Skip to content

[CIR][CIRGen][Builtin][Neon] Lower neon_vaddlv_s8 and neon_vaddlv_u8 #1671

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 1 commit into from
Jun 18, 2025
Merged

[CIR][CIRGen][Builtin][Neon] Lower neon_vaddlv_s8 and neon_vaddlv_u8 #1671

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
13 changes: 9 additions & 4 deletions clang/lib/CIR/CodeGen/CIRGenBuiltinAArch64.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4397,9 +4397,6 @@ CIRGenFunction::emitAArch64BuiltinExpr(unsigned BuiltinID, const CallExpr *E,
case NEON::BI__builtin_neon_vmul_n_f64: {
llvm_unreachable("NEON::BI__builtin_neon_vmul_n_f64 NYI");
}
case NEON::BI__builtin_neon_vaddlv_u8: {
llvm_unreachable("NEON::BI__builtin_neon_vaddlv_u8 NYI");
}
case NEON::BI__builtin_neon_vaddlvq_u8: {
llvm_unreachable("NEON::BI__builtin_neon_vaddlvq_u8 NYI");
}
Expand All @@ -4413,8 +4410,16 @@ CIRGenFunction::emitAArch64BuiltinExpr(unsigned BuiltinID, const CallExpr *E,
usgn ? "aarch64.neon.uaddlv" : "aarch64.neon.saddlv",
usgn ? UInt32Ty : SInt32Ty, getLoc(E->getExprLoc()));
}
case NEON::BI__builtin_neon_vaddlv_u8:
usgn = true;
[[fallthrough]];
case NEON::BI__builtin_neon_vaddlv_s8: {
llvm_unreachable("NEON::BI__builtin_neon_vaddlv_s8 NYI");
cir::VectorType vTy = cir::VectorType::get(usgn ? UInt8Ty : SInt8Ty, 8);
Ops.push_back(emitScalarExpr(E->getArg(0)));
Ops[0] = emitNeonCall(builder, {vTy}, Ops,
usgn ? "aarch64.neon.uaddlv" : "aarch64.neon.saddlv",
usgn ? UInt32Ty : SInt32Ty, getLoc(E->getExprLoc()));
return builder.createIntCast(Ops[0], usgn ? UInt16Ty : SInt16Ty);
}
case NEON::BI__builtin_neon_vaddlv_u16:
usgn = true;
Expand Down
23 changes: 23 additions & 0 deletions clang/test/CIR/CodeGen/AArch64/neon-arith.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -905,6 +905,29 @@ int32_t test_vaddlvq_s16(int16x8_t a) {
// LLVM: ret i32 [[VADDLV_I]]
}

int16_t test_vaddlv_s8(int8x8_t a) {
return vaddlv_s8(a);

// CIR-LABEL: vaddlv_s8
// CIR: cir.llvm.intrinsic "aarch64.neon.saddlv" {{%.*}}: (!cir.vector<!s8i x 8>) -> !s32i

// LLVM: {{.*}}test_vaddlv_s8(<8 x i8>{{.*}}[[A:%.*]])
// LLVM: [[VADDLV_I:%.*]] = call i32 @llvm.aarch64.neon.saddlv.i32.v8i8(<8 x i8> [[A]])
// LLVM-NEXT: [[TMP0:%.*]] = trunc i32 [[VADDLV_I]] to i16
// LLVM-NEXT: ret i16 [[TMP0]]
}

uint16_t test_vaddlv_u8(uint8x8_t a) {
return vaddlv_u8(a);

// CIR-LABEL: vaddlv_u8
// CIR: cir.llvm.intrinsic "aarch64.neon.uaddlv" {{%.*}}: (!cir.vector<!u8i x 8>) -> !u32i

// LLVM: {{.*}}test_vaddlv_u8(<8 x i8>{{.*}}[[A:%.*]])
// LLVM: [[VADDLV_I:%.*]] = call i32 @llvm.aarch64.neon.uaddlv.i32.v8i8(<8 x i8> [[A]])
// LLVM-NEXT: [[TMP0:%.*]] = trunc i32 [[VADDLV_I]] to i16
// LLVM-NEXT: ret i16 [[TMP0]]
}

int32_t test_vaddlv_s16(int16x4_t a) {
return vaddlv_s16(a);
Expand Down
Loading