Mark Dickinson <[EMAIL PROTECTED]> added the comment:
Timings on x86/Linux are similar: the lsum-based version is around
10% slower on average, 25% slower in the worst case, and significantly
faster for the msum worst cases.
There's probably still some snot left to optimize out, though. Some
tempting ideas are:
(1) to try using doubles instead of longs for the accumulator digits
(with 51 or 52 bits of precision), and
(2) to split each mantissa into (nearest_integer, fraction) instead
of (next_smallest_integer, fraction), using rint or lrint.
Anything else?
_______________________________________
Python tracker <[EMAIL PROTECTED]>
<http://bugs.python.org/issue2819>
_______________________________________
_______________________________________________
Python-bugs-list mailing list
Unsubscribe:
http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
