Skip to content

[AMDGPU][Clang] Support for type inferring extended image builtins for AMDGPU #164358

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 5 commits into from
Oct 30, 2025
Merged

[AMDGPU][Clang] Support for type inferring extended image builtins for AMDGPU #164358

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
b74b04b
[AMDGPU][Clang] Support for type inferring extended image builtins fo…
ranapratap55 Oct 21, 2025
901f395
Modified half 'h' to 'x' in def and adds target check tests
ranapratap55 Oct 24, 2025
d37ea02
Minor modification to format changes
ranapratap55 Oct 25, 2025
60d2b3d
Adds language sensitive mapping for float16 builtin for HIP and OpenCL
ranapratap55 Oct 29, 2025
310846d
Rename the mapping from 'x' to 'e'
ranapratap55 Oct 30, 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
41 changes: 41 additions & 0 deletions clang/include/clang/Basic/BuiltinsAMDGPU.def
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -958,6 +958,47 @@ TARGET_BUILTIN(__builtin_amdgcn_image_sample_3d_v4f32_f32, "V4fifffQtV4ibii", "n
TARGET_BUILTIN(__builtin_amdgcn_image_sample_3d_v4f16_f32, "V4hifffQtV4ibii", "nc", "image-insts")
TARGET_BUILTIN(__builtin_amdgcn_image_sample_cube_v4f32_f32, "V4fifffQtV4ibii", "nc", "image-insts")
TARGET_BUILTIN(__builtin_amdgcn_image_sample_cube_v4f16_f32, "V4hifffQtV4ibii", "nc", "image-insts")
TARGET_BUILTIN(__builtin_amdgcn_image_sample_lz_1d_v4f32_f32, "V4fifQtV4ibii", "nc", "extended-image-insts")
TARGET_BUILTIN(__builtin_amdgcn_image_sample_lz_1d_v4f16_f32, "V4hifQtV4ibii", "nc", "extended-image-insts")
TARGET_BUILTIN(__builtin_amdgcn_image_sample_lz_1darray_v4f32_f32, "V4fiffQtV4ibii", "nc", "extended-image-insts")
TARGET_BUILTIN(__builtin_amdgcn_image_sample_lz_1darray_v4f16_f32, "V4hiffQtV4ibii", "nc", "extended-image-insts")
TARGET_BUILTIN(__builtin_amdgcn_image_sample_lz_2d_f32_f32, "fiffQtV4ibii", "nc", "extended-image-insts")
TARGET_BUILTIN(__builtin_amdgcn_image_sample_lz_2d_v4f32_f32, "V4fiffQtV4ibii", "nc", "extended-image-insts")
TARGET_BUILTIN(__builtin_amdgcn_image_sample_lz_2d_v4f16_f32, "V4hiffQtV4ibii", "nc", "extended-image-insts")
TARGET_BUILTIN(__builtin_amdgcn_image_sample_lz_2darray_f32_f32, "fifffQtV4ibii", "nc", "extended-image-insts")
TARGET_BUILTIN(__builtin_amdgcn_image_sample_lz_2darray_v4f32_f32, "V4fifffQtV4ibii", "nc", "extended-image-insts")
TARGET_BUILTIN(__builtin_amdgcn_image_sample_lz_2darray_v4f16_f32, "V4hifffQtV4ibii", "nc", "extended-image-insts")
TARGET_BUILTIN(__builtin_amdgcn_image_sample_lz_3d_v4f32_f32, "V4fifffQtV4ibii", "nc", "extended-image-insts")
TARGET_BUILTIN(__builtin_amdgcn_image_sample_lz_3d_v4f16_f32, "V4hifffQtV4ibii", "nc", "extended-image-insts")
TARGET_BUILTIN(__builtin_amdgcn_image_sample_lz_cube_v4f32_f32, "V4fifffQtV4ibii", "nc", "extended-image-insts")
TARGET_BUILTIN(__builtin_amdgcn_image_sample_lz_cube_v4f16_f32, "V4hifffQtV4ibii", "nc", "extended-image-insts")
TARGET_BUILTIN(__builtin_amdgcn_image_sample_l_1d_v4f32_f32, "V4fiffQtV4ibii", "nc", "extended-image-insts")
