mersenneforum.org  

Go Back   mersenneforum.org > Great Internet Mersenne Prime Search > Data > Marin's Mersenne-aries

Reply
 
Thread Tools
Old 2020-09-22, 22:58   #12
moebius
 
moebius's Avatar
 
Jul 2009
Germany

22·5·17 Posts
Default

One thing makes little sense to me. Why does the fastest processor get the smallest exponents and not the other way around?
moebius is online now   Reply With Quote
Old 2020-09-22, 23:06   #13
mathwiz
 
Mar 2019

2·32·5 Posts
Default

Quote:
Originally Posted by moebius View Post
The difference is a human would choose the DC curtisc vs. anonymous, because he knows that anonymous often has wrong results and curtisc sometimes. Means a better chance to find a prime.
That may or may not be the case. But whatever rules might be "optimal" regarding which numbers should be double- or triple-checked first, could still be encoded into the server.
mathwiz is offline   Reply With Quote
Old 2020-09-22, 23:20   #14
moebius
 
moebius's Avatar
 
Jul 2009
Germany

22·5·17 Posts
Default

Quote:
Originally Posted by mathwiz View Post
That may or may not be the case. But whatever rules might be-..
Yes, of course, but that's probably not being done at the moment.

To my another statement. I realize that the creators of the project are entitled to create a slightly better chance of finding a prime number for themselves after 25 years of computing. Even though individual people now have twice the computing power as the rest of the users. Nevertheless, I think that overall a slightly higher throughput could be achieved if the assignment rules were different.
moebius is online now   Reply With Quote
Old 2020-09-23, 01:04   #15
Mark Rose
 
Mark Rose's Avatar
 
"/X\(‘-‘)/X\"
Jan 2013

2×1,429 Posts
Default

Quote:
Originally Posted by moebius View Post
Yes, of course, but that's probably not being done at the moment.

To my another statement. I realize that the creators of the project are entitled to create a slightly better chance of finding a prime number for themselves after 25 years of computing. Even though individual people now have twice the computing power as the rest of the users. Nevertheless, I think that overall a slightly higher throughput could be achieved if the assignment rules were different.
It's not for that reason. Before these rules were introduced, the smaller exponents assigned to slow computers (either from speed or not being powered on) were often poached by other users who wanted to reach milestones: thus wasting the effort of those slow computers. It's better to have the slow computers finish assignment before they're in danger of being poached.
Mark Rose is offline   Reply With Quote
Old 2020-09-23, 14:30   #16
Uncwilly
6809 > 6502
 
Uncwilly's Avatar
 
"""""""""""""""""""
Aug 2003
101×103 Posts

3·19·151 Posts
Default

Quote:
Originally Posted by ATH View Post
New TC: 53571863
I have it registered and it is in the front of the line.
Uncwilly is offline   Reply With Quote
Old 2020-09-23, 14:56   #17
Uncwilly
6809 > 6502
 
Uncwilly's Avatar
 
"""""""""""""""""""
Aug 2003
101×103 Posts

3·19·151 Posts
Default

Quote:
Originally Posted by mathwiz View Post
Why are we managing these checks manually? Why are any humans involved in manually doling out and coordinating assignments?
When we see a machine that may be diverging from reality or someone has has a set-up that they are tuning to get the max, it maybe a good idea for a fast turn around. Yes, the server will eventually assign them. But, dealing with these troubled results before they become the last in queue is helpful. Occasionally they need a QC. People like doing different types of work.

slaur/nomead likes to find exponents with errors and do the TC on them.
Quote:
Originally Posted by moebius View Post
One thing makes little sense to me. Why does the fastest processor get the smallest exponents and not the other way around?
It is not always the fastest, rather those that can produce at a suitable and reliable rate. Also to get the Cat 0 assignments, one needs to agree to take them. This is only the clean-up work. In an ideal world, there should be no Cat 0 or even Cat 1 being assigned. Cat 2 and 3's should be completing and none expiring. Only when what were assigned as a higher Cat expires are Cat 1 and 0 assignments available. So, to pick up the pieces and sweep this tail end up a few machines are volunteered to do this. It is the users choice.
Quote:
Originally Posted by moebius View Post
To my another statement. I realize that the creators of the project are entitled to create a slightly better chance of finding a prime number for themselves after 25 years of computing. Even though individual people now have twice the computing power as the rest of the users. I think that overall a slightly higher throughput could be achieved if the assignment rules were different.
George does not self serve any special exponents to himself. If you look at the top producers list, Ryan is doing ECM, Ben is spending money on cloud servers to run this, the next 3 are mainly TF which does not give a chance of finding a prime, Curtis has found primes. No one person is more likely to be assigned the next prime (except by total number of assignments).

There are threads about assignment rules elsewhere. You can make suggestions there.
Uncwilly is offline   Reply With Quote
Old 2020-09-23, 15:38   #18
kruoli
 
kruoli's Avatar
 
"Oliver"
Sep 2017
Porta Westfalica, DE

2·151 Posts
Default

Quote:
Originally Posted by Uncwilly View Post
George does not self serve any special exponents to himself.
That's mostly true. He self-assigns CERTs, though, if the first run was erroneous.
kruoli is offline   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
Strategic Double Clicking Madpoo Marin's Mersenne-aries 1841 2019-07-16 03:30
PRP triple checks? GP2 Marin's Mersenne-aries 4 2019-03-16 15:31
How often are triple checks released? patrik Data 5 2004-01-31 22:19
How do I get rid of the Triple Checks?? outlnder Lounge 4 2003-04-07 18:06
Triple Checks- How are they disposed of?? outlnder Lounge 6 2003-01-24 22:01

All times are UTC. The time now is 08:26.

Thu Sep 24 08:26:18 UTC 2020 up 14 days, 5:37, 0 users, load averages: 1.12, 1.20, 1.31

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.