ptrendx commented on issue #9774: mx.io.ImageRecordIter does not respect dtype argument / FP16 performance on Volta URL: https://github.com/apache/incubator-mxnet/issues/9774#issuecomment-367057821 @rahul003 Could you paste here how you invoked the benchmark script? Did you set the autotune env var to 2? Also, which cudnn version do you use? You should see many more s884 kernels in your profile... CIFAR has very small images and most of the time the dimensions of convolutions are not suitable for TensorCore.
---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services
