NVPTX: Add supported CL features

[jvesely]

Summary

Fixes libclc compilation.
Taken from clinfo on one nVidia system.

[jvesely]

add test

[yaxunl]

I have a similar, but yours is more complete. I probably will drop my patch once yours goes through.

test/Misc/nvptx.languageOptsOpenCL.cl

2 ↗

(On Diff #57693)

-Werror is unnecessary, same as below

6 ↗

(On Diff #57693)

should be like

#pragma OPENCL EXTENSION cl_xxx : enable

same for below

[jholewinski]

lib/Basic/Targets.cpp

1849 ↗

(On Diff #57693)

Do you really intend to enable subgroup support?

[jvesely]

remove subgroups support.
better tests.

note: I had to add OPENCL_VERSION defines manually, these can be removed once clang correctly reports the version

[yaxunl]

note: I had to add OPENCL_VERSION defines manually, these can be removed once clang correctly reports the version

clang correctly generates OPENCL_C_VERSION and OPENCL_VERSION_x_y macros. You can use them instead. For CL 1.1/1.2/2.0, you need to use -cl-std=CLx.y to specify OpenCL version.

[jvesely]

even better tests
use existing OPENCL_C_VERSION
use -cl-std=CL for explicit OpenCL1.0

[jvesely]

Report only clinfo exported extensions.

[Anastasia]

test/Misc/nvptx.languageOptsOpenCL.cl

1 ↗

(On Diff #58849)

Should we check for errors on unsupported extensions for this target too?

[jvesely]

test/Misc/nvptx.languageOptsOpenCL.cl

1 ↗

(On Diff #58849)

I didn't want to do that. Including unsupported extensions in these tests would imply that every OCL extension needs test in 4 places (nvptx, amdgcn, r600, extension-version.cl), possibly more if more targets are added, which seemed excessive.

[jvesely]

Test all known extensions against expected nvptx outcome (add negative tests)

[jvesely]

add back require registered target

[jvesely]

ping

@jholewinski, is the exported list of features OK with you?

@Anastasia, are the tests OK now?

[jholewinski]

Looks good to me

cfe/trunk/lib/Basic/Targets.cpp

GPU = llvm::StringSwitch<GPUKind>(Name)
               .Case("sm_37", GK_SM37)
               .Case("sm_50", GK_SM50)
               .Case("sm_52", GK_SM52)
               .Case("sm_53", GK_SM53)
               .Default(GK_NONE);
 
     return GPU != GK_NONE;
   }
+  void setSupportedOpenCLOpts() override {
+    auto &Opts = getSupportedOpenCLOpts();
+    Opts.cl_clang_storage_class_specifiers = 1;
+    Opts.cl_khr_gl_sharing = 1;
+    Opts.cl_khr_icd = 1;
+
+    Opts.cl_khr_fp64 = 1;
+    Opts.cl_khr_byte_addressable_store = 1;
+    Opts.cl_khr_global_int32_base_atomics = 1;
+    Opts.cl_khr_global_int32_extended_atomics = 1;
+    Opts.cl_khr_local_int32_base_atomics = 1;
+    Opts.cl_khr_local_int32_extended_atomics = 1;
+  }
 };
 
 const Builtin::Info NVPTXTargetInfo::BuiltinInfo[] = {
 #define BUILTIN(ID, TYPE, ATTRS)                                               \
   { #ID, TYPE, ATTRS, nullptr, ALL_LANGUAGES, nullptr },
 #define LIBBUILTIN(ID, TYPE, ATTRS, HEADER)                                    \
   { #ID, TYPE, ATTRS, HEADER, ALL_LANGUAGES, nullptr },
 #include "clang/Basic/BuiltinsNVPTX.def"

