Skip to content

[SYCL][NVPTX] Obey -fcuda-short-ptr when compiling SYCL for NVPTX #15642

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 7 commits into from
Dec 11, 2024
Merged

[SYCL][NVPTX] Obey -fcuda-short-ptr when compiling SYCL for NVPTX #15642

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
7 commits
Select commit Hold shift + click to select a range
fdc2ad6
[SYCL][NVPTX] Enable -fcuda-short-ptr by default
frasercrmck Oct 9, 2024
9ec2d4c
Merge remote-tracking branch 'origin/sycl' into sycl-nvptx-short-ptrs
frasercrmck Dec 3, 2024
9ceb7d3
reduce scope of PR
frasercrmck Dec 3, 2024
4b1ce3b
add tests
frasercrmck Dec 3, 2024
4243e07
remove requires from test
frasercrmck Dec 3, 2024
b2301c3
Merge remote-tracking branch 'origin/sycl' into sycl-nvptx-short-ptrs
frasercrmck Dec 9, 2024
516261c
update test with comments and more checks
frasercrmck Dec 9, 2024
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
5 changes: 3 additions & 2 deletions clang/lib/Driver/ToolChains/Clang.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -8537,9 +8537,10 @@ void Clang::ConstructJob(Compilation &C, const JobAction &JA,
}
}

if (IsCuda) {
if (IsCuda || (IsSYCLDevice && Triple.isNVPTX())) {
bool UseShortPtr = IsSYCLDevice && Triple.isNVPTX();
if (Args.hasFlag(options::OPT_fcuda_short_ptr,
options::OPT_fno_cuda_short_ptr, false))
options::OPT_fno_cuda_short_ptr, UseShortPtr))
CmdArgs.push_back("-fcuda-short-ptr");
}

Expand Down
9 changes: 5 additions & 4 deletions clang/lib/Driver/ToolChains/Cuda.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1010,6 +1010,11 @@ void CudaToolChain::addClangTargetOptions(
CC1Args.push_back(DriverArgs.MakeArgString(LibSpirvFile));
}

bool UseShortPtr = DeviceOffloadingKind == Action::OFK_SYCL;
if (DriverArgs.hasFlag(options::OPT_fcuda_short_ptr,
options::OPT_fno_cuda_short_ptr, UseShortPtr))
CC1Args.append({"-mllvm", "--nvptx-short-ptr"});

if (DriverArgs.hasArg(options::OPT_nogpulib))
return;

Expand All @@ -1035,10 +1040,6 @@ void CudaToolChain::addClangTargetOptions(

clang::CudaVersion CudaInstallationVersion = CudaInstallation.version();

if (DriverArgs.hasFlag(options::OPT_fcuda_short_ptr,
options::OPT_fno_cuda_short_ptr, false))
CC1Args.append({"-mllvm", "--nvptx-short-ptr"});

if (CudaInstallationVersion >= CudaVersion::UNKNOWN)
CC1Args.push_back(
DriverArgs.MakeArgString(Twine("-target-sdk-version=") +
Expand Down
7 changes: 6 additions & 1 deletion libclc/CMakeLists.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -450,10 +450,15 @@ foreach( t ${LIBCLC_TARGETS_TO_BUILD} )
list(APPEND flags -D__unix__)
endif()

set(spirv_flags ${flags})
if( ARCH STREQUAL nvptx OR ARCH STREQUAL nvptx64 )
list(APPEND spirv_flags -Xclang -fcuda-short-ptr -mllvm -nvptx-short-ptr)
Copy link

@Naghasan Naghasan Oct 22, 2024

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think this is ok, but you could end up linking the libspirv (with short ptr) to a program compiled without. As builtins don't write pointers, I don't think this is an issue but would be good to test prior to merging and document this.

Copy link
Contributor Author

@frasercrmck frasercrmck Dec 3, 2024

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Note I've now reduced the scope of this PR and the option is no longer enabled by default. Thus libclc/libspirv is no longer compiled with this option. If a user passes -fcuda-short-ptr they'll see a warning while linking libclc/libspirv which I think is correct to do.

But yes in general we need to consider whether this is okay. I also think it's okay, and I've run several benchmarks with it and not seen a problem. Ideally we'd compile two versions of libspirv for NVPTX, but I don't know if it's going to be worth it for this relatively obscure option.

endif()

add_libclc_builtin_set(libspirv-${arch_suffix}
TRIPLE ${clang_triple}
TARGET_ENV libspirv
COMPILE_OPT ${flags}
COMPILE_OPT ${spirv_flags}
OPT_FLAGS ${opt_flags}
FILES ${libspirv_files}
GEN_FILES ${libspirv_gen_files}
Expand Down
Loading