github NVIDIA/numba-cuda v0.30.1

7 hours ago

What's Changed

  • Fix CABI return type mismatch when calling Numba-ABI subroutines by @isVoid in #852
  • Fix CUDA context leak in with cuda.gpus[N]: context manager by @isVoid in #855
  • fix: support np.count_nonzero in device functions by @kaeun97 in #854
  • sync with cuda-python and add 2-gpu runners by @leofang in #814
  • Sync CI with cuda-python by @leofang in #858
  • Remove Conflicting Lowering for FP8->Int8 Conversion by @isVoid in #869
  • Flip to stable rapids release in thirdparty tests by @brandon-b-miller in #863
  • Alternate half of linux arm64 CI runners from a100 to l4 by @leofang in #864
  • build(deps): bump the actions-monthly group across 1 directory with 5 updates by @dependabot[bot] in #860
  • Fix Windows path parsing in generate_raw_ltoir.py by @rluo8 in #867
  • Bump rapidsai/sccache tag to v0.14.0-rapids.9 by @leofang in #871
  • Fix indentation bug in test_linker.py by @leofang in #875
  • Bump Version to 0.30.1 by @isVoid in #873

New Contributors

Full Changelog: v0.30.0...v0.30.1

Don't miss a new numba-cuda release

NewReleases is sending notifications on new releases.