TARGET_BUILTIN(__builtin_amdgcn_image_sample_l_1d_v4f16_f32, "V4hiffQtV4ibii", "nc", "extended-image-insts")
TARGET_BUILTIN(__builtin_amdgcn_image_sample_l_1darray_v4f32_f32, "V4fifffQtV4ibii", "nc", "extended-image-insts")
TARGET_BUILTIN(__builtin_amdgcn_image_sample_l_1darray_v4f16_f32, "V4hifffQtV4ibii", "nc", "extended-image-insts")
TARGET_BUILTIN(__builtin_amdgcn_image_sample_l_2d_f32_f32, "fifffQtV4ibii", "nc", "extended-image-insts")
TARGET_BUILTIN(__builtin_amdgcn_image_sample_l_2d_v4f32_f32, "V4fifffQtV4ibii", "nc", "extended-image-insts")
TARGET_BUILTIN(__builtin_amdgcn_image_sample_l_2d_v4f16_f32, "V4hifffQtV4ibii", "nc", "extended-image-insts")
TARGET_BUILTIN(__builtin_amdgcn_image_sample_l_2darray_f32_f32, "fiffffQtV4ibii", "nc", "extended-image-insts")
TARGET_BUILTIN(__builtin_amdgcn_image_sample_l_2darray_v4f32_f32, "V4fiffffQtV4ibii", "nc", "extended-image-insts")
TARGET_BUILTIN(__builtin_amdgcn_image_sample_l_2darray_v4f16_f32, "V4hiffffQtV4ibii", "nc", "extended-image-insts")
TARGET_BUILTIN(__builtin_amdgcn_image_sample_l_3d_v4f32_f32, "V4fiffffQtV4ibii", "nc", "extended-image-insts")
TARGET_BUILTIN(__builtin_amdgcn_image_sample_l_3d_v4f16_f32, "V4hiffffQtV4ibii", "nc", "extended-image-insts")
TARGET_BUILTIN(__builtin_amdgcn_image_sample_l_cube_v4f32_f32, "V4fiffffQtV4ibii", "nc", "extended-image-insts")
TARGET_BUILTIN(__builtin_amdgcn_image_sample_l_cube_v4f16_f32, "V4hiffffQtV4ibii", "nc", "extended-image-insts")
TARGET_BUILTIN(__builtin_amdgcn_image_sample_d_1d_v4f32_f32, "V4fifffQtV4ibii", "nc", "extended-image-insts")
TARGET_BUILTIN(__builtin_amdgcn_image_sample_d_1d_v4f16_f32, "V4hifffQtV4ibii", "nc", "extended-image-insts")
TARGET_BUILTIN(__builtin_amdgcn_image_sample_d_1darray_v4f32_f32, "V4fiffffQtV4ibii", "nc", "extended-image-insts")
TARGET_BUILTIN(__builtin_amdgcn_image_sample_d_1darray_v4f16_f32, "V4hiffffQtV4ibii", "nc", "extended-image-insts")
TARGET_BUILTIN(__builtin_amdgcn_image_sample_d_2d_f32_f32, "fiffffffQtV4ibii", "nc", "extended-image-insts")
TARGET_BUILTIN(__builtin_amdgcn_image_sample_d_2d_v4f32_f32, "V4fiffffffQtV4ibii", "nc", "extended-image-insts")
TARGET_BUILTIN(__builtin_amdgcn_image_sample_d_2d_v4f16_f32, "V4hiffffffQtV4ibii", "nc", "extended-image-insts")
TARGET_BUILTIN(__builtin_amdgcn_image_sample_d_2darray_f32_f32, "fifffffffQtV4ibii", "nc", "extended-image-insts")
TARGET_BUILTIN(__builtin_amdgcn_image_sample_d_2darray_v4f32_f32, "V4fifffffffQtV4ibii", "nc", "extended-image-insts")
TARGET_BUILTIN(__builtin_amdgcn_image_sample_d_2darray_v4f16_f32, "V4hifffffffQtV4ibii", "nc", "extended-image-insts")
TARGET_BUILTIN(__builtin_amdgcn_image_sample_d_3d_v4f32_f32, "V4fifffffffffQtV4ibii", "nc", "extended-image-insts")
TARGET_BUILTIN(__builtin_amdgcn_image_sample_d_3d_v4f16_f32, "V4hifffffffffQtV4ibii", "nc", "extended-image-insts")
TARGET_BUILTIN(__builtin_amdgcn_image_gather4_lz_2d_v4f32_f32, "V4fiffQtV4ibii", "nc", "extended-image-insts")

