[OpenCL] Add missing OpenCL 3.0 features to OpenCLExtensions.def; revert header-only macros

[wenju-he]

Adds the remaining optional feature macros from the OpenCL C 3.0 spec (section 6.2.1 table). Targets can now enable these via OpenCLFeaturesMap returned by getSupportedOpenCLOpts().

Revert a84599f (header‑only feature macros).
Header‑only macros are difficult to disable on SPIR-V targets,
and the prior undef approach (a60b8f4) does not scale.
After this PR, they can be disabled via -cl-ext=-<feature>.

KhronosGroup/OpenCL-Docs#1328 also notes that unconditional
definition of the header‑only macros in opencl-c-base.h should be removed.

[Copilot]

Pull Request Overview

This PR adds missing OpenCL C 3.0 optional feature macros to the OpenCLExtensions.def file, ensuring completeness with the OpenCL C 3.0 specification (section 6.2.1 table). This enables targets to properly advertise support for these optional features through the OpenCLFeaturesMap.

• Adds 11 missing optional feature macros for OpenCL C 3.0
• Ensures all features from the OpenCL C 3.0 spec table are represented in the codebase

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[bader]

I would expect clang/test/SemaOpenCL/features.cl tests will be extended to cover new features.
Potentially clang/test/Headers/opencl-c-header.cl and clang/test/SemaOpenCL/extension-version.cl as well.

[llvmbot]

@llvm/pr-subscribers-backend-amdgpu

Author: Wenju He (wenju-he)

Changes

Adds the remaining optional feature macros from the OpenCL C 3.0 spec (section 6.2.1 table). Targets can now enable these via OpenCLFeaturesMap returned by getSupportedOpenCLOpts().

---

Patch is 39.74 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/168016.diff

6 Files Affected:

• (modified) clang/include/clang/Basic/OpenCLExtensions.def (+46-9)
• (modified) clang/lib/Basic/Targets/AMDGPU.h (+2)
• (modified) clang/lib/Headers/opencl-c-base.h (-99)
• (modified) clang/test/Headers/opencl-c-header.cl (+3-3)
• (modified) clang/test/SemaOpenCL/extension-version.cl (+244)
• (modified) clang/test/SemaOpenCL/features.cl (+66-34)