cfe/trunk/test/Misc/nvptx.languageOptsOpenCL.cl

+// REQUIRES: nvptx-registered-target
+// RUN: %clang_cc1 -x cl -cl-std=CL %s -verify -triple nvptx-unknown-unknown
+// RUN: %clang_cc1 -x cl -cl-std=CL1.1 %s -verify -triple nvptx-unknown-unknown
+// RUN: %clang_cc1 -x cl -cl-std=CL1.2 %s -verify -triple nvptx-unknown-unknown
+// RUN: %clang_cc1 -x cl -cl-std=CL2.0 %s -verify -triple nvptx-unknown-unknown
+// RUN: %clang_cc1 -x cl -cl-std=CL %s -verify -triple nvptx-unknown-unknown -Wpedantic-core-features -DTEST_CORE_FEATURES
+// RUN: %clang_cc1 -x cl -cl-std=CL1.1 %s -verify -triple nvptx-unknown-unknown -Wpedantic-core-features -DTEST_CORE_FEATURES
+// RUN: %clang_cc1 -x cl -cl-std=CL1.2 %s -verify -triple nvptx-unknown-unknown -Wpedantic-core-features -DTEST_CORE_FEATURES
+// RUN: %clang_cc1 -x cl -cl-std=CL2.0 %s -verify -triple nvptx-unknown-unknown -Wpedantic-core-features -DTEST_CORE_FEATURES
+// RUN: %clang_cc1 -x cl -cl-std=CL %s -verify -triple nvptx64-unknown-unknown
+// RUN: %clang_cc1 -x cl -cl-std=CL1.1 %s -verify -triple nvptx64-unknown-unknown
+// RUN: %clang_cc1 -x cl -cl-std=CL1.2 %s -verify -triple nvptx64-unknown-unknown
+// RUN: %clang_cc1 -x cl -cl-std=CL2.0 %s -verify -triple nvptx64-unknown-unknown
+// RUN: %clang_cc1 -x cl -cl-std=CL %s -verify -triple nvptx64-unknown-unknown -Wpedantic-core-features -DTEST_CORE_FEATURES
+// RUN: %clang_cc1 -x cl -cl-std=CL1.1 %s -verify -triple nvptx64-unknown-unknown -Wpedantic-core-features -DTEST_CORE_FEATURES
+// RUN: %clang_cc1 -x cl -cl-std=CL1.2 %s -verify -triple nvptx64-unknown-unknown -Wpedantic-core-features -DTEST_CORE_FEATURES
+// RUN: %clang_cc1 -x cl -cl-std=CL2.0 %s -verify -triple nvptx64-unknown-unknown -Wpedantic-core-features -DTEST_CORE_FEATURES
+
+// Extensions in all versions
+#ifndef cl_clang_storage_class_specifiers
+#error "Missing cl_clang_storage_class_specifiers define"
+#endif
+#pragma OPENCL EXTENSION cl_clang_storage_class_specifiers: enable
+
+#ifdef cl_khr_fp16
+#error "Incorrect cl_khr_fp16 define"
+#endif
+#pragma OPENCL EXTENSION cl_khr_fp16: enable
+// expected-warning@-1{{unsupported OpenCL extension 'cl_khr_fp16' - ignoring}}
+
+#ifdef cl_khr_int64_base_atomics
+#error "Incorrect cl_khr_int64_base_atomics define"
+#endif
+#pragma OPENCL EXTENSION cl_khr_int64_base_atomics: enable
+// expected-warning@-1{{unsupported OpenCL extension 'cl_khr_int64_base_atomics' - ignoring}}
+
+#ifdef cl_khr_int64_extended_atomics
+#error "Incorrect cl_khr_int64_extended_atomics define"
+#endif
+#pragma OPENCL EXTENSION cl_khr_int64_extended_atomics: enable
+// expected-warning@-1{{unsupported OpenCL extension 'cl_khr_int64_extended_atomics' - ignoring}}
+
+#ifndef cl_khr_gl_sharing
+#error "Missing cl_khr_gl_sharing define"
+#endif
+#pragma OPENCL EXTENSION cl_khr_gl_sharing: enable
+
+#ifndef cl_khr_icd
+#error "Missing cl_khr_icd define"
+#endif
+#pragma OPENCL EXTENSION cl_khr_icd: enable
+
+// Core features in CL 1.1
+
+#ifndef cl_khr_byte_addressable_store
+#error "Missing cl_khr_byte_addressable_store define"
+#endif
+#pragma OPENCL EXTENSION cl_khr_byte_addressable_store: enable
