#6818: [with patch; needs review] maxima interface gets dramatically slower over
time
------------------------+---------------------------------------------------
Reporter: was | Owner: was
Type: defect | Status: new
Priority: critical | Milestone: sage-4.1.2
Component: interfaces | Keywords:
Reviewer: | Author:
Merged: |
------------------------+---------------------------------------------------
Description changed by was:
Old description:
> If you do a few hundred calls to the maxima interface, it gets way way
> slower.
>
> {{{
> sage: timeit("str(maxima.eval('1+2'))")
> 5 loops, best of 3: 758 µs per loop
> sage: timeit("str(maxima.eval('1+2'))")
> 625 loops, best of 3: 1.22 ms per loop
> sage: timeit("str(maxima.eval('1+2'))")
> 125 loops, best of 3: 2.98 ms per loop
> sage: timeit("str(maxima.eval('1+2'))")
> 125 loops, best of 3: 3.97 ms per loop
> }}}
>
> It turns out that this is caused by computation of the maxima input
> prompt number, which uses the following "clever" algorithm to compute
> "n+1":
>
> {{{
> (defmfun makelabel (x)
> (when (and $dskuse (not $nolabels) (> (incf dcount) $filesize))
> (setq dcount 0)
> (dsksave))
> (setq linelable ($concat '|| x $linenum))
> (unless $nolabels
> (when (or (null (cdr $labels))
> (when (member linelable (cddr $labels) :test #'equal)
> (setf $labels (delete linelable $labels :count 1 :test
> #'eq)) t)
> (not (eq linelable (cadr $labels))))
> (setq $labels (cons (car $labels) (cons linelable (cdr
> $labels))))))
> linelable)
> }}}
>
> More precisely, this code "checks
> that the list containing all labels does not contain the new label
> which it generates. After you create 2*35150 labels, this takes longer
> than when maxima starts.", according to Andrej Vodopivec who tracked this
> down and told us a simple fix. Put:
>
> {{{
> nolabels:true;
> }}}
>
> at the beginning of our Maxima session. This is fine for Sage, since
> Sage doesn't use the labels in any way.
New description:
If you do a few hundred calls to the maxima interface, it gets way way
slower.
{{{
sage: timeit("str(maxima.eval('1+2'))")
5 loops, best of 3: 758 µs per loop
sage: timeit("str(maxima.eval('1+2'))")
625 loops, best of 3: 1.22 ms per loop
sage: timeit("str(maxima.eval('1+2'))")
125 loops, best of 3: 2.98 ms per loop
sage: timeit("str(maxima.eval('1+2'))")
125 loops, best of 3: 3.97 ms per loop
}}}
It turns out that this is caused by computation of the maxima input prompt
number, which uses the following "clever" algorithm to compute "n+1":
{{{
(defmfun makelabel (x)
(when (and $dskuse (not $nolabels) (> (incf dcount) $filesize))
(setq dcount 0)
(dsksave))
(setq linelable ($concat '|| x $linenum))
(unless $nolabels
(when (or (null (cdr $labels))
(when (member linelable (cddr $labels) :test #'equal)
(setf $labels (delete linelable $labels :count 1 :test
#'eq)) t)
(not (eq linelable (cadr $labels))))
(setq $labels (cons (car $labels) (cons linelable (cdr $labels))))))
linelable)
}}}
More precisely, this code "checks
that the list containing all labels does not contain the new label
which it generates. After you create 2*35150 labels, this takes longer
than when maxima starts.", according to Andrej Vodopivec and Robert Dodier
who (independently) tracked this down and told us a simple fix. Put:
{{{
nolabels:true;
}}}
at the beginning of our Maxima session. This is fine for Sage, since Sage
doesn't use the labels in any way.
--
--
Ticket URL: <http://trac.sagemath.org/sage_trac/ticket/6818#comment:1>
Sage <http://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
-~----------~----~----~----~------~----~------~--~---
