[NVVM] Support - Followup enhancements #1218
Draft
+370
−2
Conversation
This file contains hidden or 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
Contributor
leofang
added
enhancement
Member
|
Thanks, @abhilash1910! Any ETA to wrap this up? |
Contributor
Author
|
pre-commit.ci autofix |
Contributor
Author
|
pre-commit.ci autofix |
leofang
reviewed
Nov 25, 2025
Member
leofang
left a comment
leofang
left a comment
There was a problem hiding this comment.
Thanks, Abhilash! Leaving a few early feedbacks.
Comment on lines
481
to
487
| # TODO: support pre-loaded headers & include names | ||
| # TODO: allow tuples once NVIDIA/cuda-python#72 is resolved | ||
|
|
||
| if options.extra_sources is not None: | ||
| raise ValueError("extra_sources is not supported by the NVRTC backend (C++ code_type)") | ||
|
|
||
| self._mnff.handle = handle_return(nvrtc.nvrtcCreateProgram(code.encode(), options._name, 0, [], [])) |
Member
There was a problem hiding this comment.
nit: organize the comments better
Suggested change
| # TODO: support pre-loaded headers & include names | |
| # TODO: allow tuples once NVIDIA/cuda-python#72 is resolved | |
| if options.extra_sources is not None: | |
| raise ValueError("extra_sources is not supported by the NVRTC backend (C++ code_type)") | |
| self._mnff.handle = handle_return(nvrtc.nvrtcCreateProgram(code.encode(), options._name, 0, [], [])) | |
| # TODO: support pre-loaded headers & include names | |
| if options.extra_sources is not None: | |
| raise ValueError("extra_sources is not supported by the NVRTC backend (C++ code_type)") | |
| # TODO: allow tuples once NVIDIA/cuda-python#72 is resolved | |
| self._mnff.handle = handle_return(nvrtc.nvrtcCreateProgram(code.encode(), options._name, 0, [], [])) |
cuda_core/tests/test_program.py
Outdated
Comment on lines
538
to
546
| bitcode_path = os.environ.get("BITCODE_NVVM_PATH") | ||
| if not bitcode_path: | ||
| pytest.skip("BITCODE_NVVM_PATH environment variable is not set.Disabling the test.") | ||
| bitcode_file = Path(bitcode_path) | ||
| if not bitcode_file.exists(): | ||
| pytest.skip(f"Bitcode file not found: {bitcode_path}") | ||
|
|
||
| if bitcode_file.suffix != ".bc": | ||
| pytest.skip(f"Expected .bc file, got: {bitcode_file.suffix}") |
Member
There was a problem hiding this comment.
Would it be possible for us to avoid having a file locally? We have bitcode in this repo already:
cuda-python/cuda_bindings/tests/test_nvvm.py
Lines 12 to 141 in b9c76b3
so I suggest that we move it to the common place, say a new file under
cuda_python_test_helpers:https://github.com/NVIDIA/cuda-python/tree/main/cuda_python_test_helpers/cuda_python_test_helpers
and have it imported in both cuda.bindings/core tests.
Contributor
Author
There was a problem hiding this comment.
Agree on this.
Contributor
Author
There was a problem hiding this comment.
This should be partially addressed now as I have yet to remove the existing invocations from cuda_bindings test
Comment on lines
532
to
533
| extra_name = f"{options.name}_extra_{i}" | ||
| nvvm.add_module_to_program(self._mnff.handle, extra_source, len(extra_source), extra_name) |
Member
There was a problem hiding this comment.
Two notes:
- I am torn if we should allow users to specify the module names, instead of us making one up for them. Could you try my "sequence of 2-tuples" suggestion from our offline chat, and see how bad the code looks like? We should have something similar already in one of the options.
- As discussed let's check if all modules can be lazily loaded.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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.
Description
Issue Link - #981
Changes to be addressed in this WIP PR:
{If / when it is possible to add multiple modules, a test with code that uses something from libdevice is probably a good idea.
It's also useful to be able to lazily add a module}
cc @leofang