mersenneforum.org  

Go Back   mersenneforum.org > New To GIMPS? Start Here! > Information & Answers

Reply
 
Thread Tools
Old 2015-07-09, 19:02   #1
Judge Hale
 
"Kenneth Hale"
May 2015
Fort Worth, Texas

2·3 Posts
Question Trial Factor Assignment Time Limits

This is my first time to ever start a thread, so please forgive me if I mess up. I was just wondering why there is a 6 month time allowance on manual trial factor assignments. It seems that there are several users who take thousands upon thousands of assignments and who then sit on them only to let them expire and never make a good faith effort to work them, thus severely holding up progress and preventing the many people who would agressively work them to get the assignments done. I am not complaining, but just seeking to try to understand this situation. It makes no sense to me. A one month window to work manual assignments would seem logical to me. Of course, George and Scott can make these decisions as they please and see fit and I will never question their wisdom on this. I merely seek to understand the situation. If this matter has been previously covered, I apologize, since I could not find it anywhere on the forum.
Judge Hale is offline   Reply With Quote
Old 2015-07-09, 23:07   #2
chalsall
If I May
 
chalsall's Avatar
 
"Chris Halsall"
Sep 2002
Barbados

2×4,657 Posts
Default

Quote:
Originally Posted by Judge Hale View Post
It seems that there are several users who take thousands upon thousands of assignments and who then sit on them only to let them expire and never make a good faith effort to work them, thus severely holding up progress and preventing the many people who would agressively work them to get the assignments done.
What ranges are you referring to?

So you know, the user "GPU Factoring" reserves TF'ing candidates ahead of the LL'ing and DC'ing "wave fronts" (but not via a manual assignment) for distribution to workers via GPU to 72. Anyone is welcome to join and be assigned work.

If you're talking about other ranges, I can't speak to that. But please be aware that such "outlier" work doesn't really help the project much.

Last fiddled with by chalsall on 2015-07-09 at 23:27 Reason: moved the "but not via a manual assignment" string to make more sense.
chalsall is offline   Reply With Quote
Old 2015-07-09, 23:41   #3
Judge Hale
 
"Kenneth Hale"
May 2015
Fort Worth, Texas

610 Posts
Default

I should have been more specific. I am not referring to the wave front. I understand what is going on there. I am referring to ranges well outside the wave front like 109 and 110 and others that seem to be totally stagnated. Thanks.
Judge Hale is offline   Reply With Quote
Old 2015-07-09, 23:51   #4
TheMawn
 
TheMawn's Avatar
 
May 2013
East. Always East.

11×157 Posts
Default

Welcome to the forums!

Like Chris said, work that could "hold up" the project is never reserved in such large batches of assignments. To be honest, a one month window WOULD be fine but six months hasn't really been an issue.

We've worked on and implemented a system to deal with "stuck" assignments in the LL and DC departments and they seem to be working very well. In theory we could do something similar with TF but we haven't noticed that to be an issue.

You're probably looking at exponents above 100 million which is sort of lawless territory. Any work done there will not be relevant for years.
TheMawn is offline   Reply With Quote
Old 2015-07-10, 01:01   #5
kladner
 
kladner's Avatar
 
"Kieren"
Jul 2011
In My Own Galaxy!

23·32·139 Posts
Default

Quote:
Originally Posted by TheMawn View Post
Welcome to the forums!
.....

You're probably looking at exponents above 100 million which is sort of lawless territory. Any work done there will not be relevant for years.
But it would make Uncwilly happy if the work is TF done in the 332M range.

I second the welcome. The general ethos here is that people do work that pleases them. There are many special foci around which I don't have the math to understand.

GPU72 has grown into a system which can filter assignments based on various metrics such as their immediacy and the processing steps still needed. You get Trial Factoring assignments by these metrics by selecting "What Makes Sense" or "Let GPU72 Decide." You also have choices which allow you to target a particular range for whatever reasons appeal to you.

The point of the setup is to keep up with the demands of work being done in GIMPS as a whole, such that assignments are optimally preprocessed before being handed out for Lucas-Lehmer testing.

