On Monday, July 13, 2015 at 10:39:17 AM UTC+2, Ralf Stephan wrote:
>
> As quick hint I would use valgrind for the search.
> http://wiki.sagemath.org/ValgrindingSage
>
Does valgrind understand python memory layout? most leaks in sage are due
to lingering references and hence are detectable by python. If I run:
import gc
from collections import Counter
gc.collect()
pre={id(c) for c in gc.get_objects()}
n = 5
m = Integer(n*(n-1)/2)
for i in IntegerRange(2^m):
d = i.digits(base=2, padto=m)
l = [[1]+[0]*(n-1)]
for j in range(n-1):
l.append( d[(j+1)*j/2:(j+1)*(j+2)/2]+ [1] + [0]*(n-j-2) )
M = Matrix(l)
eig = sorted((M*M.transpose()).eigenvalues())[0]
if eig <= 0:
print "Foobar"
gc.collect()
post=Counter(type(o) for o in gc.get_objects() if id(o) not in pre)
sorted(post.iteritems(),key=lambda t: t[1])
I find:
...
[(<type 'set'>, 1),
(<type 'generator'>, 1),
(<type 'builtin_function_or_method'>, 1),
(<class '_ast.Attribute'>, 1),
(<class 'sage.rings.qqbar.AlgebraicPolynomialTracker'>, 1),
(<class 'sage.rings.qqbar.ANRoot'>, 1),
(<type 'enumerate'>, 1),
(<class '_ast.Interactive'>, 1),
(<class '_ast.comprehension'>, 1),
(<type 'function'>, 1),
(<type 'instancemethod'>, 1),
(<class
'sage.rings.polynomial.polynomial_element_generic.Polynomial_generic_dense_field'>,
1),
(<type 'listiterator'>, 1),
(<class '_ast.Assign'>, 1),
(<type 'sage.rings.complex_interval.ComplexIntervalFieldElement'>, 1),
(<class '_ast.Compare'>, 1),
(<class '_ast.GeneratorExp'>, 1),
(<class '_ast.Module'>, 1),
(<type 'weakref'>, 3),
(<class '_ast.Call'>, 4),
(<type 'frame'>, 6),
(<type 'sage.rings.rational.Rational'>, 6),
(<type 'sage.rings.real_mpfi.RealIntervalFieldElement'>, 6),
(<class 'sage.rings.qqbar.ANRational'>, 6),
(<class 'sage.rings.qqbar.AlgebraicNumber'>, 7),
(<class '_ast.Name'>, 9),
(<type 'tuple'>, 11),
(<type 'list'>, 29),
(<type 'dict'>, 32),
(<type 'sage.rings.polynomial.polynomial_compiled.univar_pd'>, 6273)]
...
so I'd think there's a leak in constructing univar_pd. My guess would be a
cached routine that shouldn't be cached.
--
You received this message because you are subscribed to the Google Groups
"sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to sage-devel+unsubscr...@googlegroups.com.
To post to this group, send email to sage-devel@googlegroups.com.
Visit this group at http://groups.google.com/group/sage-devel.
For more options, visit https://groups.google.com/d/optout.