#undef BUILTIN
#undef TARGET_BUILTIN
133 changes: 104 additions & 29 deletions clang/lib/CodeGen/TargetBuiltins/AMDGPU.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -133,8 +133,8 @@ Value *EmitAMDGPUWorkGroupSize(CodeGenFunction &CGF, unsigned Index) {
}

llvm::MDBuilder MDHelper(CGF.getLLVMContext());
llvm::MDNode *RNode = MDHelper.createRange(APInt(16, 1),
APInt(16, CGF.getTarget().getMaxOpenCLWorkGroupSize() + 1));
llvm::MDNode *RNode = MDHelper.createRange(
APInt(16, 1), APInt(16, CGF.getTarget().getMaxOpenCLWorkGroupSize() + 1));
LD->setMetadata(llvm::LLVMContext::MD_range, RNode);
LD->setMetadata(llvm::LLVMContext::MD_noundef,
llvm::MDNode::get(CGF.getLLVMContext(), {}));
Expand Down Expand Up @@ -253,8 +253,7 @@ emitAMDGCNImageOverloadedReturnType(clang::CodeGen::CodeGenFunction &CGF,
}

// Emit an intrinsic that has 1 float or double operand, and 1 integer.
static Value *emitFPIntBuiltin(CodeGenFunction &CGF,
const CallExpr *E,
static Value *emitFPIntBuiltin(CodeGenFunction &CGF, const CallExpr *E,
unsigned IntrinsicID) {
llvm::Value *Src0 = CGF.EmitScalarExpr(E->getArg(0));
llvm::Value *Src1 = CGF.EmitScalarExpr(E->getArg(1));
Expand Down Expand Up @@ -456,8 +455,8 @@ Value *CodeGenFunction::EmitAMDGPUBuiltinExpr(unsigned BuiltinID,
llvm::Value *Y = EmitScalarExpr(E->getArg(1));
llvm::Value *Z = EmitScalarExpr(E->getArg(2));

llvm::Function *Callee = CGM.getIntrinsic(Intrinsic::amdgcn_div_scale,
X->getType());
llvm::Function *Callee =
CGM.getIntrinsic(Intrinsic::amdgcn_div_scale, X->getType());

llvm::Value *Tmp = Builder.CreateCall(Callee, {X, Y, Z});

Expand All @@ -477,8 +476,8 @@ Value *CodeGenFunction::EmitAMDGPUBuiltinExpr(unsigned BuiltinID,
llvm::Value *Src2 = EmitScalarExpr(E->getArg(2));
llvm::Value *Src3 = EmitScalarExpr(E->getArg(3));

llvm::Function *F = CGM.getIntrinsic(Intrinsic::amdgcn_div_fmas,
Src0->getType());
llvm::Function *F =
CGM.getIntrinsic(Intrinsic::amdgcn_div_fmas, Src0->getType());
llvm::Value *Src3ToBool = Builder.CreateIsNotNull(Src3);
return Builder.CreateCall(F, {Src0, Src1, Src2, Src3ToBool});
}
Expand Down Expand Up @@ -619,13 +618,13 @@ Value *CodeGenFunction::EmitAMDGPUBuiltinExpr(unsigned BuiltinID,
case AMDGPU::BI__builtin_amdgcn_frexp_expf: {
Value *Src0 = EmitScalarExpr(E->getArg(0));
Function *F = CGM.getIntrinsic(Intrinsic::amdgcn_frexp_exp,
{ Builder.getInt32Ty(), Src0->getType() });
{Builder.getInt32Ty(), Src0->getType()});
return Builder.CreateCall(F, Src0);
}
case AMDGPU::BI__builtin_amdgcn_frexp_exph: {
Value *Src0 = EmitScalarExpr(E->getArg(0));
Function *F = CGM.getIntrinsic(Intrinsic::amdgcn_frexp_exp,
{ Builder.getInt16Ty(), Src0->getType() });
{Builder.getInt16Ty(), Src0->getType()});
return Builder.CreateCall(F, Src0);
}
case AMDGPU::BI__builtin_amdgcn_fract:
Expand All @@ -646,8 +645,8 @@ Value *CodeGenFunction::EmitAMDGPUBuiltinExpr(unsigned BuiltinID,
case AMDGPU::BI__builtin_amdgcn_ballot_w64: {
llvm::Type *ResultType = ConvertType(E->getType());
llvm::Value *Src = EmitScalarExpr(E->getArg(0));
Function *F = CGM.getIntrinsic(Intrinsic::amdgcn_ballot, { ResultType });
return Builder.CreateCall(F, { Src });
Function *F = CGM.getIntrinsic(Intrinsic::amdgcn_ballot, {ResultType});
return Builder.CreateCall(F, {Src});
}
case AMDGPU::BI__builtin_amdgcn_inverse_ballot_w32:
case AMDGPU::BI__builtin_amdgcn_inverse_ballot_w64: {
Expand All @@ -671,8 +670,8 @@ Value *CodeGenFunction::EmitAMDGPUBuiltinExpr(unsigned BuiltinID,

