On Thursday 03 June 2004 00:07, George Woltman wrote:
> At 07:58 PM 6/2/2004 +0100, Tom Womack wrote:
> >I'm impressed by the speed of verification; it sounds as if the FFT itself
> >ran in parallel on the Itaniums, which I thought people on this list had
> >repeatedly claimed was impractical. Is it in fact practical but pointless
> > -- you could check a single number in eight days on a dual Opteron, but
> > just running Prime95 on both CPUs would check two numbers in fourteen.
>
> You are correct.  Glucas and now mlucas are now multithreaded.  You'll
> get better throughput running single-threaded, but verifications really
> fly multi-threaded.
>
> Due to inefficiencies in the glucas implementation, since fixed, the
> verification only used about half of the 16 Itanium CPUs.
>

This has been fixed. Now Glucas use most of the power, scale better and can 
use any number of threads. All these improvements are  side-effect of M41 
verification. See latest snapshots (code still experimental),

ftp://ftp.oxixares.com/pub/glucas/snapshots

Guillermo


-- 
Guillermo Ballester Valor
[EMAIL PROTECTED]
Ogijares, Granada  SPAIN
Linux user #117181. See http://counter.li.org/
Public GPG KEY http://www.oxixares.com/~gbv/pubgpg.html

 
_______________________________________________
Prime mailing list
[EMAIL PROTECTED]
http://hogranch.com/mailman/listinfo/prime

Reply via email to