hgcd1/2

Niels Möller nisse at lysator.liu.se
Mon Sep 16 20:20:45 UTC 2019


nisse at lysator.liu.se (Niels Möller) writes:

>>   I'm appending another iteration of the patch to add div2 function based
>>   on div1 on the high limbs. Selected via HGCD2_DIV2_METHOD. Benchmarks:
>>
>>   HGCD2_DIV2_METHOD mpn_hgcd2_1   mpn_hgcd2_2   mpn_hgcd2_3
>>                   1    #1504.47       1740.53       1513.56  div1 + unlikely case
>>                   2    #1739.66       1858.01       1753.73  current code
>>                   3    #1615.32       1736.69       1621.95  the if:ed out version
>>
>> Looks good.  Ready for repo?
>
> I think it's quite ready, but not today. I also wonder if we can delete
> the "current" code right away, to reduce number ov variants. Or if we
> should wire it all up for tuning before deciding that.

Pushed now, with minor changes and deletion of the "current" code
(method 2 above).

Regards,
/Niels

-- 
Niels Möller. PGP-encrypted email is preferred. Keyid 368C6677.
Internet email is subject to wholesale government surveillance.


More information about the gmp-devel mailing list