mersenneforum.org  

Go Back   mersenneforum.org > Other Stuff > Archived Projects > ElevenSmooth

 
 
Thread Tools
Old 2008-02-23, 02:36   #23
frmky
 
frmky's Avatar
 
Jul 2003
So Cal

26×3×11 Posts
Default

I finally got the 64-bit asm code to compile in Linux. Here are the benchmarks on a 2.4 GHz Opteron:

64-bit C: 0.41129 sec/rel
32-bit asm: 0.33758 sec/rel
64-bit asm: 0.21451 sec/rel

For those using 64-bit Linux, binaries are available at http://physics.fullerton.edu/gchilde...eve4_64.tar.gz

Greg
frmky is online now  
Old 2008-02-23, 10:01   #24
Andi47
 
Andi47's Avatar
 
Oct 2004
Austria

2·17·73 Posts
Default

May I suggest to move this thread to the "factoring" forum (and leave a post with a link here), so that Tom can edit his first post to collect reservations and finished Q-ranges in a better readable way?

Or alternatively give him special universal powers for the elevensmooth forum so that he can do the same here?
Andi47 is offline  
Old 2008-02-23, 13:54   #25
bsquared
 
bsquared's Avatar
 
"Ben"
Feb 2007

22·3·293 Posts
Default

Quote:
Originally Posted by frmky View Post
I finally got the 64-bit asm code to compile in Linux. Here are the benchmarks on a 2.4 GHz Opteron:

64-bit C: 0.41129 sec/rel
32-bit asm: 0.33758 sec/rel
64-bit asm: 0.21451 sec/rel

For those using 64-bit Linux, binaries are available at http://physics.fullerton.edu/gchilde...eve4_64.tar.gz

Greg
Would you (or anyone else) also have binaries for a core 2 processor, running red hat enterprise linus 3.5? Any lasieve binary would be better than what I have for that platform, since I don't have the priv's to install anything on the box, but I have permission to run programs. The above fail immediately with the message "FATAL: kernel too old", then a segfault.

thanks,
- ben.
bsquared is offline  
Old 2008-02-23, 17:57   #26
frmky
 
frmky's Avatar
 
Jul 2003
So Cal

211210 Posts
Default

Quote:
Originally Posted by bsquared View Post
Would you (or anyone else) also have binaries for a core 2 processor, running red hat enterprise linus 3.5? Any lasieve binary would be better than what I have for that platform, since I don't have the priv's to install anything on the box, but I have permission to run programs. The above fail immediately with the message "FATAL: kernel too old", then a segfault.

thanks,
- ben.
These will work with Core 2 in 64-bit Linux. However, they require kernel 2.6.9 or later. I don't have an older Linux computer to compile them on. Are GMP and m4 installed on the box? If so, then you probably have the privileges needed to compile your own binary. (For that matter, you can compile GMP in your own directory without having to install it systemwide, but that's another issue.)

Greg
frmky is online now  
Old 2008-02-23, 19:17   #27
bsquared
 
bsquared's Avatar
 
"Ben"
Feb 2007

66748 Posts
Default

Quote:
Originally Posted by frmky View Post
These will work with Core 2 in 64-bit Linux. However, they require kernel 2.6.9 or later. I don't have an older Linux computer to compile them on. Are GMP and m4 installed on the box? If so, then you probably have the privileges needed to compile your own binary. (For that matter, you can compile GMP in your own directory without having to install it systemwide, but that's another issue.)

Greg
GMP and m4 are not installed, which is one of the reasons I didn't attempt to build ggnfs on this box. Do I need to build GMP with these binaries? Or is it just a kernel issue?
bsquared is offline  
Old 2008-02-23, 20:48   #28
frmky
 
frmky's Avatar
 
Jul 2003
So Cal

211210 Posts
Default

Quote:
Originally Posted by bsquared View Post
GMP and m4 are not installed, which is one of the reasons I didn't attempt to build ggnfs on this box. Do I need to build GMP with these binaries? Or is it just a kernel issue?
I suspect it's just a kernel issue. You need 2.6.9 or later to use them. You can type
uname -r
on that computer to see what version it is using.

Greg
frmky is online now  
Old 2008-02-24, 15:55   #29
bsquared
 
bsquared's Avatar
 
"Ben"
Feb 2007

22×3×293 Posts
Default

27M -28M is done. reserving 28M - 29M.
bsquared is offline  
Old 2008-02-25, 07:11   #30
Andi47
 
Andi47's Avatar
 
Oct 2004
Austria

2×17×73 Posts
Default

25M - 26M is done.
Edit: total yield: 1143546 relations, 0,45488 sec/rel.

reserving 29M - 30M

Last fiddled with by Andi47 on 2008-02-25 at 07:53 Reason: added additional info
Andi47 is offline  
Old 2008-03-19, 22:53   #31
FactorEyes
 
FactorEyes's Avatar
 
Oct 2006
vomit_frame_pointer

23×32×5 Posts
Default Benchmark for Opteron 2218 2.66 GHz

Opteron 2218 (2.67 GHz), running 64-bit lasieve binary -- courtesy of frmky -- on one core:
total yield: 1096, q=50001037 (0.21362 sec/rel)

Last fiddled with by FactorEyes on 2008-03-19 at 22:57
FactorEyes is offline  
Old 2008-03-20, 00:32   #32
WraithX
 
WraithX's Avatar
 
Mar 2006

47910 Posts
Default

Dual Xeon 5335 (2.0 GHz), running 32-bit lasieve binary (in Cygwin on 64-bit Windows XP: on one core):
total yield: 1097, q=500001037 (0.37322 sec/rel)

I saw times as low as 0.344 sec/rel, but at the end of the run, it had climbed back up to 0.37322 sec/rel. Also, I saw the same results whether I ran 1,2, or 3 concurrent instances of the sieve.

I was wondering, frmky (or anyone else), can you make a 64-bit binary for windows? Since cygwin and msys are 32-bit environments, I can't run 64-bit binaries in them. (I think)
WraithX is offline  
Old 2008-03-20, 01:04   #33
fivemack
(loop (#_fork))
 
fivemack's Avatar
 
Feb 2006
Cambridge, England

2×132×19 Posts
Default

Q6600 2.4GHz, running gnfs-lasieve4I14e as distributed with ggnfs mid-2007 snapshot: total yield: 1097, q=50001037 (0.52766 sec/rel). 8m5s utime (=0.442s/rel); correcting by utime/realtime gives 0.40621 sec/rel.

running gnfs-lasieve4I14e (64-bit asm) as turbocharged by Joppe_Bos: 1019 relations in 4m41s utime (=0.275s/rel, but the ggnfs version's timer doesn't account for the setup time). Not quite sure why the relation counts are different when the ranges are the same.

This is a machine which was also running four active sieving jobs, so these are not optimistic timings.
fivemack is offline  
 

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
SNFS 200 parameters JoeCrump Factoring 3 2009-12-06 14:50
Quartic: Parameters R.D. Silverman Factoring 9 2009-02-18 23:24
M2376 reservations thread fivemack ElevenSmooth 53 2008-04-28 08:53
optimal parameters for GMP-ECM , -oe+ , -I Walter Nissen GMP-ECM 16 2007-03-20 19:35
ECM parameters for RSA Spider Factoring 24 2006-06-05 23:42

All times are UTC. The time now is 17:19.


Wed Jul 28 17:19:46 UTC 2021 up 5 days, 11:48, 0 users, load averages: 3.30, 3.61, 3.55

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.