bfloat> @llvm.sqrt.v4bf16(<4 x bfloat> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %4 = call <8 x bfloat> @llvm.sqrt.v8bf16(<8 x bfloat> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %5 = call <16 x bfloat> @llvm.sqrt.v16bf16(<16 x bfloat> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %6 = call <vscale x 2 x bfloat> @llvm.sqrt.nxv2bf16(<vscale x 2 x bfloat> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %7 = call <vscale x 4 x bfloat> @llvm.sqrt.nxv4bf16(<vscale x 4 x bfloat> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %8 = call <vscale x 8 x bfloat> @llvm.sqrt.nxv8bf16(<vscale x 8 x bfloat> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %9 = call <vscale x 16 x bfloat> @llvm.sqrt.nxv16bf16(<vscale x 16 x bfloat> undef)
; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %10 = call float @llvm.sqrt.f32(float undef)
; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %11 = call <2 x float> @llvm.sqrt.v2f32(<2 x float> undef)
; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %12 = call <4 x float> @llvm.sqrt.v4f32(<4 x float> undef)
@@ -66,17 +66,29 @@ define void @sqrt() {
}
define void @sqrt_f16() {
-; CHECK-LABEL: 'sqrt_f16'
-; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %1 = call half @llvm.sqrt.f16(half undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %2 = call <2 x half> @llvm.sqrt.v2f16(<2 x half> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %3 = call <4 x half> @llvm.sqrt.v4f16(<4 x half> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %4 = call <8 x half> @llvm.sqrt.v8f16(<8 x half> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %5 = call <16 x half> @llvm.sqrt.v16f16(<16 x half> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %6 = call <vscale x 2 x half> @llvm.sqrt.nxv2f16(<vscale x 2 x half> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %7 = call <vscale x 4 x half> @llvm.sqrt.nxv4f16(<vscale x 4 x half> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %8 = call <vscale x 8 x half> @llvm.sqrt.nxv8f16(<vscale x 8 x half> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %9 = call <vscale x 16 x half> @llvm.sqrt.nxv16f16(<vscale x 16 x half> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
+; ZVFH-LABEL: 'sqrt_f16'
+; ZVFH-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %1 = call half @llvm.sqrt.f16(half undef)
+; ZVFH-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %2 = call <2 x half> @llvm.sqrt.v2f16(<2 x half> undef)
+; ZVFH-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %3 = call <4 x half> @llvm.sqrt.v4f16(<4 x half> undef)
+; ZVFH-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %4 = call <8 x half> @llvm.sqrt.v8f16(<8 x half> undef)
+; ZVFH-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %5 = call <16 x half> @llvm.sqrt.v16f16(<16 x half> undef)
+; ZVFH-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %6 = call <vscale x 2 x half> @llvm.sqrt.nxv2f16(<vscale x 2 x half> undef)
+; ZVFH-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %7 = call <vscale x 4 x half> @llvm.sqrt.nxv4f16(<vscale x 4 x half> undef)
+; ZVFH-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %8 = call <vscale x 8 x half> @llvm.sqrt.nxv8f16(<vscale x 8 x half> undef)
+; ZVFH-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %9 = call <vscale x 16 x half> @llvm.sqrt.nxv16f16(<vscale x 16 x half> undef)
+; ZVFH-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
+;
+; ZVFHMIN-LABEL: 'sqrt_f16'
+; ZVFHMIN-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %1 = call half @llvm.sqrt.f16(half undef)
+; ZVFHMIN-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %2 = call <2 x half> @llvm.sqrt.v2f16(<2 x half> undef)
