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 2020/01/16 21:20:50 UTC
[GitHub] [incubator-tvm] icemelon9 opened a new pull request #4728: [Fix]
Fix dense x86 schedule
icemelon9 opened a new pull request #4728: [Fix] Fix dense x86 schedule
URL: https://github.com/apache/incubator-tvm/pull/4728
Remove the vectorize on `y` axis as it could be very large, causing llvm codegen hangs.
https://discuss.tvm.ai/t/relay-build-hangs-for-5d-max-pool3d-reshape-matmul/5398/5
@apivovarov @yzhliu
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
[GitHub] [incubator-tvm] icemelon9 commented on a change in pull request
#4728: [Fix] Fix dense x86 schedule
Posted by GitBox <gi...@apache.org>.
icemelon9 commented on a change in pull request #4728: [Fix] Fix dense x86 schedule
URL: https://github.com/apache/incubator-tvm/pull/4728#discussion_r368165655
##########
File path: topi/python/topi/x86/dense.py
##########
@@ -191,7 +191,6 @@ def _schedule_dense_pack_template(cfg, s, C):
z, y, x = s[packedB].op.axis
s[packedB].reorder(z, x, y)
s[packedB].parallel(z)
- s[packedB].vectorize(y)
Review comment:
Thanks for the tips. After took a look, I found out that it's not necessary to vectorize here.
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
[GitHub] [incubator-tvm] icemelon9 merged pull request #4728: [Fix] Fix
dense x86 schedule
Posted by GitBox <gi...@apache.org>.
icemelon9 merged pull request #4728: [Fix] Fix dense x86 schedule
URL: https://github.com/apache/incubator-tvm/pull/4728
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
[GitHub] [incubator-tvm] FrozenGene commented on a change in pull request
#4728: [Fix] Fix dense x86 schedule
Posted by GitBox <gi...@apache.org>.
FrozenGene commented on a change in pull request #4728: [Fix] Fix dense x86 schedule
URL: https://github.com/apache/incubator-tvm/pull/4728#discussion_r367747570
##########
File path: topi/python/topi/x86/dense.py
##########
@@ -191,7 +191,6 @@ def _schedule_dense_pack_template(cfg, s, C):
z, y, x = s[packedB].op.axis
s[packedB].reorder(z, x, y)
s[packedB].parallel(z)
- s[packedB].vectorize(y)
Review comment:
How about using `try_vec`? It could let AutoTVM to discard large configuration using timeout parameter. And we could keep `vectorize(y)` if it is possible.
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services