diff --git a/clang/include/clang/Basic/OpenCLExtensions.def b/clang/include/clang/Basic/OpenCLExtensions.def index 6f73b26137500..9fc8c31056f4d 100644 --- a/clang/include/clang/Basic/OpenCLExtensions.def +++ b/clang/include/clang/Basic/OpenCLExtensions.def @@ -78,10 +78,55 @@ OPENCL_EXTENSION(cl_khr_depth_images, true, 120) OPENCL_EXTENSION(cl_khr_gl_msaa_sharing,true, 120) // OpenCL 2.0. +OPENCL_EXTENSION(cl_ext_float_atomics, true, 200) +OPENCL_EXTENSION(cl_khr_extended_bit_ops, true, 200) +OPENCL_EXTENSION(cl_khr_integer_dot_product, true, 200) +OPENCL_EXTENSION(cl_khr_kernel_clock, true, 200) OPENCL_EXTENSION(cl_khr_mipmap_image, true, 200) OPENCL_EXTENSION(cl_khr_mipmap_image_writes, true, 200) OPENCL_EXTENSION(cl_khr_srgb_image_writes, true, 200) +OPENCL_EXTENSION(cl_khr_subgroup_ballot, true, 200) +OPENCL_EXTENSION(cl_khr_subgroup_clustered_reduce, true, 200) +OPENCL_EXTENSION(cl_khr_subgroup_extended_types, true, 200) +OPENCL_EXTENSION(cl_khr_subgroup_non_uniform_arithmetic, true, 200) +OPENCL_EXTENSION(cl_khr_subgroup_non_uniform_vote, true, 200) +OPENCL_EXTENSION(cl_khr_subgroup_rotate, true, 200) +OPENCL_EXTENSION(cl_khr_subgroup_shuffle_relative, true, 200) +OPENCL_EXTENSION(cl_khr_subgroup_shuffle, true, 200) OPENCL_EXTENSION(cl_khr_subgroups, true, 200) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_atomic_order_acq_rel, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_atomic_order_seq_cst, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_atomic_scope_all_devices, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_atomic_scope_device, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_device_enqueue, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_ext_fp16_global_atomic_add, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_ext_fp16_global_atomic_load_store, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_ext_fp16_global_atomic_min_max, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_ext_fp16_local_atomic_add, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_ext_fp16_local_atomic_load_store, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_ext_fp16_local_atomic_min_max, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_ext_fp32_global_atomic_add, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_ext_fp32_global_atomic_min_max, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_ext_fp32_local_atomic_add, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_ext_fp32_local_atomic_min_max, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_ext_fp64_global_atomic_add, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_ext_fp64_global_atomic_min_max, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_ext_fp64_local_atomic_add, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_ext_fp64_local_atomic_min_max, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_ext_image_raw10_raw12, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_ext_image_unorm_int_2_101010, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_ext_image_unsigned_10x6_12x4_14x2, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_generic_address_space, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_images, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_integer_dot_product_input_4x8bit, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_integer_dot_product_input_4x8bit_packed, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_kernel_clock_scope_device, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_kernel_clock_scope_sub_group, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_kernel_clock_scope_work_group, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_pipes, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_program_scope_global_variables, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_read_write_images, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_work_group_collective_functions, false, 200, OCL_C_20) // Clang Extensions. OPENCL_EXTENSION(cl_clang_storage_class_specifiers, true, 100) @@ -100,17 +145,9 @@ OPENCL_EXTENSION(cl_intel_subgroups_short, true, 120) OPENCL_EXTENSION(cl_intel_device_side_avc_motion_estimation, true, 120) // OpenCL C 3.0 features (6.2.1. Features) -OPENCL_OPTIONALCOREFEATURE(__opencl_c_pipes, false, 300, OCL_C_30) -OPENCL_OPTIONALCOREFEATURE(__opencl_c_generic_address_space, false, 300, OCL_C_30) -OPENCL_OPTIONALCOREFEATURE(__opencl_c_atomic_order_acq_rel, false, 300, OCL_C_30) -OPENCL_OPTIONALCOREFEATURE(__opencl_c_atomic_order_seq_cst, false, 300, OCL_C_30) -OPENCL_OPTIONALCOREFEATURE(__opencl_c_subgroups, false, 300, OCL_C_30) OPENCL_OPTIONALCOREFEATURE(__opencl_c_3d_image_writes, false, 300, OCL_C_30) -OPENCL_OPTIONALCOREFEATURE(__opencl_c_device_enqueue, false, 300, OCL_C_30) -OPENCL_OPTIONALCOREFEATURE(__opencl_c_read_write_images, false, 300, OCL_C_30) -OPENCL_OPTIONALCOREFEATURE(__opencl_c_program_scope_global_variables, false, 300, OCL_C_30) OPENCL_OPTIONALCOREFEATURE(__opencl_c_fp64, false, 300, OCL_C_30) -OPENCL_OPTIONALCOREFEATURE(__opencl_c_images, false, 300, OCL_C_30) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_subgroups, false, 300, OCL_C_30) #undef OPENCL_OPTIONALCOREFEATURE #undef OPENCL_COREFEATURE diff --git a/clang/lib/Basic/Targets/AMDGPU.h b/clang/lib/Basic/Targets/AMDGPU.h index dfcc79402257a..95a3d133c07ac 100644 --- a/clang/lib/Basic/Targets/AMDGPU.h +++ b/clang/lib/Basic/Targets/AMDGPU.h @@ -320,6 +320,8 @@ class LLVM_LIBRARY_VISIBILITY AMDGPUTargetInfo final : public TargetInfo { Opts["__opencl_c_3d_image_writes"] = true; Opts["cl_khr_3d_image_writes"] = true; Opts["__opencl_c_program_scope_global_variables"] = true; + Opts["__opencl_c_atomic_scope_all_devices"] = true; + Opts["__opencl_c_atomic_order_seq_cst"] = true; if (GPUKind >= llvm::AMDGPU::GK_GFX700) { Opts["__opencl_c_generic_address_space"] = true; diff --git a/clang/lib/Headers/opencl-c-base.h b/clang/lib/Headers/opencl-c-base.h index 414f10ad832ce..898026c66614a 100644 --- a/clang/lib/Headers/opencl-c-base.h +++ b/clang/lib/Headers/opencl-c-base.h @@ -9,105 +9,6 @@ #ifndef _OPENCL_BASE_H_ #define _OPENCL_BASE_H_ -// Define extension macros - -#if (defined(__OPENCL_CPP_VERSION__) || __OPENCL_C_VERSION__ >= 200) -// For SPIR and SPIR-V all extensions are supported. -#if defined(__SPIR__) || defined(__SPIRV__) -#define cl_khr_subgroup_extended_types 1 -#define cl_khr_subgroup_non_uniform_vote 1 -#define cl_khr_subgroup_ballot 1 -#define cl_khr_subgroup_non_uniform_arithmetic 1 -#define cl_khr_subgroup_shuffle 1 -#define cl_khr_subgroup_shuffle_relative 1 -#define cl_khr_subgroup_clustered_reduce 1 -#define cl_khr_subgroup_rotate 1 -#define cl_khr_extended_bit_ops 1 -#define cl_khr_integer_dot_product 