Register FAQ Search Today's Posts Mark Forums Read

 2021-12-22, 18:42 #991 ThomRuley     May 2003 3×97 Posts NFS on 2993654893573563190431934488554477666944329734311027266211 4 from the t2200 file Doing ECM to 1000 curves on composites in the t2200 file, starting from the bottom. Already found one factor. Last fiddled with by ThomRuley on 2021-12-22 at 18:43
2021-12-23, 07:56   #992
lavalamp

Oct 2007
Manchester, UK

2×5×137 Posts

What size factors are you looking for with ECM? I already did some work on this file:
Quote:
 Originally Posted by lavalamp Finally finished my ECM work on the t2200 file! Composites larger than 2^1018 have had 224 curves @ B1=50e3, B2=13.7e6 (25 digit level). Composites less than 2^1018 have had 1152 curves @ B1=3e6, B2=14e9 (40 digit level). The total haul is 815 factors, these have been submitted to factordb, and attached to this post.
Obviously the larger candidates could use some more love, but the smaller ones that could run stage 1 on a GPU have had more work done.

2021-12-23, 11:05   #993
henryzz
Just call me Henry

"David"
Sep 2007
Liverpool (GMT/BST)

25×11×17 Posts

Quote:
 Originally Posted by lavalamp What size factors are you looking for with ECM? I already did some work on this file:Obviously the larger candidates could use some more love, but the smaller ones that could run stage 1 on a GPU have had more work done.

Many of these numbers will have had further ecm as part of https://mersenneforum.org/showpost.p...&postcount=958
Also gpu ecm is no longer limited to a max of 2^1018.

2021-12-23, 14:07   #994
ThomRuley

May 2003

3×97 Posts

Quote:
 Originally Posted by lavalamp What size factors are you looking for with ECM? I already did some work on this file:Obviously the larger candidates could use some more love, but the smaller ones that could run stage 1 on a GPU have had more work done.
Using the Alpertron site, it looks like the 1000 curves I was running would check for factors between 10^30 and 10^35. The composites are from Pascal Ochem's t2200 list (not the most wanted - that's a different list).

Like I had mentioned earlier, I'm doing ECM on the largest numbers as those composites are simply too large for anything else.

Are you saying it would be better to run more curves on fewer composites or would it be better to start with going to say, 10^30 on a lot of the larger composites? I'm open to suggestions, but it would definitely be a benefit if there were some way to track what has and hasn't been done.

2021-12-25, 13:20   #995
lavalamp

Oct 2007
Manchester, UK

2·5·137 Posts

Quote:
 Originally Posted by ThomRuley The composites are from Pascal Ochem's t2200 list (not the most wanted - that's a different list).
Yes, I ran ECM on the t2200 list.

Quote:
 Originally Posted by ThomRuley Are you saying it would be better to run more curves on fewer composites or would it be better to start with going to say, 10^30 on a lot of the larger composites? I'm open to suggestions, but it would definitely be a benefit if there were some way to track what has and hasn't been done.
Hey, factors are factors, if they're popping out at a good rate there's no reason to change approach. Since you're starting at the high side it'll be a while before you get down to the smaller ones, but as I said for those below 2^1018 I've already searched around the 40 digit level.

Last fiddled with by lavalamp on 2021-12-25 at 13:20

 2021-12-25, 18:29 #996 ThomRuley     May 2003 12316 Posts EDIT - duplicate Last fiddled with by ThomRuley on 2021-12-25 at 18:32
2021-12-25, 18:31   #997
ThomRuley

May 2003

3×97 Posts

Quote:
 Originally Posted by lavalamp Yes, I ran ECM on the t2200 list. Hey, factors are factors, if they're popping out at a good rate there's no reason to change approach. Since you're starting at the high side it'll be a while before you get down to the smaller ones, but as I said for those below 2^1018 I've already searched around the 40 digit level.
Good news - factors are popping out about 10% of the time so far. Sounds like what I'm doing is working. For now the plan is to NFS the smaller numbers at the top of the list while doing ECM on the larger numbers going from biggest to smallest, and checking factordb to make sure the work is still needed. It seems to be working.

2021-12-25, 18:38   #998
VBCurtis

"Curtis"
Feb 2005
Riverside, CA

10100100111112 Posts

Quote:
 Originally Posted by ThomRuley Good news - factors are popping out about 10% of the time so far. Sounds like what I'm doing is working. For now the plan is to NFS the smaller numbers at the top of the list while doing ECM on the larger numbers going from biggest to smallest, and checking factordb to make sure the work is still needed. It seems to be working.
If you're still using the Alpertron java app to do your ECM, downloading / running GMP-ECM will give you a massive speedup- like 3x faster. You'll also benefit from not having to run the small curves- you can just run a batch at B1 = 250k on each of the large numbers. Something like 300 @ 25k and 300 @ 1M would find more factors in the 28 to 34 digit range than 1000 curves on the Alpertron app, in less time.

 2021-12-25, 20:05 #999 ThomRuley     May 2003 3×97 Posts That sounds great. How do I start using gmp-ecm?
 2021-12-26, 07:14 #1000 VBCurtis     "Curtis" Feb 2005 Riverside, CA 149F16 Posts Once you get the program (I see you asked about that in another thread), open a terminal window in the folder where you have the program. ecm -h will give you the list of options you can pass to the program. The way I run it (there are many methods for handling inputs and outputs) is to make a text file with a list of a few inputs, and invoke ecm like so: ecm -c 300 25e4 results1.txt This way, the results of the curves are recorded to a file rather than printed to screen. -c tells ECM how many curves to run, and 25e4 is my sample B1 value (corresponding to curves at T30 level- you might choose to start at 1e6 for T35 instead). If a factor is found but the cofactor is composite, ECM will continue to finish out the curves trying to crack the cofactor. If cofactor is prime, ECM will proceed to the next input in the file. You might also like to run ecm -v 25e4
2022-01-09, 11:56   #1001
RichD

Sep 2008
Kansas

5·23·31 Posts

Quote:
 Originally Posted by RichD Code: File Entries Queued ==== ======= ====== t550 18 4 t600 137 6 t800 2475 10
An update on some of the t-files.
Code:
File  Entries  Queued
====  =======  ======
t550       7      0
t600     114      0
t800    2285      1
NFS@Home has a couple from the MWRB file.

 Similar Threads Thread Thread Starter Forum Replies Last Post Xyzzy GPU Computing 1 2017-05-17 20:22 Mark Rose GPU Computing 52 2016-07-02 12:11 firejuggler GPU Computing 12 2016-02-23 06:55 Elhueno Homework Help 5 2008-06-12 16:37 jchein1 Factoring 30 2005-05-30 14:43

All times are UTC. The time now is 07:17.

Fri May 20 07:17:30 UTC 2022 up 36 days, 5:18, 0 users, load averages: 1.56, 1.43, 1.41