[Python-Dev] int/float freelists vs pymalloc

[Christian Heimes]

Andrew MacIntyre wrote:
> So I've been testing with the freelists ripped out and ints and floats
> reverted to fairly standard PyObject_New allocation (retaining the small
> int interning) and thus relying on PyMalloc to do any compaction.

Nice! What do you mean with int interning? Are you talking about the
small int cache?

> The results have been somewhat surprising...
> 
> The short version is that:
> - for ints, the freelist is ahead of PyMalloc by a very small margin
>    (<4%)
> - for floats, the freelist is a long way behind PyMalloc (>35% slower)
> 
> This without invoking freelist compaction by the way (though PyMalloc
> is releasing empty arenas).
> 
> I don't know what's pessimising the float freelist, but the results are
> similar on 2 boxes:
> - AMD Athlon XP-1600+, 512MB RAM, FreeBSD 6.1, gcc 3.4.4
>    (~27k pystones)
> - AMD Athlon XP-2500+, 512MB RAM, OS/2 v4 with EMX, gcc 2.8.1
>    (~38k pystones)

That's interesting. I wouldn't have thought that floats are much faster
with pymalloc. I'd bet that pymalloc is a tiny bit slower until the new
free list is filled.

> If there's interest in following this up, I can put my patches to 
> intobject.c and floatobject.c into the tracker.

I've uploaded my patch at http://bugs.python.org/issue2039. It keeps a
standard free list with 8192 elements for floats and ints. Does your
patch also use a free list or does it alloc/free directly?

Can your profile my patch and compare it to your results? It may be a
bit faster if your patch doesn't use a free list.

Christian