mersenneforum.org  

Go Back   mersenneforum.org > Factoring Projects > Operation Billion Digits

Reply
 
Thread Tools
Old 2005-04-01, 11:16   #1
ET_
Banned
 
ET_'s Avatar
 
"Luigi"
Aug 2002
Team Italia

10010100111102 Posts
Default Factor4 - news.

Factor4 is in alpha test.

It seems at least twice as fast as version 3.2, but requires 130 MB of free RAM.

Luigi
ET_ is offline   Reply With Quote
Old 2005-04-02, 16:43   #2
Joshua2
 
Joshua2's Avatar
 
Sep 2004

13·41 Posts
Default

Good! Glad its fast, personaly I don't care if it uses 800MB of memory, since I have a gig :)
When will I be able to download it? Don't rush it before you think it will find all factors, since we don't want to have to redo stuff.
Joshua2 is offline   Reply With Quote
Old 2005-04-03, 20:27   #3
ET_
Banned
 
ET_'s Avatar
 
"Luigi"
Aug 2002
Team Italia

10010100111102 Posts
Default

Quote:
Originally Posted by Joshua2
Good! Glad its fast, personaly I don't care if it uses 800MB of memory, since I have a gig :)
When will I be able to download it? Don't rush it before you think it will find all factors, since we don't want to have to redo stuff.
The factoring code seems to be working by now. I'm adding all those thingies you asked me for: a benchmark tool, an option that calculates the approximate time for a range, non integer bit depth steps.

I hope I will release it in a couple of weeks.

Luigi
ET_ is offline   Reply With Quote
Old 2005-04-11, 20:13   #4
ET_
Banned
 
ET_'s Avatar
 
"Luigi"
Aug 2002
Team Italia

2×2,383 Posts
Default

It's nearly ready!

Luigi
ET_ is offline   Reply With Quote
Old 2005-04-15, 21:18   #5
VBCurtis
 
VBCurtis's Avatar
 
"Curtis"
Feb 2005
Riverside, CA

22×23×47 Posts
Default

This is great news! It will be fun to get a number to 74 bits, now that we can split it among 2 or more machines. If you want me to test it on a batch, I can use a few project 100 million range files to make sure it finds all factors that v3.2 found.
-Curtis
VBCurtis is offline   Reply With Quote
Old 2005-04-17, 18:00   #6
ET_
Banned
 
ET_'s Avatar
 
"Luigi"
Aug 2002
Team Italia

2·2,383 Posts
Default

Quote:
Originally Posted by VBCurtis
This is great news! It will be fun to get a number to 74 bits, now that we can split it among 2 or more machines. If you want me to test it on a batch, I can use a few project 100 million range files to make sure it finds all factors that v3.2 found.
-Curtis
I can send you a beta, the problem is there won't be a "batch" version for now. The new approach is such that the overhead needed to speed the search over 59-60 bits (4.0 is up to three times as fast as 3.2 now) would rise the search time above version 3_2 timings for very small factors.

Luigi

Last fiddled with by ET_ on 2005-04-17 at 18:01
ET_ is offline   Reply With Quote
Old 2005-04-18, 00:30   #7
Joshua2
 
Joshua2's Avatar
 
Sep 2004

13×41 Posts
Default

Yeah! Who cares about factors less than 60-bits. They were really quick with the old version anyway. :) Good job. Can't wait.
Joshua2 is offline   Reply With Quote
Old 2005-04-21, 22:13   #8
ET_
Banned
 
ET_'s Avatar
 
"Luigi"
Aug 2002
Team Italia

129E16 Posts
Default Here it is!

I decioded to release a beta version of Factor4 for testing.

The attached version has been compiled under Windows (Cygwin) with GMP 4.1.4 and works fine on a 256 MB Athlon XP 2100.

I coundn't manage to compile it on a Pentium 4: the statically linked EXE file goes on segmentation fault every time I start it. Maybe I should change some compiling parameter...

Anyway, I'm going on holidays, will be back next Tuesday and did not want to keep you waiting until then. You can try both the executable and the source.

If you get it compiled and running on different platforms and systems (Windows, Linux, BSD, OS/X, Pentium II, III M 4, k6, k7, k8, MAC), I ask you to please send the binary to mc5225 AT mclink DOT it

The actual executable is nearly 3 times as fast as versiion 3.2, and has two new command line options:
factor4 -b prints a benchmark on the screen
factor4 -f asks for k and exponent and returns the equivalent factor 2kp+1 and its bit depth.

You can also use fractional bit depth, although this option won't be used for reservation purposes (i.e. you can switch the work among different machines).

I tested version 4 with all known factors, and it seems to work fine.

The only caveat is that the program requires 130 MB of RAM. If you think it's too much, I will decrease it (thus slowing the search of a couple of percent points.)

I'm waiting for your impressions, bug reports and hopefully working versions on different platforms.

Have fun

Luigi
Attached Files
File Type: zip Factor4.zip (22.7 KB, 226 views)

Last fiddled with by ET_ on 2005-04-21 at 22:16
ET_ is offline   Reply With Quote
Old 2005-04-24, 02:52   #9
Peter Nelson
 
Peter Nelson's Avatar
 
Oct 2004

232 Posts
Default

Well, I can confirm that the supplied Factor version 4 pre-compiled executable crashes pretty quickly on my Pentium 3 and Pentium 4 northwood machines which run Windows XP.

Previous version 3.2 worked fine.

We will either have to wait for the end of his holiday or some brave soul try to recompile and/or debug it.

I tried throwing the C source code at VC7 in Visual Studio 2003.NET to produce a working executable but I get quite a few errors probably because I haven't specified the correct compilation options, libraries etc.

Has anyone else got this problem too or got it working - either the supplied executable or by recompiling?
Peter Nelson is offline   Reply With Quote
Old 2005-04-24, 05:03   #10
Joshua2
 
Joshua2's Avatar
 
Sep 2004

13·41 Posts
Default

I too got it crash on my P4. After I entered my number and exponents, it crashed immediately. Previous version works fine. I don't know how to compile.
Joshua2 is offline   Reply With Quote
Old 2005-05-02, 07:29   #11
ET_
Banned
 
ET_'s Avatar
 
"Luigi"
Aug 2002
Team Italia

2·2,383 Posts
Default

Did anybody try the Athlon executable on a Windows platform?

Luigi
ET_ is offline   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
'Total time' value in Factor4 Rodrigo Operation Billion Digits 6 2010-11-14 16:41
P!=NP in the news willmore Computer Science & Computational Number Theory 48 2010-09-19 08:30
Other news Cruelty Riesel Prime Search 41 2010-03-08 18:46
The news giveth, the news taketh away... NBtarheel_33 Hardware 17 2009-05-04 15:52
Factor4 - Benchpage ET_ Operation Billion Digits 30 2005-11-28 13:55

All times are UTC. The time now is 11:44.

Sun Sep 27 11:44:18 UTC 2020 up 17 days, 8:55, 0 users, load averages: 1.62, 1.43, 1.40

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

This forum has received and complied with 0 (zero) government requests for information.

Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.
A copy of the license is included in the FAQ.