This is an archive of the discontinued LLVM Phabricator instance.

Differential D113333

[AArch64][SVE] Add FP types to the supported SVE structure load/stores vector type list
ClosedPublic

Authored by mdchen on Nov 5 2021, 11:01 PM.

Details

Reviewers
bsmith
paulwalker-arm
peterwaller-arm
david-arm
efriedma
CarolineConcatto
Commits
rG495e258fd7aa: [AArch64][SVE] Add FP types to the supported SVE structure load/stores vector…
Summary

This adds FP type support to the SVE Container type list as a supplement to D112303.

Diff Detail

Event Timeline

mdchen created this revision.Nov 5 2021, 11:01 PM
Herald added subscribers: ctetreau, psnobl, hiraditya and 2 others. · View Herald TranscriptNov 5 2021, 11:01 PM
mdchen requested review of this revision.Nov 5 2021, 11:01 PM
Herald added a project: Restricted Project. · View Herald TranscriptNov 5 2021, 11:01 PM
Herald added a subscriber: llvm-commits. · View Herald Transcript
Harbormaster completed remote builds in B132814: Diff 385239.Nov 5 2021, 11:50 PM
paulwalker-arm added a comment.Nov 6 2021, 5:14 AM

Thanks @mdchen, these additions look sensible to me.

mdchen updated this revision to Diff 385391.Nov 7 2021, 6:55 PM
mdchen edited the summary of this revision. (Show Details)

Update test.

Harbormaster completed remote builds in B132944: Diff 385391.Nov 7 2021, 7:31 PM
mdchen added a comment.Nov 7 2021, 11:46 PM

@paulwalker-arm Thanks. Would you please take another look if the test looks ok to you now?

peterwaller-arm accepted this revision.Nov 8 2021, 1:58 AM

Thanks, LGTM.

This revision is now accepted and ready to land.Nov 8 2021, 1:58 AM
paulwalker-arm accepted this revision.Nov 8 2021, 4:47 AM
Closed by commit rG495e258fd7aa: [AArch64][SVE] Add FP types to the supported SVE structure load/stores vector… (authored by mdchen). · Explain WhyNov 8 2021, 6:55 AM
This revision was automatically updated to reflect the committed changes.
mdchen added a commit: rG495e258fd7aa: [AArch64][SVE] Add FP types to the supported SVE structure load/stores vector….

Revision Contents

PathSize
llvm/
lib/
Target/
AArch64/
12 lines
test/
Transforms/
InterleavedAccess/
AArch64/
161 lines

Diff 385391

llvm/lib/Target/AArch64/AArch64ISelLowering.cpp

  • This file is larger than 256 KB, so syntax highlighting is disabled by default.
Show First 20 Lines • Show All 12,050 Lines • ▼ Show 20 Linesbool AArch64TargetLowering::isLegalInterleavedAccessType(
} }
// Ensure the total vector size is 64 or a multiple of 128. Types larger than // Ensure the total vector size is 64 or a multiple of 128. Types larger than
// 128 will be split into multiple interleaved accesses. // 128 will be split into multiple interleaved accesses.
return VecSize == 64 || VecSize % 128 == 0; return VecSize == 64 || VecSize % 128 == 0;
} }
static ScalableVectorType *getSVEContainerIRType(FixedVectorType *VTy) { static ScalableVectorType *getSVEContainerIRType(FixedVectorType *VTy) {
if (VTy->getElementType() == Type::getDoubleTy(VTy->getContext()))
return ScalableVectorType::get(VTy->getElementType(), 2);
if (VTy->getElementType() == Type::getFloatTy(VTy->getContext()))
return ScalableVectorType::get(VTy->getElementType(), 4);
if (VTy->getElementType() == Type::getBFloatTy(VTy->getContext()))
return ScalableVectorType::get(VTy->getElementType(), 8);
if (VTy->getElementType() == Type::getHalfTy(VTy->getContext()))
return ScalableVectorType::get(VTy->getElementType(), 8);
if (VTy->getElementType() == Type::getInt64Ty(VTy->getContext())) if (VTy->getElementType() == Type::getInt64Ty(VTy->getContext()))
return ScalableVectorType::get(VTy->getElementType(), 2); return ScalableVectorType::get(VTy->getElementType(), 2);
if (VTy->getElementType() == Type::getInt32Ty(VTy->getContext())) if (VTy->getElementType() == Type::getInt32Ty(VTy->getContext()))
return ScalableVectorType::get(VTy->getElementType(), 4); return ScalableVectorType::get(VTy->getElementType(), 4);
if (VTy->getElementType() == Type::getInt16Ty(VTy->getContext())) if (VTy->getElementType() == Type::getInt16Ty(VTy->getContext()))
return ScalableVectorType::get(VTy->getElementType(), 8); return ScalableVectorType::get(VTy->getElementType(), 8);
▲ Show 20 Lines • Show All 7,266 Lines • Show Last 20 Lines

llvm/test/Transforms/InterleavedAccess/AArch64/sve-interleaved-accesses.ll

Show First 20 Lines • Show All 345 Lines • ▼ Show 20 Lines
define void @store_min_ge_type(<8 x i64>* %ptr, <4 x i64> %v0, <4 x i64> %v1) #2 { define void @store_min_ge_type(<8 x i64>* %ptr, <4 x i64> %v0, <4 x i64> %v1) #2 {
; CHECK-LABEL: @store_min_ge_type( ; CHECK-LABEL: @store_min_ge_type(
; CHECK: call <vscale x 2 x i1> @llvm.aarch64.sve.ptrue.nxv2i1(i32 4) ; CHECK: call <vscale x 2 x i1> @llvm.aarch64.sve.ptrue.nxv2i1(i32 4)
%interleaved.vec = shufflevector <4 x i64> %v0, <4 x i64> %v1, <8 x i32> <i32 0, i32 4, i32 1, i32 5, i32 2, i32 6, i32 3, i32 7> %interleaved.vec = shufflevector <4 x i64> %v0, <4 x i64> %v1, <8 x i32> <i32 0, i32 4, i32 1, i32 5, i32 2, i32 6, i32 3, i32 7>
store <8 x i64> %interleaved.vec, <8 x i64>* %ptr, align 4 store <8 x i64> %interleaved.vec, <8 x i64>* %ptr, align 4
ret void ret void
} }
define void @load_double_factor4(<16 x double>* %ptr) #0 {
; CHECK-LABEL: @load_double_factor4(
; CHECK-NEXT: [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.ptrue.nxv2i1(i32 31)
; CHECK-NEXT: [[TMP2:%.*]] = bitcast <16 x double>* [[PTR:%.*]] to double*
; CHECK-NEXT: [[LDN:%.*]] = call { <vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x double> } @llvm.aarch64.sve.ld4.sret.nxv2f64(<vscale x 2 x i1> [[TMP1]], double* [[TMP2]])
; CHECK-NEXT: [[TMP3:%.*]] = extractvalue { <vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x double> } [[LDN]], 3
; CHECK-NEXT: [[TMP4:%.*]] = call <4 x double> @llvm.experimental.vector.extract.v4f64.nxv2f64(<vscale x 2 x double> [[TMP3]], i64 0)
; CHECK-NEXT: [[TMP5:%.*]] = extractvalue { <vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x double> } [[LDN]], 2
; CHECK-NEXT: [[TMP6:%.*]] = call <4 x double> @llvm.experimental.vector.extract.v4f64.nxv2f64(<vscale x 2 x double> [[TMP5]], i64 0)
; CHECK-NEXT: [[TMP7:%.*]] = extractvalue { <vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x double> } [[LDN]], 1
; CHECK-NEXT: [[TMP8:%.*]] = call <4 x double> @llvm.experimental.vector.extract.v4f64.nxv2f64(<vscale x 2 x double> [[TMP7]], i64 0)
; CHECK-NEXT: [[TMP9:%.*]] = extractvalue { <vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x double> } [[LDN]], 0
; CHECK-NEXT: [[TMP10:%.*]] = call <4 x double> @llvm.experimental.vector.extract.v4f64.nxv2f64(<vscale x 2 x double> [[TMP9]], i64 0)
; CHECK-NEXT: ret void
;
%interleaved.vec = load <16 x double>, <16 x double>* %ptr, align 4
%v0 = shufflevector <16 x double> %interleaved.vec, <16 x double> poison, <4 x i32> <i32 0, i32 4, i32 8, i32 12>
%v1 = shufflevector <16 x double> %interleaved.vec, <16 x double> poison, <4 x i32> <i32 1, i32 5, i32 9, i32 13>
%v2 = shufflevector <16 x double> %interleaved.vec, <16 x double> poison, <4 x i32> <i32 2, i32 6, i32 10, i32 14>
%v3 = shufflevector <16 x double> %interleaved.vec, <16 x double> poison, <4 x i32> <i32 3, i32 7, i32 11, i32 15>
ret void
}
define void @load_float_factor3(<24 x float>* %ptr) #0 {
; CHECK-LABEL: @load_float_factor3(
; CHECK-NEXT: [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.ptrue.nxv4i1(i32 31)
