ECM work on 4788:2549.c170

schickel · 2010-09-14, 02:28

Just in case we lose the "ECM completed" info ~~before~~ when Syd gets it transferred over to the new DB, here is what we have so far:

Code:

     B1            B2        Crvs Who
--------------------------------------------------
  11000000       (default?)    5  josh             
  11000000      35133391030   27  firejuggler      
  43000000     178426462987  100  Frank Schickel   
  43000000     240490660426  490  Andreas Schindel 
 110000000     769593044236  465  Andreas Schindel 
 110000000     769593044236   95  EdH
 110000000     769593044236   54  EdH
 110000000     761244805240   14  EdH
 110000000     776278396540   78  EdH 
 110000000     776278396540  106  EdH
 110000000    1589211473866 1000  Jayson R. King  
 110000000    1589211473870 1000  Jayson R. King 
 260000000    2597319715006  161  Andreas Schindel
 260000000    2716922571100  518  Andreas Schindel 
 260000000    2716922571100  462  Andreas Schindel
 260000000    2716922571100   35  Andreas Schindel
 260000000    2716922571100   91  Andreas Schindel
 260000000    2716922571100   77  Andreas Schindel
 260000000    3178559884516 1140  em99010pepe
 260000000    3178559884520  257  Andreas Schindel
 260000000    3178559884520    5  Frank Schickel 
 260000000    3178559884520  105  Frank Schickel    
 260000000    3178559884520 1863  Greg             
 260000000    3178559884520    4  EdH      
 260000000    6356857306816   10  Frank Schickel        
4200000000  142960881753708    4  Frank Schickel
4200000000  157651994607000    3  Frank Schickel

Progress:

Code:

09/11  44.49% of t55
09/14  44.81% of t55
09/15  51.60% of t55
09/16  58.24% of t55
09/25  60.2 % of t55
10/02  63   % of t55
10/05  85   % of t55

debrouxl · 2010-09-14, 07:54

Good lord, how much RAM did these curves at B1=4.2e9 require ?

xilman · 2010-09-14, 08:43

Quote:

Originally Posted by debrouxl

Good lord, how much RAM did these curves at B1=4.2e9 require ?

I don't know, but it could have been quite small. If using GMP-ECM the -maxmem parameter sets an upper limit to memory use.

Paul

science_man_88 · 2010-09-14, 12:03

Quote:

Originally Posted by xilman

I don't know, but it could have been quite small. If using GMP-ECM the -maxmem parameter sets an upper limit to memory use.

Paul

Paul wanna help me figure this out I can't seem to get aliquet or ecm to work for me lol.

schickel · 2010-09-14, 19:15

Quote:

Originally Posted by debrouxl

Good lord, how much RAM did these curves at B1=4.2e9 require ?

A lot....but the system actually remained more responsive than when I ran this huge P-1

Code:

9000000000 8344912590100000 1 Frank Schickel

and P+1 work

Code:

4000000000 2086226291290000 3 Frank Schickel

I couldn't even get Task Manager to come up during the P+1 work! Luckily I left it running when I went to bed and it finished overnight.

The ECM curves took ~6 hours each and >10GB of RAM (check out the page faults; this sucker was swapping like crazy during stage 2!)

PS. Andi47, here's the -v details for 4.2e9, default B2:

Code:

Expected number of curves to find a factor of n digits:
35      40      45      50      55      60      65      70      75      80
9       25      76      255     935     3718    15884   72435   350460  1790896

schickel · 2010-09-14, 19:16

Quote:

Originally Posted by xilman

I don't know, but it could have been quite small. If using GMP-ECM the -maxmem parameter sets an upper limit to memory use.

Paul

Silly me, I forgot to limit the RAM the first time!

Andi47 · 2010-09-15, 04:59

Quote:

Originally Posted by schickel

PS. Andi47, here's the -v details for 4.2e9, default B2:

Code:

Expected number of curves to find a factor of n digits:
35      40      45      50      55      60      65      70      75      80
9       25      76      255     935     3718    15884   72435   350460  1790896

Thanks. So if I have calculated correctly, we currently have 44.81% of t55.

schickel · 2010-09-15, 08:24

4 more curves (at a slightly lower B2, 950 curves needed for t55). Would have had one more, but had an unexplained shutdown on the last curve....

EdH · 2010-09-15, 18:25

Quote:

Originally Posted by science_man_88

Paul wanna help me figure this out I can't seem to get aliquet or ecm to work for me lol.

If you're interested, I have the procedure I use to install aliqueit onto an Ubuntu system, available on line at Steps to install and set up Aliqueit on an Ubuntu computer.

It is a work-in-progress at this point, but should be usable. Feedback here is welcomed.

As for ecm, check out the later posts in the 4788 thread. Also note that gmp-ecm is available through the repositories for most linux distros, but for some to use it, it is called via gmp-ecm rather than ecm. For Windows, I've only seen it as ecm.exe.

Andi47 · 2010-09-16, 05:22

462 curves at B1=26e7, B2=2716922571100

Together with EdH's 106 curves at 11e7 reported here, we have reached a total of 51.6% of t55.

jrk · 2010-09-16, 07:46

I finished 1000 more curves @ B1=11e7, B2=1589211473866 (again using -B2scale 2), making 2000 total.

I'm stopping my ECM run for now.

Weird. My B2 is 1589211473866 (same as the first time I reported) yet the DB changes it into B2=1589211473870. Hardly a difference, but it is curious.

2010-09-16, 07:46	#11
jrk May 2008 3·5·73 Posts	I finished 1000 more curves @ B1=11e7, B2=1589211473866 (again using -B2scale 2), making 2000 total. I'm stopping my ECM run for now. Weird. My B2 is 1589211473866 (same as the first time I reported) yet the DB changes it into B2=1589211473870. Hardly a difference, but it is curious. Last fiddled with by jrk on 2010-09-16 at 07:57 Reason: stopping

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
ECM work on 4788:i2765 c160	schickel	Aliquot Sequences	21	2011-12-16 07:27
ECM work on 4788:2632 c166	schickel	Aliquot Sequences	8	2011-03-05 01:11
ECM work on 4788:i2617 c172	schickel	Aliquot Sequences	42	2011-01-23 08:35
Team sieve #20: c170 from 4788:2549	schickel	Aliquot Sequences	153	2010-11-09 07:39
Factorization of 2,689+.c170: p59 * p112	dleclair	NFSNET Discussion	10	2004-12-30 01:46

2010-09-14, 07:54	#2
debrouxl Sep 2009 977 Posts	Good lord, how much RAM did these curves at B1=4.2e9 require ?

2010-09-15, 08:24	#8
schickel "Frank <^>" Dec 2004 CDP Janesville 2·1,061 Posts	4 more curves (at a slightly lower B2, 950 curves needed for t55). Would have had one more, but had an unexplained shutdown on the last curve....

2010-09-16, 05:22	#10
Andi47 Oct 2004 Austria 2×17×73 Posts	462 curves at B1=26e7, B2=2716922571100 Together with EdH's 106 curves at 11e7 reported here, we have reached a total of 51.6% of t55.