Package Details: python-numpy-mkl 1.19.1-1

Git Clone URL: https://aur.archlinux.org/python-numpy-mkl.git (read-only, click to copy)
Package Base: python-numpy-mkl
Description: Scientific tools for Python, compiled with intel mkl
Upstream URL: http://numpy.scipy.org/
Keywords: numpy
Licenses: custom
Conflicts: python-numpy
Provides: python-numpy=1.19.1
Submitter: bred
Maintainer: RunAndTumble
Last Packager: monson
Votes: 32
Popularity: 0.158959
First Submitted: 2012-08-25 16:55
Last Updated: 2020-07-26 06:56

Required by (1000)

Sources (1)

Latest Comments

1 2 3 4 5 6 ... Next › Last »

petronny commented on 2020-08-05 11:48

@RunAndTumble Do you want to add icc back?

Building numpy with icc is introduced in https://software.intel.com/content/www/us/en/develop/articles/numpyscipy-with-intel-mkl.html and I think it's an official document.

We should always use the official way to build every package unless we have to patch it.

monson commented on 2020-07-27 18:17

@RunAndTumble Sorry, I thought you were too busy for that. icc support has been dropped because it doesn't accelerate much, the heavy computing parts are already using mkl. Also we don't need to deal with icc's license.

RunAndTumble commented on 2020-07-27 01:25

Eh, that was unexpected.

RunAndTumble commented on 2019-11-18 05:29

Sorry, folks, I'm finishing my thesis and barely had enough time for anything else. v1.17.4 based on python3.8 and the latest icc. I completely removed python2 support. One of the tests fails, I observed a similar issue before, I'll try to find the problem.

birchwhitebrain commented on 2019-11-17 16:23

Since python upgraded to 3.8.0, this package no longer works. I wonder if it will be updated soon?

RunAndTumble commented on 2019-04-26 23:23

v 1.16.3 is out, all test are fine

RunAndTumble commented on 2019-02-10 02:12

I've updated the package, waiting for 1.16.2 is pointless. One of the tests has to fail, according to github, this is just a bad test, should be removed in 1.16.2. I also plan to add a patch that allows to build scipy-mkl w/o any problems. However, I'm not sure that you can use distutils properly after that.

RunAndTumble commented on 2019-02-02 13:51

Sorry, Folks, I think we should wait for v1.16.2 A few things seem to be broken, I tried python-numpy from the repo: it has the same issue.

RunAndTumble commented on 2019-01-14 12:43

1.16 is out

petronny commented on 2018-11-22 04:29

1.15.4 is out. Please update it.