[Feature] Support RoiPool with Cambricon MLU backend

[zcyKTH]

Thanks for your contribution and we appreciate it a lot. The following instructions would make your pull request more healthy and more easily get feedback. If you do not understand some items, don't worry, just make the pull request and seek help from maintainers.

Motivation

The motivation of the PR is to support running RoiPool with Cambricon MLU backend.

It includes three parts:

1. Add roi_pool_mlu_kernel.mlu src code as MLU kernel.
2. Add roi_pool_mlu.cpp to support launching kernel in PyTorch.
3. Refactor test_roi_pool.py to support testing RoiPool with MLU backend.

Modification

• MLU src code
  Add MLU src code of RoiPool in directory mmcv/ops/csrc/common/mlu/roi_pool_mlu_kernel.mlu.
• PyTorch adaptation
  Adapt RoiPool for PyTorch in mmcv/ops/csrc/pytorch/mlu/roi_pool_mlu.cpp.
• Common functions
  Add data conversion functions between fp32 and int32 in mmcv/ops/csrc/common/mlu/common_mlu_helper.hpp.
• Unit test
  Surpport test RoiPool with various backends in tests/test_ops/test_roi_pool.py.

BC-breaking (Optional)

Does the modification introduce changes that break the backward-compatibility of the downstream repositories?
If so, please describe how it breaks the compatibility and how the downstream projects should modify their code to keep compatibility with this PR.

Use cases (Optional)

If this PR introduces a new feature, it is better to list some use cases here, and update the documentation.

Checklist

Before PR:

• I have read and followed the workflow indicated in the CONTRIBUTING.md to create this PR.
• Pre-commit or linting tools indicated in CONTRIBUTING.md are used to fix the potential lint issues.
• Bug fixes are covered by unit tests, the case that causes the bug should be added in the unit tests.
• New functionalities are covered by complete unit tests. If not, please add more unit test to ensure the correctness.
• The documentation has been modified accordingly, including docstring or example tutorials.

After PR:

• If the modification has potential influence on downstream or other related projects, this PR should be tested with some of those projects, like MMDet or MMCls.
• CLA has been signed and all committers have signed the CLA in this PR.

[zhouzaida]

Please update https://github.com/open-mmlab/mmcv/blob/master/docs/en/understand_mmcv/ops.md and https://github.com/open-mmlab/mmcv/blob/master/docs/zh_cn/understand_mmcv/ops.md

[zcyKTH]

Please update https://github.com/open-mmlab/mmcv/blob/master/docs/en/understand_mmcv/ops.md and https://github.com/open-mmlab/mmcv/blob/master/docs/zh_cn/understand_mmcv/ops.md

Done

[grimoire]

There are some warning:

warning(0001):/tmp/roi_pool_mlu_kernel-Z9-TP_.s:2744:IMM for u32 is out of range

Is it expected?

[zcyKTH]

There are some warning:

warning(0001):/tmp/roi_pool_mlu_kernel-Z9-TP_.s:2744:IMM for u32 is out of range

Is it expected?

Which version of CNToolkit or CNAS?

[grimoire]

@zcyKTH How can I check the version?

[zcyKTH]

@zcyKTH How can I check the version?

Try to check version.txt in $NEUWARE_HOME directory.

[grimoire]

Neuware Version 2.6.4

The warning appears when compiling the op. The build finished with no error and the test also passed.
Can we ignore the warning?

[zcyKTH]

Neuware Version 2.6.4

The warning appears when compiling the op. The build finished with no error and the test also passed. Can we ignore the warning?

Yep. Just ignore these oversensitive warnings and they have been removed in future Neuware Version.

[grimoire]

LGTM