Conversation
Extend the cuDNN and NVSHMEM version registries with every release NVIDIA has published since the current newest pins (cuDNN 9.22.0, NVSHMEM 3.6.5): cuDNN 9.23.0, 9.23.1, 9.23.2, 9.24.0, 9.25.0 and NVSHMEM 3.7.0, 3.7.1, 3.7.2. Each manifest sha256 was computed from the manifest downloaded from developer.download.nvidia.com. The existing BUILD templates work unchanged with the new versions: the cuDNN sub-library set indexed from LIB_VERSIONS (ops, cnn, adv, graph, engines_precompiled, engines_runtime_compiled, heuristic) and the static archives are all present in the 9.23-9.25 linux-x86_64 and linux-sbsa archives for both cuda12 and cuda13 variants, as are libnvshmem_host.so and libnvshmem_device.a in the 3.7.x archives. The new top-level manifest entries (cudnn_jit, cudnn_samples, nvshmem_python) are ignored by the component registries. Bump the e2e pins for CUDA 13.2.1 to cuDNN 9.24.0 / NVSHMEM 3.7.1 and for CUDA 13.3.1 to cuDNN 9.25.0 / NVSHMEM 3.7.2 so the new pins are exercised, keeping older pins on CUDA 13.0.2 / 13.1.1 for coverage of existing versions.
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Extend the cuDNN and NVSHMEM version registries with every release NVIDIA has published since the current newest pins (cuDNN 9.22.0, NVSHMEM 3.6.5): cuDNN 9.23.0, 9.23.1, 9.23.2, 9.24.0, 9.25.0 and NVSHMEM 3.7.0, 3.7.1, 3.7.2. Each manifest sha256 was computed from the manifest downloaded from developer.download.nvidia.com.
The existing BUILD templates work unchanged with the new versions: the cuDNN sub-library set indexed from LIB_VERSIONS (ops, cnn, adv, graph, engines_precompiled, engines_runtime_compiled, heuristic) and the static archives are all present in the 9.23-9.25 linux-x86_64 and linux-sbsa archives for both cuda12 and cuda13 variants, as are libnvshmem_host.so and libnvshmem_device.a in the 3.7.x archives. The new top-level manifest entries (cudnn_jit, cudnn_samples, nvshmem_python) are ignored by the component registries.
Bump the e2e pins for CUDA 13.2.1 to cuDNN 9.24.0 / NVSHMEM 3.7.1 and for CUDA 13.3.1 to cuDNN 9.25.0 / NVSHMEM 3.7.2 so the new pins are exercised, keeping older pins on CUDA 13.0.2 / 13.1.1 for coverage of existing versions.