+; ZVFHMIN-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %3 = call <4 x half> @llvm.sqrt.v4f16(<4 x half> undef)
+; ZVFHMIN-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %4 = call <8 x half> @llvm.sqrt.v8f16(<8 x half> undef)
+; ZVFHMIN-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %5 = call <16 x half> @llvm.sqrt.v16f16(<16 x half> undef)
+; ZVFHMIN-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %6 = call <vscale x 2 x half> @llvm.sqrt.nxv2f16(<vscale x 2 x half> undef)
+; ZVFHMIN-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %7 = call <vscale x 4 x half> @llvm.sqrt.nxv4f16(<vscale x 4 x half> undef)
+; ZVFHMIN-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %8 = call <vscale x 8 x half> @llvm.sqrt.nxv8f16(<vscale x 8 x half> undef)
+; ZVFHMIN-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %9 = call <vscale x 16 x half> @llvm.sqrt.nxv16f16(<vscale x 16 x half> undef)
+; ZVFHMIN-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
call half @llvm.sqrt.f16(half undef)
call <2 x half> @llvm.sqrt.v2f16(<2 x half> undef)
>From e682e00e31a95f1ad482a3174351be00f6f84875 Mon Sep 17 00:00:00 2001
From: LiqinWeng <liqin.weng at spacemit.com>
Date: Mon, 9 Dec 2024 10:54:03 +0800
Subject: [PATCH 2/4] Remove the condition for sqrts
---
llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp | 3 +--
1 file changed, 1 insertion(+), 2 deletions(-)
diff --git a/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp b/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp
index 6298e05b9888df..1f4df4f4b82290 100644
--- a/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp
+++ b/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp
@@ -1060,8 +1060,7 @@ RISCVTTIImpl::getIntrinsicInstrCost(const IntrinsicCostAttributes &ICA,
if (ST->hasVInstructions() && LT.second.isVector()) {
SmallVector<unsigned, 3> Opcodes;
// f16 with zvfhmin and bf16 with zvfbfmin
- if (LT.second.getVectorElementType() == MVT::bf16 &&
- ST->hasVInstructionsBF16Minimal())
+ if (LT.second.getVectorElementType() == MVT::bf16)
Opcodes = {RISCV::VFWCVTBF16_F_F_V, RISCV::VFSQRT_V,
RISCV::VFNCVTBF16_F_F_W};
else if (LT.second.getVectorElementType() == MVT::f16 &&
>From 86bd2f75eabcb1484ac79755755c6a81ea324728 Mon Sep 17 00:00:00 2001
From: LiqinWeng <liqin.weng at spacemit.com>
Date: Mon, 16 Dec 2024 19:37:08 +0800
Subject: [PATCH 3/4] Promote the bf16/f16 to fp32 to calculate the cost
---
.../Target/RISCV/RISCVTargetTransformInfo.cpp | 14 ++++++++-----
.../Analysis/CostModel/RISCV/fp-sqrt-pow.ll | 20 +++++++++----------
2 files changed, 19 insertions(+), 15 deletions(-)
diff --git a/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp b/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp
index 1f4df4f4b82290..9595e3a76e8c65 100644
--- a/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp
+++ b/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp
@@ -1057,18 +1057,22 @@ RISCVTTIImpl::getIntrinsicInstrCost(const IntrinsicCostAttributes &ICA,
}
case Intrinsic::sqrt: {
auto LT = getTypeLegalizationCost(RetTy);
+ auto NVT = LT.second;
if (ST->hasVInstructions() && LT.second.isVector()) {
SmallVector<unsigned, 3> Opcodes;
// f16 with zvfhmin and bf16 with zvfbfmin
- if (LT.second.getVectorElementType() == MVT::bf16)
+ if (LT.second.getVectorElementType() == MVT::bf16) {
Opcodes = {RISCV::VFWCVTBF16_F_F_V, RISCV::VFSQRT_V,
RISCV::VFNCVTBF16_F_F_W};
- else if (LT.second.getVectorElementType() == MVT::f16 &&
- !ST->hasVInstructionsF16())
+ NVT = TLI->getTypeToPromoteTo(ISD::FSQRT, NVT);
+ } else if (LT.second.getVectorElementType() == MVT::f16 &&
+ !ST->hasVInstructionsF16()) {
