Adapt CUDA kernels and build system for CUDA 13.x compatibility #7
Conversation
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
- Update atomics.cuh: fix conditional compilation for atomAdd(double*) to properly exclude CAS-based fallback on CUDA 13.x (SM 70+ only) - Update setup_ops.py: add --allow-unsupported-compiler nvcc flag for CUDA 13.x to handle stricter host compiler version checks Co-authored-by: HydrogenSulfate <23737287+HydrogenSulfate@users.noreply.github.com>
Copilot
AI
changed the title
[WIP] Adapt repository for CUDA 13.x compatibility
Adapt CUDA kernels and build system for CUDA 13.x compatibility
Feb 6, 2026
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
CUDA 13.x drops support for SM < 70 architectures and introduces stricter host compiler checks. The existing
atomicAdd(double*)fallback guard and build flags needed updating.csrc/atomics.cuhatomAdd(double*)fallback: changed(__CUDA_ARCH__ < 600 || CUDA_VERSION < 8000)→(__CUDA_ARCH__ < 600) && (CUDA_VERSION < 13000). The original||was also semantically wrong—it would force the CAS path on SM 600+ whenCUDA_VERSION < 8000, despite native support being available. Under CUDA 13.x the entire fallback branch is now correctly excluded since SM < 60 targets cannot exist.setup_ops.py--allow-unsupported-compilerto nvcc flags whencuda_major >= 13to handle stricter host compiler certification checks in the new toolkit.Note: GPU architecture list for CUDA 13.x (
Turing, Ampere, Ada, Hopper, Blackwell) was already present. All.cukernels use standard device intrinsics (atomicCAS,__ldg,__shfl_{up,down}_sync) that remain supported—no kernel changes needed.Warning
Firewall rules blocked me from connecting to one or more addresses (expand for details)
I tried to connect to the following addresses, but was blocked by firewall rules:
developer.nvidia.com/home/REDACTED/work/_temp/ghcca-node/node/bin/node /home/REDACTED/work/_temp/ghcca-node/node/bin/node --enable-source-maps /home/REDACTED/work/_temp/copilot-developer-action-main/dist/index.js(dns block)docs.nvidia.com/home/REDACTED/work/_temp/ghcca-node/node/bin/node /home/REDACTED/work/_temp/ghcca-node/node/bin/node --enable-source-maps /home/REDACTED/work/_temp/copilot-developer-action-main/dist/index.js(dns block)If you need me to access, download, or install something from one of these locations, you can either:
Original prompt
💬 We'd love your input! Share your thoughts on Copilot coding agent in our 2 minute survey.