mersenneforum.org  

Go Back   mersenneforum.org > Factoring Projects > Aliquot Sequences

Reply
 
Thread Tools
Old 2013-10-01, 12:37   #485
gd_barnes
 
gd_barnes's Avatar
 
May 2007
Kansas; USA

237368 Posts
Default

With the newest version of Yafu, I am finding that the optimum crossover is likely > 100 and that's on a 2-year-old I7 running at 2.6 Ghz. (I don't overclock.) For now, I'm using 100 but I'll do some testing later on to figure out what is best. 94 or 95 is definitely not optimum. Yafu QS is FAST running on multiple cores!

But...if one does what LaurV suggested, i.e.:
Quote:
...use "yafu -tune()", as a separate command, to set the crossover to some optimum value depending on your machine, and delete the xover line (comment it out) in the ini. You will get better performance, as the optimum crossover depends on the machine.
Then you don't need to know what the optimum crossover is. Yafu will figure it out based on your machine.

It's not easy for me to test that one way or another. I may just try it for a while and see if I get a performance similar to or hopefully better than having the value hard-coded.

Last fiddled with by gd_barnes on 2013-10-01 at 12:39
gd_barnes is offline   Reply With Quote
Old 2013-10-01, 21:09   #486
Antonio
 
Antonio's Avatar
 
"Antonio Key"
Sep 2011
UK

32×59 Posts
Default

Quote:
Originally Posted by henryzz View Post
I am surprised that is so high. Maybe I need to check again with a later version of yafu. Once upon a time the optimal value was around 95 with 64-bit everything. Yafu will have improved since then I suppose.
Meant to say - the tune info is with Yafu v.1.34.5
Antonio is offline   Reply With Quote
Old 2013-10-01, 21:34   #487
henryzz
Just call me Henry
 
henryzz's Avatar
 
"David"
Sep 2007
Cambridge (GMT/BST)

2×2,861 Posts
Default

Quote:
Originally Posted by Antonio View Post
Meant to say - the tune info is with Yafu v.1.34.5
Last time I tuned was probably with 1.10 or something rediculous
henryzz is online now   Reply With Quote
Old 2014-01-19, 16:22   #488
EdH
 
EdH's Avatar
 
"Ed Hall"
Dec 2009
Adirondack Mtns

17×197 Posts
Default Aliqueit Crash with "WARNING: bad factor found: 5"

Darn my luck -- I get a downdriver run and Aliqueit crashes!

I have tried swapping prefer_yafu between true and false and all that does is add in a "running qs (msieve)..." line.

From my aliqueit.log:
Code:
...
[Jan 19 2014, 10:46:39] *** Starting 915684:1511 = 285784543673563430739334 (24 digits)
[Jan 19 2014, 10:46:39] *** prp1 = 2
[Jan 19 2014, 10:46:39] *** prp1 = 7
[Jan 19 2014, 10:46:39] *** prp2 = 13
[Jan 19 2014, 10:46:39] *** prp2 = 17
[Jan 19 2014, 10:46:39] *** prp3 = 223
[Jan 19 2014, 10:46:39] *** prp4 = 5237

[Jan 19 2014, 10:46:39] Cofactor 79091714583611 (14 digits)

[Jan 19 2014, 10:46:39] c14: running rho...
[Jan 19 2014, 10:46:39] *** c10 = 2435164709

[Jan 19 2014, 10:46:39] c10: running rho...

[Jan 19 2014, 10:46:39] c10: running qs (yafu)...
[Jan 19 2014, 10:46:39] WARNING: bad factor found: 5
01/19/14 10:46:39 v1.34.5 @ math34, starting smallmpqs on C10: 2435164709
01/19/14 10:46:39 v1.34.5 @ math34, prp1 = 5
01/19/14 10:46:39 v1.34.5 @ math34, prp10 = 2435164709
[Jan 19 2014, 10:46:39] *** c10 = 2435164709

[Jan 19 2014, 10:46:39] c10: running rho...
...
Similar with prefer_yafu=false:
Code:
...
[Jan 19 2014, 10:52:50] *** Starting 915684:1511 = 285784543673563430739334 (24 digits)
[Jan 19 2014, 10:52:50] *** prp1 = 2
[Jan 19 2014, 10:52:50] *** prp1 = 7
[Jan 19 2014, 10:52:50] *** prp2 = 13
[Jan 19 2014, 10:52:50] *** prp2 = 17
[Jan 19 2014, 10:52:50] *** prp3 = 223
[Jan 19 2014, 10:52:50] *** prp4 = 5237
[Jan 19 2014, 10:52:50] *** c10 = 2435164709

[Jan 19 2014, 10:52:50] c10: running rho...