1 -#define __opencl_c_integer_dot_product_input_4x8bit 1 -#define __opencl_c_integer_dot_product_input_4x8bit_packed 1 -#define cl_ext_float_atomics 1 -#ifdef cl_khr_fp16 -#define __opencl_c_ext_fp16_global_atomic_load_store 1 -#define __opencl_c_ext_fp16_local_atomic_load_store 1 -#define __opencl_c_ext_fp16_global_atomic_add 1 -#define __opencl_c_ext_fp16_local_atomic_add 1 -#define __opencl_c_ext_fp16_global_atomic_min_max 1 -#define __opencl_c_ext_fp16_local_atomic_min_max 1 -#endif -#ifdef cl_khr_fp64 -#define __opencl_c_ext_fp64_global_atomic_add 1 -#define __opencl_c_ext_fp64_local_atomic_add 1 -#define __opencl_c_ext_fp64_global_atomic_min_max 1 -#define __opencl_c_ext_fp64_local_atomic_min_max 1 -#endif -#define __opencl_c_ext_fp32_global_atomic_add 1 -#define __opencl_c_ext_fp32_local_atomic_add 1 -#define __opencl_c_ext_fp32_global_atomic_min_max 1 -#define __opencl_c_ext_fp32_local_atomic_min_max 1 -#define __opencl_c_ext_image_raw10_raw12 1 -#define __opencl_c_ext_image_unorm_int_2_101010 1 -#define __opencl_c_ext_image_unsigned_10x6_12x4_14x2 1 -#define cl_khr_kernel_clock 1 -#define __opencl_c_kernel_clock_scope_device 1 -#define __opencl_c_kernel_clock_scope_work_group 1 -#define __opencl_c_kernel_clock_scope_sub_group 1 - -#endif // defined(__SPIR__) || defined(__SPIRV__) -#endif // (defined(__OPENCL_CPP_VERSION__) || __OPENCL_C_VERSION__ >= 200) - -// Define feature macros for OpenCL C 2.0 -#if (__OPENCL_CPP_VERSION__ == 100 || __OPENCL_C_VERSION__ == 200) -#define __opencl_c_pipes 1 -#define __opencl_c_generic_address_space 1 -#define __opencl_c_work_group_collective_functions 1 -#define __opencl_c_atomic_order_acq_rel 1 -#define __opencl_c_atomic_order_seq_cst 1 -#define __opencl_c_atomic_scope_device 1 -#define __opencl_c_atomic_scope_all_devices 1 -#define __opencl_c_device_enqueue 1 -#define __opencl_c_read_write_images 1 -#define __opencl_c_program_scope_global_variables 1 -#define __opencl_c_images 1 -#endif - -// Define header-only feature macros for OpenCL C 3.0. -#if (__OPENCL_CPP_VERSION__ == 202100 || __OPENCL_C_VERSION__ == 300) -// For the SPIR and SPIR-V target all features are supported. -#if defined(__SPIR__) || defined(__SPIRV__) -#define __opencl_c_work_group_collective_functions 1 -#define __opencl_c_atomic_order_seq_cst 1 -#define __opencl_c_atomic_scope_device 1 -#define __opencl_c_atomic_scope_all_devices 1 -#define __opencl_c_read_write_images 1 -#endif // defined(__SPIR__) - -#endif // (__OPENCL_CPP_VERSION__ == 202100 || __OPENCL_C_VERSION__ == 300) - -// Undefine any feature macros that have been explicitly disabled using -// an __undef_<feature> macro. -#ifdef __undef___opencl_c_work_group_collective_functions -#undef __opencl_c_work_group_collective_functions -#endif -#ifdef __undef___opencl_c_atomic_order_seq_cst -#undef __opencl_c_atomic_order_seq_cst -#endif -#ifdef __undef___opencl_c_atomic_scope_device -#undef __opencl_c_atomic_scope_device -#endif -#ifdef __undef___opencl_c_atomic_scope_all_devices -#undef __opencl_c_atomic_scope_all_devices -#endif -#ifdef __undef___opencl_c_read_write_images -#undef __opencl_c_read_write_images -#endif -#ifdef __undef___opencl_c_integer_dot_product_input_4x8bit -#undef __opencl_c_integer_dot_product_input_4x8bit -#endif -#ifdef __undef___opencl_c_integer_dot_product_input_4x8bit_packed -#undef __opencl_c_integer_dot_product_input_4x8bit_packed -#endif - #if !defined(__opencl_c_generic_address_space) // Internal feature macro to provide named (global, local, private) address // space overloads for builtin functions that take a pointer argument. diff --git a/clang/test/Headers/opencl-c-header.cl b/clang/test/Headers/opencl-c-header.cl index 17cbb67f26038..e26f16827b20f 100644 --- a/clang/test/Headers/opencl-c-header.cl +++ b/clang/test/Headers/opencl-c-header.cl @@ -33,7 +33,7 @@ // === // Compile for OpenCL 2.0 for the first time. The module should change. // RUN: %clang_cc1 -triple spir-unknown-unknown -O0 -emit-llvm -o - -cl-std=CL2.0 -finclude-default-header -fmodules -fimplicit-module-maps -fmodules-cache-path=%t -fdisable-module-hash -ftime-report %s 2>&1 | FileCheck --check-prefix=CHECK20 --check-prefix=CHECK-MOD %s -// RUN: not diff %t/1_0.pcm %t/opencl_c.pcm +// RUN: not diff %t/1_0.pcm %t/opencl_c.pcm > /dev/null // RUN: chmod u-w %t/opencl_c.pcm // === @@ -44,10 +44,10 @@ // RUN: rm -rf %t // RUN: mkdir -p %t // RUN: %clang_cc1 -triple spir64-unknown-unknown -emit-llvm -o - -cl-std=CL1.2 -finclude-default-header -fmodules -fimplicit-module-maps -fmodules-cache-path=%t -ftime-report %s 2>&1 | FileCheck --check-prefix=CHECK --check-prefix=CHECK-MOD %s -// RUN: %clang_cc1 -triple amdgcn--amdhsa -O0 -emit-llvm -o - -cl-std=CL2.0 -finclude-default-header -fmodules -fimplicit-module-maps -fmodules-cache-path=%t -ftime-report %s 2>&1 | FileCheck --check-prefix=CHECK20 --check-prefix=CHECK-MOD %s +// RUN: %clang_cc1 -triple amdgcn--amdhsa -target-cpu gfx700 -O0 -emit-llvm -o - -cl-std=CL2.0 -finclude-default-header -fmodules -fimplicit-module-maps -fmodules-cache-path=%t -ftime-report %s 2>&1 | FileCheck --check-prefix=CHECK20 --check-prefix=CHECK-MOD %s // RUN: chmod u-w %t // RUN: %clang_cc1 -triple spir64-unknown-unknown -emit-llvm -o - -cl-std=CL1.2 -finclude-default-header -fmodules -fimplicit-module-maps -fmodules-cache-path=%t -ftime-report %s 2>&1 | FileCheck --check-prefix=CHECK --check-prefix=CHECK-MOD %s -// RUN: %clang_cc1 -triple amdgcn--amdhsa -O0 -emit-llvm -o - -cl-std=CL2.0 -finclude-default-header -fmodules -fimplicit-module-maps -fmodules-cache-path=%t -ftime-report %s 2>&1 | FileCheck --check-prefix=CHECK20 --check-prefix=CHECK-MOD %s +// RUN: %clang_cc1 -triple amdgcn--amdhsa -target-cpu gfx700 -O0 -emit-llvm -o - -cl-std=CL2.0 -finclude-default-header -fmodules -fimplicit-module-maps -fmodules-cache-path=%t -ftime-report %s 2>&1 | FileCheck --check-prefix=CHECK20 --check-prefix=CHECK-MOD %s // RUN: chmod u+w %t // Verify that called builtins occur in the generated IR. diff --git a/clang/test/SemaOpenCL/extension-version.cl b/clang/test/SemaOpenCL/extension-version.cl index 4d92eff6ae3ac..35a1818706f35 100644 --- a/clang/test/SemaOpenCL/extension-version.cl +++ b/clang/test/SemaOpenCL/extension-version.cl @@ -165,6 +165,150 @@ #endif #pragma OPENCL EXTENSION cl_khr_subgroups : enable +#if (defined(__OPENCL_CPP_VERSION__) || __OPENCL_C_VERSION__ >= 200) +#ifndef cl_ext_float_atomics +#error "Missing cl_ext_float_atomics define" +#endif +#else +#ifdef cl_ext_float_atomics +#error "Incorrect cl_ext_float_atomics define" +#endif +// expected-warning@+2{{unsupported OpenCL extension 'cl_ext_float_atomics' - ignoring}} +#endif +#pragma OPENCL EXTENSION cl_ext_float_atomics : enable + +#if (defined(__OPENCL_CPP_VERSION__) || __OPENCL_C_VERSION__ >= 200) +#ifndef cl_khr_extended_bit_ops +#error "Missing cl_khr_extended_bit_ops define" +#endif +#else +#ifdef cl_khr_extended_bit_ops +#error "Incorrect cl_khr_extended_bit_ops define" +#endif +// expected-warning@+2{{unsupported OpenCL extension 'cl_khr_extended_bit_ops' - ignoring}} +#endif +#pragma OPENCL EXTENSION cl_khr_extended_bit_ops : enable + +#if (defined(__OPENCL_CPP_VERSION__) || __OPENCL_C_VERSION__ >= 200) +#ifndef cl_khr_integer_dot_product +#error "Missing cl_khr_integer_dot_product define" +#endif +#else +#ifdef cl_khr_integer_dot_product +#error "Incorrect cl_khr_integer_dot_product define" +#endif +// expected-warning@+2{{unsupported OpenCL extension 'cl_khr_integer_dot_product' - ignoring}} +#endif +#pragma OPENCL EXTENSION cl_khr_integer_dot_product : enable + +#if (defined(__OPENCL_CPP_VERSION__) || __OPENCL_C_VERSION__ >= 200) +#ifndef cl_khr_kernel_clock +#error "Missing cl_khr_kernel_clock define" +#endif +#else +#ifdef cl_khr_kernel_clock +#error "Incorrect cl_khr_kernel_clock define" +#endif +// expected-warning@+2{{unsupported OpenCL extension 'cl_khr_kernel_clock' - ignoring}} +#endif +#pragma OPENCL EXTENSION cl_khr_kernel_clock : enable + +#if (defined(__OPENCL_CPP_VERSION__) || __OPENCL_C_VERSION__ >= 200) +#ifndef cl_khr_subgroup_ballot +#error "Missing cl_khr_subgroup_ballot define" +#endif +#else +#ifdef cl_khr_subgroup_ballot +#error "Incorrect cl_khr_subgroup_ballot define" +#endif +// expected-warning@+2{{unsupported OpenCL extension 'cl_khr_subgroup_ballot' - ignoring}} +#endif +#pragma OPENCL EXTENSION cl_khr_subgroup_ballot : enable + +#if (defined(__OPENCL_CPP_VERSION__) || __OPENCL_C_VERSION__ >= 200) +#ifndef cl_khr_subgroup_clustered_reduce +#error "Missing cl_khr_subgroup_clustered_reduce define" +#endif +#else +#ifdef cl_khr_subgroup_clustered_reduce +#error "Incorrect cl_khr_subgroup_clustered_reduce define" +#endif +// expected-warning@+2{{unsupported OpenCL extension 'cl_khr_subgroup_clustered_reduce' - ignoring}} +#endif +#pragma OPENCL EXTENSION cl_khr_subgroup_clustered_reduce : enable + +#if (defined(__OPENCL_CPP_VERSION__) || __OPENCL_C_VERSION__ >= 200) +#ifndef cl_khr_subgroup_extended_types +#error "Missing cl_khr_subgroup_extended_types define" +#endif +#else +#ifdef cl_khr_subgroup_extended_types +#error "Incorrect cl_khr_subgroup_extended_types define" +#endif +// expected-warning@+2{{unsupported OpenCL extension 'cl_khr_subgroup_extended_types' - ignoring}} +#endif +#pragma OPENCL EXTENSION cl_khr_subgroup_extended_types : enable + +#if (defined(__OPENCL_CPP_VERSION__) || __OPENCL_C_VERSION__ >= 200) +#ifndef cl_khr_subgroup_non_uniform_arithmetic +#error "Missing cl_khr_subgroup_non_uniform_arithmetic define" +#endif +#else +#ifdef cl_khr_subgroup_non_uniform_arithmetic +#error "Incorrect cl_khr_subgroup_non_uniform_arithmetic define" +#endif +// expected-warning@+2{{unsupported OpenCL extension 'cl_khr_subgroup_non_uniform_arithmetic' - ignoring}} +#endif +#pragma OPENCL EXTENSION cl_khr_subgroup_non_uniform_arithmetic : enable + +#if (defined(__OPENCL_CPP_VERSION__) || __OPENCL_C_VERSION__ >= 200) +#ifndef cl_khr_subgroup_non_uniform_vote +#error "Missing cl_khr_subgroup_non_uniform_vote define" +#endif +#else +#ifdef cl_khr_subgroup_non_uniform_vote +#error "Incorrect cl_khr_subgroup_non_uniform_vote define" +#endif +// expected-warning@+2{{unsupported OpenCL extension 'cl_khr_subgroup_non_uniform_vote' - ignoring}} +#endif +#pragma OPENCL EXTENSION cl_khr_subgroup_non_uniform_vote : enable + +#if (defined(__OPENCL_CPP_VERSION__) || __OPENCL_C_VERSION__ >= 200) +#ifndef cl_khr_subgroup_rotate +#error "Missing cl_khr_subgroup_rotate define" +#endif +#else +#ifdef cl_khr_subgroup_rotate +#error "Incorrect cl_khr_subgroup_rotate define" +#endif +// expected-warning@+2{{unsupported OpenCL extension 'cl_khr_subgroup_rotate' - ignoring}} +#endif +#pragma OPENCL EXTENSION cl_khr_subgroup_rotate : enable + +#if (defined(__OPENCL_CPP_VERSION__) || __OPENCL_C_VERSION__ >= 200) +#ifndef cl_khr_subgroup_shuffle_relative +#error "Missing cl_khr_subgroup_shuffle_relative define" +#endif +#else +#ifdef cl_khr_subgroup_shuffle_relative +#error "Incorrect cl_khr_subgroup_shuffle_relative define" +#endif +// expected-warning@+2{{unsupported OpenCL extension 'cl_khr_subgroup_shuffle_relative' - ignoring}} +#endif +#pragma OPENCL EXTENSION cl_khr_subgroup_shuffle_relative : enable + +#if (defined(__OPENCL_CPP_VERSION__) || __OPENCL_C_VERSION__ >= 200) +#ifndef cl_khr_subgroup_shuffle +#error "Missing cl_khr_subgroup_shuffle define" +#endif +#else +#ifdef cl_khr_subgroup_shuffle +#error "Incorrect cl_khr_subgroup_shuffle define" +#endif +// expected-warning@+2{{unsupported OpenCL extension 'cl_khr_subgroup_shuffle' - ignoring}} +#endif +#pragma OPENCL EXTENSION cl_khr_subgroup_shuffle : enable + #ifndef cl_amd_media_ops #error "Missing cl_amd_media_ops define" #endif @@ -224,14 +368,62 @@ //expected-warning@-1{{OpenCL extension '__opencl_c_atomic_order_acq_rel' unknown or does not require pragma - ignoring}} #pragma OPENCL EXTENSION __opencl_c_atomic_order_seq_cst : disable //expected-warning@-1{{OpenCL extension '__opencl_c_atomic_order_seq_cst' unknown or does not require pragma - ignoring}} +#pragma OPENCL EXTENSION __opencl_c_atomic_scope_all_devices : disable +//expected-warning@-1{{OpenCL extension '__opencl_c_atomic_scope_all_devices' unknown or does not require pragma - ignoring}} +#pragma OPENCL EXTENSION __opencl_c_atomic_scope_device : disable +//expected-warning@-1{{OpenCL extension '__opencl_c_atomic_scope_device' unknown or does not require pragma - ignoring}} #pragma OPENCL EXTENSION __opencl_c_device_enqueue : disable //expected-warning@-1{{OpenCL extension '__opencl_c_device_enqueue' unknown or does not require pragma - ignoring}} +#pragma OPENCL EXTENSION __opencl_c_ext_fp16_global_atomic_add : disable +//expected-warning@-1{{OpenCL extension '__opencl_c_ext_fp16_global_atomic_add' unknown or does not require pragma - ignoring}} +#pragma OPENCL EXTENSION __opencl_c_ext_fp16_global_atomic_load_store : disable +//expected-warning@-1{{OpenCL extension '__opencl_c_ext_fp16_global_atomic_load_store' unknown or does not require pragma - ignoring}} +#pragma OPENCL EXTENSION __opencl_c_ext_fp16_global_atomic_min_max : disable +//expected-warning@-1{{OpenCL extension '__opencl_c_ext_fp16_global_atomic_min_max' unknown or does not require pragma - ignoring}} +#pragma OPENCL EXTENSION __opencl_c_ext_fp16_local_atomic_add : disable +//expected-warning@-1{{OpenCL extension '__opencl_c_ext_fp16_local_atomic_add' unknown or does not require pragma - ignoring}}... [truncated] 