; CHECK-NEXT: [[TMP2:%.*]] = bitcast <24 x float>* [[PTR:%.*]] to float*
; CHECK-NEXT: [[LDN:%.*]] = call { <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float> } @llvm.aarch64.sve.ld3.sret.nxv4f32(<vscale x 4 x i1> [[TMP1]], float* [[TMP2]])
; CHECK-NEXT: [[TMP3:%.*]] = extractvalue { <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float> } [[LDN]], 2
; CHECK-NEXT: [[TMP4:%.*]] = call <8 x float> @llvm.experimental.vector.extract.v8f32.nxv4f32(<vscale x 4 x float> [[TMP3]], i64 0)
; CHECK-NEXT: [[TMP5:%.*]] = extractvalue { <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float> } [[LDN]], 1
; CHECK-NEXT: [[TMP6:%.*]] = call <8 x float> @llvm.experimental.vector.extract.v8f32.nxv4f32(<vscale x 4 x float> [[TMP5]], i64 0)
; CHECK-NEXT: [[TMP7:%.*]] = extractvalue { <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float> } [[LDN]], 0
; CHECK-NEXT: [[TMP8:%.*]] = call <8 x float> @llvm.experimental.vector.extract.v8f32.nxv4f32(<vscale x 4 x float> [[TMP7]], i64 0)
; CHECK-NEXT: ret void
;
%interleaved.vec = load <24 x float>, <24 x float>* %ptr, align 4
%v0 = shufflevector <24 x float> %interleaved.vec, <24 x float> poison, <8 x i32> <i32 0, i32 3, i32 6, i32 9, i32 12, i32 15, i32 18, i32 21>
%v1 = shufflevector <24 x float> %interleaved.vec, <24 x float> poison, <8 x i32> <i32 1, i32 4, i32 7, i32 10, i32 13, i32 16, i32 19, i32 22>
%v2 = shufflevector <24 x float> %interleaved.vec, <24 x float> poison, <8 x i32> <i32 2, i32 5, i32 8, i32 11, i32 14, i32 17, i32 20, i32 23>
ret void
}
define void @load_half_factor2(<32 x half>* %ptr) #0 {
; CHECK-LABEL: @load_half_factor2(
; CHECK-NEXT: [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.ptrue.nxv8i1(i32 31)
; CHECK-NEXT: [[TMP2:%.*]] = bitcast <32 x half>* [[PTR:%.*]] to half*
; CHECK-NEXT: [[LDN:%.*]] = call { <vscale x 8 x half>, <vscale x 8 x half> } @llvm.aarch64.sve.ld2.sret.nxv8f16(<vscale x 8 x i1> [[TMP1]], half* [[TMP2]])
; CHECK-NEXT: [[TMP3:%.*]] = extractvalue { <vscale x 8 x half>, <vscale x 8 x half> } [[LDN]], 1
; CHECK-NEXT: [[TMP4:%.*]] = call <16 x half> @llvm.experimental.vector.extract.v16f16.nxv8f16(<vscale x 8 x half> [[TMP3]], i64 0)
; CHECK-NEXT: [[TMP5:%.*]] = extractvalue { <vscale x 8 x half>, <vscale x 8 x half> } [[LDN]], 0
; CHECK-NEXT: [[TMP6:%.*]] = call <16 x half> @llvm.experimental.vector.extract.v16f16.nxv8f16(<vscale x 8 x half> [[TMP5]], i64 0)
; CHECK-NEXT: ret void
;
%interleaved.vec = load <32 x half>, <32 x half>* %ptr, align 4
%v0 = shufflevector <32 x half> %interleaved.vec, <32 x half> poison, <16 x i32> <i32 0, i32 2, i32 4, i32 6, i32 8, i32 10, i32 12, i32 14, i32 16, i32 18, i32 20, i32 22, i32 24, i32 26, i32 28, i32 30>
%v1 = shufflevector <32 x half> %interleaved.vec, <32 x half> poison, <16 x i32> <i32 1, i32 3, i32 5, i32 7, i32 9, i32 11, i32 13, i32 15, i32 17, i32 19, i32 21, i32 23, i32 25, i32 27, i32 29, i32 31>
ret void
}
define void @load_bfloat_factor2(<32 x bfloat>* %ptr) #0 {
; CHECK-LABEL: @load_bfloat_factor2(
; CHECK-NEXT: [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.ptrue.nxv8i1(i32 31)
; CHECK-NEXT: [[TMP2:%.*]] = bitcast <32 x bfloat>* [[PTR:%.*]] to bfloat*
