#15921: work around Maxima fpprintprec bug and other ARM-specific problems
-------------------------------------+-------------------------------------
Reporter: dimpase | Owner:
Type: defect | Status: needs_review
Priority: major | Milestone: sage-6.3
Component: calculus | Resolution:
Keywords: Maxima, | Merged in:
fpprintprec, ARM | Reviewers: Peter Bruin
Authors: | Work issues:
Report Upstream: Reported | Commit:
upstream. Developers acknowledge | 079bb9af4f12892268a19f0d218ac96bd72466f4
bug. | Stopgaps:
Branch: |
u/dimpase/arm_fixes_etc |
Dependencies: |
-------------------------------------+-------------------------------------
Comment (by pbruin):
Replying to [comment:27 dimpase]:
> Replying to [comment:26 pbruin]:
> > I redid a few tests after upgrading to Maxima 5.33.0 (#13973), and
nothing really changed, except the change from `.123...` to `0.123...` is
also done (upstream, differently than in
[attachment:maxfpprintprec.patch]) by this upgrade. May I propose that we
just apply the changes from [attachment:special_functions_precision.patch]
to solve the precision problems in this ticket?
>
> I wonder about the `# For ARM: rel tol 2e-15` tag.
> How does it work? Or is this just a comment?
The `rel tol 2e-15` is a magic marker telling the doctest framework not to
complain about numerical errors smaller than this tolerance; see
[http://sagemath.org/doc/developer/coding_basics.html#special-markup-to-
influence-tests Special markup to influence tests]. The comment "for ARM"
is just a comment to clarify that the error can be this large on ARM.
Normally the relative error in basic functions like the gamma function
should be roughly bounded by 2e-16 (since this is approximately 2^-52^ and
doubles have 53 bits of precision) and hence be invisible when printing
with 16 decimal digits of precision.
> >
> > I'm happy with the other changes you made, even though they don't
really seem to be related to ARM, Maxima or precision. Two trivial
remarks:
> > - The message `Doctesting files changed since the last git commit`
doesn't seem to have a full stop at the end when I run `sage -t --new`.
> > - Does the `show_default()` doctest in `graphics.py` really take a
long time?
>
> I did timings of tests on ARM and added `#long time` to these above
certain thresholds.
I wonder why it takes so long on ARM; when testing `show_default()` on an
x86_64 it only seems to take 316 microseconds for the first time (`%timeit
-n 1 -r 1 show_default()`) and an average of 11.4 microseconds when
running it many times.
> It's not 100% clear what changes you propose. Can you do a reviewer's
git branch (or a patch)?
OK, coming soon.
--
Ticket URL: <http://trac.sagemath.org/ticket/15921#comment:28>
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 unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/sage-trac.
For more options, visit https://groups.google.com/d/optout.
