|
From: | Pádraig Brady |
Subject: | bug#12350: Composites identified as primes in factor.c (when HAVE_GMP) |
Date: | Tue, 04 Sep 2012 20:26:45 +0100 |
User-agent: | Mozilla/5.0 (X11; Linux x86_64; rv:6.0) Gecko/20110816 Thunderbird/6.0 |
On 09/04/2012 07:10 PM, Torbjorn Granlund wrote:
Pádraig Brady<address@hidden> writes: On 09/04/2012 03:46 PM, Jim Meyering wrote: > There are 9008992 composites of the form with s=2 below 2^64. With 3 > Miller-Rabin test, one would expect about 9008992/4^64 = 140766 to be s/4^64/64/ ? For what it's worth I checked the million primes in the range 452,930,477 to 472,882,027 and they're now identified correctly (465658903 was included previously). Note processing time has increased with the patch. On my 2.1GHz i3-2310M, running over the above range used to take 14m, but now takes 18m. It sometimes takes more time to do things correctly.
Sure. I was just quantifying the performance change, for others who may be referencing or noticing patches. (Actually, I'd add a note to the commit message that, this increases calculations by about 25%).
As I mentioned in the original post, we will replace the current code with code that is many times faster. Your example above will run at less than a minute on your system.
I'd left my test files in place in anticipation ;) thanks, Pádraig.
[Prev in Thread] | Current Thread | [Next in Thread] |