mersenneforum.org  

Go Back   mersenneforum.org > Great Internet Mersenne Prime Search > PrimeNet

Reply
 
Thread Tools
Old 2015-12-18, 00:19   #1
dbaugh
 
dbaugh's Avatar
 
Aug 2005

11101102 Posts
Default ECM report

Where does the data for the ECM report come from? It does not match exponent status for say, just to pick a random exponent, 9007753. I do not see the 714 Bound #1 equals 3,000,000 curves listed in the ECM report in the exponent status history for 9007753.
dbaugh is offline   Reply With Quote
Old 2015-12-18, 01:06   #2
VBCurtis
 
VBCurtis's Avatar
 
"Curtis"
Feb 2005
Riverside, CA

478110 Posts
Default

The ECM individual report lists only curves reported via Prime95. Manually-reported curves are manually added to the curve totals, but not to the individual reports.
VBCurtis is online now   Reply With Quote
Old 2015-12-18, 02:06   #3
Gordon
 
Gordon's Avatar
 
Nov 2008

3·167 Posts
Default

Quote:
Originally Posted by VBCurtis View Post
The ECM individual report lists only curves reported via Prime95. Manually-reported curves are manually added to the curve totals, but not to the individual reports.
That's not actually true.
Gordon is offline   Reply With Quote
Old 2015-12-18, 03:16   #4
VBCurtis
 
VBCurtis's Avatar
 
"Curtis"
Feb 2005
Riverside, CA

478110 Posts
Default

What does that link demonstrate? Please explain.
If your most-recent submissions on that link were manually submitted, change my "are" to "were", as for quite a few years the manual submissions were not [consistently] posted. It's still the reason the individual-report counts don't match the total work done in many cases.
VBCurtis is online now   Reply With Quote
Old 2015-12-18, 03:49   #5
dbaugh
 
dbaugh's Avatar
 
Aug 2005

1668 Posts
Default

So, who did the 714 Bound #1 = 3,000,000 curves? I must have a secret helper.
dbaugh is offline   Reply With Quote
Old 2015-12-18, 06:34   #6
VBCurtis
 
VBCurtis's Avatar
 
"Curtis"
Feb 2005
Riverside, CA

10010101011012 Posts
Default

Ah, I see. Curves run at various bounds can be converted- the site uses a primitive conversion that 3 curves at 1M is "worth" 1 curve at 3M. It's not very accurate, but that's how it works. A better conversion is 1500 curves at 1M is about the same fraction of a T40 as 250-300 curves at 3M, but that's not how it works on the site.

If you bring up a list of ECM work, you'll see the lowest # of 3M curves as 527 or so; this is the number of curves at 3M equivalent to a T35. That is, the moment someone completes enough work for 1580 curves at 1M, the column changes to "complete" and the 3M column changes to 527 (or 526?). This reflects the chances that curves run at 1M find a factor in the 40-digit range.

So, your own curves at 1M were your secret helper.
VBCurtis is online now   Reply With Quote
Old 2015-12-18, 07:09   #7
dbaugh
 
dbaugh's Avatar
 
Aug 2005

2·59 Posts
Default

Thank you. That is an extremely helpful answer. So If someone did 17,100 44M curves would 40D, 45D and 50D show done with a head start on 55D? I would not actually do the 44M curves before finishing the 3M curves except for benchmark timing purposes. I am just asking so as to test my new understanding of the ECM report.
dbaugh is offline   Reply With Quote
Old 2015-12-18, 19:21   #8
VBCurtis
 
VBCurtis's Avatar
 
"Curtis"
Feb 2005
Riverside, CA

7·683 Posts
Default

Yes, definitely. Larger-bound curves will find the smaller factors, just not as efficiently (in terms of time spent to find, say, a 40-digit factor) as the "optimal" bounds.

Unfortunately, smaller-bound curves have a vanishingly small chance to find large factors (say, sixty curves at 50k have almost-no chance to find a 40+ digit factor), but the B1 conversion is done linearly for simplicity; that is, 60 curves at 50k is "worth" 3M B1-points, even though those 60 curves have no meaningful contribution to a T40.
VBCurtis is online now   Reply With Quote
Old 2015-12-19, 02:58   #9
Gordon
 
Gordon's Avatar
 
Nov 2008

3×167 Posts
Default

Quote:
Originally Posted by VBCurtis View Post
What does that link demonstrate? Please explain.
If your most-recent submissions on that link were manually submitted, change my "are" to "were", as for quite a few years the manual submissions were not [consistently] posted. It's still the reason the individual-report counts don't match the total work done in many cases.
Just that your assertion that you can't get the details updated with purely manual results wasn't correct :-)

Like this M29287
Gordon is offline   Reply With Quote
Old 2015-12-19, 04:58   #10
VBCurtis
 
VBCurtis's Avatar
 
"Curtis"
Feb 2005
Riverside, CA

7×683 Posts
Default

Cool! This encourages me to do a bit more work with GMP-ECM on the smallest composites-with-no-known-factor.
VBCurtis is online now   Reply With Quote
Old 2015-12-19, 09:07   #11
VictordeHolland
 
VictordeHolland's Avatar
 
"Victor de Hollander"
Aug 2011
the Netherlands

49816 Posts
Default

Quote:
Originally Posted by VBCurtis View Post
Cool! This encourages me to do a bit more work with GMP-ECM on the smallest composites-with-no-known-factor.
I've done GMP-ECM curves on M1277 and George has imported them to PrimeNet and they also show up in the ECM history.

There was a post where George quoted the conversion formula used by Primenet to convert the larger GMP-ECM Stage 2 bounds to the bounds used by Primenet. I can't find it at the moment, as I'm typing from a phone.
VictordeHolland is offline   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
Report for PRP houding PrimeNet 15 2018-01-10 19:24
How to report GPU results dcorbit Information & Answers 4 2012-08-13 21:53
report not updating ??? nobody Information & Answers 3 2012-04-26 14:31
How often do you report ? Axel Fox Lone Mersenne Hunters 19 2003-10-19 07:40
Bug report remaker Software 4 2002-10-17 08:18

All times are UTC. The time now is 16:16.

Thu May 13 16:16:12 UTC 2021 up 35 days, 10:57, 2 users, load averages: 2.77, 2.92, 2.76

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.