Hey Nicolas,
Thanks for investigating this! If you feel like it, I am all in favor
> of a small patch taking care of the most urgent optimizations before
> we go for the full cythonization. I am not surprised by the results of
> your timings. In MuPAD we had done a similar optimization by using the
> infinite value of IEEE floats (and its negative); it was quite fast,
> and we did not even have to handle a special case. Could you give it a
> shot using the following?
>
> sage: float('inf')
> inf
>
>
I only made the additional change to the max_part=float(infinity) default
argument to IntegerListsLex in the integer_lists_speedup-ts.patch and got
the following:
1127783 function calls (1127779 primitive calls) in 16.864 seconds
Ordered by: internal time
ncalls tottime percall cumtime percall filename:lineno(function)
5096 2.495 0.000 6.065 0.001
integer_list.py:200(rightmost_pivot)
5096 2.190 0.000 11.775 0.002 integer_list.py:44(first)
108196 2.137 0.000 6.252 0.000 integer_list.py:366(<lambda>)
156168 2.048 0.000 2.974 0.000 integer_list.py:1080(ceiling)
156168 1.916 0.000 4.890 0.000 integer_list.py:1122(<lambda>)
277562 1.619 0.000 1.619 0.000 {len}
111200 1.475 0.000 2.114 0.000 integer_list.py:1061(floor)
108196 0.778 0.000 0.778 0.000 {min}
58132 0.684 0.000 2.500 0.000 integer_list.py:359(<lambda>)
53051 0.379 0.000 0.379 0.000 {max}
5096 0.301 0.000 18.239 0.004 integer_list.py:329(next)
47956 0.292 0.000 0.292 0.000 {method 'insert' of 'list'
objects}
5097 0.195 0.000 18.490 0.004 integer_list.py:1141(__iter__)
15288 0.146 0.000 0.146 0.000 {range}
10191 0.116 0.000 0.116 0.000 {sum}
5096 0.049 0.000 0.049 0.000 integer_list.py:638(check)
1 0.033 0.033 18.522 18.522
finite_enumerated_sets.py:126(_list_from_iterator)
...
sage: time p = Partitions(150, max_slope=-1, length=15).list()
Time: CPU 2.86 s, Wall: 2.98 s
without the patch I get approximately the same timings.
I've created ticket #13840 <http://trac.sagemath.org/sage_trac/ticket/13840>
for
this and uploaded my patch (it's ready for review). I've put this patch at
the front of the combinat queue (and moved it passed the partition options
patch #13605).
Best,
Travis
--
You received this message because you are subscribed to the Google Groups
"sage-combinat-devel" group.
To view this discussion on the web visit
https://groups.google.com/d/msg/sage-combinat-devel/-/6Deh8WGVO_8J.
To post to this group, send email to sage-combinat-devel@googlegroups.com.
To unsubscribe from this group, send email to
sage-combinat-devel+unsubscr...@googlegroups.com.
For more options, visit this group at
http://groups.google.com/group/sage-combinat-devel?hl=en.