; CHECK-NEXT: [[LDN:%.*]] = call { <vscale x 8 x bfloat>, <vscale x 8 x bfloat> } @llvm.aarch64.sve.ld2.sret.nxv8bf16(<vscale x 8 x i1> [[TMP1]], bfloat* [[TMP2]])
; CHECK-NEXT: [[TMP3:%.*]] = extractvalue { <vscale x 8 x bfloat>, <vscale x 8 x bfloat> } [[LDN]], 1
; CHECK-NEXT: [[TMP4:%.*]] = call <16 x bfloat> @llvm.experimental.vector.extract.v16bf16.nxv8bf16(<vscale x 8 x bfloat> [[TMP3]], i64 0)
; CHECK-NEXT: [[TMP5:%.*]] = extractvalue { <vscale x 8 x bfloat>, <vscale x 8 x bfloat> } [[LDN]], 0
; CHECK-NEXT: [[TMP6:%.*]] = call <16 x bfloat> @llvm.experimental.vector.extract.v16bf16.nxv8bf16(<vscale x 8 x bfloat> [[TMP5]], i64 0)
; CHECK-NEXT: ret void
;
%interleaved.vec = load <32 x bfloat>, <32 x bfloat>* %ptr, align 4
%v0 = shufflevector <32 x bfloat> %interleaved.vec, <32 x bfloat> poison, <16 x i32> <i32 0, i32 2, i32 4, i32 6, i32 8, i32 10, i32 12, i32 14, i32 16, i32 18, i32 20, i32 22, i32 24, i32 26, i32 28, i32 30>
%v1 = shufflevector <32 x bfloat> %interleaved.vec, <32 x bfloat> poison, <16 x i32> <i32 1, i32 3, i32 5, i32 7, i32 9, i32 11, i32 13, i32 15, i32 17, i32 19, i32 21, i32 23, i32 25, i32 27, i32 29, i32 31>
ret void
}
define void @store_double_factor4(<16 x double>* %ptr, <4 x double> %v0, <4 x double> %v1, <4 x double> %v2, <4 x double> %v3) #0 {
; CHECK-LABEL: @store_double_factor4(
; CHECK-NEXT: [[S0:%.*]] = shufflevector <4 x double> [[V0:%.*]], <4 x double> [[V1:%.*]], <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
; CHECK-NEXT: [[S1:%.*]] = shufflevector <4 x double> [[V2:%.*]], <4 x double> [[V3:%.*]], <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
; CHECK-NEXT: [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.ptrue.nxv2i1(i32 31)
; CHECK-NEXT: [[TMP2:%.*]] = shufflevector <8 x double> [[S0]], <8 x double> [[S1]], <4 x i32> <i32 0, i32 1, i32 2, i32 3>
; CHECK-NEXT: [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.experimental.vector.insert.nxv2f64.v4f64(<vscale x 2 x double> undef, <4 x double> [[TMP2]], i64 0)
; CHECK-NEXT: [[TMP4:%.*]] = shufflevector <8 x double> [[S0]], <8 x double> [[S1]], <4 x i32> <i32 4, i32 5, i32 6, i32 7>
; CHECK-NEXT: [[TMP5:%.*]] = call <vscale x 2 x double> @llvm.experimental.vector.insert.nxv2f64.v4f64(<vscale x 2 x double> undef, <4 x double> [[TMP4]], i64 0)
; CHECK-NEXT: [[TMP6:%.*]] = shufflevector <8 x double> [[S0]], <8 x double> [[S1]], <4 x i32> <i32 8, i32 9, i32 10, i32 11>
; CHECK-NEXT: [[TMP7:%.*]] = call <vscale x 2 x double> @llvm.experimental.vector.insert.nxv2f64.v4f64(<vscale x 2 x double> undef, <4 x double> [[TMP6]], i64 0)
; CHECK-NEXT: [[TMP8:%.*]] = shufflevector <8 x double> [[S0]], <8 x double> [[S1]], <4 x i32> <i32 12, i32 13, i32 14, i32 15>
; CHECK-NEXT: [[TMP9:%.*]] = call <vscale x 2 x double> @llvm.experimental.vector.insert.nxv2f64.v4f64(<vscale x 2 x double> undef, <4 x double> [[TMP8]], i64 0)
; CHECK-NEXT: [[TMP10:%.*]] = bitcast <16 x double>* [[PTR:%.*]] to double*
; CHECK-NEXT: call void @llvm.aarch64.sve.st4.nxv2f64(<vscale x 2 x double> [[TMP3]], <vscale x 2 x double> [[TMP5]], <vscale x 2 x double> [[TMP7]], <vscale x 2 x double> [[TMP9]], <vscale x 2 x i1> [[TMP1]], double* [[TMP10]])
