Issues

• [BUG]: CUtensorMap owned storage does not honor required alignment

  bug
  NVIDIA/cuda-python#3181 · Filter by author isVoid opened on Oct 21, 2026
  1

• [CI failure] CI: Nightly optional-deps scheduled runs

  bugCI/CDci-workflow-healthtriage
  NVIDIA/cuda-python#3180 · Filter by author github-actions opened on Oct 21, 2026

• [CI failure] CI scheduled runs

  bugCI/CDci-workflow-healthtriage
  NVIDIA/cuda-python#3171 · Filter by author github-actions opened on Oct 21, 2026

• [BUG]: cuFile BAR-size test rejects zero on GB10 integrated GPUs

  triage
  NVIDIA/cuda-python#3167 · Filter by author isVoid opened on Oct 20, 2026

• [BUG]: NVML PCI discovery test requires administrator privileges on Linux

  triage
  NVIDIA/cuda-python#3166 · Filter by author isVoid opened on Oct 20, 2026

• [CI failure] CI on main

  bugCI/CDci-workflow-healthtriage
  NVIDIA/cuda-python#3161 · Filter by author github-actions opened on Oct 20, 2026
  4

• [BUG]: Timing dependent concurrency flake via test_destroying_child_node_invalidates_embedded_handles

  bug
  NVIDIA/cuda-python#3156 · Filter by author jpascucci-nv opened on Oct 19, 2026
  

• [BUG]: Windows system search does use PATH, contrary to docs

  bug
  NVIDIA/cuda-python#3137 · Filter by author yangcal opened on Oct 16, 2026
  1

• cuda.core: warn when a child is forked from a process that initialized CUDA

  cuda.coreenhancementP2
  NVIDIA/cuda-python#3127 · Filter by author juenglin opened on Oct 15, 2026 · cuda.core 1.3.0
  

• cuda.core: transfer VMM IPC allocations one at a time

  cuda.coreenhancementP2
  NVIDIA/cuda-python#3110 · Filter by author Andy-Jost opened on Oct 14, 2026 · cuda.core next
  

• test_definition_node_update_changes_future_instantiations[event-record] is racy: it expects a graph launch to mark the event pending before the host callback starts

  bugcuda.coreP1
  NVIDIA/cuda-python#3106 · Filter by author Andy-Jost opened on Oct 14, 2026
  1

• [CI failure] CI: pixi run test (source build) scheduled runs

  bugCI/CDci-workflow-healthtriage
  NVIDIA/cuda-python#3102 · Filter by author github-actions opened on Oct 14, 2026
  7

• cuda.core: import fails with undefined symbol cudaGetErrorString when _tensor_map is built with -z now (after #2940)

  bugcuda.coreP1
  NVIDIA/cuda-python#3098 · Filter by author Andy-Jost opened on Oct 13, 2026

• cuda.core: IPC pickling rebuilds memory resources with the sender's device indices

  bugcuda.coreP1
  NVIDIA/cuda-python#3097 · Filter by author Andy-Jost opened on Oct 13, 2026 · cuda.core 1.3.0
  

• cuda-bindings 13.4.3 & 12.9.10 release checklist

  cuda.bindingsP0
  NVIDIA/cuda-python#3094 · Filter by author leofang opened on Oct 13, 2026
  

• cuda-core v1.2.2 release checklist

  cuda.coreP0
  NVIDIA/cuda-python#3093 · Filter by author leofang opened on Oct 13, 2026
  

• [DOC]: 7 dead external links in samples/*/README.md fail pre-commit run --all-files (lychee 404s)

  documentation
  Documentation

  NVIDIA/cuda-python#3075 · Filter by author lijinf2 opened on Oct 12, 2026 · cuda.core 1.2.x
  

• Investigate mini-CTK cache duplication after HTTP 429 restore failures

  CI/CD
  Bug

  NVIDIA/cuda-python#3074 · Filter by author rwgk opened on Oct 11, 2026

• MNT: Switch to the "src" layout

  cuda.bindingscuda.corecuda.pathfindersupport
  Task

  NVIDIA/cuda-python#3048 · Filter by author leofang opened on Oct 7, 2026 · cuda.core next
  

• CI: Random segfaults on Windows

  bugcuda.coreP0
  Bug

  NVIDIA/cuda-python#3038 · Filter by author leofang opened on Oct 6, 2026 · cuda.core 1.3.0
  2

  

• Define the isolation guarantee for installed-wheel CI

  CI/CD
  Task

  NVIDIA/cuda-python#3016 · Filter by author rwgk opened on Oct 4, 2026 · cuda.bindings 13.5.0 & 12.9.10
  3

  

• CI: Reduce traffic and slow down size-bloating of gh-pages

  CI/CDP1triage
  NVIDIA/cuda-python#3014 · Filter by author leofang opened on Oct 3, 2026 · cuda.core 1.3.0
  1

  

• CI: Consolidate the two release workflows

  CI/CDP1
  NVIDIA/cuda-python#3012 · Filter by author leofang opened on Oct 2, 2026 · cuda.pathfinder next
  4

  

• cuda.core: release synchronous allocations in stream order through a host callback instead of blocking

  cuda.coreenhancement
  NVIDIA/cuda-python#2989 · Filter by author Andy-Jost opened on Oct 1, 2026 · cuda.core next
  

• EPIC: VMM feature completeness

  cuda.corefeatureP0
  NVIDIA/cuda-python#2980 · Filter by author leofang opened on Sep 30, 2026 · cuda.core 1.3.0
  2