mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   Operation Billion Digits (https://www.mersenneforum.org/forumdisplay.php?f=50)
-   -   "LateX" version of factorX? (https://www.mersenneforum.org/showthread.php?t=17528)

kracker 2012-12-04 16:32

"LateX" version of factorX?
 
Hi, I was just wondering, is there a updated version/binary for factor5 ? [URL="http://www.moregimps.it/billion/download.php"]here[/URL], the latest intel build seems like it is made for core2, made in 2009. Or maybe do I have to compile myself?
Just curious, :smile: Thanks.

ET_ 2012-12-04 16:46

[QUOTE=kracker;320455]Hi, I was just wondering, is there a updated version/binary for factor5 ? [URL="http://www.moregimps.it/billion/download.php"]here[/URL], the latest intel build seems like it is made for core2, made in 2009. Or maybe do I have to compile myself?
Just curious, :smile: Thanks.[/QUOTE]

No way, the "latest" version is the one you are using...

I planned to do a slightly faster, GUI enabled version, but had no time for it :sad:

Luigi

kracker 2012-12-04 17:08

Ah, hehe, I see. :smile: Thanks!

Just that on my K10 AMD cpu, I'm doing a [URL="http://mersenne.ca/credit.php?worktype=TF&exponent=3321935213&f_exponent=&b1=&b2=&numcurves=&factor=&frombits=74&tobits=75&submitbutton=Calculate"]exponent from 74 to 75[/URL], after 24 hours, it's at right now (41%), while a regular "low" factor exponent that is about the same on P95 (1.1x GHZ/days) I can do about one in a single day, are these different on it?

(Sorry for my bad English)

ET_ 2012-12-04 17:42

[QUOTE=kracker;320459]Ah, hehe, I see. :smile: Thanks!

Just that on my K10 AMD cpu, I'm doing a [URL="http://mersenne.ca/credit.php?worktype=TF&exponent=3321935213&f_exponent=&b1=&b2=&numcurves=&factor=&frombits=74&tobits=75&submitbutton=Calculate"]exponent from 74 to 75[/URL], after 24 hours, it's at right now (41%), while a regular "low" factor exponent that is about the same on P95 (1.1x GHZ/days) I can do about one in a single day, are these different on it?

(Sorry for my bad English)[/QUOTE]

If you are saying that a lower exponent on P95 is proceeding at about the same speed of an OBD exponent with Factor5, you are in the good ballpark.

Are you using just one thread?

Luigi

kracker 2012-12-04 17:58

[QUOTE=ET_;320464]If you are saying that a lower exponent on P95 is proceeding at about the same speed of an OBD exponent with Factor5, you are in the good ballpark.

Are you using just one thread?

Luigi[/QUOTE]

Yes one thread. Sorry, probably misunderstood me, at 24 hours, it is at 40%(OBD), while one thread of P95 is done by 24 hours or less (about same GHZ/days)

ET_ 2012-12-04 18:04

[QUOTE=kracker;320468]Yes one thread. Sorry, probably misunderstood me, at 24 hours, it is at 40%(OBD), while one thread of P95 is done by 24 hours or less (about same GHZ/days)[/QUOTE]

Don't be sorry, I'm not a native English speaker either... :smile:

Factor5 is slower than Prime95, even with bigger exponents. That's why I decided to multi-thread it. The bad news are that once you start Factor5 with 1 thread, you can't change the thread number on the fly: you should either finish the exponent with 1 thread or restart it from scratch.

Luigi

kracker 2012-12-04 18:16

:explode:
Sorry, I get your meaning now, thanks :smile: Just curious, how efficient is multi-threaded? (I have a quad core 2.9 GHZ at the moment)

ET_ 2012-12-04 18:30

[QUOTE=kracker;320472]:explode:
Sorry, I get your meaning now, thanks :smile: Just curious, how efficient is multi-threaded? (I have a quad core 2.9 GHZ at the moment)[/QUOTE]

It multithreads very well. On a Q6600 (quad core) I reached 90%-95%.

Luigi

kracker 2012-12-04 18:35

[QUOTE=ET_;320476]It multithreads very well. On a Q6600 (quad core) I reached 90%-95%.

Luigi[/QUOTE]

Ah I see, thanks, I guess that answers all my questions :smile:


All times are UTC. The time now is 12:16.

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2020, Jelsoft Enterprises Ltd.