Skip to content
This repository has been archived by the owner on Nov 17, 2023. It is now read-only.

[v1.x] Static build for mxnet-cu110 #19272

Merged
merged 5 commits into from
Oct 20, 2020
Merged

[v1.x] Static build for mxnet-cu110 #19272

merged 5 commits into from
Oct 20, 2020

Conversation

waytrue17
Copy link
Contributor

@waytrue17 waytrue17 commented Oct 2, 2020 •
edited

Description

Added static build MXNet with CUDA 11.0

Checklist

Essentials

  • Changes are complete (i.e. I finished coding on this PR)

Changes

  • added script to statically build MXNet with CUDA 11.0

@waytrue17 waytrue17 requested a review from szha as a code owner October 2, 2020 03:04
@mxnet-bot
Copy link

Hey @waytrue17 , Thanks for submitting the PR
All tests are already queued to run once. If tests fail, you can trigger one or more tests again with the following commands:

  • To trigger all jobs: @mxnet-bot run ci [all]
  • To trigger specific jobs: @mxnet-bot run ci [job1, job2]

CI supported jobs: [sanity, miscellaneous, windows-cpu, unix-gpu, centos-gpu, website, windows-gpu, unix-cpu, clang, edge, centos-cpu]

Note:
Only following 3 categories can trigger CI :PR Author, MXNet Committer, Jenkins Admin.
All CI tests must pass before the PR can be merged.

@waytrue17 waytrue17 mentioned this pull request Oct 2, 2020
Merged
2 tasks
@waytrue17
Copy link
Contributor Author

@leezu @ChaiBapchya Can you please review this PR?

@sandeep-krishnamurthy
Copy link
Contributor

Thank you so much @waytrue17

ChaiBapchya
ChaiBapchya reviewed Oct 2, 2020
View reviewed changes
tools/pip/setup.py Outdated Show resolved Hide resolved
@ChaiBapchya
Copy link
Contributor

How & Where has this change been tested?

@waytrue17
Copy link
Contributor Author

waytrue17 commented Oct 2, 2020 •
edited

How & Where has this change been tested?

I am testing it by calling tools/staticbuild/build.sh cu110 to build libmxnet.so on a gpu instance, binding it with python language. Then run a simple script in python with ctx=mx.gpu(0).

ChaiBapchya
ChaiBapchya approved these changes Oct 2, 2020
View reviewed changes
Copy link
Contributor

@ChaiBapchya ChaiBapchya left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Sure. thanks & hope we confirmed ldd libmxnet.so is linked to cuda11 so

@waytrue17
Copy link
Contributor Author

Sure. thanks & hope we confirmed ldd libmxnet.so is linked to cuda11 so

@ChaiBapchya Thanks for the suggestion. The following is the ldd libmxnet.so outputs:
linux-vdso.so.1 (0x00007ffda2bb9000) libdl.so.2 => /lib/x86_64-linux-gnu/libdl.so.2 (0x00007f5f8c986000) libcudart.so.11.0 => /usr/local/cuda-11.0/targets/x86_64-linux/lib/libcudart.so.11.0 (0x00007f5f8c708000) libcublas.so.11 => /usr/local/cuda-11.0/targets/x86_64-linux/lib/libcublas.so.11 (0x00007f5f868b8000) libcurand.so.10 => /usr/local/cuda-10.1/targets/x86_64-linux/lib/libcurand.so.10 (0x00007f5f82857000) libcusolver.so.10 => /usr/local/cuda-10.1/targets/x86_64-linux/lib/libcusolver.so.10 (0x00007f5f77d4c000) librt.so.1 => /lib/x86_64-linux-gnu/librt.so.1 (0x00007f5f77b44000) libnvToolsExt.so.1 => /usr/local/cuda/lib64/libnvToolsExt.so.1 (0x00007f5f7793b000) libgfortran.so.4 (0x00007f5f7755c000) libcufft.so.10 => /usr/local/cuda-10.1/targets/x86_64-linux/lib/libcufft.so.10 (0x00007f5f6ef22000) libcuda.so.1 => /usr/lib/x86_64-linux-gnu/libcuda.so.1 (0x00007f5f6da6f000) libnvrtc.so.11.0 => /usr/local/cuda-11.0/targets/x86_64-linux/lib/libnvrtc.so.11.0 (0x00007f5f6c284000) libstdc++.so.6 => /usr/lib/x86_64-linux-gnu/libstdc++.so.6 (0x00007f5f6befb000) libm.so.6 => /lib/x86_64-linux-gnu/libm.so.6 (0x00007f5f6bb5d000) libgomp.so.1 => /usr/lib/x86_64-linux-gnu/libgomp.so.1 (0x00007f5f6b92e000) libgcc_s.so.1 => /lib/x86_64-linux-gnu/libgcc_s.so.1 (0x00007f5f6b716000) libpthread.so.0 => /lib/x86_64-linux-gnu/libpthread.so.0 (0x00007f5f6b4f7000) libc.so.6 => /lib/x86_64-linux-gnu/libc.so.6 (0x00007f5f6b106000) /lib64/ld-linux-x86-64.so.2 (0x00007f5fe6faa000) libcublasLt.so.11 => /usr/local/cuda-11.0/targets/x86_64-linux/lib/libcublasLt.so.11 (0x00007f5f5ff77000) libquadmath.so.0 (0x00007f5f5fd37000)

