Skip to content

[clang] Update SVE load and store intrinsics to have FP8 variants #126726

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 2 commits into from
Feb 28, 2025
Merged

[clang] Update SVE load and store intrinsics to have FP8 variants #126726

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
f41eb54
Update SVE load and store intrinsics to have FP8 variants
virginia-cangelosi Feb 11, 2025
ec2a2ac
Add multi loads and stores
virginia-cangelosi Feb 27, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
34 changes: 17 additions & 17 deletions clang/include/clang/Basic/arm_sve.td
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -19,7 +19,7 @@ include "arm_sve_sme_incl.td"
// Loads
Copy link
Contributor

@Lukacma Lukacma Feb 26, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

You missed some intrinsics in here I think:
Line 2044: MultiVecLoad class should also support fp8.
Same for MultiVecStore on line 2069


// Load one vector (scalar base)
def SVLD1 : MInst<"svld1[_{2}]", "dPc", "csilUcUsUiUlhfd", [IsLoad, VerifyRuntimeMode], MemEltTyDefault, "aarch64_sve_ld1">;
def SVLD1 : MInst<"svld1[_{2}]", "dPc", "csilUcUsUiUlhfdm", [IsLoad, VerifyRuntimeMode], MemEltTyDefault, "aarch64_sve_ld1">;
def SVLD1SB : MInst<"svld1sb_{d}", "dPS", "silUsUiUl", [IsLoad, VerifyRuntimeMode], MemEltTyInt8, "aarch64_sve_ld1">;
def SVLD1UB : MInst<"svld1ub_{d}", "dPW", "silUsUiUl", [IsLoad, IsZExtReturn, VerifyRuntimeMode], MemEltTyInt8, "aarch64_sve_ld1">;
def SVLD1SH : MInst<"svld1sh_{d}", "dPT", "ilUiUl", [IsLoad, VerifyRuntimeMode], MemEltTyInt16, "aarch64_sve_ld1">;
Expand All @@ -33,7 +33,7 @@ let SVETargetGuard = "sve,bf16", SMETargetGuard = "sme,bf16" in {
}

// Load one vector (scalar base, VL displacement)
def SVLD1_VNUM : MInst<"svld1_vnum[_{2}]", "dPcl", "csilUcUsUiUlhfd", [IsLoad, VerifyRuntimeMode], MemEltTyDefault, "aarch64_sve_ld1">;
def SVLD1_VNUM : MInst<"svld1_vnum[_{2}]", "dPcl", "csilUcUsUiUlhfdm", [IsLoad, VerifyRuntimeMode], MemEltTyDefault, "aarch64_sve_ld1">;
def SVLD1SB_VNUM : MInst<"svld1sb_vnum_{d}", "dPSl", "silUsUiUl", [IsLoad, VerifyRuntimeMode], MemEltTyInt8, "aarch64_sve_ld1">;
def SVLD1UB_VNUM : MInst<"svld1ub_vnum_{d}", "dPWl", "silUsUiUl", [IsLoad, IsZExtReturn, VerifyRuntimeMode], MemEltTyInt8, "aarch64_sve_ld1">;
def SVLD1SH_VNUM : MInst<"svld1sh_vnum_{d}", "dPTl", "ilUiUl", [IsLoad, VerifyRuntimeMode], MemEltTyInt16, "aarch64_sve_ld1">;
Expand Down Expand Up @@ -247,10 +247,10 @@ let SVETargetGuard = "sve,bf16", SMETargetGuard = InvalidMode in {
}

// Load one vector, unextended load, non-temporal (scalar base)
def SVLDNT1 : MInst<"svldnt1[_{2}]", "dPc", "csilUcUsUiUlhfd", [IsLoad, VerifyRuntimeMode], MemEltTyDefault, "aarch64_sve_ldnt1">;
def SVLDNT1 : MInst<"svldnt1[_{2}]", "dPc", "csilUcUsUiUlhfdm", [IsLoad, VerifyRuntimeMode], MemEltTyDefault, "aarch64_sve_ldnt1">;