; CHECK-NEXT: ret void
;
%s0 = shufflevector <4 x double> %v0, <4 x double> %v1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
%s1 = shufflevector <4 x double> %v2, <4 x double> %v3, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
%interleaved.vec = shufflevector <8 x double> %s0, <8 x double> %s1, <16 x i32> <i32 0, i32 4, i32 8, i32 12, i32 1, i32 5, i32 9, i32 13, i32 2, i32 6, i32 10, i32 14, i32 3, i32 7, i32 11, i32 15>
store <16 x double> %interleaved.vec, <16 x double>* %ptr, align 4
ret void
}
define void @store_float_factor3(<24 x float>* %ptr, <8 x float> %v0, <8 x float> %v1, <8 x float> %v2) #0 {
; CHECK-LABEL: @store_float_factor3(
; CHECK-NEXT: [[S0:%.*]] = shufflevector <8 x float> [[V0:%.*]], <8 x float> [[V1:%.*]], <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
; CHECK-NEXT: [[S1:%.*]] = shufflevector <8 x float> [[V2:%.*]], <8 x float> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
; CHECK-NEXT: [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.ptrue.nxv4i1(i32 31)
; CHECK-NEXT: [[TMP2:%.*]] = shufflevector <16 x float> [[S0]], <16 x float> [[S1]], <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
; CHECK-NEXT: [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.experimental.vector.insert.nxv4f32.v8f32(<vscale x 4 x float> undef, <8 x float> [[TMP2]], i64 0)
; CHECK-NEXT: [[TMP4:%.*]] = shufflevector <16 x float> [[S0]], <16 x float> [[S1]], <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
; CHECK-NEXT: [[TMP5:%.*]] = call <vscale x 4 x float> @llvm.experimental.vector.insert.nxv4f32.v8f32(<vscale x 4 x float> undef, <8 x float> [[TMP4]], i64 0)
; CHECK-NEXT: [[TMP6:%.*]] = shufflevector <16 x float> [[S0]], <16 x float> [[S1]], <8 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23>
; CHECK-NEXT: [[TMP7:%.*]] = call <vscale x 4 x float> @llvm.experimental.vector.insert.nxv4f32.v8f32(<vscale x 4 x float> undef, <8 x float> [[TMP6]], i64 0)
; CHECK-NEXT: [[TMP8:%.*]] = bitcast <24 x float>* [[PTR:%.*]] to float*
; CHECK-NEXT: call void @llvm.aarch64.sve.st3.nxv4f32(<vscale x 4 x float> [[TMP3]], <vscale x 4 x float> [[TMP5]], <vscale x 4 x float> [[TMP7]], <vscale x 4 x i1> [[TMP1]], float* [[TMP8]])
; CHECK-NEXT: ret void
;
%s0 = shufflevector <8 x float> %v0, <8 x float> %v1, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7,
i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
%s1 = shufflevector <8 x float> %v2, <8 x float> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7,
i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
%interleaved.vec = shufflevector <16 x float> %s0, <16 x float> %s1, <24 x i32> <i32 0, i32 8, i32 16, i32 1, i32 9, i32 17, i32 2, i32 10, i32 18, i32 3, i32 11, i32 19,
i32 4, i32 12, i32 20, i32 5, i32 13, i32 21, i32 6, i32 14, i32 22, i32 7, i32 15, i32 23>
store <24 x float> %interleaved.vec, <24 x float>* %ptr, align 4
ret void
}
define void @store_half_factor2(<32 x half>* %ptr, <16 x half> %v0, <16 x half> %v1) #0 {
; CHECK-LABEL: @store_half_factor2(
; CHECK-NEXT: [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.ptrue.nxv8i1(i32 31)
