[GitHub] asmushetzel commented on issue #7625: entire codebase build with mshadow_use_clas=0

[gi...@git.apache.org]

asmushetzel commented on issue #7625: entire codebase build with mshadow_use_clas=0
URL: https://github.com/apache/incubator-mxnet/pull/7625#issuecomment-325162710
 
 
   Hi Dick, 
   much better that way. Thanks a lot.
   As you have looked quite a bit into this file, I wonder whether you are the best person to talk to about one other issue: Is there any way to make the GPU-batch versions of operators as trmm/potrf/potri where cuBlas/cuSolver do not supply batch mode operations more efficient? Is there anyone at NVidia who can take a look at this? It is important as a very common use case of these operations is batch mode processing with a lot of small matrices. And that won't be that great in performance on GPU w/ the current way of naive batch processing. 
   
 
----------------------------------------------------------------
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