Shilei Tian bde0a8cd63

[Clang][HIP][CUDA] Add __cluster_dims__ and __no_cluster__ attribute (#156686) ...

This PR adds basic frontend support for `__cluster_dims__` and
`__no_cluster__` attribute.

In CUDA/HIP programming, the ``__cluster_dims__`` attribute can be
applied to a kernel function to set the dimensions of a thread block
cluster. The ``__no_cluster__`` attribute can be applied to a kernel
function to indicate that the thread block cluster feature will not be
enabled at both compile time and kernel launch time. Note that
`__no_cluster__` is a LLVM/Clang only attribute.

Co-authored-by: Yaxun (Sam) Liu <yaxun.liu@amd.com>
Co-authored-by: Jay Foad <jay.foad@amd.com>

2025-10-17 11:39:43 -04:00

..

Inputs

[Clang][HIP][CUDA] Add __cluster_dims__ and __no_cluster__ attribute (#156686)

2025-10-17 11:39:43 -04:00

address-spaces.cu

[cuda][HIP] __constant__ should imply constant (#110182)

2024-09-29 01:22:52 +01:00

alias.cu

…

amdgpu-alias-undef-symbols.cu

…

amdgpu-atomic-ops.cu

Add clang atomic control options and attribute (#114841)

2025-02-27 10:41:04 -05:00

amdgpu-bf16.cu

[Clang] Treat ext_vector_type as a regular type attribute (#130177)

2025-03-07 10:09:06 -06:00

amdgpu-code-object-version.cu

Reapply "[AMDGPU] Use COV6 by default (#118515)" (#130963)

2025-03-21 15:26:45 -04:00

amdgpu-kernel-arg-pointer-type.cu

[IR] Convert from nocapture to captures(none) (#123181)

2025-01-29 16:56:47 +01:00

amdgpu-kernel-attrs.cu

[AMDGPU] fix amdgpu_max_num_work_groups in templates (#141633)

2025-05-28 08:05:03 -04:00

amdgpu-visibility.cu

[cuda][HIP] __constant__ should imply constant (#110182)

2024-09-29 01:22:52 +01:00

amdgpu-workgroup-size.cu

[Clang][AMDGPU] Remove special handling for COV4 libraries (#132870)

2025-03-28 07:35:16 -05:00

anon-ns.cu

[cuda][HIP] __constant__ should imply constant (#110182)

2024-09-29 01:22:52 +01:00

atomic-ops.cu

Add clang atomic control options and attribute (#114841)

2025-02-27 10:41:04 -05:00

atomic-options.hip

Add clang atomic control options and attribute (#114841)

2025-02-27 10:41:04 -05:00

atomics-remarks-gfx90a.cu

…

bf16.cu

[NVPTX] Fix v2i8 call lowering, use generic ld/st nodes for call params (#146930)

2025-07-28 10:41:51 -07:00

bug-kerner-registration-reuse.cu

…

builtin-mangled-name.cu

…

builtins-amdgcn.cu

Fix codegen of consteval functions returning an empty class, and related issues (#93115)

2024-08-01 16:18:20 -07:00

builtins-sm90.cu

[NVPTX] Add support for Shared Cluster Memory address space [2/2] (#136768)

2025-04-22 16:50:45 -07:00

builtins-spirv-amdgcn.cu

Fix codegen of consteval functions returning an empty class, and related issues (#93115)

2024-08-01 16:18:20 -07:00

builtins-unsafe-atomics-gfx90a.cu

clang/AMDGPU: Emit atomicrmw from ds_fadd builtins (#95395)

2024-06-18 20:51:14 +02:00

builtins-unsafe-atomics-spirv-amdgcn-gfx90a.cu

clang/AMDGPU: Emit atomicrmw from ds_fadd builtins (#95395)

2024-06-18 20:51:14 +02:00

cluster_dims.cu

[Clang][HIP][CUDA] Add __cluster_dims__ and __no_cluster__ attribute (#156686)

2025-10-17 11:39:43 -04:00

const-var.cu

…

constexpr-variables.cu

…

convergent.cu

[clang][CUDA] Add 'noconvergent' function and statement attribute

2024-07-31 11:30:48 -04:00

correctly-rounded-div.cu

cuda clang: Move nvptx-surface.cu test to CodeGenCUDA (#134758)

2025-04-22 12:36:39 -07:00

cuda-builtin-vars.cu

fixup cuda-builtin-vars.cu broken in IntrRange change (#94639)

2024-06-06 10:10:00 -07:00

cxx-call-kernel.cpp

…

debug-info-address-class.cu

[RemoveDIs] Print IR with debug records by default (#91724)

2024-06-14 15:07:27 +01:00

debug-info-template.cu

…

deferred-diag.cu

…

dependent-libs.cu

…

device-fun-linkage.cu

[NVPTX] Switch front-ends and tests to ptx_kernel cc (#120806)

2025-01-07 18:24:50 -08:00

device-init-fun.cu

[clang][HIP] Don't use the OpenCLKernel CC when targeting AMDGCNSPIRV (#110447)

2024-10-22 17:16:46 +01:00

device-stub.cu

[HIP][Clang][CodeGen] Handle hip bin symbols properly. (#107458)

2024-09-11 18:46:46 -04:00

device-use-host-var.cu

…

device-var-init.cu

[cuda][HIP] __constant__ should imply constant (#110182)

2024-09-29 01:22:52 +01:00

device-var-linkage.cu

[cuda][HIP] __constant__ should imply constant (#110182)

2024-09-29 01:22:52 +01:00

device-vtable.cu

…

dft-func-attr-skip-intrinsic.hip

…

filter-decl.cu

[cuda][HIP] __constant__ should imply constant (#110182)

2024-09-29 01:22:52 +01:00

flush-denormals.cu

…

fp-contract.cu

[NVPTX] use untyped loads and stores where ever possible (#137698)

2025-05-10 08:26:26 -07:00

function-overload.cu

…

global-initializers.cu

…

grid-constant.cu

[NVPTX] Auto-upgrade nvvm.grid_constant to param attribute (#155489)

2025-08-27 16:32:28 -07:00

host-used-device-var.cu

[HIP] Allow partial linking for -fgpu-rdc (#81700)

2024-02-22 13:51:31 -05:00

host-used-extern-determinism.cu

[CUDA] Emit used function list in deterministic order. (#102661)

2024-08-12 10:21:23 -07:00

host-used-extern.cu

[HIP] fix host-used external kernel (#83870)

2024-03-08 10:50:38 -05:00

implicit-host-device-fun.cu

…

incomplete-func-ptr-type.cu

…

increment-index-for-thunks.cu

[CUDA] Increment VTable index for device thunks (#124989)

2025-02-19 23:47:22 -05:00

kernel-amdgcn.cu

[clang][HIP] Don't use the OpenCLKernel CC when targeting AMDGCNSPIRV (#110447)

2024-10-22 17:16:46 +01:00

kernel-arg-name-metadata.cu

…

kernel-args-alignment.cu

…

kernel-args.cu

[clang][CodeGen][SPIR-V][AMDGPU] Tweak AMDGCNSPIRV ABI to allow for the correct handling of aggregates passed to kernels / functions. (#102776)

2024-08-21 13:16:59 +01:00

kernel-call.cu

…

kernel-dbg-info.cu

…

kernel-stub-name.cu

[CUDA] make kernel stub ICF-proof (#90155)

2024-05-01 10:24:23 -04:00

lambda-constexpr-capture.cu

[CUDA][HIP] capture possible ODR-used var (#136645)

2025-04-23 12:50:28 -04:00

lambda-noinline.cu

…

lambda-reference-var.cu

[IRBuilder] Generate nuw GEPs for struct member accesses (#99538)

