Description of problem: Please BuildRequire: openblas instead of atlas where available (x86_64 %{ix86} armv7hl %{power64} aarch64 s390x, see %{openblas_arches} macro defined in openblas-srpm-macros package). Currently, it's available on all primary Fedora arches, so you can technically skip using the macro. Version-Release number of selected component (if applicable): iml-1.0.5-8.fc29 Additional information: OpenBLAS selects best code path for the target architecture at runtime unlike ATLAS. A lot of benchmarks around the net suggest openblas is faster than ATLAS in most workloads: https://github.com/tmolteno/necpp/issues/18 http://blog.nguyenvq.com/blog/2014/11/10/optimized-r-and-python-standard-blas-vs-atlas-vs-openblas-vs-mkl/ https://freddie.witherden.org/pages/blas-gemm-bench/ http://nghiaho.com/?p=1726 Many Fedora packages have switched to openblas already (e.g. scalapack and all dependent ones and numpy).
Same comment on a bunch of packages, but I'll just make it here: we need to switch the entire Macaulay2 / sagemath stack at once, or else we'll wind up with binaries that are linked with both atlas and openblas. Unfortunately, the fflas-ffpack testsuite is segfaulting on s390x when built with openblas. This may be a bug in the fflas-ffpack code, but it is suspicous that no such segfault occurs with atlas, nor with the reference blas, and that with openblas it only occurs on s390x. See the referenced bug. That will have to be fixed prior to converting the rest of the stack to openblas. See https://koji.fedoraproject.org/koji/taskinfo?taskID=29190683 for a scratch build demonstrating the problem.
Built in Rawhide.