fix compile error in debug mode

[NEWPLAN]

Description

fix compiling error in debug mode
i.e. the latest 'BufferEntry' do not contain ctx function calls, and it will result in a compiling error.
inside of BufferEntry is an object of NDArray, which is the expected data.

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

[ChaiBapchya]

Looks like the CI is failing. Weird that its in the website.
Possible case of #13833
Might want to retrigger the CI

[stu1130]

@mxnet-label-bot add [pr-awaiting-review]
Thanks for your contribution @NEWPLAN

[NEWPLAN]

Looks like the CI is failing. Weird that its in the website.
Possible case of #13833
Might want to retrigger the CI

Hi, thanks for your check.
So, I don't know what is going on about the website and cannot see the running logs in details, how can we put this issue forward now?

[ChaiBapchya]

@NEWPLAN Thanks for your contribution.

This issue seems to be website flakiness (not related to your change)
It is being monitored separately already on the repo as #13833

For your PR, best way ahead is to retrigger the CI using -
git commit --allow-empty -m "trigger CI"

[KellenSunderland]

Thanks for the catch. Would you be able to either add a test that shows when this fails, or to give some minimal steps to reproduce (i.e. the build flags and build systems your using)? Ideally we'd like to make sure that we don't regress on this once it's fixed.

[NEWPLAN]

Thanks for the catch. Would you be able to either add a test that shows when this fails, or to give some minimal steps to reproduce (i.e. the build flags and build systems your using)? Ideally we'd like to make sure that we don't regress on this once it's fixed.

Thanks for your reply.
In my personal environment, I didn't get any error associated with the website issues. However, my fixed issues is the compiling error in enabling DEBUG.
The related code is in the src/kvstore/kvstore_nccl.h lines 345, i.e., "assert(root == buf.ctx().dev_id);"
the buf is an objection of BufferEntry, which contains an objection of NDArray, namely, merged.
It is the merged implements the ctx() function calls, rather than the buf.
The key configurations in my environment are described as follows.
OS: Ubuntu-1604 x64
GCC: 5.4.0
CUDA: 9.0.176
NCCL-2.1.15
cuDNN: 7.2

[KellenSunderland]

Thanks for the info @NEWPLAN. You can ignore the website failure, that's a problem on our side.

One more question, were you building with cmake or make?

[NEWPLAN]

Thanks for the info @NEWPLAN. You can ignore the website failure, that's a problem on our side.

One more question, were you building with cmake or make?

Yes, I built it via make.

[ptrendx]

Yes, I noticed this same issue with building in DEBUG mode some time ago, but never actually make a PR for that - thank you! The changes look good.

[sandeep-krishnamurthy]

@eric-haibin-lin - Can you please take a look at this PR? Thanks.