Opcodes = {RISCV::VFWCVT_F_F_V, RISCV::VFSQRT_V, RISCV::VFNCVT_F_F_W};
- else
+ NVT = TLI->getTypeToPromoteTo(ISD::FSQRT, NVT);
+ } else {
Opcodes = {RISCV::VFSQRT_V};
- return LT.first * getRISCVInstructionCost(Opcodes, LT.second, CostKind);
+ }
+ return LT.first * getRISCVInstructionCost(Opcodes, NVT, CostKind);
}
break;
}
diff --git a/llvm/test/Analysis/CostModel/RISCV/fp-sqrt-pow.ll b/llvm/test/Analysis/CostModel/RISCV/fp-sqrt-pow.ll
index d3ceb162c13364..88832c988d4d0b 100644
--- a/llvm/test/Analysis/CostModel/RISCV/fp-sqrt-pow.ll
+++ b/llvm/test/Analysis/CostModel/RISCV/fp-sqrt-pow.ll
@@ -7,12 +7,12 @@ define void @sqrt() {
; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %1 = call bfloat @llvm.sqrt.bf16(bfloat undef)
; CHECK-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %2 = call <2 x bfloat> @llvm.sqrt.v2bf16(<2 x bfloat> undef)
; CHECK-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %3 = call <4 x bfloat> @llvm.sqrt.v4bf16(<4 x bfloat> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %4 = call <8 x bfloat> @llvm.sqrt.v8bf16(<8 x bfloat> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %5 = call <16 x bfloat> @llvm.sqrt.v16bf16(<16 x bfloat> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %4 = call <8 x bfloat> @llvm.sqrt.v8bf16(<8 x bfloat> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %5 = call <16 x bfloat> @llvm.sqrt.v16bf16(<16 x bfloat> undef)
; CHECK-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %6 = call <vscale x 2 x bfloat> @llvm.sqrt.nxv2bf16(<vscale x 2 x bfloat> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %7 = call <vscale x 4 x bfloat> @llvm.sqrt.nxv4bf16(<vscale x 4 x bfloat> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %8 = call <vscale x 8 x bfloat> @llvm.sqrt.nxv8bf16(<vscale x 8 x bfloat> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %9 = call <vscale x 16 x bfloat> @llvm.sqrt.nxv16bf16(<vscale x 16 x bfloat> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %7 = call <vscale x 4 x bfloat> @llvm.sqrt.nxv4bf16(<vscale x 4 x bfloat> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %8 = call <vscale x 8 x bfloat> @llvm.sqrt.nxv8bf16(<vscale x 8 x bfloat> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 24 for instruction: %9 = call <vscale x 16 x bfloat> @llvm.sqrt.nxv16bf16(<vscale x 16 x bfloat> undef)
; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %10 = call float @llvm.sqrt.f32(float undef)
; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %11 = call <2 x float> @llvm.sqrt.v2f32(<2 x float> undef)
; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %12 = call <4 x float> @llvm.sqrt.v4f32(<4 x float> undef)
@@ -82,12 +82,12 @@ define void @sqrt_f16() {
; ZVFHMIN-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %1 = call half @llvm.sqrt.f16(half undef)
; ZVFHMIN-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %2 = call <2 x half> @llvm.sqrt.v2f16(<2 x half> undef)
; ZVFHMIN-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %3 = call <4 x half> @llvm.sqrt.v4f16(<4 x half> undef)
-; ZVFHMIN-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %4 = call <8 x half> @llvm.sqrt.v8f16(<8 x half> undef)
-; ZVFHMIN-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %5 = call <16 x half> @llvm.sqrt.v16f16(<16 x half> undef)
+; ZVFHMIN-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %4 = call <8 x half> @llvm.sqrt.v8f16(<8 x half> undef)
+; ZVFHMIN-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %5 = call <16 x half> @llvm.sqrt.v16f16(<16 x half> undef)
; ZVFHMIN-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %6 = call <vscale x 2 x half> @llvm.sqrt.nxv2f16(<vscale x 2 x half> undef)
