Skip to content

update and fix templates in cuDNN + NCCL easyconfigs#23856

Merged
Micket merged 1 commit intoeasybuilders:developfrom
Flamefire:cudnn-templates
Nov 7, 2025
Merged

update and fix templates in cuDNN + NCCL easyconfigs#23856
Micket merged 1 commit intoeasybuilders:developfrom
Flamefire:cudnn-templates

Conversation

@Flamefire
Copy link
Copy Markdown
Contributor

@Flamefire Flamefire commented Sep 12, 2025

Use the correct templates for CUDA and software versions to avoid local variables and double-templates

This is just a cleanup for easier updating going forward. Tests with --fetch --force-redownload suffice for this

@Flamefire Flamefire force-pushed the cudnn-templates branch 4 times, most recently from 5efd303 to 4353391 Compare September 12, 2025 15:21
Use the correct templates for CUDA and software versions to avoid local
variables and double-templates
@Flamefire
Copy link
Copy Markdown
Contributor Author

Flamefire commented Sep 15, 2025

Test report by @Flamefire
FAILED
Build succeeded for 22 out of 25 (25 easyconfigs in total)
login2.romeo.hpc.tu-dresden.de - Linux Rocky Linux 9.6, x86_64, AMD EPYC 7702 64-Core Processor (zen2), Python 3.9.21
See https://gist.github.com/Flamefire/8296fff4501b2033485e6b537e116b29 for a full test report.

The 3 NCCL ECs fail as they need to be downloaded manually.

But works:

WARNING: Found file nccl_2.8.3-1+cuda11.1_x86_64.txz at /sources/n/NCCL/nccl_2.8.3-1+cuda11.1_x86_64.txz, but re-downloading it anyway...
WARNING: Found file nccl_2.9.9-1+cuda11.3_x86_64.txz at /sources/n/NCCL/nccl_2.9.9-1+cuda11.3_x86_64.txz, but re-downloading it anyway...

Copy link
Copy Markdown
Contributor

@Micket Micket left a comment

Choose a reason for hiding this comment

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

lgtm

@Micket Micket merged commit 3808e0c into easybuilders:develop Nov 7, 2025
8 checks passed
@Micket Micket added this to the next release (5.2.0?) milestone Nov 7, 2025
@Flamefire Flamefire deleted the cudnn-templates branch November 7, 2025 11:19
@boegel boegel changed the title Update and fix templates in cuDNN update and fix templates in cuDNN + NCCL easyconfigs Dec 23, 2025
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.

3 participants