mersenneforum.org  

Go Back   mersenneforum.org > Factoring Projects > Aliquot Sequences

Reply
 
Thread Tools
Old 2009-03-24, 18:15   #56
mklasson
 
Feb 2004

25810 Posts
Default

Quote:
Originally Posted by mklasson View Post
Curiously, the original Floyd cycle finder yields better speed than the Brent version when I benchmark on real sequence data. "In the lab" Brent seems to perform significantly faster though... I've probably just overlooked something.
Too low iteration cap... Upping it to 6k gives Brent the upper hand in real life as well.

Doing the first 1225 lines of 10212 takes 48s now.

Incidentally, running the entire sequence for 840 (one of the original Lehmer Six) takes just 30 seconds. Isn't Moore's Law great?
mklasson is offline   Reply With Quote
Old 2009-03-24, 18:18   #57
10metreh
 
10metreh's Avatar
 
Nov 2008

2·33·43 Posts
Default

Quote:
Originally Posted by mklasson View Post
Too low iteration cap... Upping it to 6k gives Brent the upper hand in real life as well.

Doing the first 1225 lines of 10212 takes 48s now.

Incidentally, running the entire sequence for 840 (one of the original Lehmer Six) takes just 30 seconds. Isn't Moore's Law great?
If Moore's Law continues, sequence 840 will take 0.00002 seconds by 2050.
10metreh is offline   Reply With Quote
Old 2009-03-24, 18:41   #58
Phil MjX
 
Phil MjX's Avatar
 
Sep 2004

B916 Posts
Default

Quote:
Originally Posted by mklasson View Post
Ah, the problem is that the 2924 line has no space between "*" and the cofactor. Even so, it certainly isn't polite to crash...

Just add a space for now and it'll work ok.

Thanks!
Thanks a lot to you !
Politeness isn't a bug so all is OK !
Phil MjX is offline   Reply With Quote
Old 2009-03-24, 18:57   #59
mklasson
 
Feb 2004

2·3·43 Posts
Default

Quote:
Originally Posted by mklasson View Post
Incidentally, running the entire sequence for 840 (one of the original Lehmer Six) takes just 30 seconds. Isn't Moore's Law great?
Ack... I should've said "Dr Lehmer sure was impatient..." instead. What a lost opportunity!

Quote:
Originally Posted by Phil MjX View Post
Thanks a lot to you !
Politeness isn't a bug so all is OK !
I was aiming for "Being buggy isn't polite", but I like your version as well.
mklasson is offline   Reply With Quote
Old 2009-03-29, 11:37   #60
hhh
 
hhh's Avatar
 
Jun 2005

373 Posts
Default

Can we please have the subordinated commands started at low priority? For the long jobs, I adjust it by hand, but that's no solution.

And could there be a switch to skip ecm for the first line? Like when you stop the program and restart it. Even better would be to keep track of the ecm done.

When I stop the program, restart, stop, restart, there are too many factors stored, which leads to "Warning, 2 is no factor." of so messages.

Finally, imagine aliqueit is working on some c100. after some ecm, a p20 is found, leaving a c80. Then, ecm starts over at zero, even if one would better continue directly at the 20 digit level.

For everything else: nice work, and thank you! H.
hhh is offline   Reply With Quote
Old 2009-03-29, 11:57   #61
mklasson
 
Feb 2004

2×3×43 Posts
Default

Quote:
Originally Posted by hhh View Post
Can we please have the subordinated commands started at low priority? For the long jobs, I adjust it by hand, but that's no solution.
I suppose I could, but it's very easy to do yourself with "start" or "nice" depending on your system. For windows, "start /b /low /wait aliqueit.exe 162126" starts aliqueit.exe and all it's subprocesses with low priority.

Quote:
Originally Posted by hhh View Post
And could there be a switch to skip ecm for the first line? Like when you stop the program and restart it. Even better would be to keep track of the ecm done.
There already is. Sander beat you to it. Try "-e". Yeah, keeping track of work done would be neat.

Quote:
Originally Posted by hhh View Post
When I stop the program, restart, stop, restart, there are too many factors stored, which leads to "Warning, 2 is no factor." of so messages.
I know, it's bugged me a little too. All in all though, it's just a minor cosmetic thing.

Quote:
Originally Posted by hhh View Post
Finally, imagine aliqueit is working on some c100. after some ecm, a p20 is found, leaving a c80. Then, ecm starts over at zero, even if one would better continue directly at the 20 digit level.
Yeah... I should do something about that.

