CUDA 13.2 and torch 2.12.1 compatible Dockerfile. Volta (SM70) support dropped#110
Merged
Conversation
…t dropped Signed-off-by: Thorsten Kurth <tkurth@nvidia.com>
azrael417
force-pushed
the
tkurth/cuda-13-upgrade
branch
from
a519140 to
bd68e86
Compare
Signed-off-by: Thorsten Kurth <tkurth@nvidia.com>
Signed-off-by: Thorsten Kurth <tkurth@nvidia.com>
Signed-off-by: Thorsten Kurth <tkurth@nvidia.com>
…u 24.04 (#111) Mirror the CUDA 13.2 + torch 2.12.1 upgrade into Dockerfile_gnu (CUDA base image 13.2.0, HPCX cuda13 build, updated pip deps). Move both Dockerfile and Dockerfile_gnu to ubuntu24.04, which requires: - --break-system-packages on pip installs (PEP 668 / externally managed py3.12) - python3.10 -> python3.12 dist-packages paths Signed-off-by: Thorsten Kurth <tkurth@nvidia.com> Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Match the other Dockerfiles: bump base to ubuntu:24.04 and CPU torch to 2.12.1. Ubuntu 24.04 ships Python 3.12 (PEP 668 externally managed), so add --break-system-packages to pip installs and update the python3.10 -> python3.12 dist-packages path. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Signed-off-by: Thorsten Kurth <tkurth@nvidia.com>
The CUDA -devel base image ships no NCCL, so cmake's fallback system search for nccl.h failed. Install nvidia-nccl-cu13 via pip, add the libnccl.so symlink, and set TORCHFORT_NCCL_ROOT to it -- matching the nvhpc Dockerfile. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Signed-off-by: Thorsten Kurth <tkurth@nvidia.com>
torch 2.12.1/cu132 requires libnvshmem_host.so.3, which the plain CUDA devel image does not ship (the NVHPC SDK does, which is why the nvhpc Dockerfile works without it). Without it `import torch` fails, breaking the cmake CMAKE_PREFIX_PATH query. Install nvidia-nvshmem-cu13==3.4.5 (torch's pin) and add its lib dir to LD_LIBRARY_PATH. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Signed-off-by: Thorsten Kurth <tkurth@nvidia.com>
Signed-off-by: Thorsten Kurth <tkurth@nvidia.com>
romerojosh
approved these changes
Jul 8, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
This MR updates the Dockerfile to support CUDA 13.2.