bartekkuncer commented on pull request #20962: URL: https://github.com/apache/incubator-mxnet/pull/20962#issuecomment-1074309552
Checked on clx-8280 machine that using oneDNN slows down these scalar operators by ~2% for Albert large model. Model taken from gluon-nlp. Command used to run the model: OMP_NUM_THREADS=28 KMP_AFFINITY=compact,1,0,granularity=fine numactl -l sh commands/run_squad2_albert_large.sh. Results: no dnnl:  dnnl:  -- 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]
