NVIDIA / NVIDIA/cuda-python

[MNT] cuda.core: device-attribute catch-up for CUDA 13.1/13.3

Open
#2,367 0 comments 0 reactions 1 assignee View on GitHub

@juenglin is already working on this.

Since Jul 24, 2026.

cuda.core enhancement
Dominant language
Cython
Stars
3.4k
Forks
329
Avg merge
1d 23h
Merged PRs (30d)
116

Description

Summary

DeviceProperties is current through the CUDA 12.9/13.0 attribute additions, but the 13.1 and
13.3 attributes are missing:

Attribute Version Note
CU_DEVICE_ATTRIBUTE_ATOMIC_REDUCTION_SUPPORTED 13.1 companion of #2361
CU_DEVICE_ATTRIBUTE_DMA_BUF_MMAP_SUPPORTED 13.3
CU_DEVICE_ATTRIBUTE_D3D12_CIG_STREAMS_SUPPORTED 13.3
CU_DEVICE_ATTRIBUTE_LOGICAL_ENDPOINT_UNICAST_SUPPORTED 13.3 companion of #2358
CU_DEVICE_ATTRIBUTE_LOGICAL_ENDPOINT_MULTICAST_SUPPORTED 13.3 companion of #2358
CU_DEVICE_ATTRIBUTE_LOGICAL_ENDPOINT_COUNTED_OPS_SUPPORTED 13.3 companion of #2358
CU_DEVICE_ATTRIBUTE_LOGICAL_ENDPOINT_UNICAST_ACCESS_ON_OWNER_DEVICE_SUPPORTED 13.3 companion of #2358

Proposed approach

Mechanical: add properties to DeviceProperties
(_device.pyx)
following the existing naming conventions, with the usual graceful handling on older
drivers/bindings; update the attribute coverage tests. Same shape as the (closed) #675 pass
for earlier toolkits. Doesn't need to wait for #2358/#2361 — the boolean queries are useful on
their own.

References

  • Found during the CUDA 12.8 → 13.3 bindings vs. cuda.core gap sweep (2026-07-14)

-- Leo's bot

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.