mersenneforum.org Factoring projects algorithms - What is Aliqueit ?
 Register FAQ Search Today's Posts Mark Forums Read

2016-08-07, 09:15   #23
xilman
Bamboozled!

"𒉺𒌌𒇷𒆷𒀭"
May 2003
Down not across

2·5,827 Posts

Quote:
 Originally Posted by Romuald ... I see. The main difference between your installation and potentially mine is that I would use the python script (i have the bases of python but i am ignorant of PL/I).
I also do not speak PL/I.

ITYM Perl

 2016-08-07, 09:56 #24 Romuald     Oct 2015 France 3F16 Posts Yes, but for some reasons I have more confidence in python than PL/I, I know how python works, i know how to install it, etc. PL/I is totally unknown for me. I don't even know if it's installed on my laptop.
 2016-08-07, 10:19 #25 Dubslow Basketry That Evening!     "Bunslow the Bold" Jun 2011 40
2016-08-07, 10:42   #26
xilman
Bamboozled!

"𒉺𒌌𒇷𒆷𒀭"
May 2003
Down not across

2×5,827 Posts

Quote:
 Originally Posted by Dubslow I have no idea what PL/I is.
You should find out, as an awful warning. Luckily it is much easier to avoid PL/I now than it used to be.

 2016-08-07, 11:27 #27 Romuald     Oct 2015 France 32×7 Posts Here is my architecture: Code: matthias@matthias-Aspire-V3-771G:~\$ screenfetch ./+o+- matthias@matthias-Aspire-V3-771G yyyyy- -yyyyyy+ OS: Ubuntu 16.04 xenial ://+//////-yyyyyyo Kernel: x86_64 Linux 4.4.0-34-generic .++ .:/++++++/-.+sss/ Uptime: 3d 2h 4m .:++o: /++++++++/:--:/- Packages: 4090 o:+o+:++....-/oo+++++/ Shell: bash 4.3.46 .:+o:+o/. +sssoo+/ Resolution: 1600x900 .++/+:+oo+o: /sssooo. DE: XFCE /+++//+:oo+o /::--:. WM: Xfwm4 \+/+o+++o++o ++////. WM Theme: Numix .++.o+++oo+: /dddhhh. GTK Theme: Greybird [GTK2] .+.o+oo:. oddhhhh+ Icon Theme: elementary-xfce-dark \+.++o+o-.:ohdhhhhh+ Font: Noto Sans 9 :o+++ ohhhhhhhhyo++os: CPU: Intel Core i5-3230M CPU @ 3.2GHz .o:.syhhhhhhh/.oo++o GPU: GeForce 710M /osyyyyyyo++ooo+++/ RAM: 4114MiB / 7803MiB  +oo+++o\: oo++.
 2016-08-07, 12:21 #28 Dubslow Basketry That Evening!     "Bunslow the Bold" Jun 2011 40
2016-08-07, 14:12   #29
henryzz
Just call me Henry

"David"
Sep 2007
Liverpool (GMT/BST)

136178 Posts

Quote:
 Originally Posted by Dubslow Try these from my Sandy Bridge. Err, attachment is too large? My bziped archive is 3.5MiB, larger than the 2MiB limit. But I know I've seen similar archives floating around here Edit: I just recalled an old link that B^2 has: https://sites.google.com/site/bbuhrow/lasieve4_64.7z That's probably out of date for modern computers but for Ivy Bridge it might well be the best you can get.
There is very little difference. Most of the time is spent in asm which hasn't been updated since core 2. This means that you can get a decent amount of extra performance on a hyperthreaded cpu unlike most programs.
As long as the binary outputs an eta for that block of sieving then it is recent enough for me. There are also lasieve5 binaries about but there is very little performance difference and the extra features are only really applicable to very large factorisations.

2016-08-07, 17:09   #30
VBCurtis

"Curtis"
Feb 2005
Riverside, CA

3·1,877 Posts

Quote:
 Originally Posted by Romuald Here is my architecture: :o+++ `ohhhhhhhhyo++os: CPU: Intel Core i5-3230M CPU @ 3.2GHz
You have the script set to 12 threads (6 cores two hyperthreads per core), too many for your actual CPU. Your CPU is a dual-core with HT, so 4 threads is appropriate for sieving.

The linear algebra step (run by msieve after sieving completes) does not usually benefit from HT, which is why the script has the ability to use a different number of threads for linear algebra than sieving.

Fixing this isn't strictly necessary to get a factorization to run, but running 12 threads on 4-thread silicon isn't very efficient.

2016-08-07, 18:38   #31
Dubslow

"Bunslow the Bold"
Jun 2011
40<A<43 -89<O<-88

3·29·83 Posts

Quote:
 Originally Posted by henryzz There is very little difference. Most of the time is spent in asm which hasn't been updated since core 2.
Oh right lol. Shows how much compiling I've not been doing lately.

 2016-08-09, 09:06 #32 Romuald     Oct 2015 France 32·7 Posts Ok, so I downloaded the binaries of GGNFS from Dubslow's link. I guess I just have to set the path in yafu.ini file to use it. Let's see if it works. Last fiddled with by Romuald on 2016-08-09 at 09:07
2016-08-09, 10:44   #33
Romuald

Oct 2015
France

778 Posts

Quote:
 Originally Posted by VBCurtis You have the script set to 12 threads (6 cores two hyperthreads per core), too many for your actual CPU. Your CPU is a dual-core with HT, so 4 threads is appropriate for sieving. The linear algebra step (run by msieve after sieving completes) does not usually benefit from HT, which is why the script has the ability to use a different number of threads for linear algebra than sieving. Fixing this isn't strictly necessary to get a factorization to run, but running 12 threads on 4-thread silicon isn't very efficient.
Yes. But where do I modify this parameter? I didn't find neither in the yafu.ini nor in the makefile before compilation.
However it's written in the docfile to add '-thread <number>' such as
Code:
,
(note that number is not what I especially want to factorize, it's completely arbitrary as an example, don't tell me "To factorize $2^{127}-1$ rather use this or that, you don't need yafu...")

returns
Code:
This is boring, I follow exactly what it's expected me to do, no way to run YAFU properly.

I note that this problem only occurs when I precisely use GMP & msieve at compiling time, and then I use GMP-ECM & ggnfs binaries added in the yafu.ini file.
I mean when I compile YAFU with simply
Code:
make x86_64 [NFS=1] [PROFILE=1] [TIMING=1] [USE_SSE41=1]
as said in README file, I get no problem.

Last fiddled with by Romuald on 2016-08-09 at 11:02

 Similar Threads Thread Thread Starter Forum Replies Last Post rogue Factoring 20 2014-11-19 01:08 Raman Hobbies 45 2009-05-11 05:11 ThiloHarich Factoring 0 2007-09-02 20:32 koders333 Factoring 14 2006-01-25 14:08 koders333 Factoring 1 2006-01-19 20:04

All times are UTC. The time now is 00:01.

Tue Jan 31 00:01:44 UTC 2023 up 165 days, 21:30, 1 user, load averages: 0.50, 0.79, 0.87