#12313: Fix yet another memory leak caused by caching of coercion data
--------------------------------------------------------+-------------------
Reporter: SimonKing | Owner:
Type: defect | Status:
needs_work
Priority: major | Milestone:
sage-5.7
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):
Here is the outcome of
{{{
sage: from sage.combinat.sf.k_dual import DualkSchurFunctions
sage: Sym = SymmetricFunctions(QQ['t'].fraction_field())
sage: dks4 = DualkSchurFunctions(Sym.kBoundedQuotient(4))
sage: %prun TestSuite(dks4).run()
}}}
First, sage-5.6.rc0 with #12215:
{{{
9365 7.254 0.001 11.881 0.001 {method 'subs' of
'sage.structure.element.Element' objects}
30786 2.596 0.000 3.067 0.000
fraction_field.py:271(<lambda>)
5358/84 1.781 0.000 26.087 0.311
{sage.combinat.dict_addition.dict_linear_combination}
1778 1.295 0.001 13.615 0.008 sfa.py:988(_from_cache)
54541 1.061 0.000 2.015 0.000
infinity.py:883(_element_constructor_)
456782 1.051 0.000 1.051 0.000 {isinstance}
27795 0.815 0.000 1.931 0.000
partition.py:3510(Partitions)
124520 0.754 0.000 0.754 0.000 fraction_field.py:430(ring)
1715 0.518 0.000 2.106 0.001
integer_list.py:192(rightmost_pivot)
3951 0.433 0.000 0.547 0.000
free_module.py:411(_coefficient_fast)
4015 0.416 0.000 1.096 0.000
polynomial_ring.py:290(_element_constructor_)
54541 0.352 0.000 0.352 0.000 infinity.py:942(__init__)
2236 0.348 0.000 0.448 0.000 free_module.py:562(support)
80417 0.347 0.000 0.347 0.000 combinat.py:840(__hash__)
37514 0.340 0.000 0.340 0.000 {method 'sub' of
'_sre.SRE_Pattern' objects}
37514 0.275 0.000 0.938 0.000 re.py:144(sub)
51922 0.275 0.000 0.275 0.000 partition.py:4003(__repr__)
19970 0.262 0.000 2.203 0.000 partition.py:2716(parent)
16195 0.259 0.000 0.515 0.000
free_module.py:2142(_from_dict)
37514 0.244 0.000 0.322 0.000 re.py:226(_compile)
4015 0.238 0.000 1.933 0.000
fraction_field.py:456(_element_constructor_)
39940 0.232 0.000 0.580 0.000 {repr}
9365 0.230 0.000 0.292 0.000 fraction_field.py:532(gen)
35636 0.186 0.000 0.305 0.000 infinity.py:285(__cmp__)
1715 0.186 0.000 0.634 0.000 integer_list.py:38(first)
}}}
Second, adding #13378
{{{
9365 7.525 0.001 12.378 0.001 {method 'subs' of
'sage.structure.element.Element' objects}
30786 2.666 0.000 3.166 0.000
fraction_field.py:271(<lambda>)
5363/85 1.908 0.000 29.846 0.351
{sage.combinat.dict_addition.dict_linear_combination}
1778 1.340 0.001 15.319 0.009 sfa.py:988(_from_cache)
494409 1.269 0.000 1.269 0.000 {isinstance}
55877 1.196 0.000 2.254 0.000
infinity.py:883(_element_constructor_)
27906 0.876 0.000 2.117 0.000
partition.py:3510(Partitions)
127929 0.809 0.000 0.809 0.000 fraction_field.py:430(ring)
1855 0.561 0.000 2.359 0.001
integer_list.py:192(rightmost_pivot)
4248/4191 0.493 0.000 1.256 0.000
polynomial_ring.py:290(_element_constructor_)
4070 0.461 0.000 0.591 0.000
free_module.py:411(_coefficient_fast)