// Load one vector, unextended load, non-temporal (scalar base, VL displacement)
def SVLDNT1_VNUM : MInst<"svldnt1_vnum[_{2}]", "dPcl", "csilUcUsUiUlhfd", [IsLoad, VerifyRuntimeMode], MemEltTyDefault, "aarch64_sve_ldnt1">;
def SVLDNT1_VNUM : MInst<"svldnt1_vnum[_{2}]", "dPcl", "csilUcUsUiUlhfdm", [IsLoad, VerifyRuntimeMode], MemEltTyDefault, "aarch64_sve_ldnt1">;

let SVETargetGuard = "sve,bf16", SMETargetGuard = "sme,bf16" in {
def SVLDNT1_BF : MInst<"svldnt1[_{2}]", "dPc", "b", [IsLoad, VerifyRuntimeMode], MemEltTyDefault, "aarch64_sve_ldnt1">;
Expand All @@ -265,7 +265,7 @@ let SVETargetGuard = "sve,bf16", SMETargetGuard = "sme,bf16" in {
}

multiclass StructLoad<string name, string proto, string i> {
def : SInst<name, proto, "csilUcUsUiUlhfd", MergeNone, i, [IsStructLoad, VerifyRuntimeMode]>;
def : SInst<name, proto, "csilUcUsUiUlhfdm", MergeNone, i, [IsStructLoad, VerifyRuntimeMode]>;
let SVETargetGuard = "sve,bf16", SMETargetGuard = "sme,bf16" in {
def: SInst<name, proto, "b", MergeNone, i, [IsStructLoad, VerifyRuntimeMode]>;
}
Expand Down Expand Up @@ -314,11 +314,11 @@ let SVETargetGuard = "sve2p1", SMETargetGuard = InvalidMode in {
def SVLD1UDQ_VNUM : MInst<"svld1udq_vnum[_{d}]", "dPcl", "lUld", [IsLoad], MemEltTyInt64, "aarch64_sve_ld1udq">;

// Load one vector (vector base + scalar offset)
def SVLD1Q_GATHER_U64BASE_OFFSET : MInst<"svld1q_gather[_{2}base]_offset_{d}", "dPgl", "cUcsUsiUilUlfhdb", [IsGatherLoad, IsByteIndexed], MemEltTyDefault, "aarch64_sve_ld1q_gather_scalar_offset">;
def SVLD1Q_GATHER_U64BASE : MInst<"svld1q_gather[_{2}base]_{d}", "dPg", "cUcsUsiUilUlfhdb", [IsGatherLoad, IsByteIndexed], MemEltTyDefault, "aarch64_sve_ld1q_gather_scalar_offset">;
def SVLD1Q_GATHER_U64BASE_OFFSET : MInst<"svld1q_gather[_{2}base]_offset_{d}", "dPgl", "cUcsUsiUilUlfhdbm", [IsGatherLoad, IsByteIndexed], MemEltTyDefault, "aarch64_sve_ld1q_gather_scalar_offset">;
def SVLD1Q_GATHER_U64BASE : MInst<"svld1q_gather[_{2}base]_{d}", "dPg", "cUcsUsiUilUlfhdbm", [IsGatherLoad, IsByteIndexed], MemEltTyDefault, "aarch64_sve_ld1q_gather_scalar_offset">;

// Load one vector (scalar base + vector offset)
def SVLD1Q_GATHER_U64OFFSET : MInst<"svld1q_gather_[{3}]offset[_{d}]", "dPcg", "cUcsUsiUilUlfhdb", [IsGatherLoad, IsByteIndexed], MemEltTyDefault, "aarch64_sve_ld1q_gather_vector_offset">;
def SVLD1Q_GATHER_U64OFFSET : MInst<"svld1q_gather_[{3}]offset[_{d}]", "dPcg", "cUcsUsiUilUlfhdbm", [IsGatherLoad, IsByteIndexed], MemEltTyDefault, "aarch64_sve_ld1q_gather_vector_offset">;

// Load N-element structure into N vectors (scalar base)
defm SVLD2Q : StructLoad<"svld2q[_{2}]", "2Pc", "aarch64_sve_ld2q_sret">;
Expand All @@ -341,7 +341,7 @@ let SVETargetGuard = "sve2p1", SMETargetGuard = InvalidMode in {
// Stores