-; ZVFHMIN-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %7 = call <vscale x 4 x half> @llvm.sqrt.nxv4f16(<vscale x 4 x half> undef)
-; ZVFHMIN-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %8 = call <vscale x 8 x half> @llvm.sqrt.nxv8f16(<vscale x 8 x half> undef)
-; ZVFHMIN-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %9 = call <vscale x 16 x half> @llvm.sqrt.nxv16f16(<vscale x 16 x half> undef)
+; ZVFHMIN-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %7 = call <vscale x 4 x half> @llvm.sqrt.nxv4f16(<vscale x 4 x half> undef)
+; ZVFHMIN-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %8 = call <vscale x 8 x half> @llvm.sqrt.nxv8f16(<vscale x 8 x half> undef)
+; ZVFHMIN-NEXT: Cost Model: Found an estimated cost of 24 for instruction: %9 = call <vscale x 16 x half> @llvm.sqrt.nxv16f16(<vscale x 16 x half> undef)
; ZVFHMIN-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
call half @llvm.sqrt.f16(half undef)
>From 1fe46551fd0af7d47f8928422b4eb4034d774315 Mon Sep 17 00:00:00 2001
From: LiqinWeng <liqin.weng at spacemit.com>
Date: Fri, 27 Dec 2024 10:36:47 +0800
Subject: [PATCH 4/4] Add cost for fsqrt of type nxv32[b]f16
---
.../Target/RISCV/RISCVTargetTransformInfo.cpp | 25 +++++--
.../Analysis/CostModel/RISCV/fp-sqrt-pow.ll | 72 +++++++++++--------
2 files changed, 60 insertions(+), 37 deletions(-)
diff --git a/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp b/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp
index 9595e3a76e8c65..8db3265627a67a 100644
--- a/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp
+++ b/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp
@@ -1060,15 +1060,28 @@ RISCVTTIImpl::getIntrinsicInstrCost(const IntrinsicCostAttributes &ICA,
auto NVT = LT.second;
if (ST->hasVInstructions() && LT.second.isVector()) {
SmallVector<unsigned, 3> Opcodes;
- // f16 with zvfhmin and bf16 with zvfbfmin
+ // f16 with zvfhmin and bf16 with zvfbfmin and the type of nxv32[b]f16
+ // will be spilt.
if (LT.second.getVectorElementType() == MVT::bf16) {
- Opcodes = {RISCV::VFWCVTBF16_F_F_V, RISCV::VFSQRT_V,
- RISCV::VFNCVTBF16_F_F_W};
- NVT = TLI->getTypeToPromoteTo(ISD::FSQRT, NVT);
+ if (LT.second == MVT::nxv32bf16) {
+ Opcodes = {RISCV::VFWCVTBF16_F_F_V, RISCV::VFWCVTBF16_F_F_V,
+ RISCV::VFSQRT_V, RISCV::VFSQRT_V,
+ RISCV::VFNCVTBF16_F_F_W, RISCV::VFNCVTBF16_F_F_W};
+ } else {
+ Opcodes = {RISCV::VFWCVTBF16_F_F_V, RISCV::VFSQRT_V,
+ RISCV::VFNCVTBF16_F_F_W};
+ NVT = TLI->getTypeToPromoteTo(ISD::FSQRT, NVT);
+ }
} else if (LT.second.getVectorElementType() == MVT::f16 &&
!ST->hasVInstructionsF16()) {
- Opcodes = {RISCV::VFWCVT_F_F_V, RISCV::VFSQRT_V, RISCV::VFNCVT_F_F_W};
- NVT = TLI->getTypeToPromoteTo(ISD::FSQRT, NVT);
+ if (LT.second == MVT::nxv32f16) {
+ Opcodes = {RISCV::VFWCVT_F_F_V, RISCV::VFWCVT_F_F_V,
+ RISCV::VFSQRT_V, RISCV::VFSQRT_V,
+ RISCV::VFNCVT_F_F_W, RISCV::VFNCVT_F_F_W};
+ } else {
+ Opcodes = {RISCV::VFWCVT_F_F_V, RISCV::VFSQRT_V, RISCV::VFNCVT_F_F_W};
+ NVT = TLI->getTypeToPromoteTo(ISD::FSQRT, NVT);
+ }
} else {
Opcodes = {RISCV::VFSQRT_V};
}
diff --git a/llvm/test/Analysis/CostModel/RISCV/fp-sqrt-pow.ll b/llvm/test/Analysis/CostModel/RISCV/fp-sqrt-pow.ll
index 88832c988d4d0b..73df85478e7591 100644
--- a/llvm/test/Analysis/CostModel/RISCV/fp-sqrt-pow.ll
+++ b/llvm/test/Analysis/CostModel/RISCV/fp-sqrt-pow.ll
@@ -9,29 +9,31 @@ define void @sqrt() {
; CHECK-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %3 = call <4 x bfloat> @llvm.sqrt.v4bf16(<4 x bfloat> undef)
; CHECK-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %4 = call <8 x bfloat> @llvm.sqrt.v8bf16(<8 x bfloat> undef)
