Static alloc for hybridblock

[piiswrong]

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 $JIRA_ID refers to the relevant JIRA issue created (except PRs with tiny changes)
• Changes are complete (i.e. I finished coding on this PR)
• All changes have test coverage:
• Unit tests are added for small changes to verify correctness (e.g. adding a new operator)
• Nightly tests are added for complicated/long-running ones (e.g. changing distributed kvstore)
• Build tests will be added for build configuration changes (e.g. adding a new build option with NCCL)
• Code is well-documented:
• For user-facing API changes, API doc string has been updated.
• For new C++ functions in header files, their functionalities and arguments are documented.
• For new examples, README.md is added to explain the what the example does, the source of the dataset, expected performance on test set and reference to the original paper if applicable
• Check the API doc at http://mxnet-ci-doc.s3-accelerate.dualstack.amazonaws.com/PR-$PR_ID/$BUILD_ID/index.html
• To the my best 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

[marcoabreu]

Please provide proof that #11171 is not the case in this PR and that the cache-hit-miss bug is covered as well

https://github.com/apache/incubator-mxnet/pull/11313/files#diff-962bd5bb7248659d7eb3be37ee8a4c6bR1288

[piiswrong]

This is a different PR. If you have any specific concerns on the code please utter them.

If you don't understand the code please let other people who do review it.

[marcoabreu]

Sorry if I mixed it up. I tried to read the description, but it was not very helpful.

[zheng-da]

@marcoabreu I have been intensively using CachedOp with static memory allocation this week. I also read the code in the past week. Other than the small bug reported previously, I don't find problems and it works very well with me in my PR. #10451
I observe the memory allocation behavior and it works exactly as I expected.

What are you looking for? If you are looking for test code, my entire PR can serve as test code. If you are looking for review, I have been reading the code for the entire week. I don't think it's reasonable to block other people's work for no reason.

[marcoabreu]

Hi Da, thanks for your comments. I'm sure that CachedOp is behaving as expected. The problems are described at #11318

[zheng-da]

I don't understand what the problem is. Because you veto, the PR should never be merged?

We should open the PR and get proper reviews and merge it. Blocking other people to review it isn't a way of solving the problem.

[marcoabreu]

I was not the one who closed this PR, it was @tqchen. He was the one who dismissed my review and blocked other people from reviewing it. So why are you targeting this towards me?

[tqchen]

just to be fair, I dismissed the your comment with specific comment that addresses your issue https://github.com/apache/incubator-mxnet/pull/11313/files#diff-962bd5bb7248659d7eb3be37ee8a4c6bR1288

I agree that maybe more time should have been given to do the reviews, and I might be a bit aggressive, but never the less, vetoing things without understand what is going on seems to be not so constructive