// Store one vector (scalar base)
def SVST1 : MInst<"svst1[_{d}]", "vPpd", "csilUcUsUiUlhfd", [IsStore, VerifyRuntimeMode], MemEltTyDefault, "aarch64_sve_st1">;
def SVST1 : MInst<"svst1[_{d}]", "vPpd", "csilUcUsUiUlhfdm", [IsStore, VerifyRuntimeMode], MemEltTyDefault, "aarch64_sve_st1">;
def SVST1B_S : MInst<"svst1b[_{d}]", "vPAd", "sil", [IsStore, VerifyRuntimeMode], MemEltTyInt8, "aarch64_sve_st1">;
def SVST1B_U : MInst<"svst1b[_{d}]", "vPEd", "UsUiUl", [IsStore, VerifyRuntimeMode], MemEltTyInt8, "aarch64_sve_st1">;
def SVST1H_S : MInst<"svst1h[_{d}]", "vPBd", "il", [IsStore, VerifyRuntimeMode], MemEltTyInt16, "aarch64_sve_st1">;
Expand All @@ -350,7 +350,7 @@ def SVST1W_S : MInst<"svst1w[_{d}]", "vPCd", "l", [IsStore, Verify
def SVST1W_U : MInst<"svst1w[_{d}]", "vPGd", "Ul", [IsStore, VerifyRuntimeMode], MemEltTyInt32, "aarch64_sve_st1">;

// Store one vector (scalar base, VL displacement)
def SVST1_VNUM : MInst<"svst1_vnum[_{d}]", "vPpld", "csilUcUsUiUlhfd", [IsStore, VerifyRuntimeMode], MemEltTyDefault, "aarch64_sve_st1">;
def SVST1_VNUM : MInst<"svst1_vnum[_{d}]", "vPpld", "csilUcUsUiUlhfdm", [IsStore, VerifyRuntimeMode], MemEltTyDefault, "aarch64_sve_st1">;
def SVST1B_VNUM_S : MInst<"svst1b_vnum[_{d}]", "vPAld", "sil", [IsStore, VerifyRuntimeMode], MemEltTyInt8, "aarch64_sve_st1">;
def SVST1B_VNUM_U : MInst<"svst1b_vnum[_{d}]", "vPEld", "UsUiUl", [IsStore, VerifyRuntimeMode], MemEltTyInt8, "aarch64_sve_st1">;
def SVST1H_VNUM_S : MInst<"svst1h_vnum[_{d}]", "vPBld", "il", [IsStore, VerifyRuntimeMode], MemEltTyInt16, "aarch64_sve_st1">;
Expand Down Expand Up @@ -435,7 +435,7 @@ def SVST1W_SCATTER_INDEX_S : MInst<"svst1w_scatter[_{2}base]_index[_{d}]", "v
} // let SVETargetGuard = "sve"

multiclass StructStore<string name, string proto, string i> {
def : SInst<name, proto, "csilUcUsUiUlhfd", MergeNone, i, [IsStructStore, VerifyRuntimeMode]>;
def : SInst<name, proto, "csilUcUsUiUlhfdm", MergeNone, i, [IsStructStore, VerifyRuntimeMode]>;
let SVETargetGuard = "sve,bf16", SMETargetGuard = "sme,bf16" in {
def: SInst<name, proto, "b", MergeNone, i, [IsStructStore, VerifyRuntimeMode]>;
}
Expand All @@ -451,10 +451,10 @@ defm SVST3_VNUM : StructStore<"svst3_vnum[_{d}]", "vPpl3", "aarch64_sve_st3">;
defm SVST4_VNUM : StructStore<"svst4_vnum[_{d}]", "vPpl4", "aarch64_sve_st4">;

// Store one vector, with no truncation, non-temporal (scalar base)
def SVSTNT1 : MInst<"svstnt1[_{d}]", "vPpd", "csilUcUsUiUlhfd", [IsStore, VerifyRuntimeMode], MemEltTyDefault, "aarch64_sve_stnt1">;
def SVSTNT1 : MInst<"svstnt1[_{d}]", "vPpd", "csilUcUsUiUlhfdm", [IsStore, VerifyRuntimeMode], MemEltTyDefault, "aarch64_sve_stnt1">;