; CHECK-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %5 = call <16 x bfloat> @llvm.sqrt.v16bf16(<16 x bfloat> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %6 = call <vscale x 2 x bfloat> @llvm.sqrt.nxv2bf16(<vscale x 2 x bfloat> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %7 = call <vscale x 4 x bfloat> @llvm.sqrt.nxv4bf16(<vscale x 4 x bfloat> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %8 = call <vscale x 8 x bfloat> @llvm.sqrt.nxv8bf16(<vscale x 8 x bfloat> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 24 for instruction: %9 = call <vscale x 16 x bfloat> @llvm.sqrt.nxv16bf16(<vscale x 16 x bfloat> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %10 = call float @llvm.sqrt.f32(float undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %11 = call <2 x float> @llvm.sqrt.v2f32(<2 x float> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %12 = call <4 x float> @llvm.sqrt.v4f32(<4 x float> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %13 = call <8 x float> @llvm.sqrt.v8f32(<8 x float> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %14 = call <16 x float> @llvm.sqrt.v16f32(<16 x float> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %15 = call <vscale x 1 x float> @llvm.sqrt.nxv1f32(<vscale x 1 x float> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %16 = call <vscale x 2 x float> @llvm.sqrt.nxv2f32(<vscale x 2 x float> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %17 = call <vscale x 4 x float> @llvm.sqrt.nxv4f32(<vscale x 4 x float> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %18 = call <vscale x 8 x float> @llvm.sqrt.nxv8f32(<vscale x 8 x float> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %19 = call <vscale x 16 x float> @llvm.sqrt.nxv16f32(<vscale x 16 x float> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %20 = call double @llvm.sqrt.f64(double undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %21 = call <2 x double> @llvm.sqrt.v2f64(<2 x double> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %22 = call <4 x double> @llvm.sqrt.v4f64(<4 x double> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %23 = call <8 x double> @llvm.sqrt.v8f64(<8 x double> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %24 = call <16 x double> @llvm.sqrt.v16f64(<16 x double> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %25 = call <vscale x 1 x double> @llvm.sqrt.nxv1f64(<vscale x 1 x double> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %26 = call <vscale x 2 x double> @llvm.sqrt.nxv2f64(<vscale x 2 x double> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %27 = call <vscale x 4 x double> @llvm.sqrt.nxv4f64(<vscale x 4 x double> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %28 = call <vscale x 8 x double> @llvm.sqrt.nxv8f64(<vscale x 8 x double> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 24 for instruction: %6 = call <32 x bfloat> @llvm.sqrt.v32bf16(<32 x bfloat> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %7 = call <vscale x 2 x bfloat> @llvm.sqrt.nxv2bf16(<vscale x 2 x bfloat> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %8 = call <vscale x 4 x bfloat> @llvm.sqrt.nxv4bf16(<vscale x 4 x bfloat> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %9 = call <vscale x 8 x bfloat> @llvm.sqrt.nxv8bf16(<vscale x 8 x bfloat> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 24 for instruction: %10 = call <vscale x 16 x bfloat> @llvm.sqrt.nxv16bf16(<vscale x 16 x bfloat> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 48 for instruction: %11 = call <vscale x 32 x bfloat> @llvm.sqrt.nxv32bf16(<vscale x 32 x bfloat> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %12 = call float @llvm.sqrt.f32(float undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %13 = call <2 x float> @llvm.sqrt.v2f32(<2 