Skip to content

Use build dir for big files/folders while building TensorFlow#3371

Merged
branfosj merged 3 commits intoeasybuilders:developfrom
Flamefire:tensorflow-tmp-dir
Jul 3, 2024
Merged

Use build dir for big files/folders while building TensorFlow#3371
branfosj merged 3 commits intoeasybuilders:developfrom
Flamefire:tensorflow-tmp-dir

Conversation

@Flamefire
Copy link
Copy Markdown
Contributor

@Flamefire Flamefire commented Jun 20, 2024

build_pip_package (executed after Bazel finishes) uses a temporary folder which gets large (2 GB for TF 2.9).

It uses $TMPDIR (via mktemp) which we set to the temporary folder created by Easybuild in PythonPackage but that might have only enough space for logs etc.
As this folder is part of the build put it inside the build directory.

build_pip_package (executed after Bazel finishes) uses a temporary
folder which gets large (2 GB for TF 2.9).
It uses `$TMPDIR` (via mktemp) which we set to the temporary folder
created by Easybuild in PythonPackage but that might have only enough
space for logs etc.
As this folder is part of the build put it inside the build directory.
@Flamefire Flamefire force-pushed the tensorflow-tmp-dir branch from 9152a10 to 1a297ea Compare June 26, 2024 07:13
The build already creates files in $TMPDIR, so set that for the whole
build and reset afterwards.
Same issue here, use a context manager to avoid duplicate code
@Flamefire Flamefire changed the title Use build dir for building TensorFlow pip package Use build dir for big files/folders while building TensorFlow Jun 28, 2024
@Flamefire
Copy link
Copy Markdown
Contributor Author

Test report by @Flamefire

Overview of tested easyconfigs (in order)

  • SUCCESS TensorFlow-2.7.1-foss-2021b-CUDA-11.4.1.eb

Build succeeded for 1 out of 1 (1 easyconfigs in total)
n1336 - Linux RHEL 8.7 (Ootpa), x86_64, Intel(R) Xeon(R) Platinum 8470 (icelake), Python 3.8.13
See https://gist.github.com/Flamefire/2d317abd7d09ebf2caf5f9c3e9ffefef for a full test report.

@branfosj
Copy link
Copy Markdown
Member

branfosj commented Jul 2, 2024

Test report by @branfosj

Overview of tested easyconfigs (in order)

  • SUCCESS TensorFlow-2.2.3-foss-2020b.eb
  • SUCCESS TensorFlow-2.7.1-foss-2021b.eb

Build succeeded for 2 out of 2 (2 easyconfigs in total)
bear-pg0211u03a.bear.cluster - Linux RHEL 8.6, x86_64, Intel(R) Xeon(R) Gold 6248 CPU @ 2.50GHz (cascadelake), Python 3.6.8
See https://gist.github.com/branfosj/128a6e8c810c8ffee72c11d121197523 for a full test report.

@branfosj branfosj added this to the release after 4.9.2 milestone Jul 3, 2024
@branfosj
Copy link
Copy Markdown
Member

branfosj commented Jul 3, 2024

Going in, thanks @Flamefire!

@branfosj branfosj merged commit 11230bc into easybuilders:develop Jul 3, 2024
@Flamefire Flamefire deleted the tensorflow-tmp-dir branch July 3, 2024 09:03
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants