[clang] Update SVE load and store intrinsics to have FP8 variants #126726
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
llvmbot
added
clang
Member
|
@llvm/pr-subscribers-clang Author: Virginia Cangelosi (virginia-cangelosi) ChangesPatch is 97.92 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/126726.diff 18 Files Affected:
diff --git a/clang/include/clang/Basic/arm_sve.td b/clang/include/clang/Basic/arm_sve.td
index b20383e72e66a3..4ffe587e044e53 100644
--- a/clang/include/clang/Basic/arm_sve.td
+++ b/clang/include/clang/Basic/arm_sve.td
@@ -19,7 +19,7 @@ include "arm_sve_sme_incl.td"
// Loads
// 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">;
@@ -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">;
@@ -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">;
@@ -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]>;
}
@@ -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">;
@@ -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">;
@@ -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">;
@@ -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]>;
}
@@ -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">;
@@ -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">;
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index 361e4c4bf2e2ed..b04becf0065694 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -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:
@@ -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
@@ -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
diff --git a/clang/lib/CodeGen/CodeGenTypes.cpp b/clang/lib/CodeGen/CodeGenTypes.cpp
index 405242e97e75cb..0244334d3716d4 100644
--- a/clang/lib/CodeGen/CodeGenTypes.cpp
+++ b/clang/lib/CodeGen/CodeGenTypes.cpp
@@ -107,6 +107,9 @@ llvm::Type *CodeGenTypes::ConvertTypeForMem(QualType T) {
MT->getNumRows() * MT->getNumColumns());
}
+ if (T->isMFloat8Type())
+ return llvm::Type::getIntNTy(getLLVMContext(), 8);
+
llvm::Type *R = ConvertType(T);
// Check for the boolean vector case.
diff --git a/clang/test/CodeGen/AArch64/fp8-intrinsics/acle_sve2_fp8_fdot.c b/clang/test/CodeGen/AArch64/fp8-intrinsics/acle_sve2_fp8_fdot.c
index 2f3994df037848..0b355db4b20734 100644
--- a/clang/test/CodeGen/AArch64/fp8-intrinsics/acle_sve2_fp8_fdot.c
+++ b/clang/test/CodeGen/AArch64/fp8-intrinsics/acle_sve2_fp8_fdot.c
@@ -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]])
@@ -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]])
@@ -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]])
@@ -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]])
diff --git a/clang/test/CodeGen/AArch64/fp8-intrinsics/acle_sve2_fp8_fmla.c b/clang/test/CodeGen/AArch64/fp8-intrinsics/acle_sve2_fp8_fmla.c
index 425e6a57ffe3ca..0daeeec9e7dd7d 100644
--- a/clang/test/CodeGen/AArch64/fp8-intrinsics/acle_sve2_fp8_fmla.c
+++ b/clang/test/CodeGen/AArch64/fp8-intrinsics/acle_sve2_fp8_fmla.c
@@ -49,8 +49,8 @@ svfloat16_t test_svmlalb_f16_mf8(svfloat16_t zda, svmfloat8_t zn, svmfloat8_t zm
// CHECK-LABEL: define dso_local <vscale x 8 x half> @test_svmlalb_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.fmlalb.nxv8f16(<vscale x 8 x half> [[ZDA]], <vscale x 16 x i8> [[ZN]], <vscale x 16 x i8> [[DOTSPLAT]])
@@ -59,8 +59,8 @@ svfloat16_t test_svmlalb_f16_mf8(svfloat16_t zda, svmfloat8_t zn, svmfloat8_t zm
// CHECK-CXX-LABEL: define dso_local <vscale x 8 x half> @_Z22test_svmlalb_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.fmlalb.nxv8f16(<vscale x 8 x half> [[ZDA]], <vscale x 16 x i8> [[ZN]], <vscale x 16 x i8> [[DOTSPLAT]])
@@ -91,8 +91,8 @@ svfloat16_t test_svmlalt_f16_mf8(svfloat16_t zda, svmfloat8_t zn, svmfloat8_t zm
// CHECK-LABEL: define dso_local <vscale x 8 x half> @test_svmlalt_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.fmlalt.nxv8f16(<vscale x 8 x half> [[ZDA]], <vscale x 16 x i...
[truncated]
|
Member
|
@llvm/pr-subscribers-clang-codegen Author: Virginia Cangelosi (virginia-cangelosi) ChangesPatch is 97.92 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/126726.diff 18 Files Affected:
diff --git a/clang/include/clang/Basic/arm_sve.td b/clang/include/clang/Basic/arm_sve.td
index b20383e72e66a3..4ffe587e044e53 100644
--- a/clang/include/clang/Basic/arm_sve.td
+++ b/clang/include/clang/Basic/arm_sve.td
@@ -19,7 +19,7 @@ include "arm_sve_sme_incl.td"
// Loads
// 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">;
@@ -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">;
@@ -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">;
@@ -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]>;
}
@@ -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">;
@@ -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">;
@@ -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">;
@@ -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]>;
}
@@ -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">;
@@ -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">;
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index 361e4c4bf2e2ed..b04becf0065694 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -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:
@@ -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
@@ -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
diff --git a/clang/lib/CodeGen/CodeGenTypes.cpp b/clang/lib/CodeGen/CodeGenTypes.cpp
index 405242e97e75cb..0244334d3716d4 100644
--- a/clang/lib/CodeGen/CodeGenTypes.cpp
+++ b/clang/lib/CodeGen/CodeGenTypes.cpp
@@ -107,6 +107,9 @@ llvm::Type *CodeGenTypes::ConvertTypeForMem(QualType T) {
MT->getNumRows() * MT->getNumColumns());
}
+ if (T->isMFloat8Type())
+ return llvm::Type::getIntNTy(getLLVMContext(), 8);
+
llvm::Type *R = ConvertType(T);
// Check for the boolean vector case.