Quote:
Originally Posted by hhh View Post
For everything else: nice work, and thank you! H.
Thanks for the kind words and your comments.
mklasson is offline   Reply With Quote
Old 2009-03-29, 11:58   #62
10metreh
 
10metreh's Avatar
 
Nov 2008

2×33×43 Posts
Default

Quote:
Originally Posted by hhh View Post
Can we please have the subordinated commands started at low priority? For the long jobs, I adjust it by hand, but that's no solution.

And could there be a switch to skip ecm for the first line? Like when you stop the program and restart it. Even better would be to keep track of the ecm done.

When I stop the program, restart, stop, restart, there are too many factors stored, which leads to "Warning, 2 is no factor." of so messages.

Finally, imagine aliqueit is working on some c100. after some ecm, a p20 is found, leaving a c80. Then, ecm starts over at zero, even if one would better continue directly at the 20 digit level.

For everything else: nice work, and thank you! H.
There is a switch to skip ecm for the first line, and it's -e. Also, I often don't want jobs started at low priority, so could there be a -n switch for low priority?

Last fiddled with by 10metreh on 2009-03-29 at 11:58
10metreh is offline   Reply With Quote
Old 2009-03-29, 12:11   #63
mklasson
 
Feb 2004

25810 Posts
Default

Quote:
Originally Posted by 10metreh View Post
There is a switch to skip ecm for the first line, and it's -e. Also, I often don't want jobs started at low priority, so could there be a -n switch for low priority?
Yeah, if I do implement it it'll be in the form of a switch.
mklasson is offline   Reply With Quote
Old 2009-03-29, 18:30   #64
mklasson
 
Feb 2004

2·3·43 Posts
Default

Some improvements for now: aliqueit v1.05.

+ prints the size of the starting composite for each line.
+ runs pollard rho (internal) right after the trial factoring. Nice speedup for small numbers.
+ runs auto-increasing ecm forever instead of exiting if something goes wrong and ggnfs fails to find a factor.
+ detects some malformed elf files instead of crashing while verifying... Thanks Phil MjX.
+ "-h" prints a little help text on cmdline args.
+ logs cofactors explicitly.
+ the expressions used to determine maximum ecm level can now be customised with <qs_k>, <qs_m>, <gnfs_k>, and <gnfs_m>.
+ "-p" runs aliqueit and all spawned processes at idle priority.

I'm hoping it still compiles under linux. The priority code is ripped from a 5-year old proth_sieve, but I highly doubt it's become invalid since then.
mklasson is offline   Reply With Quote
Old 2009-03-30, 04:30   #65
Andi47
 
Andi47's Avatar
 
Oct 2004
Austria

2·17·73 Posts
Default

Just a question (I have not yet tested it): Does aliqueit print some statistical data for GNFS to a logfile (siever used, how many relations needed, size of special Q range)?
Andi47 is offline   Reply With Quote
Old 2009-03-30, 10:43   #66
mklasson
 
Feb 2004

2·3·43 Posts
Default

Quote:
Originally Posted by Andi47 View Post
Just a question (I have not yet tested it): Does aliqueit print some statistical data for GNFS to a logfile (siever used, how many relations needed, size of special Q range)?
It creates a new directory for each ggnfs run, runs ggnfs, cleans up the work files afterwards, parses ggnfs.log to find the factors, and continues working on the next sequence line. So no, and yes. It doesn't write any of that stuff to aliqueit.log, but the original ggnfs.log and summary file is saved so you'll find the info there. You can also customise what files will be deleted upon completion so it's possible to save other files you may want as well.
mklasson is offline   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
Resuming aliqueit johnadam74 Aliquot Sequences 4 2016-03-28 12:32
Apparent aliqueit issue with specifying factors pakaran Aliquot Sequences 2 2015-09-12 23:10
Using Several Instances of Aliqueit for a large gnfs job EdH Aliquot Sequences 6 2011-12-13 18:58
Setting up aliqueit science_man_88 Aliquot Sequences 185 2011-11-08 12:18
Tried out aliqueit.exe: ggnfs failing Greebley Aliquot Sequences 35 2010-02-13 15:23

All times are UTC. The time now is 14:27.

Thu Aug 13 14:27:45 UTC 2020 up 11:03, 1 user, load averages: 1.51, 1.53, 1.57

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.