+#if (__OPENCL_C_VERSION__ >= 110) && defined TEST_CORE_FEATURES
+// expected-warning@-2{{OpenCL extension 'cl_khr_byte_addressable_store' is core feature or supported optional core feature - ignoring}}
+#endif
+
+#ifndef cl_khr_global_int32_base_atomics
+#error "Missing cl_khr_global_int32_base_atomics define"
+#endif
+#pragma OPENCL EXTENSION cl_khr_global_int32_base_atomics: enable
+#if (__OPENCL_C_VERSION__ >= 110) && defined TEST_CORE_FEATURES
+// expected-warning@-2{{OpenCL extension 'cl_khr_global_int32_base_atomics' is core feature or supported optional core feature - ignoring}}
+#endif
+
+#ifndef cl_khr_global_int32_extended_atomics
+#error "Missing cl_khr_global_int32_extended_atomics define"
+#endif
+#pragma OPENCL EXTENSION cl_khr_global_int32_extended_atomics: enable
+#if (__OPENCL_C_VERSION__ >= 110) && defined TEST_CORE_FEATURES
+// expected-warning@-2{{OpenCL extension 'cl_khr_global_int32_extended_atomics' is core feature or supported optional core feature - ignoring}}
+#endif
+
+#ifndef cl_khr_local_int32_base_atomics
+#error "Missing cl_khr_local_int32_base_atomics define"
+#endif
+#pragma OPENCL EXTENSION cl_khr_local_int32_base_atomics: enable
+#if (__OPENCL_C_VERSION__ >= 110) && defined TEST_CORE_FEATURES
+// expected-warning@-2{{OpenCL extension 'cl_khr_local_int32_base_atomics' is core feature or supported optional core feature - ignoring}}
+#endif
+
+#ifndef cl_khr_local_int32_extended_atomics
+#error "Missing cl_khr_local_int32_extended_atomics define"
+#endif
+#pragma OPENCL EXTENSION cl_khr_local_int32_extended_atomics: enable
+#if (__OPENCL_C_VERSION__ >= 110) && defined TEST_CORE_FEATURES
+// expected-warning@-2{{OpenCL extension 'cl_khr_local_int32_extended_atomics' is core feature or supported optional core feature - ignoring}}
+#endif
+
+#if (__OPENCL_C_VERSION__ < 110)
+// Deprecated above 1.0
+#ifdef cl_khr_select_fprounding_mode
+#error "Incorrect cl_khr_select_fprounding_mode define"
+#endif
+#pragma OPENCL EXTENSION cl_khr_select_fprounding_mode: enable
+// expected-warning@-1{{unsupported OpenCL extension 'cl_khr_select_fprounding_mode' - ignoring}}
+#endif
+
+
+// Core feature in CL 1.2
+#ifndef cl_khr_fp64
+#error "Missing cl_khr_fp64 define"
+#endif
+#pragma OPENCL EXTENSION cl_khr_fp64: enable
+#if (__OPENCL_C_VERSION__ >= 120) && defined TEST_CORE_FEATURES
+// expected-warning@-2{{OpenCL extension 'cl_khr_fp64' is core feature or supported optional core feature - ignoring}}
+#endif
+
+// Core feature in CL 2.0, but not supported on nvptx
+#ifdef cl_khr_3d_image_writes
+#error "Incorrect cl_khr_3d_image_writes define"
+#endif
+#pragma OPENCL EXTENSION cl_khr_3d_image_writes: enable
+// expected-warning@-1{{unsupported OpenCL extension 'cl_khr_3d_image_writes' - ignoring}}
+
+
+
+#ifdef cl_khr_gl_event
+#error "Incorrect cl_khr_gl_event define"
+#endif
+#pragma OPENCL EXTENSION cl_khr_gl_event: enable
+// expected-warning@-1{{unsupported OpenCL extension 'cl_khr_gl_event' - ignoring}}
+
+#ifdef cl_khr_d3d10_sharing
+#error "Incorrect cl_khr_d3d10_sharing define"
+#endif
+#pragma OPENCL EXTENSION cl_khr_d3d10_sharing: enable
