zxybazh opened a new pull request, #14182: URL: https://github.com/apache/tvm/pull/14182
This PR introduces a default schedule pass that creates default thread binding for PrimFuncs, including symbolic shape functions so that these PrimFuncs could be built and run on Cuda device. The pass picks out all the blocks inside the PrimFunc and does loop fusion, split and reorder based on the loop extent and target information (max thread block number and max thread per block). Co-authored-by: Josh Fromm <<[email protected]>> CC: @sunggg @vinx13 @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]