[Jan 19 2014, 10:52:50] c10: running qs (msieve)...

[Jan 19 2014, 10:52:50] c10: running qs (yafu)...
[Jan 19 2014, 10:52:50] WARNING: bad factor found: 5
01/19/14 10:52:50 v1.34.5 @ math34, starting smallmpqs on C10: 2435164709
01/19/14 10:52:50 v1.34.5 @ math34, prp1 = 5
01/19/14 10:52:50 v1.34.5 @ math34, prp10 = 2435164709
[Jan 19 2014, 10:52:50] *** c10 = 2435164709

[Jan 19 2014, 10:52:50] c10: running rho...
...
c10 = 2435164709 = 21587 * 112807

But, I see it being called prp above. Any help?

BTW, all programs are up-to-date and this isn't one of my ancient machines:
core 2 quad core 4G RAM

Thanks for any help...

Last fiddled with by EdH on 2014-01-19 at 16:31 Reason: Machine data correction...
EdH is offline   Reply With Quote
Old 2014-01-19, 19:58   #489
VBCurtis
 
VBCurtis's Avatar
 
"Curtis"
Feb 2005
Riverside, CA

52·173 Posts
Default

Can you create the new line in the .elf manually, and let aliqueit just continue onto the next line?
VBCurtis is offline   Reply With Quote
Old 2014-01-19, 21:38   #490
EdH
 
EdH's Avatar
 
"Ed Hall"
Dec 2009
Adirondack Mtns

D1516 Posts
Default

Quote:
Originally Posted by VBCurtis View Post
Can you create the new line in the .elf manually, and let aliqueit just continue onto the next line?
Actually, I fed Aliqueit one of the factors, via the -f option and it stepped to the next line and did the same thing. It kept doing it until I took the .elf to another machine, where it added some more lines. Interestingly, the second machine has displayed the identical behaviour in the past, but I just worked around it that time. When I moved back to the original machine, it added many lines and then did the same, again. I have kept a copy of the .elf at that point to experiment with, as desired.

Here's the aliqueit.log section for that run:
Code:
[Jan 19 2014, 16:23:52] *** Starting 915684:2394 = 486308988132605250820828474054756417783557491729493124060168984953778543236406694425640 (87 digits)
[Jan 19 2014, 16:23:52] *** prp1 = 2
[Jan 19 2014, 16:23:52] *** prp1 = 2
[Jan 19 2014, 16:23:52] *** prp1 = 2
[Jan 19 2014, 16:23:52] *** prp1 = 5

[Jan 19 2014, 16:23:52] Cofactor 12157724703315131270520711851368910444588937293237328101504224623844463580910167360641 (86 digits)

[Jan 19 2014, 16:23:52] c86: running rho...
[Jan 19 2014, 16:23:52] *** c10 = 3012104413

[Jan 19 2014, 16:23:52] c10: running rho...

[Jan 19 2014, 16:23:52] c10: running qs (yafu)...
[Jan 19 2014, 16:23:52] WARNING: bad factor found: 13
01/19/14 16:23:52 v1.34.5 @ math34, starting smallmpqs on C10: 3012104413
01/19/14 16:23:52 v1.34.5 @ math34, prp2 = 13
01/19/14 16:23:52 v1.34.5 @ math34, prp10 = 3012104413
[Jan 19 2014, 16:23:52] *** c10 = 3012104413

[Jan 19 2014, 16:23:52] c10: running rho...
I have also run mprime stress testing with no failures. And, after I have worked well past the area on another machine, it is now running well, further on.

I have discovered something interesting, though and will probably move this discussion to a YAFU thread (unless B2 pops in here):
Code:
01/19/14 16:25:24 v1.34.5 @ math34, System/Build Info: 
Using GMP-ECM, Powered by GMP
detected Intel(R) Core(TM)2 Quad CPU    Q9400  @ 2.66GHz
detected L1 = 32768 bytes, L2 = 3145728 bytes, CL = 64 bytes
measured cpu frequency ~= 2659.670720
using 20 random witnesses for Rabin-Miller PRP checks

===============================================================
======= Welcome to YAFU (Yet Another Factoring Utility) =======
=======             bbuhrow@gmail.com                   =======
=======     Type help at any time, or quit to quit      =======
===============================================================
cached 78498 primes. pmax = 999983


>> siqs(3012104413)



***factors found***

P2 = 13
C10 = 3012104413
1
>>
Unfortunately, this all seems familiar about YAFU having some kind of trouble at some digit value, but my memory is too poor to find that discussion and I thought it had been taken care of.

Thanks for the help...

Last fiddled with by EdH on 2014-01-19 at 22:01
EdH is offline   Reply With Quote
Old 2014-01-19, 21:50   #491
EdH
 
EdH's Avatar
 
"Ed Hall"
Dec 2009
Adirondack Mtns

17×197 Posts
Default

I'm adding this as a separate post, so it isn't missed by someone:

Anyway, the reason why msieve didn't work when I tried prefer_yafu false is because I'm using an msieve binary with MPI and it returns the factors in .log.mpi00 instead of .log where Aliqueit searches. I can remedy that easily in Aliqueit, but that doesn't address the YAFU issue...

Edit: The above seems incorrect as in that I cannot make msieve (siqs) work at all in finding the factors for Aliqueit - another puzzle to frustrate me.

Edit2: I have the prefer_yafu = false version working now with a copy of msieve without MPI. I don't know how, but the msieve.log file was hidden along with a 1.6G msieve.dat file. They appeared when I commented out the delete msieve.log line in aliqueit.cc. The msieve.log file had all the previous runs in it. I deleted both and reinstated the original line in aliqueit.cc and msieve now works. prefer_yafu = true still does not.

Last fiddled with by EdH on 2014-01-19 at 22:32
EdH is offline   Reply With Quote
Old 2014-01-20, 00:46   #492
EdH
 
EdH's Avatar
 
"Ed Hall"
Dec 2009
Adirondack Mtns

334910 Posts
Default

It appears to be one of the upgrades that broke it. Upgrading other machines broke them, too I've posted in the YAFU thread and, if I get a chance, I'll try to see which upgrade broke it.

I'd be interested if anyone else has this trouble...

Last fiddled with by EdH on 2014-01-20 at 00:48
EdH is offline   Reply With Quote
Old 2014-01-20, 02:55   #493
LaurV
Romulan Interpreter
 
LaurV's Avatar
 
Jun 2011
Thailand

2·54·7 Posts
Default

Here 1.34.3, with MPIR under Win64, it is working well. Didn't try other combinations. If it works, don't fix it!

Code:
01/20/14 09:49:47 v1.34.3 @ NB-12906, System/Build Info:
Using GMP-ECM 6.4.2, Powered by MPIR 2.6.0
detected        Intel(R) Core(TM) i5-3340M CPU @ 2.70GHz
detected L1 = 32768 bytes, L2 = 3145728 bytes, CL = 64 bytes
measured cpu frequency ~= 2690.490420
using 20 random witnesses for Rabin-Miller PRP checks

===============================================================
======= Welcome to YAFU (Yet Another Factoring Utility) =======
=======             bbuhrow@gmail.com                   =======
=======     Type help at any time, or quit to quit      =======
===============================================================
cached 78498 primes. pmax = 999983

>> siqs(3012104413)

***factors found***

P5 = 97577
P5 = 30869

ans = 1

>>
LaurV is online now   Reply With Quote
Old 2014-01-20, 05:06   #494
EdH
 
EdH's Avatar
 
"Ed Hall"
Dec 2009
Adirondack Mtns

17×197 Posts
Default

I've isolated it to YAFU. See this post.
EdH is offline   Reply With Quote
Old 2016-09-19, 01:41   #495
EdH
 
EdH's Avatar
 
"Ed Hall"
Dec 2009
Adirondack Mtns

64258 Posts
Default

The current version (1.12 linux) is in need of updating. I've had to remove the b2scale portions from the source code and change the factorization.ath.cx reference to factordb.com to report factors.

Is no one else having these issues? Is anyone else using Aliqueit on linux?

I would prefer an "official" version with updates over my local modifications...
EdH 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 11:33.

Tue Sep 29 11:33:17 UTC 2020 up 19 days, 8:44, 0 users, load averages: 1.87, 1.49, 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.