mersenneforum.org  

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

Reply
 
Thread Tools
Old 2016-06-13, 17:25   #1
Mark Rose
 
Mark Rose's Avatar
 
"/X\(‘-‘)/X\"
Jan 2013

292910 Posts
Default What percentage of CPUs/GPUs have done a double check?

Now that we've had the double-check-once-per-year setting in place for a while, I'm curious how that has affected the number of machines that have done at least one double check. Are we finding a lot more bad machines?
Mark Rose is offline   Reply With Quote
Old 2016-06-16, 22:44   #2
kladner
 
kladner's Avatar
 
"Kieren"
Jul 2011
In My Own Galaxy!

22·2,539 Posts
Default

Quote:
Originally Posted by Mark Rose View Post
Now that we've had the double-check-once-per-year setting in place for a while, I'm curious how that has affected the number of machines that have done at least one double check. Are we finding a lot more bad machines?
GPU double checks led me to abandon CUDALucas. I could probably have gotten it to work if I had slowed everything some more. However, now, with only on card running, I am reluctant to stop TF, which seems very reliable.
My CPU does almost all DCs, after a brief flirtation with P-1 on one worker. chalsall's worries about having enough TF'd assignments for habitual P-1 people sent me back to all DC.

Last fiddled with by kladner on 2016-06-16 at 22:44 Reason: ,
kladner is offline   Reply With Quote
Old 2016-06-17, 05:27   #3
Madpoo
Serpentine Vermin Jar
 
Madpoo's Avatar
 
Jul 2014

2×5×7×47 Posts
Default

Quote:
Originally Posted by Mark Rose View Post
Now that we've had the double-check-once-per-year setting in place for a while, I'm curious how that has affected the number of machines that have done at least one double check. Are we finding a lot more bad machines?
I should have done a check earlier, but as of right now there are 19,569 machines that have zero good, zero bad results.

I don't know where we were at earlier this year... any change may be from first-time check only machines doing a DC and showing they were reliable (or maybe not).

I also just gave AirSquirrels a batch of exponents to test from systems with no good/no bad, just so we can get something on the board for them.

Many of these systems haven't checked anything in recently, they only have a handful of results from years ago, so people double-checking their stuff is the only way we're going to know.

AirSquirrels and I have also been going through the backlog of exponents that need a triple-check... I think there's something like 5000 total, of which 2000 or so are currently unassigned.

Sometimes I feel like that's a "two steps forward, one step back" because just when I think I'm gaining ground, a bunch of those assigned exponents will expire. But then again I think when I started on this it was somewhere around 3000 unassigned, so it's definitely chugging along. Right now, just about everything below 40M is in progress, and between 40-48M is mostly being done but it's building up again from those expired assignments.

Someone else will have to do a bunch of triple-checks that I can't, because I turned in one of the results, but at least in those cases I'm just assuming mine was right.
Madpoo is offline   Reply With Quote
Old 2016-06-17, 13:57   #4
Mark Rose
 
Mark Rose's Avatar
 
"/X\(‘-‘)/X\"
Jan 2013

29·101 Posts
Default

All the way up to 48M? Nice!

And less than 20k machines shouldn't take more than a year to validate, especially if a lot of those machines self-validate when they first get a DC.
Mark Rose is offline   Reply With Quote
Old 2016-06-17, 14:38   #5
petrw1
1976 Toyota Corona years forever!
 
petrw1's Avatar
 
"Wayne"
Nov 2006
Saskatchewan, Canada

5·911 Posts
Default

These are the year- to-date user counts today.
From here; erasing the personal numbers and only reporting the count in the "of" column.
http://www.mersenne.org/account/?details=1

Code:
Type    of
TF	1,794
P-1	630	
LL	2,553	
LL-D	4,457	
ECM	846	
ECMF	105
Overall	6,068
I don't have the post handy but I do recall a few years ago noting that P-1, LL and DC were all almost exactly 2,000.

Today DC is well over double that.
petrw1 is offline   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
How should we compare GPUs to CPUs? chalsall GPU to 72 56 2012-07-15 19:24
Double Check Unregistered Information & Answers 3 2011-10-01 04:38
First check and double check llrnet servers. opyrt Prime Sierpinski Project 3 2009-01-02 01:50
Double-check check? M0CZY Software 15 2008-10-30 14:20
Double precision GPUs coming very soon dsouza123 Hardware 4 2007-10-15 02:20

All times are UTC. The time now is 01:18.

Mon Mar 8 01:18:05 UTC 2021 up 94 days, 21:29, 1 user, load averages: 2.03, 2.11, 2.18

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.