x float> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %14 = call <4 x float> @llvm.sqrt.v4f32(<4 x float> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %15 = call <8 x float> @llvm.sqrt.v8f32(<8 x float> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %16 = call <16 x float> @llvm.sqrt.v16f32(<16 x float> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %17 = call <vscale x 1 x float> @llvm.sqrt.nxv1f32(<vscale x 1 x float> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %18 = call <vscale x 2 x float> @llvm.sqrt.nxv2f32(<vscale x 2 x float> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %19 = call <vscale x 4 x float> @llvm.sqrt.nxv4f32(<vscale x 4 x float> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %20 = call <vscale x 8 x float> @llvm.sqrt.nxv8f32(<vscale x 8 x float> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %21 = call <vscale x 16 x float> @llvm.sqrt.nxv16f32(<vscale x 16 x float> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %22 = call double @llvm.sqrt.f64(double undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %23 = call <2 x double> @llvm.sqrt.v2f64(<2 x double> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %24 = call <4 x double> @llvm.sqrt.v4f64(<4 x double> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %25 = call <8 x double> @llvm.sqrt.v8f64(<8 x double> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %26 = call <16 x double> @llvm.sqrt.v16f64(<16 x double> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %27 = call <vscale x 1 x double> @llvm.sqrt.nxv1f64(<vscale x 1 x double> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %28 = call <vscale x 2 x double> @llvm.sqrt.nxv2f64(<vscale x 2 x double> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %29 = call <vscale x 4 x double> @llvm.sqrt.nxv4f64(<vscale x 4 x double> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %30 = call <vscale x 8 x double> @llvm.sqrt.nxv8f64(<vscale x 8 x double> undef)
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
call bfloat @llvm.sqrt.bf16(bfloat undef)
@@ -39,10 +41,12 @@ define void @sqrt() {
call <4 x bfloat> @llvm.sqrt.v4bf16(<4 x bfloat> undef)
call <8 x bfloat> @llvm.sqrt.v8bf16(<8 x bfloat> undef)
call <16 x bfloat> @llvm.sqrt.v16bf16(<16 x bfloat> undef)
+ call <32 x bfloat> @llvm.sqrt.v32bf16(<32 x bfloat> undef)
call <vscale x 2 x bfloat> @llvm.sqrt.nxv2bf16(<vscale x 2 x bfloat> undef)
call <vscale x 4 x bfloat> @llvm.sqrt.nxv4bf16(<vscale x 4 x bfloat> undef)
call <vscale x 8 x bfloat> @llvm.sqrt.nxv8bf16(<vscale x 8 x bfloat> undef)
call <vscale x 16 x bfloat> @llvm.sqrt.nxv16bf16(<vscale x 16 x bfloat> undef)
+ call <vscale x 32 x bfloat> @llvm.sqrt.nxv32bf16(<vscale x 32 x bfloat> undef)
call float @llvm.sqrt.f32(float undef)
call <2 x float> @llvm.sqrt.v2f32(<2 x float> undef)
call <4 x float> @llvm.sqrt.v4f32(<4 x float> undef)
@@ -72,10 +76,12 @@ define void @sqrt_f16() {
; ZVFH-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %3 = call <4 x half> @llvm.sqrt.v4f16(<4 x half> undef)
; ZVFH-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %4 = call <8 x half> @llvm.sqrt.v8f16(<8 x half> undef)
; ZVFH-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %5 = call <16 x half> @llvm.sqrt.v16f16(<16 x half> undef)
-; ZVFH-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %6 = call <vscale x 2 x half> @llvm.sqrt.nxv2f16(<vscale x 2 x half> undef)
-; ZVFH-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %7 = call <vscale x 4 x half> @llvm.sqrt.nxv4f16(<vscale x 4 x half> undef)
-; ZVFH-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %8 = call <vscale x 8 x half> @llvm.sqrt.nxv8f16(<vscale x 8 x half> undef)
