#15367: Empty lists while creating parents
---------------------------+-------------------------
Reporter: roed | Owner:
Type: defect | Status: new
Priority: major | Milestone: sage-5.13
Component: memleak | Resolution:
Keywords: | Merged in:
Authors: | Reviewers:
Report Upstream: N/A | Work issues:
Branch: | Commit:
Dependencies: | Stopgaps:
---------------------------+-------------------------
Comment (by nbruin):
These are the empty buckets of the various `TripleDict` and `MonoDict`
data structures:
{{{
sage: import gc
sage: pre = gc.get_objects()
sage: K = Qp(7,2)
sage: post = gc.get_objects()
sage: len(post) - len(pre)
682
sage: pre_id=set( id(p) for p in pre )
sage: new=[p for p in post if id(p) not in pre_id and isinstance(p,list)
and len(p) ==0]
}}}
OK, we have now all the new lists of length 1. Let's see who's referring
to them
{{{
sage: refs=[]
sage: for p in new: refs.extend(gc.get_referrers(p))
}}}
and count, among the referrers that are lists themselves, what the lengths
are.
{{{
sage: from collections import Counter
sage: Counter(len(l) for l in refs if isinstance(l,list))
Counter({525: 525, 117431: 525, 53: 265, 23: 228})
}}}
The first one must be `new` itself and the one of length `117431` also
seems to contain all 525 of them. The other ones have conspicuous lengths:
53 and 23:
{{{
sage: search_src("TripleDict\(23\)")
structure/parent.pyx:684: self._action_hash = TripleDict(23)
structure/parent_old.pyx:87: self._action_hash = TripleDict(23)
sage: search_src("MonoDict\(23\)")
structure/parent_gens.pyx:254: self._has_coerce_map_from =
MonoDict(23)
structure/parent.pyx:682: self._coerce_from_hash = MonoDict(23)
structure/parent_old.pyx:85: self._coerce_from_hash = MonoDict(23)
structure/parent_old.pyx:100: self._has_coerce_map_from =
MonoDict(23)
structure/parent_old.pyx:343: self._has_coerce_map_from =
MonoDict(23)
sage: search_src("MonoDict\(53\)")
structure/parent.pyx:686: self._convert_from_hash =
MonoDict(53)
}}}
so it looks like we have 265/53=5 (empty) `self._convert_from_hash`
dictionaries and 228/23=9.9, so probably 10 (mostly) empty
`self._coerce_from_hash` and `self._action_hash` dictionaries. Did we
create 5 parents, perhaps?
If you think this is a real issue, we could open-code the hashtables for
`MonoDict` and `TripleDict` in the same way that `dict` is done. That's
quite a bit of work to program correctly, though! In the mean time, we
could reduce the default sizes of the dictionaries a bit. Perhaps 53 and
23 is a bit on the large side for dictionaries that tend to not fill up so
much?
--
Ticket URL: <http://trac.sagemath.org/ticket/15367#comment:1>
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/groups/opt_out.
