Graham Markall comments

Results 776 comments of


                                            Graham Markall

Use `omp_set_max_active_levels()` instead of `omp_set_nested()`

@ziofil I'd be surprised if the message were related to the crash - I'd suggest starting a thread on https://numba.discourse.group describing your problem, and hopefully you can get some more...

Use `omp_set_max_active_levels()` instead of `omp_set_nested()`

> What do you think @gmarkall ? That sounds good to me. I'll update this PR to do as suggested.

Use `omp_set_max_active_levels()` instead of `omp_set_nested()`

Note to self: reproducing the fail may be possible on an x86 box with many cores.

literal_unroll leads to segfault on M1 when called with bool

If LLVM came from homebrew, then it probably doesn't have this patch applied: https://github.com/numba/llvmlite/blob/52970defcdd82c68a420e1dc182a814c7c4b16e3/conda-recipes/llvm_11_consecutive_registers.patch - this seems to cause this kind of failure. There are more details in https://github.com/gmarkall/cf-llvmlite-0.39-issue-repro

CUDA: Fix Issue #8309 - atomics don't work on complex components

gpuci run tests

CUDA: Fix Issue #8309 - atomics don't work on complex components

@stuartarchibald Thanks for the review - typo fix committed. gpuci run tests

CUDA: Fix Issue #8309 - atomics don't work on complex components

Main is now merged, so hopefully that will fix the recipes.

CUDA: Fix Issue #8309 - atomics don't work on complex components

`main` has been merged in again, since #8356 was merged - hopefully this is now good for the CPU buildfarm @stuartarchibald / @esc? gpuci run tests

presence of while loop breaks literal_unroll compilation

Thanks for the report - I can reproduce the compilation failure.

presence of while loop breaks literal_unroll compilation

A smaller reproducer: ```python from numba import njit, literal_unroll @njit('(Tuple((int64, float64)),)') def f2(t): for t_i in literal_unroll(t): pass while True: x = 0 ```