-; ZVFH-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %9 = call <vscale x 16 x half> @llvm.sqrt.nxv16f16(<vscale x 16 x half> undef)
+; ZVFH-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %6 = call <32 x half> @llvm.sqrt.v32f16(<32 x half> undef)
+; ZVFH-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %7 = call <vscale x 2 x half> @llvm.sqrt.nxv2f16(<vscale x 2 x half> undef)
+; ZVFH-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %8 = call <vscale x 4 x half> @llvm.sqrt.nxv4f16(<vscale x 4 x half> undef)
+; ZVFH-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %9 = call <vscale x 8 x half> @llvm.sqrt.nxv8f16(<vscale x 8 x half> undef)
+; ZVFH-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %10 = call <vscale x 16 x half> @llvm.sqrt.nxv16f16(<vscale x 16 x half> undef)
+; ZVFH-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %11 = call <vscale x 32 x half> @llvm.sqrt.nxv32f16(<vscale x 32 x half> undef)
; ZVFH-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; ZVFHMIN-LABEL: 'sqrt_f16'
@@ -84,10 +90,12 @@ define void @sqrt_f16() {
; ZVFHMIN-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %3 = call <4 x half> @llvm.sqrt.v4f16(<4 x half> undef)
; ZVFHMIN-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %4 = call <8 x half> @llvm.sqrt.v8f16(<8 x half> undef)
; ZVFHMIN-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %5 = call <16 x half> @llvm.sqrt.v16f16(<16 x half> undef)
-; ZVFHMIN-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %6 = call <vscale x 2 x half> @llvm.sqrt.nxv2f16(<vscale x 2 x half> undef)
-; ZVFHMIN-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %7 = call <vscale x 4 x half> @llvm.sqrt.nxv4f16(<vscale x 4 x half> undef)
-; ZVFHMIN-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %8 = call <vscale x 8 x half> @llvm.sqrt.nxv8f16(<vscale x 8 x half> undef)
-; ZVFHMIN-NEXT: Cost Model: Found an estimated cost of 24 for instruction: %9 = call <vscale x 16 x half> @llvm.sqrt.nxv16f16(<vscale x 16 x half> undef)
+; ZVFHMIN-NEXT: Cost Model: Found an estimated cost of 24 for instruction: %6 = call <32 x half> @llvm.sqrt.v32f16(<32 x half> undef)
+; ZVFHMIN-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %7 = call <vscale x 2 x half> @llvm.sqrt.nxv2f16(<vscale x 2 x half> undef)
+; ZVFHMIN-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %8 = call <vscale x 4 x half> @llvm.sqrt.nxv4f16(<vscale x 4 x half> undef)
+; ZVFHMIN-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %9 = call <vscale x 8 x half> @llvm.sqrt.nxv8f16(<vscale x 8 x half> undef)
+; ZVFHMIN-NEXT: Cost Model: Found an estimated cost of 24 for instruction: %10 = call <vscale x 16 x half> @llvm.sqrt.nxv16f16(<vscale x 16 x half> undef)
+; ZVFHMIN-NEXT: Cost Model: Found an estimated cost of 48 for instruction: %11 = call <vscale x 32 x half> @llvm.sqrt.nxv32f16(<vscale x 32 x half> undef)
; ZVFHMIN-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
call half @llvm.sqrt.f16(half undef)
@@ -95,10 +103,12 @@ define void @sqrt_f16() {
call <4 x half> @llvm.sqrt.v4f16(<4 x half> undef)
call <8 x half> @llvm.sqrt.v8f16(<8 x half> undef)
call <16 x half> @llvm.sqrt.v16f16(<16 x half> undef)
+ call <32 x half> @llvm.sqrt.v32f16(<32 x half> undef)
call <vscale x 2 x half> @llvm.sqrt.nxv2f16(<vscale x 2 x half> undef)
call <vscale x 4 x half> @llvm.sqrt.nxv4f16(<vscale x 4 x half> undef)
call <vscale x 8 x half> @llvm.sqrt.nxv8f16(<vscale x 8 x half> undef)
call <vscale x 16 x half> @llvm.sqrt.nxv16f16(<vscale x 16 x half> undef)
+ call <vscale x 32 x half> @llvm.sqrt.nxv32f16(<vscale x 32 x half> undef)
ret void
}
More information about the llvm-commits
mailing list