ptrendx commented on issue #17767: Fix and optimize handling of vectorized 
memory accesses
URL: https://github.com/apache/incubator-mxnet/pull/17767#issuecomment-615471587
 
 
   @sxjscience Frankly I did not look in to much detail in the numpy ops but 
from what I saw they use only partially the previous implementation, as they 
need to deal with potentially different types etc, which makes them use special 
kernels. Also, they do seem to be mostly broadcasting ops (which scares me a 
little in the context of implementing fusions for example, as those would not 
qualify). So currently I would not expect it impacting numpy ops, but would 
like to work on extending this work to be able to unify and optimize both numpy 
and non-numpy versions.

----------------------------------------------------------------
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.
 
For queries about this service, please contact Infrastructure at:
[email protected]


With regards,
Apache Git Services

Reply via email to