mersenneforum.org > Data Strategic Double Clicking
 User Name Remember Me? Password
 Register FAQ Search Today's Posts Mark Forums Read

2016-09-23, 01:02   #1222
Mark Rose

"/X\(‘-‘)/X\"
Jan 2013

37·79 Posts

Quote:
 Originally Posted by bgbeuning I would like to know if any of my machines are producing bad results. Is there a way to find out?
Go to http://www.mersenne.org/cpus/ and look at the reliability rating. A CPU with no verified results will get a reliability rating of 0.98, as that's the average. If you want to test your CPUs, configure your machines to run some double checks.

2016-09-23, 01:33   #1223
GP2

Sep 2003

5×11×47 Posts

Quote:
 Originally Posted by bgbeuning I would like to know if any of my machines are producing bad results. Is there a way to find out?
Your only verified bad result in the database is M39761383. The machine was "opt24".

You can log into mersenne.org and look at http://www.mersenne.org/results/ to see your own results. Any mismatches are prominently flagged; if so, obviously it could be your result that's bad, or the other result, or both. To see them all you might need to increase the default of "100" in the "Show up to _____ results" field.

2016-09-23, 03:50   #1224

"Kieren"
Jul 2011
In My Own Galaxy!

27AE16 Posts

Quote:
 Originally Posted by bgbeuning I would like to know if any of my machines are producing bad results. Is there a way to find out?
Run double checks. Add up the preponderance of results. Some of this may require triple checks to resolve. If you get verified results (matching someone else's results), and no proven bad results (two others match, but not you), you are in pretty good shape.

If some machine not only mismatches, but is proven wrong, it is time to start testing and adjusting.
[Cross posted with GP2. That information is informed, while my response is entirely general.]

Last fiddled with by kladner on 2016-09-23 at 03:53

2016-09-23, 05:49   #1225
dragonbud20

Mar 2014

5016 Posts

I'll do these it's all the triples below 44M from that list

Quote:
 Originally Posted by Madpoo Code: exponent worktodo 42905069 DoubleCheck=42905069,72,1 42917263 DoubleCheck=42917263,72,1 42929317 DoubleCheck=42929317,72,1 42958469 DoubleCheck=42958469,72,1 42966571 DoubleCheck=42966571,72,1 42985493 DoubleCheck=42985493,72,1 42994879 DoubleCheck=42994879,72,1 43000721 DoubleCheck=43000721,72,1 43021193 DoubleCheck=43021193,72,1 43029247 DoubleCheck=43029247,72,1 43029307 DoubleCheck=43029307,72,1 43030081 DoubleCheck=43030081,72,1 43030921 DoubleCheck=43030921,72,1 43045459 DoubleCheck=43045459,72,1 43065563 DoubleCheck=43065563,72,1 43110401 DoubleCheck=43110401,72,1 43273631 DoubleCheck=43273631,72,1 43287953 DoubleCheck=43287953,72,1 43292497 DoubleCheck=43292497,72,1 43297061 DoubleCheck=43297061,72,1 43312639 DoubleCheck=43312639,72,1 43384801 DoubleCheck=43384801,72,1 43414927 DoubleCheck=43414927,72,1 43423663 DoubleCheck=43423663,72,1 43727513 DoubleCheck=43727513,72,1 43775617 DoubleCheck=43775617,72,1 43803289 DoubleCheck=43803289,72,1 43832177 DoubleCheck=43832177,72,1 43906003 DoubleCheck=43906003,72,1 43963379 DoubleCheck=43963379,72,1

2016-09-23, 18:37   #1226
bgbeuning

Dec 2014

111111112 Posts

Quote:
 Originally Posted by Mark Rose Go to http://www.mersenne.org/cpus/ and look at the reliability rating. A CPU with no verified results will get a reliability rating of 0.98, as that's the average.
Do not see reliability rating.
Attached Thumbnails

Last fiddled with by bgbeuning on 2016-09-23 at 19:03

2016-09-23, 19:30   #1227
sdbardwick

Aug 2002
North San Diego County

2B116 Posts

Quote:
 Originally Posted by bgbeuning Do not see reliability rating.
Click on a computer name for computer properties.

