All pull requests

Pull requests

• Support interop with active green contexts

  #853·caugonnet opened Apr 9, 2026Contributor·

  ·
  3

• Map types.NoneType return type to LLVM void instead of i8*

  #847·isVoid opened Apr 3, 2026Contributor·6 tasks done·

  ·
  1

  5

  

• add fma arg to disable compiling with FMA

  #839·edoyango opened Mar 24, 2026·

  ·
  1

  1

• Delegate all path discovery in cuda_paths.py to pathfinder

  #810·cpcloud opened Feb 26, 2026Contributor·

  ·

  · numba-cuda next
  1

  5

  

• Find libdevice via pathfinder

  0 - Blocked
  #798·brandon-b-miller opened Feb 19, 2026Contributor·

  ·

  · numba-cuda next
  4

  

• Create Highlevel Bindings for FP8 Datatype

  3 - Ready for Review
  #788·isVoid opened Feb 11, 2026Contributor·

  ·
  1

  10

• [WIP] Second pass at LaunchConfig

  2 - In Progress
  #727·tpn opened Jan 20, 2026Contributor·

  ·
  8

• perf: Reduce copy overhead in inline passes

  4 - Waiting on reviewer
  #689·ccam80 opened Jan 2, 2026Contributor·

  ·
  15

• {WIP] Follow-on from #612: try to align with NumPy

  3 - Ready for Review
  #613·gmarkall opened Nov 27, 2025Contributor·

  ·
  8

• Enable global lists within kernels

  4 - Waiting on author
  #584·brandon-b-miller opened Nov 6, 2025Contributor·

  ·
  1

  6

• Vendor numpy median, percentile, and quantile reduction tests

  2 - In Progress
  #567·brandon-b-miller opened Oct 31, 2025Contributor·

  ·
  6

• Deprecate the DeviceNDArray class and public APIs that return instances

  2 - In Progress
  #546·brandon-b-miller opened Oct 24, 2025Contributor·

  ·

  · numba-cuda backlog
  84

• [Testing] Add some numpy array testing for the CUDA target

  3 - Ready for Review
  #535·atmnp opened Oct 20, 2025Contributor·

  ·
  4

  

• Allow registering an ArgHandler

  4 - Waiting on author
  #504·brandon-b-miller opened Oct 3, 2025Contributor·

  ·
  37

• Fix packaging quirks

  2 - In Progress
  #489·kkraus14 opened Sep 24, 2025Contributor·

  ·
  1

  2

• Feature: allow return array

  4 - Waiting on reviewer
  #335·ZzEeKkAa opened Jul 21, 2025Contributor·

  ·
  1

  30

  

• Port change and test from numba/numba#9595 to Numba-CUDA

  0 - Blocked
  #292·gmarkall opened Jun 12, 2025Contributor·

  ·
  4

• Implement a thread-local means to access kernel launch config.

  2 - In Progress
  #288·tpn opened Jun 10, 2025Contributor·

  ·
  14

• [WIP] Add CUDAArray type and implementation with addresspace information

  2 - In Progressbreakingimprovement
  #236·ZzEeKkAa opened May 1, 2025Contributor·

  ·
  11

• Add Compiler timer to Dispatcher.compile and Dispatcher.compile_device

  4 - Waiting on reviewer
  #227·isVoid opened Apr 18, 2025Contributor·

  ·
  1

  5

• NRT Doc update

  2 - In Progress
  #195·brandon-b-miller opened Apr 11, 2025Contributor·

  ·
  2

• Add CUDA C ABI arg packer ensuring compatibility of argument types with CUDA C ABI

  4 - Waiting on author4 - Waiting on reviewer
  #179·shwina opened Mar 26, 2025Contributor·

  ·
  1

  18

• Add col/row-major layout to shared and local memory

  4 - Waiting on author
  #176·ZzEeKkAa opened Mar 25, 2025Contributor·

  ·
  6

• Make block, thread, and warp indices unsigned

  2 - In Progress
  #110·gmarkall opened Jan 9, 2025Contributor·

  ·
  1