[GitHub] elabeca commented on issue #9087: MXNET 1.0.0 - marginal performance improvement Titan V (Volta) with half precision cuda 9.0 + cudnn 7.0.5

[GitBox <gi...@apache.org>]

elabeca commented on issue #9087: MXNET 1.0.0 - marginal performance improvement Titan V (Volta) with half precision cuda 9.0 + cudnn 7.0.5
URL: https://github.com/apache/incubator-mxnet/issues/9087#issuecomment-352187739
 
 
   Thanks for the suggestion. Unfortunately at this time the train_imagenet script gives an error.
   
   Only options for MXNET_CUDNN_AUTOTUNE_DEFAULT are 1 (true) or 0 (false) I believe.
   
   Here are the latest results with train_cifar10.py:
   Titan X (Pascal) MXNET_CUDNN_AUTOTUNE_DEFAULT=0 --dtype float32 ==> ~2020 samples/second
   Titan X (Pascal) MXNET_CUDNN_AUTOTUNE_DEFAULT=0 --dtype float16 ==> ~2330 samples/second
   Titan X (Pascal) MXNET_CUDNN_AUTOTUNE_DEFAULT=1 --dtype float32 ==> ~2230 samples/second
   Titan X (Pascal) MXNET_CUDNN_AUTOTUNE_DEFAULT=1 --dtype float16 ==> ~2600 samples/second
   
   Titan V (Volta) MXNET_CUDNN_AUTOTUNE_DEFAULT=0 --dtype float32 ==> ~2200 samples/second
   Titan V (Volta) MXNET_CUDNN_AUTOTUNE_DEFAULT=0 --dtype float16 ==> ~2400 samples/second
   Titan V (Volta) MXNET_CUDNN_AUTOTUNE_DEFAULT=1 --dtype float32 ==> ~2900 samples/second
   Titan V (Volta) MXNET_CUDNN_AUTOTUNE_DEFAULT=1 --dtype float16 ==> ~2740 samples/second
   
   As you can see, when autotune is on, half-precision performance is worse on Volta.
   
   That's not the case for Pascal.
   
   Is there a way to enforce the use of the TensorCores?
   
   Thanks
   

----------------------------------------------------------------
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:
users@infra.apache.org


With regards,
Apache Git Services