vinx13 opened a new pull request, #16823: URL: https://github.com/apache/tvm/pull/16823
This PR introduce VM extension, which allow users to define target related runtime extensions. The VM holds the reference to the extension once created to ensure it is destructed at VM destruction. CUDA graph builtins are refactored as VM extension to make the cuda graph cache per-VM instead of per-thread, to allow VM to be run on different threads (non-concurrently). This PR need rebase after #16815 cc @tqchen -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