2024-08-09 13:25:04 +01:00

lambda.cu

…

launch-bounds.cu

[NVPTX] Convert vector function nvvm.annotations to attributes (#127736)

2025-02-26 08:45:27 -08:00

library-builtin.cu

…

link-builtin-bitcode-denormal-fp-mode.cu

…

link-builtin-bitcode-gpu-attrs-preserved.cu

…

link-device-bitcode.cu

…

llvm-used.cu

…

long-double.cu

[clang][SPIR-V] Add support for AMDGCN flavoured SPIRV (#89796)

2024-06-07 11:50:23 +01:00

managed-var.cu

[IRBuilder] Generate nuw GEPs for struct member accesses (#99538)

2024-08-09 13:25:04 +01:00

member-init.cu

…

memcpy-libcall.cu

[NVPTX] use untyped loads and stores where ever possible (#137698)

2025-05-10 08:26:26 -07:00

ms-linker-options.cu

…

noinline.cu

…

norecurse.cu

…

nosanitize-cuid.hip

[HIP] disable sanitizer for __hip_cuid (#141581)

2025-05-27 16:58:04 -04:00

nothrow.cu

…

nvptx-surface.cu

cuda clang: Move nvptx-surface.cu test to CodeGenCUDA (#134758)

2025-04-22 12:36:39 -07:00

offload_via_llvm.cu

[NVPTX] Switch front-ends and tests to ptx_kernel cc (#120806)

2025-01-07 18:24:50 -08:00

offloading-entries.cu

cuda clang: Move nvptx-surface.cu test to CodeGenCUDA (#134758)

2025-04-22 12:36:39 -07:00

openmp-parallel.cu

…

openmp-target.cu

…

printf-aggregate.cu

…

printf-builtin.cu

[clang][AMDGPU][CUDA] Handle __builtin_printf for device printf (#68515)

2024-02-05 23:23:13 +05:30

printf.cu

[IRBuilder] Generate nuw GEPs for struct member accesses (#99538)

2024-08-09 13:25:04 +01:00

profile-coverage-mapping.cu

cuda clang: Move nvptx-surface.cu test to CodeGenCUDA (#134758)

2025-04-22 12:36:39 -07:00

propagate-attributes.cu

[clang][CodeGen] Remove "unsafe-fp-math" attribute support (#162779)

2025-10-10 15:56:29 +08:00

ptx-kernels.cu

[NVPTX] Switch front-ends and tests to ptx_kernel cc (#120806)

2025-01-07 18:24:50 -08:00

record-layout.cu

[IRBuilder] Generate nuw GEPs for struct member accesses (#99538)

2024-08-09 13:25:04 +01:00

redux-builtins.cu

…

redux-f32-builtins.cu

[NVPTX] Add intrinsics for redux.sync f32 instructions (#126664)

2025-02-14 11:11:44 +05:30

shuffle_long_long.cu

…

spirv-amdgcn-bf16.cu

[clang][SPIR-V] Add support for AMDGCN flavoured SPIRV (#89796)

2024-06-07 11:50:23 +01:00

static-device-var-no-rdc.cu

[cuda][HIP] __constant__ should imply constant (#110182)

2024-09-29 01:22:52 +01:00

static-device-var-rdc.cu

[cuda][HIP] __constant__ should imply constant (#110182)

2024-09-29 01:22:52 +01:00

struct-mangling-number.cu

…

surface.cu

…

template-class-static-member.cu

[cuda][HIP] __constant__ should imply constant (#110182)

2024-09-29 01:22:52 +01:00

texture.cu

…

types.cu

…

unnamed-types.cu

…

unused-global-var.cu

…

usual-deallocators.cu

[NVPTX] Switch front-ends and tests to ptx_kernel cc (#120806)

2025-01-07 18:24:50 -08:00

vtbl.cu

…