You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@mxnet.apache.org by GitBox <gi...@apache.org> on 2022/01/25 23:22:04 UTC
[GitHub] [incubator-mxnet] bartekkuncer opened a new pull request #20847: Make convolution operator fully work with oneDNN v2.4+
bartekkuncer opened a new pull request #20847:
URL: https://github.com/apache/incubator-mxnet/pull/20847
## Description ##
With the upgrade of oneDNN to version 2.4+ tests/python/dnnl/subgraphs/test_conv_subgraph.py::test_pos_conv_add[True-data_shape1] started failing. During investigation of the problem it turned out that problem regards only cases with amount of input channels lower than 4 and switching away from primitive with weight dnnl::format_tag ABcd4b16a4b fixes the issue. This change implements the switch in MXNet and restores the original shape in the failing test (adjusted here: https://github.com/apache/incubator-mxnet/pull/20662).
This is only temporary solution until the full fix arrives.
Problem tracking issue: https://github.com/apache/incubator-mxnet/issues/20826.
## Checklist ##
### Changes ###
- [x] Implement switching away from problematic primitive description
- [x] Restore original shape in failing test
--
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@mxnet.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [incubator-mxnet] mxnet-bot commented on pull request #20847: Make convolution operator fully work with oneDNN v2.4+
Posted by GitBox <gi...@apache.org>.
mxnet-bot commented on pull request #20847:
URL: https://github.com/apache/incubator-mxnet/pull/20847#issuecomment-1021701810
Hey @bartekkuncer , Thanks for submitting the PR
All tests are already queued to run once. If tests fail, you can trigger one or more tests again with the following commands:
- To trigger all jobs: @mxnet-bot run ci [all]
- To trigger specific jobs: @mxnet-bot run ci [job1, job2]
***
**CI supported jobs**: [windows-gpu, windows-cpu, centos-cpu, website, centos-gpu, edge, unix-gpu, sanity, unix-cpu, miscellaneous, clang]
***
_Note_:
Only following 3 categories can trigger CI :PR Author, MXNet Committer, Jenkins Admin.
All CI tests must pass before the PR can be merged.
--
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@mxnet.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [incubator-mxnet] bgawrych commented on a change in pull request #20847: Make convolution operator fully work with oneDNN v2.4+
Posted by GitBox <gi...@apache.org>.
bgawrych commented on a change in pull request #20847:
URL: https://github.com/apache/incubator-mxnet/pull/20847#discussion_r792449419
##########
File path: src/operator/nn/dnnl/dnnl_convolution.cc
##########
@@ -116,10 +116,18 @@ std::shared_ptr<dnnl::convolution_forward::primitive_desc> GetConvFwdImpl(
// suboptimal kernel for computation that has the expected memory size requirements
auto conv_pd =
std::make_shared<dnnl::convolution_forward::primitive_desc>(desc, attr, engine);
- while (conv_pd->dst_desc().get_size() != GetArraySize(output) ||
- conv_pd->src_desc().get_size() != GetArraySize(data) ||
- (!param.dnnl_param.quantized &&
- conv_pd->weights_desc().get_size() != GetArraySize(weights))) {
+ while (
+ conv_pd->dst_desc().get_size() != GetArraySize(output) ||
+ conv_pd->src_desc().get_size() != GetArraySize(data) ||
+ (!param.dnnl_param.quantized &&
+ conv_pd->weights_desc().get_size() != GetArraySize(weights)) ||
+ // With the upgrade of oneDNN to version 2.4+
+ // tests/python/dnnl/subgraphs/test_conv_subgraph.py::test_pos_conv_add[True-data_shape1]
+ // started failing. Switching away from primitive with weight dnnl::format_tag
+ // ABcd4b16a4b is in place to temporairly fix the issue until full fix arrives.
+ // Tracking issue: https://github.com/apache/incubator-mxnet/issues/20826.
Review comment:
```suggestion
// With the upgrade of oneDNN to version 2.4+
// tests/python/dnnl/subgraphs/test_conv_subgraph.py::test_pos_conv_add[True-data_shape1]
// started failing. Switching away from primitive with weight dnnl::format_tag
// ABcd4b16a4b in order to temporarily fix the issue until full fix arrives.
// Tracking issue: https://github.com/apache/incubator-mxnet/issues/20826.
```
--
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@mxnet.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [incubator-mxnet] bartekkuncer commented on pull request #20847: Make convolution operator fully work with oneDNN v2.4+
Posted by GitBox <gi...@apache.org>.
bartekkuncer commented on pull request #20847:
URL: https://github.com/apache/incubator-mxnet/pull/20847#issuecomment-1022134828
@mxnet-bot run ci [centos-gpu]
--
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@mxnet.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [incubator-mxnet] mxnet-bot commented on pull request #20847: Make convolution operator fully work with oneDNN v2.4+
Posted by GitBox <gi...@apache.org>.
mxnet-bot commented on pull request #20847:
URL: https://github.com/apache/incubator-mxnet/pull/20847#issuecomment-1022009725
Jenkins CI successfully triggered : [unix-cpu, centos-gpu]
--
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@mxnet.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [incubator-mxnet] bgawrych merged pull request #20847: Make convolution operator fully work with oneDNN v2.4+
Posted by GitBox <gi...@apache.org>.
bgawrych merged pull request #20847:
URL: https://github.com/apache/incubator-mxnet/pull/20847
--
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@mxnet.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [incubator-mxnet] mxnet-bot commented on pull request #20847: Make convolution operator fully work with oneDNN v2.4+
Posted by GitBox <gi...@apache.org>.
mxnet-bot commented on pull request #20847:
URL: https://github.com/apache/incubator-mxnet/pull/20847#issuecomment-1022134881
Jenkins CI successfully triggered : [centos-gpu]
--
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@mxnet.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [incubator-mxnet] bartekkuncer commented on pull request #20847: Make convolution operator fully work with oneDNN v2.4+
Posted by GitBox <gi...@apache.org>.
bartekkuncer commented on pull request #20847:
URL: https://github.com/apache/incubator-mxnet/pull/20847#issuecomment-1022009645
@mxnet-bot run ci [centos-gpu, unix-cpu]
--
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@mxnet.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org