#12313: Fix yet another memory leak caused by caching of coercion data
--------------------------------------------------+-------------------------
Reporter: SimonKing | Owner:
Type: defect | Status:
needs_review
Priority: major | Milestone: sage-5.3
Component: memleak | Resolution:
Keywords: coercion weak dictionary | Work issues:
Report Upstream: N/A | Reviewers: Simon King,
Jean-Pierre Flori, John Perry
Authors: Simon King, Jean-Pierre Flori | Merged in:
Dependencies: #11521, #11599, #12969, #12215 | Stopgaps:
--------------------------------------------------+-------------------------
Comment (by SimonKing):
Replying to [comment:163 nbruin]:
> Observation: The test above is entirely unrepresentative because there
is no resizing whatsoever (default `threshold=0` leads to no resizing). So
the test above was just using `53` buckets, meaning it was mainly doing
linear search! Impressive how fast that is, of course, but more
representative is (`2/3` is the threshold ratio that Python's `dict` uses
by default)
Good point! By your comment, I have now threshold=0.7 as default in my
experimental code.
> That's why `MonoDict` won for membership. It was using a properly
calibrated `dict` versus a linear search in `MonoDictNoRefCache`.
Yes, that would explain it.
> Concerning `TupleDict`: Sure we can implement it but if our only use
cases are `MonoDict` and `TripleDict` anyway, we're better off with the
unrolled loops.
Last night I already wrote experimental code (calling the thing
"`IdKeyDict`", because the keys are compared by identity). I am now doing
some tests - perhaps we are lucky and it is fast...
`MonoDict` and `TripleDict` are the only use cases ''yet''. Until not so
long time ago, `TripleDict` has been the only use case. One could imagine
that an `IdKeyDict` being useful for a weak_cached_function, as in #12215:
If you have a function with a fixed list of arguments (hence, no *args and
**keys), and if you are happy to compare the arguments by identity,
`IdKeyDict` may be exactly the thing you want.
Plus, one avoids code duplication. And aren't modern compilers intelligent
enough to do some loop unrolling automatically?
Anyway, since I already wrote code, I'll first test the performance,
rather than immediately dropping the idea of unifying `MonoDict` and
`TripleDict`.
--
Ticket URL: <http://trac.sagemath.org/sage_trac/ticket/12313#comment:165>
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.
