altanh opened a new pull request, #10969: URL: https://github.com/apache/tvm/pull/10969
When the target shape of `broadcast_to` is an `int64` Constant, something breaks in lowering for CUDA (the thread extent ends up being `int64` and causes a mismatch error with the iteration variable). I could only get the error to show up when following up with a `sum` op. This is a hotfix to get around the problem for now while we dig into the root cause. cc @Lunderberg @mbrookhart -- 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]
