mersenneforum.org  

Go Back   mersenneforum.org > Factoring Projects > Lone Mersenne Hunters > LMH > 100M

Reply
 
Thread Tools
Old 2004-03-16, 22:37   #1
ThomRuley
 
ThomRuley's Avatar
 
May 2003

26210 Posts
Default Speed Issue

I have been doing AdvancedFactor on a range for LMH (240-250M), and others on this forum hinted that I may be experiencing speed problems. I am using Prime95 on a 900MHz Athlon with ~750MB RAM. I am using WellBehavedWork=1 and not using the computer for anything else. Also, the worktodo.ini line is
AdvancedFactor=240000000,250000000,0,32.

My computer has progressed from 240000000 to 240075000 in 24 days. Is there anything I can do to improve the speed?

On the plus side, I am noticing a 50-60% elimination rate for prime exponents. Spot-checking has shown them to be legitimate candidates.

ThomRuley
ThomRuley is offline   Reply With Quote
Old 2004-03-16, 23:47   #2
hbock
 
hbock's Avatar
 
Feb 2003

101000112 Posts
Default

Sounds very strange. On my Athlon 1.2G it took 768 s to do AdvancedFactor=250000000,250010000,0,55 (571 prime exponents tested, 319 eliminated, ie. 55.%). Extrapolated to a range of 75000 for your system it should take about 2-2.5 h. If you factor only to 32 bit it takes just few minutes but the expected rate of exponents eliminated in your range I guess is then around 10 %. So I think something is wrong. Have you factored maybe to 62 bit ? My timings have been obtained under Win with version 21.4 (same timings as 21.3). Another reason might be another process consuming all the CPU time (Prime95 runs with lowest priority), but OTOH this is contradictory to the percentage of exponents eliminated.

Last fiddled with by hbock on 2004-03-17 at 00:02
hbock is offline   Reply With Quote
Old 2004-03-17, 04:12   #3
ThomRuley
 
ThomRuley's Avatar
 
May 2003

2·131 Posts
Default

Something else I failed to mention. My entire worktodo.ini file:

AdvancedFactor=240000000,250000000,0,32
AdvancedFactor=240000000,250000000,33,50
AdvancedFactor=240000000,250000000,51,60
AdvancedFactor=240000000,250000000,61,70

Could the other lines be causing a problem? I had noticed that some of the factors found were larger, much larger than 2^32.

ThomRuley
ThomRuley is offline   Reply With Quote
Old 2004-03-17, 22:08   #4
hbock
 
hbock's Avatar
 
Feb 2003

163 Posts
Default

Ok, this explains a lot. Your factoring levels are very high. 50-60 bit maybe finished or not but factoring to 70 bit is too high I guess. If you really want to continue (AFAIK nobody else did that in these ranges) test smaller steps like 61-62 etc. or change to another range.
Maybe there is also a problem with your result file, size could be 100M meanwhile or higher. The first two steps (0-32 and 33-50bit) should be finished anyway and result in a file of at least 60 MB, so it's better to separate the results in several files. Regarding speed, remember every bit factoring depth more costs at least twice the time so you can estimate a practical limit.
hbock is offline   Reply With Quote
Old 2004-03-18, 04:15   #5
ThomRuley
 
ThomRuley's Avatar
 
May 2003

2·131 Posts
Default

I cut out those other lines already. I haven't had a chance to check whether that helped.

I haven't noticed that it slowed down as the results file expanded, but I'll try moving those lines to another file.

Any other suggestions will be greatly appreciated.

Thanks

ThomRuley
ThomRuley is offline   Reply With Quote
Old 2004-03-19, 00:50   #6
ThomRuley
 
ThomRuley's Avatar
 
May 2003

4068 Posts
Default

I tried cutting out the additional lines in worktodo.ini and moving results to another file, but these didn't seem to help.

I just tried reinstalling Prime95 to see if that helps. On a possibly related note, I have been having trouble with Windows on that machine. Could that be causing some of the problem?



ThomRuley
ThomRuley is offline   Reply With Quote
Old 2004-03-19, 01:02   #7
nfortino
 
nfortino's Avatar
 
Nov 2003

3·5·11 Posts
Default

Quote:
Originally Posted by ThomRuley
I tried cutting out the additional lines in worktodo.ini and moving results to another file, but these didn't seem to help.

I just tried reinstalling Prime95 to see if that helps. On a possibly related note, I have been having trouble with Windows on that machine. Could that be causing some of the problem?



ThomRuley
This is probably because you have WellBehavedWork=1 in prime.ini. Remove it (close/restart the client), and see what happens.
nfortino is offline   Reply With Quote
Old 2004-03-19, 15:12   #8
ThomRuley
 
ThomRuley's Avatar
 
May 2003

10616 Posts
Default

Thanks, I'll try that. I also found another solution. I replaced all my lines in worktodo.ini with AdvancedFactor=240000000,250000000,0,60. I also handed some of the range to my newer, faster computer. That should speed things up a bit.
ThomRuley is offline   Reply With Quote
Old 2004-03-21, 23:46   #9
hbock
 
hbock's Avatar
 
Feb 2003

163 Posts
Default

Sorry, I can't understand what happend. Can you tell us the size of your first results file after 24 days of running. What are the last lines of this file (hope you saved it), ie. how much was the progress (Mxxxxxxx has no factor from ....to 2^yy). You told us that some factors found have been much larger than 2^32, so factoring to 50 bit or maybe 60 bit is already finished !
If you start now again from 0-60 bit much less factors will be found because if a small factor is found on a distinct level no further search, ie. to 60 bit, will be carried out and next exponent is tested. OTOH by this means Advancedfactor 0-60 might be faster as factoring from 50-60 bit with much less factors found but the number of exponents eliminated remains the same.
hbock is offline   Reply With Quote
Old 2004-03-22, 00:00   #10
ThomRuley
 
ThomRuley's Avatar
 
May 2003

10616 Posts
Default

Strangely enough, I didn't notice a change in speed when I condensed all my worktodo.ini lines to 0-60.
ThomRuley is offline   Reply With Quote
Old 2005-04-26, 22:18   #11
moo
 
moo's Avatar
 
Jul 2004
Nowhere

809 Posts
Default

60 is rather high it will take about 1 minute per factor at 60 alone susgest going to 50
moo is offline   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
DC performance issue cuBerBruce Information & Answers 2 2017-07-10 13:43
GWNUM issue ET_ Software 4 2017-06-15 09:52
32/64 bit gmp-ecm issue... WraithX GMP-ECM 15 2016-12-19 17:42
That's odd ... ubuntu-13.10 ecm issue fivemack GMP-ECM 12 2015-06-05 02:02
PauseWhileRunning issue Kevin Software 1 2011-06-16 05:33

All times are UTC. The time now is 13:19.


Tue Dec 7 13:19:56 UTC 2021 up 137 days, 7:48, 0 users, load averages: 1.27, 1.35, 1.39

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2021, 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.