Support CUDA 13 cuBLAS libraries - #896
Open
jeongseok-meta wants to merge 1 commit into
Open
Conversation
8 tasks
jeongseok-meta
marked this pull request as ready for review
August 30, 2026 19:08
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Why: CUDA 13 provides
libcublas.so.13/cublas64_13.dll, while the runtime loader currently tries only cuBLAS 11, 12, and 10. As a result, Quadrants' optional CUDA linear-algebra paths cannot load cuBLAS from a CUDA 13-only environment. This surfaced while packaging Quadrants 1.3 in conda-forge/staged-recipes#34661.What: Accept the CUDA 13 cuBLAS SONAME while preserving the existing CUDA 11/12 preference used for PyTorch-compatible environments.
How:
13after versions11and12inCUBLASDriver::load_cublas().Evidence: The identical source patch was built into a conda package and exercised on an NVIDIA RTX 3080 with
cuda-version 13.0, cuBLAS 13.1.1, cuSPARSE 12.6.3, and cuSOLVER 12.0.4. A real sparse conjugate-gradient solve completed and matched NumPy; details are recorded in the staged-recipes validation comment.Test plan:
Run URL: local only (no run URL emitted)
No public API changes.