Move aot_cudagraphs backend here by ezyang · Pull Request #757 · pytorch/torchdynamo (original) (raw)
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service andprivacy statement. We’ll occasionally send you account related emails.
Already on GitHub?Sign in to your account
Conversation2 Commits3 Checks0 Files changed
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
[ Show hidden characters]({{ revealButtonHref }})
Previously it was in pytorch/pytorch but it depends on torchdynamo
code more closely, so this seems like the logical place.
Previously at pytorch/pytorch#80566
Signed-off-by: Edward Z. Yang ezyang@fb.com
Previously it was in pytorch/pytorch but it depends on torchdynamo code more closely, so this seems like the logical place.
Previously at pytorch/pytorch#80566
Signed-off-by: Edward Z. Yang ezyang@fb.com
Signed-off-by: Edward Z. Yang ezyang@fb.com
Signed-off-by: Edward Z. Yang ezyang@fb.com
@ezyang Should we also add a aot_nvfuser_cudagraphs i.e. AOTAutograd + Ts-nvfuser with Cudagraphs
We have aot nvfuser, but I'm not sure how the cudagraphs integration with nvfuser works. @Chillee might know more, he said he hacked in inductor's cuda graphs support to nvfuser. Note that cuda graphs is antipattern for dynamic shapes, and also there is quite a bit of logic needed to get it right (which I think is being duped between this backend and inductors' cudagraphs support)
pytorchmergebot pushed a commit to pytorch/pytorch that referenced this pull request