[llvmbot]

@llvm/pr-subscribers-backend-x86

Author: Wenju He (wenju-he)

Changes

Adds the remaining optional feature macros from the OpenCL C 3.0 spec (section 6.2.1 table). Targets can now enable these via OpenCLFeaturesMap returned by getSupportedOpenCLOpts().

---

Patch is 39.74 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/168016.diff

6 Files Affected:

• (modified) clang/include/clang/Basic/OpenCLExtensions.def (+46-9)
• (modified) clang/lib/Basic/Targets/AMDGPU.h (+2)
• (modified) clang/lib/Headers/opencl-c-base.h (-99)
• (modified) clang/test/Headers/opencl-c-header.cl (+3-3)
• (modified) clang/test/SemaOpenCL/extension-version.cl (+244)
• (modified) clang/test/SemaOpenCL/features.cl (+66-34)

diff --git a/clang/include/clang/Basic/OpenCLExtensions.def b/clang/include/clang/Basic/OpenCLExtensions.def index 6f73b26137500..9fc8c31056f4d 100644 --- a/clang/include/clang/Basic/OpenCLExtensions.def +++ b/clang/include/clang/Basic/OpenCLExtensions.def @@ -78,10 +78,55 @@ OPENCL_EXTENSION(cl_khr_depth_images, true, 120) OPENCL_EXTENSION(cl_khr_gl_msaa_sharing,true, 120) // OpenCL 2.0. +OPENCL_EXTENSION(cl_ext_float_atomics, true, 200) +OPENCL_EXTENSION(cl_khr_extended_bit_ops, true, 200) +OPENCL_EXTENSION(cl_khr_integer_dot_product, true, 200) +OPENCL_EXTENSION(cl_khr_kernel_clock, true, 200) OPENCL_EXTENSION(cl_khr_mipmap_image, true, 200) OPENCL_EXTENSION(cl_khr_mipmap_image_writes, true, 200) OPENCL_EXTENSION(cl_khr_srgb_image_writes, true, 200) +OPENCL_EXTENSION(cl_khr_subgroup_ballot, true, 200) +OPENCL_EXTENSION(cl_khr_subgroup_clustered_reduce, true, 200) +OPENCL_EXTENSION(cl_khr_subgroup_extended_types, true, 200) +OPENCL_EXTENSION(cl_khr_subgroup_non_uniform_arithmetic, true, 200) +OPENCL_EXTENSION(cl_khr_subgroup_non_uniform_vote, true, 200) +OPENCL_EXTENSION(cl_khr_subgroup_rotate, true, 200) +OPENCL_EXTENSION(cl_khr_subgroup_shuffle_relative, true, 200) +OPENCL_EXTENSION(cl_khr_subgroup_shuffle, true, 200) OPENCL_EXTENSION(cl_khr_subgroups, true, 200) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_atomic_order_acq_rel, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_atomic_order_seq_cst, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_atomic_scope_all_devices, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_atomic_scope_device, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_device_enqueue, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_ext_fp16_global_atomic_add, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_ext_fp16_global_atomic_load_store, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_ext_fp16_global_atomic_min_max, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_ext_fp16_local_atomic_add, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_ext_fp16_local_atomic_load_store, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_ext_fp16_local_atomic_min_max, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_ext_fp32_global_atomic_add, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_ext_fp32_global_atomic_min_max, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_ext_fp32_local_atomic_add, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_ext_fp32_local_atomic_min_max, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_ext_fp64_global_atomic_add, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_ext_fp64_global_atomic_min_max, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_ext_fp64_local_atomic_add, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_ext_fp64_local_atomic_min_max, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_ext_image_raw10_raw12, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_ext_image_unorm_int_2_101010, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_ext_image_unsigned_10x6_12x4_14x2, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_generic_address_space, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_images, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_integer_dot_product_input_4x8bit, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_integer_dot_product_input_4x8bit_packed, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_kernel_clock_scope_device, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_kernel_clock_scope_sub_group, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_kernel_clock_scope_work_group, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_pipes, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_program_scope_global_variables, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_read_write_images, false, 200, OCL_C_20) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_work_group_collective_functions, false, 200, OCL_C_20) // Clang Extensions. OPENCL_EXTENSION(cl_clang_storage_class_specifiers, true, 100) @@ -100,17 +145,9 @@ OPENCL_EXTENSION(cl_intel_subgroups_short, true, 120) OPENCL_EXTENSION(cl_intel_device_side_avc_motion_estimation, true, 120) // OpenCL C 3.0 features (6.2.1. Features) -OPENCL_OPTIONALCOREFEATURE(__opencl_c_pipes, false, 300, OCL_C_30) -OPENCL_OPTIONALCOREFEATURE(__opencl_c_generic_address_space, false, 300, OCL_C_30) -OPENCL_OPTIONALCOREFEATURE(__opencl_c_atomic_order_acq_rel, false, 300, OCL_C_30) -OPENCL_OPTIONALCOREFEATURE(__opencl_c_atomic_order_seq_cst, false, 300, OCL_C_30) -OPENCL_OPTIONALCOREFEATURE(__opencl_c_subgroups, false, 300, OCL_C_30) OPENCL_OPTIONALCOREFEATURE(__opencl_c_3d_image_writes, false, 300, OCL_C_30) -OPENCL_OPTIONALCOREFEATURE(__opencl_c_device_enqueue, false, 300, OCL_C_30) -OPENCL_OPTIONALCOREFEATURE(__opencl_c_read_write_images, false, 300, OCL_C_30) -OPENCL_OPTIONALCOREFEATURE(__opencl_c_program_scope_global_variables, false, 300, OCL_C_30) OPENCL_OPTIONALCOREFEATURE(__opencl_c_fp64, false, 300, OCL_C_30) -OPENCL_OPTIONALCOREFEATURE(__opencl_c_images, false, 300, OCL_C_30) +OPENCL_OPTIONALCOREFEATURE(__opencl_c_subgroups, false, 300, OCL_C_30) #undef OPENCL_OPTIONALCOREFEATURE #undef OPENCL_COREFEATURE diff --git a/clang/lib/Basic/Targets/AMDGPU.h b/clang/lib/Basic/Targets/AMDGPU.h index dfcc79402257a..95a3d133c07ac 100644 --- a/clang/lib/Basic/Targets/AMDGPU.h +++ b/clang/lib/Basic/Targets/AMDGPU.h @@ -320,6 +320,8 @@ class LLVM_LIBRARY_VISIBILITY AMDGPUTargetInfo final : public TargetInfo { Opts["__opencl_c_3d_image_writes"] = true; Opts["cl_khr_3d_image_writes"] = true; Opts["__opencl_c_program_scope_global_variables"] = true; + Opts["__opencl_c_atomic_scope_all_devices"] = true; + Opts["__opencl_c_atomic_order_seq_cst"] = true; if (GPUKind >= llvm::AMDGPU::GK_GFX700) { Opts["__opencl_c_generic_address_space"] = true; diff --git a/clang/lib/Headers/opencl-c-base.h b/clang/lib/Headers/opencl-c-base.h index 414f10ad832ce..898026c66614a 100644 --- a/clang/lib/Headers/opencl-c-base.h +++ b/clang/lib/Headers/opencl-c-base.h @@ -9,105 +9,6 @@ #ifndef _OPENCL_BASE_H_ #define _OPENCL_BASE_H_ -// Define extension macros - -#if (defined(__OPENCL_CPP_VERSION__) || __OPENCL_C_VERSION__ >= 200) -// For SPIR and SPIR-V all extensions are supported. -#if defined(__SPIR__) || defined(__SPIRV__) -#define cl_khr_subgroup_extended_types 1 -#define cl_khr_subgroup_non_uniform_vote 1 -#define cl_khr_subgroup_ballot 1 -#define cl_khr_subgroup_non_uniform_arithmetic 1 -#define cl_khr_subgroup_shuffle 1 -#define cl_khr_subgroup_shuffle_relative 1 -#define cl_khr_subgroup_clustered_reduce 1 -#define cl_khr_subgroup_rotate 1 -#define cl_khr_extended_bit_ops 1 -#define cl_khr_integer_dot_product 1 -#define __opencl_c_integer_dot_product_input_4x8bit 1 -#define __opencl_c_integer_dot_product_input_4x8bit_packed 1 -#define cl_ext_float_atomics 1 -#ifdef cl_khr_fp16 -#define __opencl_c_ext_fp16_global_atomic_load_store 1 -#define __opencl_c_ext_fp16_local_atomic_load_store 1 -#define __opencl_c_ext_fp16_global_atomic_add 1 -#define __opencl_c_ext_fp16_local_atomic_add 1 -#define __opencl_c_ext_fp16_global_atomic_min_max 1 -#define __opencl_c_ext_fp16_local_atomic_min_max 1 -#endif -#ifdef cl_khr_fp64 -#define __opencl_c_ext_fp64_global_atomic_add 1 -#define __opencl_c_ext_fp64_local_atomic_add 1 -#define __opencl_c_ext_fp64_global_atomic_min_max 1 -#define __opencl_c_ext_fp64_local_atomic_min_max 1 -#endif -#define __opencl_c_ext_fp32_global_atomic_add 1 -#define __opencl_c_ext_fp32_local_atomic_add 1 -#define __opencl_c_ext_fp32_global_atomic_min_max 1 -#define __opencl_c_ext_fp32_local_atomic_min_max 1 -#define __opencl_c_ext_image_raw10_raw12 1 -#define __opencl_c_ext_image_unorm_int_2_101010 1 -#define __opencl_c_ext_image_unsigned_10x6_12x4_14x2 1 -#define cl_khr_kernel_clock 1 -#define __opencl_c_kernel_clock_scope_device 1 -#define __opencl_c_kernel_clock_scope_work_group 1 -#define __opencl_c_kernel_clock_scope_sub_group 1 - -#endif // defined(__SPIR__) || defined(__SPIRV__) -#endif // (defined(__OPENCL_CPP_VERSION__) || __OPENCL_C_VERSION__ >= 200) - -// Define feature macros for OpenCL C 2.0 -#if (__OPENCL_CPP_VERSION__ == 100 || __OPENCL_C_VERSION__ == 200) -#define __opencl_c_pipes 1 -#define __opencl_c_generic_address_space 1 -#define __opencl_c_work_group_collective_functions 1 -#define __opencl_c_atomic_order_acq_rel 1 -#define __opencl_c_atomic_order_seq_cst 1 -#define __opencl_c_atomic_scope_device 1 -#define __opencl_c_atomic_scope_all_devices 1 -#define __opencl_c_device_enqueue 1 -#define __opencl_c_read_write_images 1 -#define __opencl_c_program_scope_global_variables 1 -#define __opencl_c_images 1 -#endif - -// Define header-only feature macros for OpenCL C 3.0. -#if (__OPENCL_CPP_VERSION__ == 202100 || __OPENCL_C_VERSION__ == 300) -// For the SPIR and SPIR-V target all features are supported. -#if defined(__SPIR__) || defined(__SPIRV__) -#define __opencl_c_work_group_collective_functions 1 -#define __opencl_c_atomic_order_seq_cst 1 -#define __opencl_c_atomic_scope_device 1 -#define __opencl_c_atomic_scope_all_devices 1 -#define __opencl_c_read_write_images 1 -#endif // defined(__SPIR__) - -#endif // (__OPENCL_CPP_VERSION__ == 202100 || __OPENCL_C_VERSION__ == 300) - -// Undefine any feature macros that have been explicitly disabled using -// an __undef_<feature> macro. -#ifdef __undef___opencl_c_work_group_collective_functions -#undef __opencl_c_work_group_collective_functions -#endif -#ifdef __undef___opencl_c_atomic_order_seq_cst -#undef __opencl_c_atomic_order_seq_cst -#endif -#ifdef __undef___opencl_c_atomic_scope_device -#undef __opencl_c_atomic_scope_device -#endif -#ifdef __undef___opencl_c_atomic_scope_all_devices -#undef __opencl_c_atomic_scope_all_devices -#endif -#ifdef __undef___opencl_c_read_write_images -#undef __opencl_c_read_write_images -#endif -#ifdef __undef___opencl_c_integer_dot_product_input_4x8bit -#undef __opencl_c_integer_dot_product_input_4x8bit -#endif -#ifdef __undef___opencl_c_integer_dot_product_input_4x8bit_packed -#undef __opencl_c_integer_dot_product_input_4x8bit_packed -#endif - #if !defined(__opencl_c_generic_address_space) // Internal feature macro to provide named (global, local, private) address // space overloads for builtin functions that take a pointer argument. diff --git a/clang/test/Headers/opencl-c-header.cl b/clang/test/Headers/opencl-c-header.cl index 17cbb67f26038..e26f16827b20f 100644 --- a/clang/test/Headers/opencl-c-header.cl +++ b/clang/test/Headers/opencl-c-header.cl @@ -33,7 +33,7 @@ // === // Compile for OpenCL 2.0 for the first time. The module should change. // RUN: %clang_cc1 -triple spir-unknown-unknown -O0 -emit-llvm -o - -cl-std=CL2.0 -finclude-default-header -fmodules -fimplicit-module-maps -fmodules-cache-path=%t -fdisable-module-hash -ftime-report %s 2>&1 | FileCheck --check-prefix=CHECK20 --check-prefix=CHECK-MOD %s -// RUN: not diff %t/1_0.pcm %t/opencl_c.pcm +// RUN: not diff %t/1_0.pcm %t/opencl_c.pcm > /dev/null // RUN: chmod u-w %t/opencl_c.pcm // === @@ -44,10 +44,10 @@ // RUN: rm -rf %t // RUN: mkdir -p %t // RUN: %clang_cc1 -triple spir64-unknown-unknown -emit-llvm -o - -cl-std=CL1.2 -finclude-default-header -fmodules -fimplicit-module-maps -fmodules-cache-path=%t -ftime-report %s 2>&1 | FileCheck --check-prefix=CHECK --check-prefix=CHECK-MOD %s -// RUN: %clang_cc1 -triple amdgcn--amdhsa -O0 -emit-llvm -o - -cl-std=CL2.0 -finclude-default-header -fmodules -fimplicit-module-maps -fmodules-cache-path=%t -ftime-report %s 2>&1 | FileCheck --check-prefix=CHECK20 --check-prefix=CHECK-MOD %s +// RUN: %clang_cc1 -triple amdgcn--amdhsa -target-cpu gfx700 -O0 -emit-llvm -o - -cl-std=CL2.0 -finclude-default-header -fmodules -fimplicit-module-maps -fmodules-cache-path=%t -ftime-report %s 2>&1 | FileCheck --check-prefix=CHECK20 --check-prefix=CHECK-MOD %s // RUN: chmod u-w %t // RUN: %clang_cc1 -triple spir64-unknown-unknown -emit-llvm -o - -cl-std=CL1.2 -finclude-default-header -fmodules -fimplicit-module-maps -fmodules-cache-path=%t -ftime-report %s 2>&1 | FileCheck --check-prefix=CHECK --check-prefix=CHECK-MOD %s -// RUN: %clang_cc1 -triple amdgcn--amdhsa -O0 -emit-llvm -o - -cl-std=CL2.0 -finclude-default-header -fmodules -fimplicit-module-maps -fmodules-cache-path=%t -ftime-report %s 2>&1 | FileCheck --check-prefix=CHECK20 --check-prefix=CHECK-MOD %s +// RUN: %clang_cc1 -triple amdgcn--amdhsa -target-cpu gfx700 -O0 -emit-llvm -o - -cl-std=CL2.0 -finclude-default-header -fmodules -fimplicit-module-maps -fmodules-cache-path=%t -ftime-report %s 2>&1 | FileCheck --check-prefix=CHECK20 --check-prefix=CHECK-MOD %s // RUN: chmod u+w %t // Verify that called builtins occur in the generated IR. diff --git a/clang/test/SemaOpenCL/extension-version.cl b/clang/test/SemaOpenCL/extension-version.cl index 4d92eff6ae3ac..35a1818706f35 100644 --- a/clang/test/SemaOpenCL/extension-version.cl +++ b/clang/test/SemaOpenCL/extension-version.cl @@ -165,6 +165,150 @@ #endif #pragma OPENCL EXTENSION cl_khr_subgroups : enable +#if (defined(__OPENCL_CPP_VERSION__) || __OPENCL_C_VERSION__ >= 200) +#ifndef cl_ext_float_atomics +#error "Missing cl_ext_float_atomics define" +#endif +#else +#ifdef cl_ext_float_atomics +#error "Incorrect cl_ext_float_atomics define" +#endif +// expected-warning@+2{{unsupported OpenCL extension 'cl_ext_float_atomics' - ignoring}} +#endif +#pragma OPENCL EXTENSION cl_ext_float_atomics : enable + +#if (defined(__OPENCL_CPP_VERSION__) || __OPENCL_C_VERSION__ >= 200) +#ifndef cl_khr_extended_bit_ops +#error "Missing cl_khr_extended_bit_ops define" +#endif +#else +#ifdef cl_khr_extended_bit_ops +#error "Incorrect cl_khr_extended_bit_ops define" +#endif +// expected-warning@+2{{unsupported OpenCL extension 'cl_khr_extended_bit_ops' - ignoring}} +#endif +#pragma OPENCL EXTENSION cl_khr_extended_bit_ops : enable + +#if (defined(__OPENCL_CPP_VERSION__) || __OPENCL_C_VERSION__ >= 200) +#ifndef cl_khr_integer_dot_product +#error "Missing cl_khr_integer_dot_product define" +#endif +#else +#ifdef cl_khr_integer_dot_product +#error "Incorrect cl_khr_integer_dot_product define" +#endif +// expected-warning@+2{{unsupported OpenCL extension 'cl_khr_integer_dot_product' - ignoring}} +#endif +#pragma OPENCL EXTENSION cl_khr_integer_dot_product : enable + +#if (defined(__OPENCL_CPP_VERSION__) || __OPENCL_C_VERSION__ >= 200) +#ifndef cl_khr_kernel_clock +#error "Missing cl_khr_kernel_clock define" +#endif +#else +#ifdef cl_khr_kernel_clock +#error "Incorrect cl_khr_kernel_clock define" +#endif +// expected-warning@+2{{unsupported OpenCL extension 'cl_khr_kernel_clock' - ignoring}} +#endif +#pragma OPENCL EXTENSION cl_khr_kernel_clock : enable + +#if (defined(__OPENCL_CPP_VERSION__) || __OPENCL_C_VERSION__ >= 200) +#ifndef cl_khr_subgroup_ballot +#error "Missing cl_khr_subgroup_ballot define" +#endif +#else +#ifdef cl_khr_subgroup_ballot +#error "Incorrect cl_khr_subgroup_ballot define" +#endif +// expected-warning@+2{{unsupported OpenCL extension 'cl_khr_subgroup_ballot' - ignoring}} +#endif +#pragma OPENCL EXTENSION cl_khr_subgroup_ballot : enable + +#if (defined(__OPENCL_CPP_VERSION__) || __OPENCL_C_VERSION__ >= 200) +#ifndef cl_khr_subgroup_clustered_reduce +#error "Missing