// FIXME-GFX10: How should 32 bit mask be handled?
Function *F = CGM.getIntrinsic(Intrinsic::amdgcn_icmp,
{ Builder.getInt64Ty(), Src0->getType() });
return Builder.CreateCall(F, { Src0, Src1, Src2 });
{Builder.getInt64Ty(), Src0->getType()});
return Builder.CreateCall(F, {Src0, Src1, Src2});
}
case AMDGPU::BI__builtin_amdgcn_fcmp:
case AMDGPU::BI__builtin_amdgcn_fcmpf: {
Expand All @@ -682,8 +681,8 @@ Value *CodeGenFunction::EmitAMDGPUBuiltinExpr(unsigned BuiltinID,

// FIXME-GFX10: How should 32 bit mask be handled?
Function *F = CGM.getIntrinsic(Intrinsic::amdgcn_fcmp,
{ Builder.getInt64Ty(), Src0->getType() });
return Builder.CreateCall(F, { Src0, Src1, Src2 });
{Builder.getInt64Ty(), Src0->getType()});
return Builder.CreateCall(F, {Src0, Src1, Src2});
}
case AMDGPU::BI__builtin_amdgcn_class:
case AMDGPU::BI__builtin_amdgcn_classf:
Expand All @@ -695,11 +694,12 @@ Value *CodeGenFunction::EmitAMDGPUBuiltinExpr(unsigned BuiltinID,
Intrinsic::amdgcn_fmed3);
case AMDGPU::BI__builtin_amdgcn_ds_append:
case AMDGPU::BI__builtin_amdgcn_ds_consume: {
Intrinsic::ID Intrin = BuiltinID == AMDGPU::BI__builtin_amdgcn_ds_append ?
Intrinsic::amdgcn_ds_append : Intrinsic::amdgcn_ds_consume;
Intrinsic::ID Intrin = BuiltinID == AMDGPU::BI__builtin_amdgcn_ds_append
? Intrinsic::amdgcn_ds_append
: Intrinsic::amdgcn_ds_consume;
Value *Src0 = EmitScalarExpr(E->getArg(0));
Function *F = CGM.getIntrinsic(Intrin, { Src0->getType() });
return Builder.CreateCall(F, { Src0, Builder.getFalse() });
Function *F = CGM.getIntrinsic(Intrin, {Src0->getType()});
return Builder.CreateCall(F, {Src0, Builder.getFalse()});
}
case AMDGPU::BI__builtin_amdgcn_global_load_tr_b64_i32:
case AMDGPU::BI__builtin_amdgcn_global_load_tr_b64_v2i32:
Expand Down Expand Up @@ -919,12 +919,10 @@ Value *CodeGenFunction::EmitAMDGPUBuiltinExpr(unsigned BuiltinID,

// The builtins take these arguments as vec4 where the last element is
// ignored. The intrinsic takes them as vec3.
RayOrigin = Builder.CreateShuffleVector(RayOrigin, RayOrigin,
{0, 1, 2});
RayDir =
Builder.CreateShuffleVector(RayDir, RayDir, {0, 1, 2});
RayInverseDir = Builder.CreateShuffleVector(RayInverseDir, RayInverseDir,
{0, 1, 2});
RayOrigin = Builder.CreateShuffleVector(RayOrigin, RayOrigin, {0, 1, 2});
RayDir = Builder.CreateShuffleVector(RayDir, RayDir, {0, 1, 2});
RayInverseDir =
Builder.CreateShuffleVector(RayInverseDir, RayInverseDir, {0, 1, 2});