82386 0.397 0.000 0.397 0.000 combinat.py:840(__hash__)
2236 0.374 0.000 0.498 0.000 free_module.py:562(support)
55877 0.372 0.000 0.372 0.000 infinity.py:942(__init__)
37516 0.367 0.000 0.369 0.000 {method 'sub' of
'_sre.SRE_Pattern' objects}
46893 0.326 0.000 0.764 0.000 {repr}
56811 0.318 0.000 0.318 0.000 partition.py:4003(__repr__)
37514 0.304 0.000 1.020 0.000 re.py:144(sub)
16322 0.301 0.000 0.601 0.000
free_module.py:2142(_from_dict)
19970 0.290 0.000 2.418 0.000 partition.py:2716(parent)
37517 0.265 0.000 0.362 0.000 re.py:226(_compile)
4015 0.261 0.000 2.149 0.001
fraction_field.py:456(_element_constructor_)
9426 0.239 0.000 0.312 0.000 fraction_field.py:532(gen)
17385 0.223 0.000 0.365 0.000 combinat.py:667(__init__)
36359 0.216 0.000 0.349 0.000 infinity.py:285(__cmp__)
}}}
And finally adding #12313:
{{{
952735/392465 59.690 0.000 204.150 0.001 {method
'has_coerce_map_from' of 'sage.structure.parent.Parent' objects}
3818389 44.077 0.000 91.652 0.000 {repr}
1626 27.430 0.017 243.763 0.150 k_dual.py:1277(<genexpr>)
475630 24.414 0.000 71.594 0.000 {method 'coerce_map_from' of
'sage.structure.parent.Parent' objects}
471632 23.574 0.000 204.044 0.000
fraction_field.py:194(_coerce_map_from_)
1909173 15.683 0.000 111.239 0.000 combinat.py:1051(__cmp__)
1425904 15.023 0.000 23.084 0.000
polynomial_ring.py:1369(_repr_)
475629/475628 13.746 0.000 90.469 0.000
polynomial_ring.py:468(_coerce_map_from_)
3286022 12.918 0.000 17.094 0.000
unique_representation.py:531(__hash__)
558 11.842 0.021 73.743 0.132 k_dual.py:1234(<genexpr>)
475332 11.785 0.000 19.538 0.000
fraction_field.py:380(_repr_)
1925108 11.632 0.000 11.632 0.000 partition.py:4003(__repr__)
1439041 8.110 0.000 8.110 0.000
polynomial_ring.py:716(__hash__)
1425904 8.061 0.000 8.061 0.000
polynomial_ring.py:725(_repr_)
9365 7.681 0.001 12.918 0.001 {method 'subs' of
'sage.structure.element.Element' objects}
16080 7.001 0.000 9.528 0.001 homset.py:353(__init__)
1934768 5.774 0.000 5.774 0.000 {isinstance}
564 5.764 0.010 33.683 0.060 k_dual.py:1110(<genexpr>)
3334205 4.241 0.000 4.241 0.000 {id}
1927477 3.937 0.000 3.937 0.000 {cmp}
1450944 3.416 0.000 3.416 0.000
rational_field.py:217(__hash__)
479620 3.123 0.000 24.605 0.000
rational_field.py:261(_coerce_map_from_)
30786 3.106 0.000 3.593 0.000
fraction_field.py:271(<lambda>)
16060 2.354 0.000 13.278 0.001 homset.py:80(Hom)
315 2.335 0.007 50.641 0.161 k_dual.py:1076(retract)
}}}
I'm afraid I have to leave now. But the fact that repr is called so often
makes me guess that there is indeed caching involved - namely with a hash
function that relies on the string representation. But if that was
correct, neither `MonoDict` nor `TripleDict` are to blame: They compare by
identity and thus don't rely on the hash of the keys.
--
Ticket URL: <http://trac.sagemath.org/sage_trac/ticket/12313#comment:279>
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.