cl_khr_subgroup_clustered_reduce define" +#endif +#else +#ifdef cl_khr_subgroup_clustered_reduce +#error "Incorrect cl_khr_subgroup_clustered_reduce define" +#endif +// expected-warning@+2{{unsupported OpenCL extension 'cl_khr_subgroup_clustered_reduce' - ignoring}} +#endif +#pragma OPENCL EXTENSION cl_khr_subgroup_clustered_reduce : enable + +#if (defined(__OPENCL_CPP_VERSION__) || __OPENCL_C_VERSION__ >= 200) +#ifndef cl_khr_subgroup_extended_types +#error "Missing cl_khr_subgroup_extended_types define" +#endif +#else +#ifdef cl_khr_subgroup_extended_types +#error "Incorrect cl_khr_subgroup_extended_types define" +#endif +// expected-warning@+2{{unsupported OpenCL extension 'cl_khr_subgroup_extended_types' - ignoring}} +#endif +#pragma OPENCL EXTENSION cl_khr_subgroup_extended_types : enable + +#if (defined(__OPENCL_CPP_VERSION__) || __OPENCL_C_VERSION__ >= 200) +#ifndef cl_khr_subgroup_non_uniform_arithmetic +#error "Missing cl_khr_subgroup_non_uniform_arithmetic define" +#endif +#else +#ifdef cl_khr_subgroup_non_uniform_arithmetic +#error "Incorrect cl_khr_subgroup_non_uniform_arithmetic define" +#endif +// expected-warning@+2{{unsupported OpenCL extension 'cl_khr_subgroup_non_uniform_arithmetic' - ignoring}} +#endif +#pragma OPENCL EXTENSION cl_khr_subgroup_non_uniform_arithmetic : enable + +#if (defined(__OPENCL_CPP_VERSION__) || __OPENCL_C_VERSION__ >= 200) +#ifndef cl_khr_subgroup_non_uniform_vote +#error "Missing cl_khr_subgroup_non_uniform_vote define" +#endif +#else +#ifdef cl_khr_subgroup_non_uniform_vote +#error "Incorrect cl_khr_subgroup_non_uniform_vote define" +#endif +// expected-warning@+2{{unsupported OpenCL extension 'cl_khr_subgroup_non_uniform_vote' - ignoring}} +#endif +#pragma OPENCL EXTENSION cl_khr_subgroup_non_uniform_vote : enable + +#if (defined(__OPENCL_CPP_VERSION__) || __OPENCL_C_VERSION__ >= 200) +#ifndef cl_khr_subgroup_rotate +#error "Missing cl_khr_subgroup_rotate define" +#endif +#else +#ifdef cl_khr_subgroup_rotate +#error "Incorrect cl_khr_subgroup_rotate define" +#endif +// expected-warning@+2{{unsupported OpenCL extension 'cl_khr_subgroup_rotate' - ignoring}} +#endif +#pragma OPENCL EXTENSION cl_khr_subgroup_rotate : enable + +#if (defined(__OPENCL_CPP_VERSION__) || __OPENCL_C_VERSION__ >= 200) +#ifndef cl_khr_subgroup_shuffle_relative +#error "Missing cl_khr_subgroup_shuffle_relative define" +#endif +#else +#ifdef cl_khr_subgroup_shuffle_relative +#error "Incorrect cl_khr_subgroup_shuffle_relative define" +#endif +// expected-warning@+2{{unsupported OpenCL extension 'cl_khr_subgroup_shuffle_relative' - ignoring}} +#endif +#pragma OPENCL EXTENSION cl_khr_subgroup_shuffle_relative : enable + +#if (defined(__OPENCL_CPP_VERSION__) || __OPENCL_C_VERSION__ >= 200) +#ifndef cl_khr_subgroup_shuffle +#error "Missing cl_khr_subgroup_shuffle define" +#endif +#else +#ifdef cl_khr_subgroup_shuffle +#error "Incorrect cl_khr_subgroup_shuffle define" +#endif +// expected-warning@+2{{unsupported OpenCL extension 'cl_khr_subgroup_shuffle' - ignoring}} +#endif +#pragma OPENCL EXTENSION cl_khr_subgroup_shuffle : enable + #ifndef cl_amd_media_ops #error "Missing cl_amd_media_ops define" #endif @@ -224,14 +368,62 @@ //expected-warning@-1{{OpenCL extension '__opencl_c_atomic_order_acq_rel' unknown or does not require pragma - ignoring}} #pragma OPENCL EXTENSION __opencl_c_atomic_order_seq_cst : disable //expected-warning@-1{{OpenCL extension '__opencl_c_atomic_order_seq_cst' unknown or does not require pragma - ignoring}} +#pragma OPENCL EXTENSION __opencl_c_atomic_scope_all_devices : disable +//expected-warning@-1{{OpenCL extension '__opencl_c_atomic_scope_all_devices' unknown or does not require pragma - ignoring}} +#pragma OPENCL EXTENSION __opencl_c_atomic_scope_device : disable +//expected-warning@-1{{OpenCL extension '__opencl_c_atomic_scope_device' unknown or does not require pragma - ignoring}} #pragma OPENCL EXTENSION __opencl_c_device_enqueue : disable //expected-warning@-1{{OpenCL extension '__opencl_c_device_enqueue' unknown or does not require pragma - ignoring}} +#pragma OPENCL EXTENSION __opencl_c_ext_fp16_global_atomic_add : disable +//expected-warning@-1{{OpenCL extension '__opencl_c_ext_fp16_global_atomic_add' unknown or does not require pragma - ignoring}} +#pragma OPENCL EXTENSION __opencl_c_ext_fp16_global_atomic_load_store : disable +//expected-warning@-1{{OpenCL extension '__opencl_c_ext_fp16_global_atomic_load_store' unknown or does not require pragma - ignoring}} +#pragma OPENCL EXTENSION __opencl_c_ext_fp16_global_atomic_min_max : disable +//expected-warning@-1{{OpenCL extension '__opencl_c_ext_fp16_global_atomic_min_max' unknown or does not require pragma - ignoring}} +#pragma OPENCL EXTENSION __opencl_c_ext_fp16_local_atomic_add : disable +//expected-warning@-1{{OpenCL extension '__opencl_c_ext_fp16_local_atomic_add' unknown or does not require pragma - ignoring}}... [truncated] 

[wenju-he]

I would expect clang/test/SemaOpenCL/features.cl tests will be extended to cover new features. Potentially clang/test/Headers/opencl-c-header.cl and clang/test/SemaOpenCL/extension-version.cl as well.

updated clang tests in da29d83 and 8118731

[github-actions]

🐧 Linux x64 Test Results

• 111323 tests passed
• 4417 tests skipped

[svenvh]

LGTM with some formatting nits.

cc: @arsenm for the AMDGPU.h changes.