Am Tue, 26 Sep 2017 11:33:04 -0500 schrieb Jason Bacon <bacon4000%gmail.com@localhost>: > Anyone interested please test wip/openblas on your favorite platform(s) How is the relation of that one to wip/OpenBLAS? I patched that up for local use here and intended to push my patches that also make use of it in other packages: http://src.rrz.uni-hamburg.de/extra/pkgsrc-patches/wip-OpenBLAS-enhancement.patch http://src.rrz.uni-hamburg.de/extra/pkgsrc-patches/wip-suitesparse-openblas.patch http://src.rrz.uni-hamburg.de/extra/pkgsrc-patches/math-R-openblas.patch http://src.rrz.uni-hamburg.de/extra/pkgsrc-patches/math-py-numpy-openblas.patch http://src.rrz.uni-hamburg.de/extra/pkgsrc-patches/math-octave-moredeps.patch Did you start from scratch? I'm all for merging the efforts into wip/openblas, as the CamelCase annoys me in a package name. I pushed my patch to wip/OpenBLAS itself now. Maybe you can have a look and port things over into wip/openblas? Then we can add the dependencies of the other packages with the new name. So far I tested my version of OpenBLAS on CentOS 7, including rather nice performance comparison to Intel MKL with R, for example. Alrighty then, Thomas -- Dr. Thomas Orgis Universität Hamburg RRZ / Basis-Infrastruktur / HPC Schlüterstr. 70 20146 Hamburg Tel.: 040/42838 8826 Fax: 040/428 38 6270
Attachment:
smime.p7s
Description: S/MIME cryptographic signature