Skip to content

More documentation for CUDA stream functions.#4756

Merged
apaszke merged 3 commits intopytorch:masterfrom
yongjik:0119-doc
Jan 21, 2018
Merged

More documentation for CUDA stream functions.#4756
apaszke merged 3 commits intopytorch:masterfrom
yongjik:0119-doc

Conversation

@yongjik
Copy link
Copy Markdown
Contributor

@yongjik yongjik commented Jan 20, 2018

No description provided.

Comment thread torch/cuda/streams.py Outdated
create a (non-default) stream (or explicitly request asynchronous operation
in, e.g., :meth:`~torch.Tensor.copy_`), your code will run as if every
operation was executed synchronously. However, when using non-default
streams, it is the user's responsibility to ensure proper synchronization.

This comment was marked as off-topic.

This comment was marked as off-topic.


By default, GPU operations are asynchronous. When you call a function that
involves GPU, the operations are *enqueued* to the particular device, but not
uses the GPU, the operations are *enqueued* to the particular device, but not

This comment was marked as off-topic.

@apaszke apaszke merged commit dd5c195 into pytorch:master Jan 21, 2018
@apaszke
Copy link
Copy Markdown
Contributor

apaszke commented Jan 21, 2018

Thanks @yongjik!

@yongjik yongjik deleted the 0119-doc branch February 5, 2018 18:04
@soumith soumith added the 0.3.1 label Feb 5, 2018
laurentdupin pushed a commit to laurentdupin/pytorch that referenced this pull request Apr 24, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants