On Jun 7, 2010, at 12:10 PM, Simon King wrote:
On 7 Jun., 19:28, Mike Hansen <mhan...@gmail.com> wrote:
Hi folks,
Sage 4.4.4.alpha0 has been released. With the exception of a few
more
patches, I think this should be pretty close to final.
I wonder whether #8972 has a chance to make it into sage-4.4.4. There
is a partial review of Robert Bradshaw, but further changes had been
necessary afterwards.
Perhaps a quick review is possible?
The patches that I had posted provided an improvement for power series
and Laurent series. But due to some "subtle" changes (a quotient of
two elements of a power series ring shall now always belong to the
fraction field, which now is a Laurent series ring), some code for
elliptic curves broke.
This is now fixed. I am afraid I only tested with sage-4.4.2, but
there, sage -testall passes.
Moreover, the Monsky-Washnitzer code (not my field of expertise, but
it seems to be important for elliptic curves) seems to become faster
by about 33%.
Before:
sage -t "devel/sage/sage/schemes/elliptic_curves/
monsky_washnitzer.py"
[3.9 s]
With my patches
sage -t "devel/sage/sage/schemes/elliptic_curves/
monsky_washnitzer.py"
[2.3 s]
So, it would be a nice improvement.
Nice. Did you do multiple runs? (Sometimes the variance is pretty high
for me, at least at first.)
However, I wonder whether a 'quick
review' is acceptable -- after all, a change in power series will
impact many parts of Sage.
Given my initial reviews were not thorough enough (not that I didn't
bring up quite a few issues) I don't think a "quick" (in terms of
depth, not timeline) review is a good idea. (This is not a statement
about the quality of the code, it's just that it's quite a big
change.) I'll try to look at this some more, but it wouldn't hurt to
have another set of eyeballs comment as well.
- Robert
--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to
sage-devel+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org
