Hi Nathaniel,
Will you be providing portable Linux wheels aka manylinux1?
https://www.python.org/dev/peps/pep-0513/
Does this also open up the door to releasing wheels for SciPy
too?
While speeding up "pip install" would be of benefit in itself,
I am particularly keen to see this for use within automated
testing frameworks like TravisCI where currently having to
install NumPy (and SciPy) from source is an unreasonable
overhead.
Many thanks to everyone working on this,
Peter
On Tue, Mar 15, 2016 at 11:33 PM, Nathaniel Smith
Hi all,
Just a heads-up that we're planning to upload Linux wheels for numpy to PyPI soon. Unless there's some objection, these will be using ATLAS, just like the current Windows wheels, for the same reasons -- moving to something faster like OpenBLAS would be good, but given the concerns about OpenBLAS's reliability we want to get something working first and then worry about making it fast. (Plus it doesn't make sense to ship different BLAS libraries on Windows versus Linux -- that just multiplies our support burden for no reason.)
-n
-- Nathaniel J. Smith -- https://vorpus.org _______________________________________________ NumPy-Discussion mailing list NumPy-Discussion@scipy.org https://mail.scipy.org/mailman/listinfo/numpy-discussion