Memory Leak in mutipolynomial evaluation

[Gabriel Furstenheim Milerud]

There is a memory leak in the evaluation of multivariable polynomials:

-C.<x,y,z>=GF(2)[]

-f=x^4+x*y^3+z^6

-get_memory_usage()

.....1014.47265625

- for i in xrange(1000000):

    a=f(1,0,0)

-get_memory_usage()

.....1052.47265625

I think the problem is in the implementation of the __call__ method in rings/polynomials/multi_polynomial_element.py . It has:

 def __call__(self,*x,**kwds): 

[...]

 try:

    K = x[0].parent()

  except AttributeError:

    K = self.parent().base_ring()

  y = K(0)

  for (m,c) in self.element().dict().iteritems():

     y += c*misc.mul([ x[i]**m[i] for i in range(n) if m[i] != 0])

  return y

So in every loop a new ring K is created and saved

[Simon King]

Hi Gabriel,

On 2014-05-25, Gabriel Furstenheim Milerud <furst...@gmail.com> wrote:
> There is a memory leak in the evaluation of multivariable polynomials:
>
> -C.<x,y,z>=GF(2)[]
> -f=x^4+x*y^3+z^6
>
> -get_memory_usage()
> .....1014.47265625
> - for i in xrange(1000000):
> a=f(1,0,0)
> -get_memory_usage()
> .....1052.47265625

Note that you first request the memory usage before assigning
"a=f(1,0,0)". But to some extend, I can confirm an increase of memory
usage (with Sage 6.2):

sage: C.<x,y,z> = GF(2)[]
sage: f = x^4+x*y^3+z^6
sage: a = f(1,0,0)
sage: get_memory_usage()
176.08984375
sage: for i in xrange(1000000):
....: a = f(1,0,0)
....:
sage: get_memory_usage()
198.08984375
sage: for i in xrange(1000000):
....: a = f(1,0,0)
....:
sage: get_memory_usage()
222.08984375


We have an increase of 22-24 megabytes in 10^6 rounds, thus 23-25 byte per round.
Actually I wonder how such a tiny leak is even possible! In any case, it
should be fixed.

> try:
> K = x[0].parent()
> except AttributeError:
> K = self.parent().base_ring()
> y = K(0)
> for (m,c) in self.element().dict().iteritems():
> y += c*misc.mul([ x[i]**m[i] for i in range(n) if m[i] != 0])
> return y
>
> So in every loop a new ring K is created and saved

No, K is already there, it is either x[0].parent() or self.parent().base_ring().
Actually it is K=GF(2), if I am not mistaken. Moreover, I don't think this branch
of code is evaluated in your example.

Anyway, thank you for the example!
Best regards,
Simon

[leif]

I'm getting more:

sage: import gc
sage: gc.collect()
182
sage: get_memory_usage()
292.03515625
sage: C.<x,y,z>=GF(2)[]
sage: f=x^4+x*y^3+z^6
sage: get_memory_usage()
292.19921875

sage: for i in xrange(1000000):
....: a=f(1,0,0)
....:
sage: get_memory_usage()

338.19921875
sage: gc.collect()
11
sage: get_memory_usage()
338.19921875

(Sage 6.2, MacOS X 10.6 x86_64)


-leif

--
() The ASCII Ribbon Campaign
/\ Help Cure HTML E-Mail

[Jori Mantysalo]

On Sun, 25 May 2014, Gabriel Furstenheim Milerud wrote:

> There is a memory leak in the evaluation of multivariable polynomials:

I wrote about same problem on sage-devel at 2014-05-08.

> -C.<x,y,z>=GF(2)[]

My example did not use GF, just QQ.

--
Jori Mäntysalo

[Gabriel Furstenheim Milerud]

Hi Simon,

You're completely right. Actually I messed up the libraries, the __call__ method used should be the one from Mpolynomial_libsingular.