Disable Default CUDA Build, main branch (2024.08.13.) #672
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
krasznaa
force-pushed
the
DisableDefaultCUDA-main-20240813
branch
from
1e02937 to
3b9731c
Compare
|
As a little palate cleanser, what about this proposal? |
|
I don't use preset but do old-school cmake command like ancient people |
|
For that case this change means that you'd need to add Right now I need to edit |
beomki-yeo
approved these changes
Aug 16, 2024
To be consistent with how all other GPU languages are handled, the CUDA code of the repository now only gets built if the user explicitly asks for it.
krasznaa
force-pushed
the
DisableDefaultCUDA-main-20240813
branch
from
3b9731c to
5aa08e1
Compare
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Outdated suggestions cannot be applied.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.
This is a bit of a subjective proposal, so I'm interested in the feedback...
To make my life just a little easier, I'd like to stop the project from automatically enabling the build of its CUDA code when it thinks that it found a functional CUDA version. And so be consistent with how the SYCL, Alpaka, etc. sources need to be "turned on" for the build.
My reason is a bit complicated. It stems from the issue described under: https://discourse.cmake.org/t/checklanguage-does-not-take-the-c-standard-into-account When I set up the latest version of oneAPI with the latest version of CUDA, right now I both have to explicitly turn on the build of the SYCL code (for the NVIDIA backend, that's why I also have CUDA set up), and turn off the build of the CUDA code. Unfortunately the
syclpreset inCMakePresets.jsondoesn't do this at the moment. So I can't use VSCode for SYCL code development with the latest oneAPI version at the moment, without modifying the CMake code a little.Instead of doing that, I thought that this specific update could be acceptable by everyone. This way I could easily use the
syclpreset (with VSCode) with the latest oneAPI version, and thecudapreset in a different environment in which I do not set up oneAPI at all.But this does mean that if one of you didn't use any presets, and didn't set
TRACCC_BUILD_CUDAso far when doing CUDA developments, you'd need to have a little longer configuration command from now on. 🤔