Last fiddled with by kladner on 2015-07-10 at 01:02
kladner is offline   Reply With Quote
Old 2015-07-10, 01:29   #6
Judge Hale
 
"Kenneth Hale"
May 2015
Fort Worth, Texas

1102 Posts
Default

I understand what you guys are saying and I totally agree that each should work any exponents that please him. What I am having difficulty understanding is why anyone snaps up thousands of exponents and then does virtually no work on them at all and lets them expire only for them to then be snapped up again and not worked again. Then a year has gone by, ad nauseam. I often take over 20,000 assignments at one time, but, to my knowledge, I have never let any expire. I was just wondering why that was being allowed to happen when it could so easily be avoided. The Mawn sort of answered my question. At any rate, it does not affect me directly since I have charted out trial factoring assignments that do not appear to be in any conflict with others in their choices. It was simply a rhetorical question. I really do appreciate your responses.
Judge Hale is offline   Reply With Quote
Old 2015-07-10, 18:54   #7
chalsall
If I May
 
chalsall's Avatar
 
"Chris Halsall"
Sep 2002
Barbados

2·4,657 Posts
Default

Quote:
Originally Posted by Judge Hale View Post
What I am having difficulty understanding is why anyone snaps up thousands of exponents and then does virtually no work on them at all and lets them expire only for them to then be snapped up again and not worked again. Then a year has gone by, ad nauseam.
That is an interesting question. And I can't really answer as to the motivation.

What I /can/ say is this is an issue we still deal with in the P-1 space. Several workers manually reserve hundreds of candidates for P-1 work, and then just let them expire 180 days later with little or no work done. This is a more immediate issue for the project as we're currently having some difficulty "feeding" the P-1'ers at the appropriate bit level of 75.

As an aside, if you'd like to help the project's root goal (finding the next Mersenne Prime), perhaps consider joining GPU72 and do some TF'ing in the wave front. Even going from 71 to 72 bits in the LL range would be immediately helpful. We have pretty graphs and everything....
chalsall is offline   Reply With Quote
Old 2015-07-10, 23:21   #8
Judge Hale
 
"Kenneth Hale"
May 2015
Fort Worth, Texas

2·3 Posts
Default

I joined GPU72 yesterday. I believe I have set up the work fetch (through GPU72) to begin at 80M with lowest exponent, to start when my current batch of exponents runs out. I'm running some decent equipment here, so, hopefully, it will make as much a difference to the project as it has to my electrical bill. I just hate that doing this will cause a very marked drop in my throughput. Oh, well, I can see where it is much better for project goals. I will always be of the opinion that people should not reserve exponents that are going to be allowed to expire, no matter where the exponent lies.
Judge Hale is offline   Reply With Quote
Old 2015-07-11, 00:44   #9
chalsall
If I May
 
chalsall's Avatar
 
"Chris Halsall"
Sep 2002
Barbados

2×4,657 Posts
Default

Quote:
Originally Posted by Judge Hale View Post
I joined GPU72 yesterday. I believe I have set up the work fetch (through GPU72) to begin at 80M with lowest exponent, to start when my current batch of exponents runs out.
Ummm... GPU72 doesn't hold any candidates above 79M (inclusive). Except, of course, for the 332.2M range.

Again, this little side project is trying to help with the overall goal of GIMPS: finding the next Mersenne Prime.

Last fiddled with by chalsall on 2015-07-11 at 00:49 Reason: Edit: I often forget about 332.2....
chalsall is offline   Reply With Quote
Old 2015-07-11, 01:09   #10
Judge Hale
 
"Kenneth Hale"
May 2015
Fort Worth, Texas

2×3 Posts
Default

Yeah, Chris, I finally figured that out after it told me no work several times. I adjusted it to go back to 79M about 5 minutes ago after reading your reply. We'll see how that works. Once I can get the dodo out of me, I'll figure it all out. I'll try factoring to 75 for awhile to see what happens. Thanks for the info. Oh, by the way, I'm "Ken".
Judge Hale is offline   Reply With Quote
Old 2015-07-11, 05:20   #11
petrw1
1976 Toyota Corona years forever!
 
petrw1's Avatar
 
"Wayne"
Nov 2006
Saskatchewan, Canada

10001001000112 Posts
Default

Quote:
Originally Posted by Judge Hale View Post
Yeah, Chris, I finally figured that out after it told me no work several times. I adjusted it to go back to 79M about 5 minutes ago after reading your reply. We'll see how that works. Once I can get the dodo out of me, I'll figure it all out. I'll try factoring to 75 for awhile to see what happens. Thanks for the info. Oh, by the way, I'm "Ken".
I think you need to pick under 79M...that is 78M or less....
petrw1 is online now   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
Trial Factoring by GPU my LL assignment Luis PrimeNet 3 2016-06-28 18:41
Trial Factor Bit Depth lavalamp Operation Billion Digits 8 2010-08-02 18:49
trial division over a factor base Peter Hackman Factoring 7 2009-10-26 18:27
new options - factor to low limits, etc Jud McCranie Information & Answers 11 2008-12-01 18:27
Shortest time to complete a 2^67 trial factor (no factor) dsouza123 Software 12 2003-08-21 18:38

All times are UTC. The time now is 18:06.

Mon Sep 28 18:06:35 UTC 2020 up 18 days, 15:17, 1 user, load averages: 2.13, 1.93, 1.94

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2020, 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.