Re: MPQS behavior

[Bill Allombert <allomber@math.u-bordeaux.fr>]

On Fri, Mar 24, 2006 at 10:10:01PM -0800, Justin C. Walker wrote:
> PowerPC:
> 186:  
> 489374912706076771525100405333124116411054140808019375469349301329808748 
> 04891650866221688438
> 
>   *** factor: Warning: MPQS: factoring this number will take several  
> hours.
> [2, 1; 3, 2; 7, 1; 11, 1; 557810501066166798560937148699439400223, 1;  
> 63298262031087966002932642338205110362839613755121, 1]
> 
> Core Duo:
> 186:  
> 489374912706076771525100405333124116411054140808019375469349301329808748 
> 04891650866221688438
>   *** factor: Warning: MPQS: factoring this number will take several  
> hours:
> N =  
> 353084352601786992442352384800233850224425787018772998174133695043152054 
> 86934813034791983.
> [2, 1; 3, 2; 7, 1; 11, 1; 557810501066166798560937148699439400223, 1;  
> 63298262031087966002932642338205110362839613755121, 1]
> 
> In order to get to the factorization, it took the PowerPC box much  
> longer.
> 
> A side question: what is the significance of the "N=" line from the  
> Core Duo?  And why do I get it only on the Core Duo?

You are not using the same version of PARI/GP on both machine.

The N= was added in PARI/GP 2.2.11.

I would suggest you restart your timing with the same version on
both machines. MPQS has been improved in recent versions.

> At this point, the PowerPC has been chewing on this for roughly twice  
> as long as the Core Duo (the Core Duo has been sleeping a lot :-}).   
> However, the PowerPC is up to n=193, while the Core Duo is up to n=204.

MPQS is very sensible to the cache size:
What is the cache size of theses two box ?

Cheers,
Bill.