pengzhao-intel commented on issue #13598: More fine-grained operator 
implementation dispatch & memory planning flow 
URL: 
https://github.com/apache/incubator-mxnet/issues/13598#issuecomment-446544319
 
 
   Regarding memory planning, there are different requirements from GPU and CPU.
   Looks like the GPU wants to support both NCHW and NHWC, but CPU with MKL-DNN 
prefer the padded memory (GPU still needs this in some cases).
   
   @eric-haibin-lin do you mind elaborate the possible change of "memory 
planning"?  How the new change cover the diversity of memory requirement from 
the different backend?
   
   We're open to the changes and happy to work together :)
   

----------------------------------------------------------------
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

Reply via email to