+// expected-warning@-1{{unsupported OpenCL extension 'cl_khr_d3d10_sharing' - ignoring}}
+
+#ifdef cl_khr_context_abort
+#error "Incorrect cl_context_abort define"
+#endif
+#pragma OPENCL EXTENSION cl_khr_context_abort: enable
+// expected-warning@-1{{unsupported OpenCL extension 'cl_khr_context_abort' - ignoring}}
+
+#ifdef cl_khr_d3d11_sharing
+#error "Incorrect cl_khr_d3d11_sharing define"
+#endif
+#pragma OPENCL EXTENSION cl_khr_d3d11_sharing: enable
+// expected-warning@-1{{unsupported OpenCL extension 'cl_khr_d3d11_sharing' - ignoring}}
+
+#ifdef cl_khr_dx9_media_sharing
+#error "Incorrect cl_khr_dx9_media_sharing define"
+#endif
+#pragma OPENCL EXTENSION cl_khr_dx9_media_sharing: enable
+// expected-warning@-1{{unsupported OpenCL extension 'cl_khr_dx9_media_sharing' - ignoring}}
+
+#ifdef cl_khr_image2d_from_buffer
+#error "Incorrect cl_khr_image2d_from_buffer define"
+#endif
+#pragma OPENCL EXTENSION cl_khr_image2d_from_buffer: enable
+// expected-warning@-1{{unsupported OpenCL extension 'cl_khr_image2d_from_buffer' - ignoring}}
+
+#ifdef cl_khr_initialize_memory
+#error "Incorrect cl_khr_initialize_memory define"
+#endif
+#pragma OPENCL EXTENSION cl_khr_initialize_memory: enable
+// expected-warning@-1{{unsupported OpenCL extension 'cl_khr_initialize_memory' - ignoring}}
+
+#ifdef cl_khr_gl_depth_images
+#error "Incorrect cl_khr_gl_depth_images define"
+#endif
+#pragma OPENCL EXTENSION cl_khr_gl_depth_images: enable
+// expected-warning@-1{{unsupported OpenCL extension 'cl_khr_gl_depth_images' - ignoring}}
+
+#ifdef cl_khr_gl_msaa_sharing
+#error "Incorrect cl_khr_gl_msaa_sharing define"
+#endif
+#pragma OPENCL EXTENSION cl_khr_gl_msaa_sharing: enable
+// expected-warning@-1{{unsupported OpenCL extension 'cl_khr_gl_msaa_sharing' - ignoring}}
+
+#ifdef cl_khr_spir
+#error "Incorrect cl_khr_spir define"
+#endif
+#pragma OPENCL EXTENSION cl_khr_spir: enable
+// expected-warning@-1{{unsupported OpenCL extension 'cl_khr_spir' - ignoring}}
+
+#ifdef cl_khr_egl_event
+#error "Incorrect cl_khr_egl_event define"
+#endif
+#pragma OPENCL EXTENSION cl_khr_egl_event: enable
+// expected-warning@-1{{unsupported OpenCL extension 'cl_khr_egl_event' - ignoring}}
+
+#ifdef cl_khr_egl_image
+#error "Missing cl_khr_egl_image define"
+#endif
+#pragma OPENCL EXTENSION cl_khr_egl_image: enable
+// expected-warning@-1{{unsupported OpenCL extension 'cl_khr_egl_image' - ignoring}}
+
+#ifdef cl_khr_srgb_image_writes
+#error "Incorrect cl_khr_srgb_image_writes define"
+#endif
+#pragma OPENCL EXTENSION cl_khr_srgb_image_writes: enable
+// expected-warning@-1{{unsupported OpenCL extension 'cl_khr_srgb_image_writes' - ignoring}}
+
+#ifdef cl_khr_subgroups
+#error "Incorrect cl_khr_subgroups define"
+#endif
+#pragma OPENCL EXTENSION cl_khr_subgroups: enable
+// expected-warning@-1{{unsupported OpenCL extension 'cl_khr_subgroups' - ignoring}}
+
+#ifdef cl_khr_terminate_context
+#error "Incorrect cl_khr_terminate_context define"
+#endif
+#pragma OPENCL EXTENSION cl_khr_terminate_context: enable
+// expected-warning@-1{{unsupported OpenCL extension 'cl_khr_terminate_context' - ignoring}}