[flang-rt] Add experimental support for GPU build #131826
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
jhuber6
requested review from
clementval,
jeanPerier,
jplehr,
jsjodin,
klausler,
Meinersbur,
mjklemm,
ronlieb and
sscalpone
Meinersbur
reviewed
Mar 21, 2025
Comment on lines
+90
to
+91
| # Sources that can be compiled directly for the GPU. | ||
| set(gpu_sources |
There was a problem hiding this comment.
Can we avoid duplicating the list of sources? What makes theses gpu-compatible?
There was a problem hiding this comment.
Mostly this is just the list of files that didn't error. There's a few different ones from the existing offloading builds, mostly due to the fact that we don't borrow the system's headers. I was thinking of a creative way to split these up but figured this was easiest.
Most of these fail due to missing headers, mostly POSIX related things because as much as we try pthreads and sema are not going to work on a GPU.
Meinersbur
approved these changes
Mar 21, 2025
Meinersbur
approved these changes
Mar 21, 2025
Meinersbur
approved these changes
Mar 21, 2025
What I figured, could you look at #131695 as well? I think in the future we'll need to think of how to run tests for this and investigating setting up the missing files. |
Summary: This patch adds initial support for compiling `flang-rt` directly for the GPU. The method used here matches what's already done for `libc` and `libc++` for the GPU and builds off of those projects. Mainly this requires setting up some flags and setting the sources that currently work. This will deposit the resulting library in the appropriate directory. These files are then intended to be linked via `-Xoffload-linker` support in the offloading driver. ``` lib/clang/21/lib/nvptx64-nvidia-cuda/libflang_rt.runtime.a lib/clang/21/lib/amdgcn-amd-amdhsa/libflang_rt.runtime.a ``` This is obviously missing a lot of functions, mainly the `io` support. Most of what we cannot support is due to using POSIX things that just don't make sense on the GPU. Stuff like `pthreads` or `sema`. Getting unit tests to run on this will also be a challenge. We could run tests the same way we do with `libc`, but the problem there is that the `libc` test suite is freestanding while `gtest` currently doesn't compile on the GPU bcause it uses a lot of weird stuff. If the unit tests were simply `int main` then it would work. I don't understand the actual runtime code very well, I'd appreciate some guidance on how to actually support Fortran IO from this interface. As I understand it, Fortran IO requires a stack-like operation, which conflicts with the SIMT model GPUs use. Worst case scenario we could burn some LDS to keep a stack, or serialize it somehow since we can always just iterate over all the active lanes. Building this right now looks like this, which depends on the arguments added in llvm#131695. ``` -DRUNTIMES_nvptx64-nvidia-cuda_LLVM_ENABLE_RUNTIMES=compiler-rt;libc;libcxx;libcxxabi;flang-rt \ -DRUNTIMES_amdgcn-amd-amdhsa_LLVM_ENABLE_RUNTIMES=compiler-rt;libc;libcxx;libcxxabi;flang-rt \ -DRUNTIMES_nvptx64-nvidia-cuda_FLANG_RT_LIBC_PROVIDER=llvm \ -DRUNTIMES_nvptx64-nvidia-cuda_FLANG_RT_LIBCXX_PROVIDER=llvm \ -DRUNTIMES_amdgcn-amd-amdhsa_FLANG_RT_LIBC_PROVIDER=llvm \ -DRUNTIMES_amdgcn-amd-amdhsa_FLANG_RT_LIBCXX_PROVIDER=llvm ```
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Summary:
This patch adds initial support for compiling
flang-rtdirectly forthe GPU. The method used here matches what's already done for
libcandlibc++for the GPU and builds off of those projects.Mainly this requires setting up some flags and setting the sources that
currently work. This will deposit the resulting library in the
appropriate directory. These files are then intended to be linked via
-Xoffload-linkersupport in the offloading driver.This is obviously missing a lot of functions, mainly the
iosupport.Most of what we cannot support is due to using POSIX things that just
don't make sense on the GPU. Stuff like
pthreadsorsema.Getting unit tests to run on this will also be a challenge. We could run
tests the same way we do with
libc, but the problem there is that thelibctest suite is freestanding whilegtestcurrently doesn'tcompile on the GPU bcause it uses a lot of weird stuff. If the unit
tests were simply
int mainthen it would work.I don't understand the actual runtime code very well, I'd appreciate
some guidance on how to actually support Fortran IO from this interface.
As I understand it, Fortran IO requires a stack-like operation, which
conflicts with the SIMT model GPUs use. Worst case scenario we could
burn some LDS to keep a stack, or serialize it somehow since we can
always just iterate over all the active lanes.
Building this right now looks like this, which depends on the arguments
added in #131695.