Disable USE_GPERFTOOLS by default

[ZhennanQin]

Description

Because of #8968, mxnet will random crash after USE_GPERFTOOLS = 1. Some end-user may get trouble if gperf-tools is pre-installed. So before fix this, we'd better to disable it by default.

@pengzhao-intel @szha

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

[pengzhao-intel]

@szha @eric-haibin-lin is there any reason that the gperf is ON by default?
Our customer encountered the crash w/ gperf ON, so we prefer to make it OFF by default.

[pengzhao-intel]

cc @lebeg

[wkcn]

LGTM. I also met the issue about GPERFTOOLS.

[larroy]

LGTM

[pengzhao-intel]

LGTM

Will merge in 24 hours if there are no further comments :)

[szha]

I guess it was decided by an individual as there is significant performance benefit in using alternative malloc. It wasn't thoroughly evaluated and I actually noticed similar issues with tcmalloc, so I'm actually surprised that this is on by default in cmake.

[pengzhao-intel]

Merging now since the BOSS agree :)

[anirudh2290]

I guess it was decided by an individual as there is significant performance benefit in using alternative malloc. It wasn't thoroughly evaluated and I actually noticed similar issues with tcmalloc, so I'm actually surprised that this is on by default in cmake.

An "individual" by himself/herself cannot make decisions in an apache incubating project.

[cjolivier01]

Interesting theory szha, although also incorrect. But disable if you want. Jemalloc is faster most of the time anyway.

[szha]

@cjolivier01 thanks for stepping in. I'd like to hear what you would like clairfy and what option you recommend?