leezu edited a comment on issue #20675: URL: https://github.com/apache/incubator-mxnet/issues/20675#issuecomment-1008020898
@RafLit thank you for investigating the issue. Would you like to open a PR to error out on the known bad build settings or at least not use it by default? For example, you can change the default by modifying the https://github.com/apache/incubator-mxnet/blob/9fa75b470b8f0238a98635f20f5af941feb60929/CMakeLists.txt#L352 or error out via ``` if((NOT USE_ONEDNN) AND USE_INT64_TENSOR_SIZE) message(FATAL_ERROR "INT64 Tensor without OneDNN is unsupported https://github.com/apache/incubator-mxnet/issues/20675";) endif() ``` Does the observation that VC2017 is much slower than VC2019 still hold, or was the root-cause here also the ILP/onednn setting? cc @barry-jin -- 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] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
