access2rohit commented on issue #16104: Faster Transpose 2D
URL: https://github.com/apache/incubator-mxnet/pull/16104#issuecomment-540239351
@anirudh2290 can you merge this PR
This is an automated message from the Apache Git
access2rohit commented on issue #16104: Faster Transpose 2D
URL: https://github.com/apache/incubator-mxnet/pull/16104#issuecomment-539701107
> What are the performances of the input sizes that are not divisible by
blocksize?
@sxjscience 1,1 is not divisible by 32. Is there
access2rohit commented on issue #16104: Faster Transpose 2D
URL: https://github.com/apache/incubator-mxnet/pull/16104#issuecomment-539599930
@ChaiBapchya also paste the results of new unittest run here.
This is an automated
access2rohit commented on issue #16104: Faster Transpose 2D
URL: https://github.com/apache/incubator-mxnet/pull/16104#issuecomment-528686527
> Will the GPU side be accelerated?
@sxjscience not yet! work will be done on that after CPU one is accelerated
first
access2rohit commented on issue #16104: Faster Transpose 2D
URL: https://github.com/apache/incubator-mxnet/pull/16104#issuecomment-528606194
Can you add more description of how are you achieving the faster transpose ?
You can either paste a link that describes your approach or write it