samskalicky pushed a commit that referenced this pull request Oct 3, 2020
@waytrue17
[v1.8.x] Backport PR #19272 to v1.8.x (#19273)
20830a3 
* static build with cuda 11.0

* add newline at the end of files, add set -e

* add new line

Co-authored-by: Wei Chu <weichu@amazon.com>
@szha
Copy link
Member

szha commented Oct 3, 2020 •
edited

@waytrue17 could you also port this to master once merged? It will help resolve #18657 (comment)

@szha szha requested a review from leezu October 3, 2020 21:22
szha
szha reviewed Oct 3, 2020
View reviewed changes
Copy link
Member

@szha szha left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

A couple more missing items:

@szha
Copy link
Member

szha commented Oct 3, 2020 •
edited

Remember to also configure the CD pipeline for a nightly build of this variant for testing. https://jenkins.mxnet-ci.amazon-ml.com/job/restricted-mxnet-cd/

szha
szha reviewed Oct 3, 2020
View reviewed changes
tools/setup_gpu_build_tools.sh Outdated Show resolved Hide resolved
@szha
Copy link
Member

szha commented Oct 3, 2020

Looks like CUDA 11.1 is available now and we should consider adding that too.

@waytrue17
Copy link
Contributor Author

@waytrue17 could you also port this to master once merged? It will help resolve #18657 (comment)

opened a separate PR for master branch:#19295

@waytrue17
Copy link
Contributor Author

waytrue17 commented Oct 7, 2020 •
edited

A couple more missing items:

@szha Added the missing items and also configured the cd tests. Please review again, thanks

@waytrue17
Copy link
Contributor Author

@mxnet-bot run ci [centos-gpu]

@mxnet-bot
Copy link

Jenkins CI successfully triggered : [centos-gpu]

@lanking520 lanking520 added pr-awaiting-testing PR is reviewed and waiting CI build and test pr-work-in-progress PR is still work in progress and removed pr-awaiting-testing PR is reviewed and waiting CI build and test labels Oct 19, 2020
@waytrue17
Copy link
Contributor Author

@mxnet-bot run ci [sanity]

@mxnet-bot
Copy link

Jenkins CI successfully triggered : [sanity]

@lanking520 lanking520 added pr-awaiting-testing PR is reviewed and waiting CI build and test pr-work-in-progress PR is still work in progress and removed pr-work-in-progress PR is still work in progress pr-awaiting-testing PR is reviewed and waiting CI build and test labels Oct 19, 2020
mseth10
mseth10 approved these changes Oct 20, 2020
View reviewed changes
Copy link
Contributor

@mseth10 mseth10 left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

@lanking520 lanking520 added pr-work-in-progress PR is still work in progress pr-awaiting-testing PR is reviewed and waiting CI build and test pr-awaiting-merge Review and CI is complete. Ready to Merge and removed pr-awaiting-testing PR is reviewed and waiting CI build and test pr-work-in-progress PR is still work in progress labels Oct 20, 2020
@samskalicky samskalicky merged commit 0bc01e9 into apache:v1.x Oct 20, 2020
@samskalicky samskalicky mentioned this pull request Oct 21, 2020
Merged
6 tasks
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Reviewers

@szha szha szha approved these changes

@josephevans josephevans josephevans approved these changes

@ChaiBapchya ChaiBapchya ChaiBapchya approved these changes

@mseth10 mseth10 mseth10 approved these changes

@leezu leezu Awaiting requested review from leezu

Assignees
No one assigned
Labels
pr-awaiting-merge Review and CI is complete. Ready to Merge
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.

None yet
9 participants
@waytrue17 @mxnet-bot @sandeep-krishnamurthy @ChaiBapchya @szha @josephevans @mseth10 @samskalicky @lanking520