CUDA: add a flag "GGML_CUDA_JETSON_DEVICE" for optimization #13861
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
github-actions
bot
added
documentation
Contributor
Author
|
Hi @JohannesGaessler ,could I ask for your expertise in reviewing this PR when convenient?I am currently studying this Integrated GPU architecture(Jetson) in cuda. If you think my PR is inappropriate, I will revoke it immediately 0.0 |
Collaborator
|
I think this is not a good solution. Is there no way to detect iGPUs automatically? |
Contributor
Author
Thank you very much for your reply. Your opinion is correct. I am aware of the problem with this commit. I will review the cuda runtime API to find a more appropriate way to solve this problem. |
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Outdated suggestions cannot be applied.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.
As described in (#13856), this PR introduces a build flag (GGML_CUDA_JETSON_DEVICE) for CUDA Jetson-related optimizations and modifies the logic of cuda_device_support_buft when the flag is enabled.
The specific modification logic of “cuda_device_support_buft” is to add support for host_buffer (pinned memory) when cuda_device is jetson
According to the nvidia official website manual：https://docs.nvidia.com/cuda/cuda-for-tegra-appnote/index.html#memory-types-table