[v1.x][submodule] Upgrade to oneDNN v1.6

[bartekkuncer]

Description

(Brief description on what this PR is about)

Checklist

Essentials

Please feel free to remove inapplicable items for your PR.

• The PR title starts with [MXNET-$JIRA_ID], where $JIRA_ID refers to the relevant JIRA issue created (except PRs with tiny changes)
• Changes are complete (i.e. I finished coding on this PR)
• All changes have test coverage:
• Unit tests are added for small changes to verify correctness (e.g. adding a new operator)
• Nightly tests are added for complicated/long-running ones (e.g. changing distributed kvstore)
• Build tests will be added for build configuration changes (e.g. adding a new build option with NCCL)
• Code is well-documented:
• For user-facing API changes, API doc string has been updated.
• For new C++ functions in header files, their functionalities and arguments are documented.
• For new examples, README.md is added to explain the what the example does, the source of the dataset, expected performance on test set and reference to the original paper if applicable
• Check the API doc at https://mxnet-ci-doc.s3-accelerate.dualstack.amazonaws.com/PR-$PR_ID/$BUILD_ID/index.html
• To the best of my knowledge, examples are either not affected by this change, or have been fixed to be compatible with this change

Changes

• Feature1, tests, (and when applicable, API doc)
• Feature2, tests, (and when applicable, API doc)

Comments

• If this change is a backward incompatible change, why must this change be made.
• Interesting edge cases to note here

[mxnet-bot]

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: [edge, windows-gpu, unix-cpu, clang, centos-cpu, centos-gpu, miscellaneous, unix-gpu, sanity, website, windows-cpu]

---

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.

[ChaiBapchya]

Thanks @bartekkuncer
Looks like 1 test

FAILED  ] 1 test, listed below:
[2020-07-29T10:50:39.039Z] [  FAILED  ] MKLDNN_UTIL_FUNC.MemFormat
``` CPP for unix-gpu pipeline.

[pengzhao-intel]

@bartekkuncer we may add the new mkldnn format in the file :)

[bartekkuncer]

@ChaiBapchya @pengzhao-intel all tests are passing now.

[TaoLv]

@bartekkuncer let's update the library to the formal release: https://github.com/oneapi-src/oneDNN/releases/tag/v1.6

[pengzhao-intel]

LGTM, thanks for the fix.

[ChaiBapchya]

LGTM! Thanks!

[ChaiBapchya]

@mxnet-label-bot add [pr-awaiting-merge]

[ChaiBapchya]

@szha Thanks for the merge
@bartekkuncer Can you cherry-pick this PR into v1.6.x branch?

[bartekkuncer]

New PR with update to official 1.6 oneDNN release: #18867.
@ChaiBapchya I will cherry-pick the change with official release to v1.6.x branch as soon as its merged.

[samskalicky]

New PR with update to official 1.6 oneDNN release: #18867.
@ChaiBapchya I will cherry-pick the change with official release to v1.6.x branch as soon as its merged.

and v1.7.x too?

[TaoLv]

@bartekkuncer @ChaiBapchya No, we cannot upgrade v1.6.x to the new release. Currently MXNet v1.6.x is using a version has LSTMP INT8 which is not present in oneDNN v1.6.

@samskalicky Yes, we can upgrade it to v1.7.x if needed. And yes, we will add it to the scope of v1.8 release.

[samskalicky]

I was only suggesting we do 1.7 if we were gonna do 1.6. But if not we can just do 1.8

[ChaiBapchya]

Currently MXNet v1.6.x is using a version has LSTMP INT8 which is not present in oneDNN v1.6.

@TaoLv which version of oneDNN will have both those fixes?

1. MKL optimization for GEMM kernels [AVX512 specific]
2. LSTMP int8

[bartekkuncer]

@ChaiBapchya v1.7 of oneDNN is supposed to have both these fixes