This issue tracker has been migrated to GitHub, and is currently read-only.
For more information, see the GitHub FAQs in the Python's Developer Guide.

Author mark.dickinson
Recipients MrJean1, mark.dickinson, rhettinger, tim.peters
Date 2008-07-31.15:55:22
SpamBayes Score 0.00451738
Marked as misclassified No
Message-id <1217519723.22.0.97208848367.issue2819@psf.upfronthosting.co.za>
In-reply-to
Content
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?
History
Date User Action Args
2008-07-31 15:55:23mark.dickinsonsetrecipients: + mark.dickinson, tim.peters, rhettinger, MrJean1
2008-07-31 15:55:23mark.dickinsonsetmessageid: <1217519723.22.0.97208848367.issue2819@psf.upfronthosting.co.za>
2008-07-31 15:55:22mark.dickinsonlinkissue2819 messages
2008-07-31 15:55:22mark.dickinsoncreate