samskalicky commented on issue #14570: add a compiler flag to use int64 as 
tensor size
URL: https://github.com/apache/incubator-mxnet/pull/14570#issuecomment-484614702
 
 
   @larroy we dont have any data that guarantees only the transpose operator is 
affected. In reality it could be all mshadow implemented operators that are 
affected. We dont have performance tests for each operator that we could run 
and verify. There are efforts to look into putting together some operator 
performance tests, but i dont think we should wait on this fix for that. 
   
   This PR is important to restore "balance to the force" until we can work on 
a better solution. Since new functionality for large tensors was already added 
and is being used for the Deep Graph Library effort we need to provide a way 
for them to use the large tensor feature. At the same time, we need to fix the 
performance degradation. So this is fix (adding the flag) the common middle 
ground between the two. 

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