merrymercy commented on pull request #7250: URL: https://github.com/apache/tvm/pull/7250#issuecomment-758212157
Although it is buggy for x86 cpu. This feature is still useful for some targets like metal according to @antinucleon . CUDA runtime always throws runtime errors if the GPU kernels are invalid but metal runtime may just return wrong results without any runtime error. So we need this check_correctness for metal. ---------------------------------------------------------------- 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. For queries about this service, please contact Infrastructure at: us...@infra.apache.org