diff --git a/clang/test/CodeGen/AArch64/fp8-intrinsics/acle_sve2_fp8_fdot.c b/clang/test/CodeGen/AArch64/fp8-intrinsics/acle_sve2_fp8_fdot.c
index 2f3994df037848..0b355db4b20734 100644
--- a/clang/test/CodeGen/AArch64/fp8-intrinsics/acle_sve2_fp8_fdot.c
+++ b/clang/test/CodeGen/AArch64/fp8-intrinsics/acle_sve2_fp8_fdot.c
@@ -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]])
@@ -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]])
@@ -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]])
@@ -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]])
diff --git a/clang/test/CodeGen/AArch64/fp8-intrinsics/acle_sve2_fp8_fmla.c b/clang/test/CodeGen/AArch64/fp8-intrinsics/acle_sve2_fp8_fmla.c
index 425e6a57ffe3ca..0daeeec9e7dd7d 100644
--- a/clang/test/CodeGen/AArch64/fp8-intrinsics/acle_sve2_fp8_fmla.c
+++ b/clang/test/CodeGen/AArch64/fp8-intrinsics/acle_sve2_fp8_fmla.c
@@ -49,8 +49,8 @@ svfloat16_t test_svmlalb_f16_mf8(svfloat16_t zda, svmfloat8_t zn, svmfloat8_t zm
// CHECK-LABEL: define dso_local <vscale x 8 x half> @test_svmlalb_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.fmlalb.nxv8f16(<vscale x 8 x half> [[ZDA]], <vscale x 16 x i8> [[ZN]], <vscale x 16 x i8> [[DOTSPLAT]])
@@ -59,8 +59,8 @@ svfloat16_t test_svmlalb_f16_mf8(svfloat16_t zda, svmfloat8_t zn, svmfloat8_t zm
// CHECK-CXX-LABEL: define dso_local <vscale x 8 x half> @_Z22test_svmlalb_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.fmlalb.nxv8f16(<vscale x 8 x half> [[ZDA]], <vscale x 16 x i8> [[ZN]], <vscale x 16 x i8> [[DOTSPLAT]])
@@ -91,8 +91,8 @@ svfloat16_t test_svmlalt_f16_mf8(svfloat16_t zda, svmfloat8_t zn, svmfloat8_t zm
// CHECK-LABEL: define dso_local <vscale x 8 x half> @test_svmlalt_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.fmlalt.nxv8f16(<vscale x 8 x half> [[ZDA]], <vscale x 16 x i...
[truncated]
|
Contributor
Author
|
Please can anyone give feedback on the effect of the changes in CGBuiltin.cpp to arm-mfp8.c |
efriedma-quic
requested review from
CarolineConcatto,
SpencerAbson and
momchil-velikov
Lukacma
reviewed
Feb 26, 2025
| include "arm_sve_sme_incl.td" | ||
|
|
||
| //////////////////////////////////////////////////////////////////////////////// | ||
| // Loads |
Contributor
There was a problem hiding this comment.
You missed some intrinsics in here I think:
Line 2044: MultiVecLoad class should also support fp8.
Same for MultiVecStore on line 2069
clang/lib/CodeGen/CodeGenTypes.cpp
Outdated
| } | ||
|
|
||
| if (T->isMFloat8Type()) | ||
| return llvm::Type::getIntNTy(getLLVMContext(), 8); |
Contributor
There was a problem hiding this comment.
Suggested change
| return llvm::Type::getIntNTy(getLLVMContext(), 8); | |
| return llvm::Type::getInt8Ty(getLLVMContext()); |
Contributor
|
I think the changes in arm-mfp8.c are fine. These unnecessary loads and stores just for the purpose of changing type, get captured by various optimization passes and disapper in the end. |
Lukacma
approved these changes
Feb 28, 2025
Collaborator
|
LLVM Buildbot has detected a new failure on builder Full details are available at: https://lab.llvm.org/buildbot/#/builders/65/builds/12965 Here is the relevant piece of the build log for the reference |
Collaborator
|
LLVM Buildbot has detected a new failure on builder Full details are available at: https://lab.llvm.org/buildbot/#/builders/190/builds/15447 Here is the relevant piece of the build log for the reference |
Collaborator
|
LLVM Buildbot has detected a new failure on builder Full details are available at: https://lab.llvm.org/buildbot/#/builders/60/builds/20721 Here is the relevant piece of the build log for the reference |
Contributor
|
I think this just needed a rebase after #125097 was landed. |
Lukacma
pushed a commit
that referenced
this pull request
Feb 28, 2025
Fix error in fp8-init-list.c introduced by PR #126726
Collaborator
|
LLVM Buildbot has detected a new failure on builder Full details are available at: https://lab.llvm.org/buildbot/#/builders/153/builds/24243 Here is the relevant piece of the build log for the reference |
cheezeburglar
pushed a commit
to cheezeburglar/llvm-project
that referenced
this pull request
Feb 28, 2025
cheezeburglar
pushed a commit
to cheezeburglar/llvm-project
that referenced
this pull request
Feb 28, 2025
Fix error in fp8-init-list.c introduced by PR llvm#126726
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
5 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
No description provided.