; CHECK-NEXT: [[TMP2:%.*]] = shufflevector <16 x half> [[V0:%.*]], <16 x half> [[V1:%.*]], <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
; CHECK-NEXT: [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.experimental.vector.insert.nxv8f16.v16f16(<vscale x 8 x half> undef, <16 x half> [[TMP2]], i64 0)
; CHECK-NEXT: [[TMP4:%.*]] = shufflevector <16 x half> [[V0]], <16 x half> [[V1]], <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
; CHECK-NEXT: [[TMP5:%.*]] = call <vscale x 8 x half> @llvm.experimental.vector.insert.nxv8f16.v16f16(<vscale x 8 x half> undef, <16 x half> [[TMP4]], i64 0)
; CHECK-NEXT: [[TMP6:%.*]] = bitcast <32 x half>* [[PTR:%.*]] to half*
; CHECK-NEXT: call void @llvm.aarch64.sve.st2.nxv8f16(<vscale x 8 x half> [[TMP3]], <vscale x 8 x half> [[TMP5]], <vscale x 8 x i1> [[TMP1]], half* [[TMP6]])
; CHECK-NEXT: ret void
;
%interleaved.vec = shufflevector <16 x half> %v0, <16 x half> %v1, <32 x i32> <i32 0, i32 16, i32 1, i32 17, i32 2, i32 18, i32 3, i32 19, i32 4, i32 20, i32 5, i32 21, i32 6, i32 22, i32 7, i32 23,
i32 8, i32 24, i32 9, i32 25, i32 10, i32 26, i32 11, i32 27, i32 12, i32 28, i32 13, i32 29, i32 14, i32 30, i32 15, i32 31>
store <32 x half> %interleaved.vec, <32 x half>* %ptr, align 4
ret void
}
define void @store_bfloat_factor2(<32 x bfloat>* %ptr, <16 x bfloat> %v0, <16 x bfloat> %v1) #0 {
; CHECK-LABEL: @store_bfloat_factor2(
; CHECK-NEXT: [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.ptrue.nxv8i1(i32 31)
; CHECK-NEXT: [[TMP2:%.*]] = shufflevector <16 x bfloat> [[V0:%.*]], <16 x bfloat> [[V1:%.*]], <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
; CHECK-NEXT: [[TMP3:%.*]] = call <vscale x 8 x bfloat> @llvm.experimental.vector.insert.nxv8bf16.v16bf16(<vscale x 8 x bfloat> undef, <16 x bfloat> [[TMP2]], i64 0)
; CHECK-NEXT: [[TMP4:%.*]] = shufflevector <16 x bfloat> [[V0]], <16 x bfloat> [[V1]], <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
; CHECK-NEXT: [[TMP5:%.*]] = call <vscale x 8 x bfloat> @llvm.experimental.vector.insert.nxv8bf16.v16bf16(<vscale x 8 x bfloat> undef, <16 x bfloat> [[TMP4]], i64 0)
; CHECK-NEXT: [[TMP6:%.*]] = bitcast <32 x bfloat>* [[PTR:%.*]] to bfloat*
; CHECK-NEXT: call void @llvm.aarch64.sve.st2.nxv8bf16(<vscale x 8 x bfloat> [[TMP3]], <vscale x 8 x bfloat> [[TMP5]], <vscale x 8 x i1> [[TMP1]], bfloat* [[TMP6]])
; CHECK-NEXT: ret void
;
%interleaved.vec = shufflevector <16 x bfloat> %v0, <16 x bfloat> %v1, <32 x i32> <i32 0, i32 16, i32 1, i32 17, i32 2, i32 18, i32 3, i32 19, i32 4, i32 20, i32 5, i32 21, i32 6, i32 22, i32 7, i32 23,
i32 8, i32 24, i32 9, i32 25, i32 10, i32 26, i32 11, i32 27, i32 12, i32 28, i32 13, i32 29, i32 14, i32 30, i32 15, i32 31>
store <32 x bfloat> %interleaved.vec, <32 x bfloat>* %ptr, align 4
ret void
}
attributes #0 = { vscale_range(2,2) "target-features"="+sve" } attributes #0 = { vscale_range(2,2) "target-features"="+sve" }
attributes #1 = { vscale_range(2,4) "target-features"="+sve" } attributes #1 = { vscale_range(2,4) "target-features"="+sve" }
attributes #2 = { vscale_range(4,4) "target-features"="+sve" } attributes #2 = { vscale_range(4,4) "target-features"="+sve" }