Add support for multi-arch CUDA binaries - #2001
Open
Sbozzolo wants to merge 1 commit into
Open
Conversation
This PR follows MAGMA's example and adds `CUDA_TARGETS`. `CUDA_TARGETS` can be used to bake multiple `CUDA_ARCHs` in a single binary. In particular, like in MAGMA, I build real targets (SASS) for all the architectures as well as a virtual target (PTX) for the last one. This allows architectures that were not in CUDA_ARCH to run by compiling the PTX directly.
Contributor
Author
Thanks for the super quick response! |
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
This PR follows MAGMA's example and adds
CUDA_TARGETS.CUDA_TARGETScan be used to bake multipleCUDA_ARCHsin a single binary.In particular, like in MAGMA, I build real targets (SASS) for all the architectures as well as a virtual target (PTX) for the last one. This allows architectures that were not in CUDA_ARCH to run by compiling the PTX directly.
LLM/GenAI Disclosure:
A large language model pointed me to the Magma makefile. The large language model reviewed the implementation, and suggested changing from using
NV_COMP_LAST := $(wordlist $(nwords_1),$(nwords),$(NV_COMP))to$(call cuda_gencode_compute,$(lastword $(CUDA_SMS))), which is simpler.The LLM and drove testing and implemented the change in the spack recipe.
Related Spack PR: spack/spack-packages#5888