Disable test coverage of C++ codebase on CI

[stu1130]

Description

Related to #15971

Root Cause

When linker link the object files, it uses relative addressing method. The error happens when the project is too large for linker to link the target objects using relative addressing with 32-bit address.

Things we have tried

1. add -fPIC flag to make sure the shared library can be loaded at any address in memory. It end up with the same error message.
2. add -mcmodel=large. GCC default use -mcmodel=small where the program and symbols must be linked within 2 GB which is 32 bit address space. When it is set to large, there is no 2GB limitation. Unfortunately, all the code MXNet statically link including MKLDNN will need to build this way. It also ended up with same failure. So it leaded to our final solution.
3. Removing the test_coverage flag which injects monitoring code every if else branch, while loop, etc.It dramatically increase the size of libmxnet.a with test coverage. After removing the test coverage flag, it passed the Clang MKLDNN CI job.

After discussion with @marcoabreu, we agreed to disable C++ backend test coverage monitoring, which will also boost the CI.

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

Comments

@marcoabreu could you help check if some runtime functions are missing or some functions should not disable? Thanks

[anirudhacharya]

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

[marcoabreu]

@anirudhacharya I already approved

[yuxihu]

@stu1130 It seems centos-gpu failed to build. I have re-triggered it for multiple times. Could you please take a look?

[stu1130]

@yuxihu Sure

[anirudhacharya]

@marcoabreu apologies, did not see your review.
@mxnet-label-bot add [pr-awaiting-merge] remove [pr-awaiting-review]