Skip to content

Conversation

@isuruf
Copy link
Contributor

@isuruf isuruf commented Jul 14, 2021

When building with -fdefault-integer=8, the type changes, but
only some compilers like gfortran implements omp_set_num_threads
with integer*8 input. When swapping the openmp library with
LLVM openmp, this doesn't work yet. Until LLVM openmp is fixed
explicitly setting the type fixes the build issue.

Description

Checklist

  • The documentation has been updated.
  • If the PR solves a specific issue, it is set to be closed on merge.

cc @martin-frbg, @h-vetinari

When building with -fdefault-integer=8, the type changes, but
only some compiler like gfortran implements `omp_set_num_threads`
with integer*8 input. When swapping the openmp library with
LLVM openmp, this doesn't work yet. Until LLVM openmp is fixed
explicitly setting the type fixes the build issue.
@h-vetinari
Copy link
Contributor

Thanks @isuruf!

Some references for where this is coming from:
conda-forge/blas-feedstock#74
conda-forge/openblas-feedstock#122 (particularly this comment ff.)
OpenMathLib/OpenBLAS#3308

martin-frbg
martin-frbg previously approved these changes Jul 14, 2021
@martin-frbg
Copy link
Collaborator

appveyor failure is unrelated conda update problem - some kind of circular dependency between conda and tqdm. Suggest setting conda config --set auto_update_conda false in appveyor.yml for now (which is what I did for OpenBLAS)

@weslleyspereira weslleyspereira merged commit 3b26987 into Reference-LAPACK:master Jul 21, 2021
@isuruf isuruf deleted the n_threads branch July 21, 2021 22:02
@julielangou julielangou added this to the LAPACK 3.10.1 milestone Nov 12, 2022
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

5 participants