// Store one vector, with no truncation, non-temporal (scalar base, VL displacement)
def SVSTNT1_VNUM : MInst<"svstnt1_vnum[_{d}]", "vPpld", "csilUcUsUiUlhfd", [IsStore, VerifyRuntimeMode], MemEltTyDefault, "aarch64_sve_stnt1">;
def SVSTNT1_VNUM : MInst<"svstnt1_vnum[_{d}]", "vPpld", "csilUcUsUiUlhfdm", [IsStore, VerifyRuntimeMode], MemEltTyDefault, "aarch64_sve_stnt1">;

let SVETargetGuard = "sve,bf16", SMETargetGuard = "sme,bf16" in {
def SVSTNT1_BF : MInst<"svstnt1[_{d}]", "vPpd", "b", [IsStore, VerifyRuntimeMode], MemEltTyDefault, "aarch64_sve_stnt1">;
Expand All @@ -470,12 +470,12 @@ let SVETargetGuard = "sve2p1", SMETargetGuard = InvalidMode in {
def SVST1UDQ_VNUM : MInst<"svst1dq_vnum[_{d}]", "vPpld", "lUld", [IsStore], MemEltTyInt64, "aarch64_sve_st1dq">;

// Store one vector (vector base + scalar offset)
def SVST1Q_SCATTER_U64BASE_OFFSET : MInst<"svst1q_scatter[_{2}base]_offset[_{d}]", "vPgld", "cUcsUsiUilUlfhdb", [IsScatterStore, IsByteIndexed], MemEltTyDefault, "aarch64_sve_st1q_scatter_scalar_offset">;
def SVST1Q_SCATTER_U64BASE : MInst<"svst1q_scatter[_{2}base][_{d}]", "vPgd", "cUcsUsiUilUlfhdb", [IsScatterStore, IsByteIndexed], MemEltTyDefault, "aarch64_sve_st1q_scatter_scalar_offset">;
def SVST1Q_SCATTER_U64BASE_OFFSET : MInst<"svst1q_scatter[_{2}base]_offset[_{d}]", "vPgld", "cUcsUsiUilUlfhdbm", [IsScatterStore, IsByteIndexed], MemEltTyDefault, "aarch64_sve_st1q_scatter_scalar_offset">;
def SVST1Q_SCATTER_U64BASE : MInst<"svst1q_scatter[_{2}base][_{d}]", "vPgd", "cUcsUsiUilUlfhdbm", [IsScatterStore, IsByteIndexed], MemEltTyDefault, "aarch64_sve_st1q_scatter_scalar_offset">;

// Store one vector (scalar base + vector offset)
def SVST1Q_SCATTER_OFFSETS_U : MInst<"svst1q_scatter_[{3}]offset[_{d}]", "vPpgd", "cUcsUsiUilUlfhdb", [IsScatterStore, IsByteIndexed], MemEltTyDefault, "aarch64_sve_st1q_scatter_vector_offset">;
def SVST1Q_SCATTER_OFFSETS_S : MInst<"svst1q_scatter_[{3}]offset[_{d}]", "vPp#d", "cUcsUsiUilUlfhdb", [IsScatterStore, IsByteIndexed], MemEltTyDefault, "aarch64_sve_st1q_scatter_vector_offset">;
def SVST1Q_SCATTER_OFFSETS_U : MInst<"svst1q_scatter_[{3}]offset[_{d}]", "vPpgd", "cUcsUsiUilUlfhdbm", [IsScatterStore, IsByteIndexed], MemEltTyDefault, "aarch64_sve_st1q_scatter_vector_offset">;
def SVST1Q_SCATTER_OFFSETS_S : MInst<"svst1q_scatter_[{3}]offset[_{d}]", "vPp#d", "cUcsUsiUilUlfhdbm", [IsScatterStore, IsByteIndexed], MemEltTyDefault, "aarch64_sve_st1q_scatter_vector_offset">;

// Store N vectors into N-element structure (scalar base)
defm SVST2Q : StructStore<"svst2q[_{d}]", "vPc2", "aarch64_sve_st2q">;
Expand Down
5 changes: 3 additions & 2 deletions clang/lib/CodeGen/CGBuiltin.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -10199,6 +10199,7 @@ llvm::Type *CodeGenFunction::getEltType(const SVETypeFlags &TypeFlags) {
default:
llvm_unreachable("Invalid SVETypeFlag!");

case SVETypeFlags::EltTyMFloat8:
case SVETypeFlags::EltTyInt8:
return Builder.getInt8Ty();
case SVETypeFlags::EltTyInt16:
Expand Down Expand Up @@ -10627,7 +10628,7 @@ Value *CodeGenFunction::EmitSVEMaskedLoad(const CallExpr *E,
unsigned IntrinsicID,
bool IsZExtReturn) {
QualType LangPTy = E->getArg(1)->getType();
llvm::Type *MemEltTy = CGM.getTypes().ConvertType(
llvm::Type *MemEltTy = CGM.getTypes().ConvertTypeForMem(
LangPTy->castAs<PointerType>()->getPointeeType());

// The vector type that is returned may be different from the
Expand Down Expand Up @@ -10674,7 +10675,7 @@ Value *CodeGenFunction::EmitSVEMaskedStore(const CallExpr *E,
SmallVectorImpl<Value *> &Ops,
unsigned IntrinsicID) {
QualType LangPTy = E->getArg(1)->getType();
llvm::Type *MemEltTy = CGM.getTypes().ConvertType(
llvm::Type *MemEltTy = CGM.getTypes().ConvertTypeForMem(
LangPTy->castAs<PointerType>()->getPointeeType());

// The vector type that is stored may be different from the
Expand Down
3 changes: 3 additions & 0 deletions clang/lib/CodeGen/CodeGenTypes.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -107,6 +107,9 @@ llvm::Type *CodeGenTypes::ConvertTypeForMem(QualType T) {
MT->getNumRows() * MT->getNumColumns());
}

if (T->isMFloat8Type())
return llvm::Type::getIntNTy(getLLVMContext(), 8);
Copy link
Contributor

@Lukacma Lukacma Feb 26, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
return llvm::Type::getIntNTy(getLLVMContext(), 8);
return llvm::Type::getInt8Ty(getLLVMContext());


llvm::Type *R = ConvertType(T);

// Check for the boolean vector case.
Expand Down
8 changes: 4 additions & 4 deletions clang/test/CodeGen/AArch64/fp8-intrinsics/acle_sve2_fp8_fdot.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -49,8 +49,8 @@ svfloat32_t test_svdot_f32_mf8(svfloat32_t zda, svmfloat8_t zn, svmfloat8_t zm,
// CHECK-LABEL: define dso_local <vscale x 4 x float> @test_svdot_n_f32_mf8(
// CHECK-SAME: <vscale x 4 x float> [[ZDA:%.*]], <vscale x 16 x i8> [[ZN:%.*]], <1 x i8> [[ZM:%.*]], i64 noundef [[FPM:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
// CHECK-NEXT: tail call void @llvm.aarch64.set.fpmr(i64 [[FPM]])
// CHECK-NEXT: [[TMP0:%.*]] = extractelement <1 x i8> [[ZM]], i64 0
// CHECK-NEXT: tail call void @llvm.aarch64.set.fpmr(i64 [[FPM]])
// CHECK-NEXT: [[DOTSPLATINSERT:%.*]] = insertelement <vscale x 16 x i8> poison, i8 [[TMP0]], i64 0
// CHECK-NEXT: [[DOTSPLAT:%.*]] = shufflevector <vscale x 16 x i8> [[DOTSPLATINSERT]], <vscale x 16 x i8> poison, <vscale x 16 x i32> zeroinitializer
// CHECK-NEXT: [[TMP1:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.fp8.fdot.nxv4f32(<vscale x 4 x float> [[ZDA]], <vscale x 16 x i8> [[ZN]], <vscale x 16 x i8> [[DOTSPLAT]])
Expand All @@ -59,8 +59,8 @@ svfloat32_t test_svdot_f32_mf8(svfloat32_t zda, svmfloat8_t zn, svmfloat8_t zm,
// CHECK-CXX-LABEL: define dso_local <vscale x 4 x float> @_Z20test_svdot_n_f32_mf8u13__SVFloat32_tu13__SVMfloat8_tu6__mfp8m(
// CHECK-CXX-SAME: <vscale x 4 x float> [[ZDA:%.*]], <vscale x 16 x i8> [[ZN:%.*]], <1 x i8> [[ZM:%.*]], i64 noundef [[FPM:%.*]]) #[[ATTR0]] {
// CHECK-CXX-NEXT: [[ENTRY:.*:]]
// CHECK-CXX-NEXT: tail call void @llvm.aarch64.set.fpmr(i64 [[FPM]])
