CUDA: Introduce arm64 subarchitecture. #8622
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
c05a8a9
to
f264e8f
Compare
|
@maleadt Thanks for the ping - highly appreciated. It would be nice to keep 10.2 functional to to be able build "ancient" stuff like Torch_jll v1.10.2 (using the original binary dependencies, CUDA 10.2 and 11.3): FluxML/Torch.jl#61 ... and yeah, I know - 11.3 is also not available since CUDA_SDK... :-) |
|
I've added the old SDKs etc back, just only providing SBSA. 10.2 is special, only providing Tegra. @stemann I'll leave the generalization to SBSA up to you, if you're interested. |
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Outdated suggestions cannot be applied.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 PR introduces yet another CUDA-specific platform tag,
"cuda_platform", to differentiate between SBSA and Jetson on aarch64. Jetson binaries are only available on the redist server starting with CUDA 11.8, so I've removed support for older versions.@stemann Sorry for removing CUDA 10.2 support again, but I haven't seen any changes in CUDA.jl so I figured you weren't interested in that anymore. If you are, it would be relatively easy to get your 10.2 build support compatible with this change: It only requires adding a regular runfile path for SBSA in the 10.2 SDK JLL, keeping your apt-based ones for the Jetson platform.
Changes like this are always fragile and very hard to test, so I expect breakage...