Skip to content
esc edited this page Apr 18, 2024 · 1 revision

Numba Meeting: 2024-04-09

Attendees: Da Li, Todd Anderson, Siu Kwan Lam, Graham Markall FPOC (last week): Graham FPOC (incoming): Siu

NOTE: All communication is subject to the Numba Code of Conduct.

Please refer to this calendar for the next meeting date.

0. Discussion

  • Triage tasks:

    • numba 0.60.0 is planned to be released ASAP now that NumPy 2.0 conda packages are available in testing channels.
    • completed 0.60.0 milestone pruning; moved non-critical items to 0.61.0 milestone
  • numba-scipy release 0.4 in progress :: Questions

    • What range of Python to support
    • What range of NumPy to support
    • What range of SciPy to support
    • Azure vs GHA -- can we drop azure?
    • Update to 3.12 going to come in 0.5 release due to complexity of upgrade
    • NumPy 2.0 support?
  • Review ping for llvmlite #1025: skip raise basic blocks in verifyFanoutBackward

  • Scope for rolling llvmlite PR #979 (Update llvmdev to build compiler-rt) into #1036 (LLVM 15 conda recipe)

  • LLVM 15 upgrade update: New Pass Manager, etc.

  • Short term roadmap: https://hackmd.io/XoqNGgLhR06icOumaybC2g

New "Ready for Review" PRs

1. New Issues

  • numba#9522 - Jitclass + numpy functions can't compile with attribute as second parameter
  • numba#9525 - Unable to convert generator returning 1D arrays into a 2D array using fromiter/vstack
  • numba#9526 - CUDA: Error handling variables not added to the @llvm.used list
  • numba#9527 - Support for Advanced Numpy dtypes

Closed Issues

2. New PRs

  • numba#9521 - Support binop reduction.
  • numba#9523 - Test Codecov
  • numba#9524 - Introduce context.call_overload and raise dynamic exception in guard_match_core_dims

Closed PRs

3. Short-term Roadmap

2024-gantt: TBD 2023-gantt: https://github.com/numba/numba/issues/8971

Clone this wiki locally