You are viewing a plain text version of this content. The canonical link for it is here.

Posted to commits@tvm.apache.org by GitBox <gi...@apache.org> on 2022/01/18 21:35:13 UTC

[GitHub] [tvm] masahi commented on a change in pull request #9873: [TOPI] Add support for groupped conv3d

masahi commented on a change in pull request #9873:
URL: https://github.com/apache/tvm/pull/9873#discussion_r787169459



##########
File path: python/tvm/relay/op/strategy/cuda.py
##########
@@ -641,7 +642,7 @@ def conv3d_strategy_cuda(attrs, inputs, out_type, target):
                     (N % 16 == 0 and CI % 16 == 0 and CO % 16 == 0)
                     or (N % 8 == 0 and CI % 16 == 0 and CO % 32 == 0)
                     or (N % 32 == 0 and CI % 16 == 0 and CO % 8 == 0)
-                ):
+                ) and out_type == "float16":

Review comment:
       How about making this change to conv2d as well?. Perhaps better to send a separate PR for disabling fp32 tensorcore use since that would be a breaking change and potentially controversial.




-- 
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.

To unsubscribe, e-mail: commits-unsubscribe@tvm.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org