#12313: Fix yet another memory leak caused by caching of coercion data
--------------------------------------------------------+-------------------
Reporter: SimonKing | Owner:
Type: defect | Status:
positive_review
Priority: major | Milestone:
sage-5.8
Component: memleak | Resolution:
Keywords: coercion weak dictionary | Work issues:
Report Upstream: N/A | Reviewers: Simon
King, Jean-Pierre Flori, John Perry, Nils Bruin
Authors: Simon King, Jean-Pierre Flori | Merged in:
Dependencies: #715, #11521, #12215, #13746, #13378 | Stopgaps:
--------------------------------------------------------+-------------------
Comment (by SimonKing):
Replying to [comment:304 nbruin]:
> Replying to [comment:302 SimonKing]:
> > {{{
> > ...
> > 1000000 loops, best of 3: 348 ns per loop
> > sage: D[ZZ]=2
> > sage: M[ZZ]=2
> > sage: timeit("_=M[R]", number=10^6)
> > 1000000 loops, best of 3: 307 ns per loop
> > sage: timeit("_=D[R]", number=10^6)
> > 1000000 loops, best of 3: 351 ns per loop
> > }}}
> Those timing lines don't relate to `ZZ`.
I didn't claim they are. This was in order to show that under some not so
unlikely circumstances `MonoDict` is faster than a usual dict-
> Since your other timings are in line with what I found doing exactly the
same example above, you'd probably also see that lookup of `ZZ` in D is
about 3 times as fast as it is in `M`.
Yep, I posted these timings as well, didn't I?
> A more serious issue is the limited use of `MonoDict` in this case.
Essentially, this `MonoDict` only has an effect for the cases where we
cache that there is NO coercion: In the other cases, we have an entry
> {{{
> M[Domain] = coercion map from Domain to This Parent
> }}}
> i.e., the value in the weak key dict will hold a strong reference to the
key.
I am not so sure, because I think this was one of the memleaks I
explicitly addressed somewhere. But since my Sage install currently is
''without'' #12313, I need to recompile now -- which takes an eternity. I
hope I'll be able to show you an example.
Anyway.
Here is a timing that shows that this patch has a bad effect ''both'' on
looking up the coercion map ''and'' applying the coercion:
Without #12313
{{{
sage: timeit("phi=QQ.coerce_map_from(ZZ)", number=10^6)
1000000 loops, best of 3: 503 ns per loop
sage: phi=QQ.coerce_map_from(ZZ)
sage: x = -20
sage: timeit("phi(x)", number=10^6)
1000000 loops, best of 3: 563 ns per loop
}}}
With #12313:
{{{
sage: timeit("phi=QQ.coerce_map_from(ZZ)", number=10^6)
1000000 loops, best of 3: 738 ns per loop
sage: phi=QQ.coerce_map_from(ZZ)
sage: x = -20
sage: timeit("phi(x)", number=10^6)
1000000 loops, best of 3: 616 ns per loop
}}}
--
Ticket URL: <http://trac.sagemath.org/sage_trac/ticket/12313#comment:305>
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?hl=en.
For more options, visit https://groups.google.com/groups/opt_out.