Function *F = CGM.getIntrinsic(Intrinsic::amdgcn_image_bvh_intersect_ray,
{NodePtr->getType(), RayDir->getType()});
Expand Down Expand Up @@ -998,8 +996,8 @@ Value *CodeGenFunction::EmitAMDGPUBuiltinExpr(unsigned BuiltinID,
Value *Rtn = Builder.CreateExtractValue(Call, 0);
Value *A = Builder.CreateExtractValue(Call, 1);
llvm::Type *RetTy = ConvertType(E->getType());
Value *I0 = Builder.CreateInsertElement(PoisonValue::get(RetTy), Rtn,
(uint64_t)0);
Value *I0 =
Builder.CreateInsertElement(PoisonValue::get(RetTy), Rtn, (uint64_t)0);
// ds_bvh_stack_push8_pop2_rtn returns {i64, i32} but the builtin returns
// <2 x i64>, zext the second value.
if (A->getType()->getPrimitiveSizeInBits() <
Expand Down Expand Up @@ -1138,6 +1136,83 @@ Value *CodeGenFunction::EmitAMDGPUBuiltinExpr(unsigned BuiltinID,
case AMDGPU::BI__builtin_amdgcn_image_sample_cube_v4f16_f32:
return emitAMDGCNImageOverloadedReturnType(
*this, E, Intrinsic::amdgcn_image_sample_cube, false);
case clang::AMDGPU::BI__builtin_amdgcn_image_sample_lz_1d_v4f32_f32:
case clang::AMDGPU::BI__builtin_amdgcn_image_sample_lz_1d_v4f16_f32:
return emitAMDGCNImageOverloadedReturnType(
*this, E, Intrinsic::amdgcn_image_sample_lz_1d, false);
case clang::AMDGPU::BI__builtin_amdgcn_image_sample_l_1d_v4f32_f32:
case clang::AMDGPU::BI__builtin_amdgcn_image_sample_l_1d_v4f16_f32:
return emitAMDGCNImageOverloadedReturnType(
*this, E, Intrinsic::amdgcn_image_sample_l_1d, false);
case clang::AMDGPU::BI__builtin_amdgcn_image_sample_d_1d_v4f32_f32:
case clang::AMDGPU::BI__builtin_amdgcn_image_sample_d_1d_v4f16_f32:
return emitAMDGCNImageOverloadedReturnType(
*this, E, Intrinsic::amdgcn_image_sample_d_1d, false);
case clang::AMDGPU::BI__builtin_amdgcn_image_sample_lz_2d_v4f32_f32:
case clang::AMDGPU::BI__builtin_amdgcn_image_sample_lz_2d_v4f16_f32:
case clang::AMDGPU::BI__builtin_amdgcn_image_sample_lz_2d_f32_f32:
return emitAMDGCNImageOverloadedReturnType(
*this, E, Intrinsic::amdgcn_image_sample_lz_2d, false);
case clang::AMDGPU::BI__builtin_amdgcn_image_sample_l_2d_v4f32_f32:
case clang::AMDGPU::BI__builtin_amdgcn_image_sample_l_2d_v4f16_f32:
case clang::AMDGPU::BI__builtin_amdgcn_image_sample_l_2d_f32_f32:
return emitAMDGCNImageOverloadedReturnType(
*this, E, Intrinsic::amdgcn_image_sample_l_2d, false);
case clang::AMDGPU::BI__builtin_amdgcn_image_sample_d_2d_v4f32_f32:
case clang::AMDGPU::BI__builtin_amdgcn_image_sample_d_2d_v4f16_f32:
case clang::AMDGPU::BI__builtin_amdgcn_image_sample_d_2d_f32_f32:
return emitAMDGCNImageOverloadedReturnType(
*this, E, Intrinsic::amdgcn_image_sample_d_2d, false);
case clang::AMDGPU::BI__builtin_amdgcn_image_sample_lz_3d_v4f32_f32:
case clang::AMDGPU::BI__builtin_amdgcn_image_sample_lz_3d_v4f16_f32:
return emitAMDGCNImageOverloadedReturnType(
*this, E, Intrinsic::amdgcn_image_sample_lz_3d, false);
case clang::AMDGPU::BI__builtin_amdgcn_image_sample_l_3d_v4f32_f32:
case clang::AMDGPU::BI__builtin_amdgcn_image_sample_l_3d_v4f16_f32:
return emitAMDGCNImageOverloadedReturnType(
*this, E, Intrinsic::amdgcn_image_sample_l_3d, false);
case clang::AMDGPU::BI__builtin_amdgcn_image_sample_d_3d_v4f32_f32:
case clang::AMDGPU::BI__builtin_amdgcn_image_sample_d_3d_v4f16_f32:
return emitAMDGCNImageOverloadedReturnType(
*this, E, Intrinsic::amdgcn_image_sample_d_3d, false);
case clang::AMDGPU::BI__builtin_amdgcn_image_sample_lz_cube_v4f32_f32:
case clang::AMDGPU::BI__builtin_amdgcn_image_sample_lz_cube_v4f16_f32:
return emitAMDGCNImageOverloadedReturnType(
*this, E, Intrinsic::amdgcn_image_sample_lz_cube, false);
case clang::AMDGPU::BI__builtin_amdgcn_image_sample_l_cube_v4f32_f32:
case clang::AMDGPU::BI__builtin_amdgcn_image_sample_l_cube_v4f16_f32:
return emitAMDGCNImageOverloadedReturnType(
*this, E, Intrinsic::amdgcn_image_sample_l_cube, false);
case clang::AMDGPU::BI__builtin_amdgcn_image_sample_lz_1darray_v4f32_f32:
case clang::AMDGPU::BI__builtin_amdgcn_image_sample_lz_1darray_v4f16_f32:
return emitAMDGCNImageOverloadedReturnType(
*this, E, Intrinsic::amdgcn_image_sample_lz_1darray, false);
case clang::AMDGPU::BI__builtin_amdgcn_image_sample_l_1darray_v4f32_f32:
case clang::AMDGPU::BI__builtin_amdgcn_image_sample_l_1darray_v4f16_f32:
return emitAMDGCNImageOverloadedReturnType(
*this, E, Intrinsic::amdgcn_image_sample_l_1darray, false);
case clang::AMDGPU::BI__builtin_amdgcn_image_sample_d_1darray_v4f32_f32:
case clang::AMDGPU::BI__builtin_amdgcn_image_sample_d_1darray_v4f16_f32:
return emitAMDGCNImageOverloadedReturnType(
*this, E, Intrinsic::amdgcn_image_sample_d_1darray, false);
case clang::AMDGPU::BI__builtin_amdgcn_image_sample_lz_2darray_v4f32_f32:
case clang::AMDGPU::BI__builtin_amdgcn_image_sample_lz_2darray_v4f16_f32:
case clang::AMDGPU::BI__builtin_amdgcn_image_sample_lz_2darray_f32_f32:
return emitAMDGCNImageOverloadedReturnType(
*this, E, Intrinsic::amdgcn_image_sample_lz_2darray, false);
case clang::AMDGPU::BI__builtin_amdgcn_image_sample_l_2darray_v4f32_f32:
case clang::AMDGPU::BI__builtin_amdgcn_image_sample_l_2darray_v4f16_f32:
case clang::AMDGPU::BI__builtin_amdgcn_image_sample_l_2darray_f32_f32:
return emitAMDGCNImageOverloadedReturnType(
*this, E, Intrinsic::amdgcn_image_sample_l_2darray, false);
case clang::AMDGPU::BI__builtin_amdgcn_image_sample_d_2darray_v4f32_f32:
case clang::AMDGPU::BI__builtin_amdgcn_image_sample_d_2darray_v4f16_f32:
case clang::AMDGPU::BI__builtin_amdgcn_image_sample_d_2darray_f32_f32:
return emitAMDGCNImageOverloadedReturnType(
*this, E, Intrinsic::amdgcn_image_sample_d_2darray, false);
case clang::AMDGPU::BI__builtin_amdgcn_image_gather4_lz_2d_v4f32_f32:
return emitAMDGCNImageOverloadedReturnType(
*this, E, Intrinsic::amdgcn_image_gather4_lz_2d, false);
case AMDGPU::BI__builtin_amdgcn_mfma_scale_f32_16x16x128_f8f6f4:
case AMDGPU::BI__builtin_amdgcn_mfma_scale_f32_32x32x64_f8f6f4: {
llvm::FixedVectorType *VT = FixedVectorType::get(Builder.getInt32Ty(), 8);
Expand Down Expand Up @@ -1627,7 +1702,7 @@ Value *CodeGenFunction::EmitAMDGPUBuiltinExpr(unsigned BuiltinID,
llvm::Value *Src1 = EmitScalarExpr(E->getArg(1));
llvm::Value *Src2 = EmitScalarExpr(E->getArg(2));
Function *F = CGM.getIntrinsic(Intrinsic::fshr, Src0->getType());
return Builder.CreateCall(F, { Src0, Src1, Src2 });
return Builder.CreateCall(F, {Src0, Src1, Src2});
}
case AMDGPU::BI__builtin_amdgcn_fence: {
ProcessOrderScopeAMDGCN(EmitScalarExpr(E->getArg(0)),
Expand Down
Loading