// CHECK-CXX-NEXT: [[TMP0:%.*]] = extractelement <1 x i8> [[ZM]], i64 0
// CHECK-CXX-NEXT: tail call void @llvm.aarch64.set.fpmr(i64 [[FPM]])
// CHECK-CXX-NEXT: [[DOTSPLATINSERT:%.*]] = insertelement <vscale x 16 x i8> poison, i8 [[TMP0]], i64 0
// CHECK-CXX-NEXT: [[DOTSPLAT:%.*]] = shufflevector <vscale x 16 x i8> [[DOTSPLATINSERT]], <vscale x 16 x i8> poison, <vscale x 16 x i32> zeroinitializer
// CHECK-CXX-NEXT: [[TMP1:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.fp8.fdot.nxv4f32(<vscale x 4 x float> [[ZDA]], <vscale x 16 x i8> [[ZN]], <vscale x 16 x i8> [[DOTSPLAT]])
Expand Down Expand Up @@ -91,8 +91,8 @@ svfloat16_t test_svdot_f16_mf8(svfloat16_t zda, svmfloat8_t zn, svmfloat8_t zm,
// CHECK-LABEL: define dso_local <vscale x 8 x half> @test_svdot_n_f16_mf8(
// CHECK-SAME: <vscale x 8 x half> [[ZDA:%.*]], <vscale x 16 x i8> [[ZN:%.*]], <1 x i8> [[ZM:%.*]], i64 noundef [[FPM:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
// CHECK-NEXT: tail call void @llvm.aarch64.set.fpmr(i64 [[FPM]])
// CHECK-NEXT: [[TMP0:%.*]] = extractelement <1 x i8> [[ZM]], i64 0
// CHECK-NEXT: tail call void @llvm.aarch64.set.fpmr(i64 [[FPM]])
// CHECK-NEXT: [[DOTSPLATINSERT:%.*]] = insertelement <vscale x 16 x i8> poison, i8 [[TMP0]], i64 0
// CHECK-NEXT: [[DOTSPLAT:%.*]] = shufflevector <vscale x 16 x i8> [[DOTSPLATINSERT]], <vscale x 16 x i8> poison, <vscale x 16 x i32> zeroinitializer
// CHECK-NEXT: [[TMP1:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.fp8.fdot.nxv8f16(<vscale x 8 x half> [[ZDA]], <vscale x 16 x i8> [[ZN]], <vscale x 16 x i8> [[DOTSPLAT]])
Expand All @@ -101,8 +101,8 @@ svfloat16_t test_svdot_f16_mf8(svfloat16_t zda, svmfloat8_t zn, svmfloat8_t zm,
// CHECK-CXX-LABEL: define dso_local <vscale x 8 x half> @_Z20test_svdot_n_f16_mf8u13__SVFloat16_tu13__SVMfloat8_tu6__mfp8m(
// CHECK-CXX-SAME: <vscale x 8 x half> [[ZDA:%.*]], <vscale x 16 x i8> [[ZN:%.*]], <1 x i8> [[ZM:%.*]], i64 noundef [[FPM:%.*]]) #[[ATTR0]] {
// CHECK-CXX-NEXT: [[ENTRY:.*:]]
// CHECK-CXX-NEXT: tail call void @llvm.aarch64.set.fpmr(i64 [[FPM]])
// CHECK-CXX-NEXT: [[TMP0:%.*]] = extractelement <1 x i8> [[ZM]], i64 0
// CHECK-CXX-NEXT: tail call void @llvm.aarch64.set.fpmr(i64 [[FPM]])
// CHECK-CXX-NEXT: [[DOTSPLATINSERT:%.*]] = insertelement <vscale x 16 x i8> poison, i8 [[TMP0]], i64 0
// CHECK-CXX-NEXT: [[DOTSPLAT:%.*]] = shufflevector <vscale x 16 x i8> [[DOTSPLATINSERT]], <vscale x 16 x i8> poison, <vscale x 16 x i32> zeroinitializer
// CHECK-CXX-NEXT: [[TMP1:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.fp8.fdot.nxv8f16(<vscale x 8 x half> [[ZDA]], <vscale x 16 x i8> [[ZN]], <vscale x 16 x i8> [[DOTSPLAT]])
Expand Down
Loading