#6456: Upgrade cvxopt in sage from 0.9 to 1.1.3
-------------------------------------------------+--------------------------
Reporter: was | Owner: mabshoff
Type: defect | Status: needs_info
Priority: major | Milestone: sage-4.6.1
Component: packages | Keywords:
Author: Harald Schilly, Dmitrii Pasechnik | Upstream: Completely
fixed; Fix reported upstream
Reviewer: | Merged:
Work_issues: licence |
-------------------------------------------------+--------------------------
Changes (by drkirkby):
* status: needs_review => needs_info
Comment:
Dima,
I don't have the maths knowledge to fully understand all this, but you are
an expert in this area, so I'm going to accept all what you say.
I'm not convinced the GPLK doctest will necessarily pass on Solaris 10
SPARC systems, due to different rounding issues. As you note, GPLK does
give an exact answer on my Intel system and your system. However, it's far
from obvious this will give the exact answer on SPARC, due to differences
in the rounding of the floating point processors. The Intel/AMD use
80-bits for intermediate calculations, whereas SPARC uses 64. All
processors return their answers using 64-bits.
If you can test this on a Solaris 10 SPARC system, showing the
* The doctests pass
* The package's set-test pass (set {{{SAGE_CHECK=yes}}} when building
cvxopt)
then I'm willing to give this a positive review! You can use 'mark',
'mark2' or t2.math for testing on Solaris 10 SPARC.
If you use t2.math, you can set {{{SAGE_ATLAS_LIB=/usr/local/ATLAS32}}}
which should mean that ATLAS does not have to be built, so the build time
will be significantly reduced. (I've just built the whole of Sage on
t2.math, and have copied the necessary 32-bit ATLAS files to
/usr/local/ATLAS32). I have not built Sage using those files, but I'm
pretty confident that setting {{{SAGE_ATLAS_LIB}}} will work. If not,
ATLAS will have to be built, but I doubt that will be necessary. So it
should be possible to build Sage on t2.math in a few hours.
Note Mercurial is not currently installed on t2.math, so it might be
easier if you create a patched tar file and copy that over. I will install
a new Python and Mercurial at some point on t2.math, but this is not yet
set up. I only got t2.math working today after William messed it up.
Sorry, I don't have time to test this myself, but if you can show the
packages self-tests and the doctests pass on Solaris 10 SPARC, I'll give
this a positive review.
The ticket has been open for 17 months, but it should be possible to get a
positive review in well under 17 hours!
Dave
--
Ticket URL: <http://trac.sagemath.org/sage_trac/ticket/6456#comment:120>
Sage <http://www.sagemath.org>
Sage: Creating a Viable Open Source Alternative to Magma, Maple, Mathematica,
and MATLAB
--
You received this message because you are subscribed to the Google Groups
"sage-trac" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to
[email protected].
For more options, visit this group at
http://groups.google.com/group/sage-trac?hl=en.