2016-09-25, 06:07   #1228
Serpentine Vermin Jar

Jul 2014

3,313 Posts

Quote:
 Originally Posted by GP2 Your only verified bad result in the database is M39761383. The machine was "opt24". You can log into mersenne.org and look at http://www.mersenne.org/results/ to see your own results. Any mismatches are prominently flagged; if so, obviously it could be your result that's bad, or the other result, or both. To see them all you might need to increase the default of "100" in the "Show up to _____ results" field.
That's the only one I saw also (and was delighted to see that my result was the tie-breaker in that 3-way race, but then I tried to snag all the quad checks that I could).

I see a total of 32 different systems from him that had returned an LL result at some point or another.

That particular system with the bad result has 160 good (one is "assumed" good based on the other system being sketchier), and 24 still unknown.

2016-09-25, 16:13   #1229
bgbeuning

Dec 2014

3·5·17 Posts

Quote:
 Originally Posted by GP2 Your only verified bad result in the database is M39761383. The machine was "opt24".
I checked the results.txt for that exponent but there were no special case messages.
Some exponents get this message

Quote:
 Trying 1000 iterations for exponent 37059551 using 1920K FFT. If average roundoff error is above 0.24245, then a larger FFT will be used. Final average roundoff error is 0.22443, using 1920K FFT for exponent 37059551.
Thanks everyone for the tips on checking machine stability.

2016-09-26, 00:07   #1230
GP2

Sep 2003

A1916 Posts

Quote:
 Originally Posted by bgbeuning I checked the results.txt for that exponent but there were no special case messages.
That's because it was not known at the time that the result was wrong.

If you look at the report for M39761383, you can see that you checked in your result on 2016-01-21, and it was not until Madpoo checked in his quadruple-check result on 2016-02-22 that it was known that your result was wrong.

However, if you log in to mersenne.org and look at http://www.mersenne.org/results/ (not the results.txt file on your disk), then it should flag any mismatches or bad results. Again, to see all your old results you might need to increase the default of "100" in the "Show up to _____ results" field.

 2016-09-26, 18:30 #1231 Mark Rose     "/X\(‘-‘)/X\" Jan 2013 55538 Posts Everything below 50M from the triple check list has been assigned.
2016-09-27, 20:52   #1232
endless mike

Jan 2004
Milwaukee, WI

22×5×7 Posts

It looks like these all finished within the last few days while I was at work. I matched you on all of them.

Quote:
 Originally Posted by endless mike I grabbed the rest of the list. You (among others) have helped whittle down my list of exponents needing a triple check, so I can help with some of yours.
Quote:
 Originally Posted by Mark Rose I have 25 exponents that need a triple check, if anyone is interested: DoubleCheck=61756243,74,1 DoubleCheck=61841803,74,1 DoubleCheck=61843973,74,1 DoubleCheck=61844329,74,1 DoubleCheck=61846193,74,1 DoubleCheck=61995611,74,1 DoubleCheck=62012707,74,1 DoubleCheck=62796527,74,1 DoubleCheck=63773497,74,1 DoubleCheck=64071221,74,1 DoubleCheck=67876027,74,1 DoubleCheck=68804809,74,1 DoubleCheck=70869973,75,1 DoubleCheck=71069681,75,1 DoubleCheck=71642387,75,1 DoubleCheck=76417333,75,1 DoubleCheck=77133689,75,1

 Similar Threads Thread Thread Starter Forum Replies Last Post Mysticial Software 52 2021-04-23 06:51 marigonzes Information & Answers 2 2017-02-14 16:56 jasong jasong 7 2015-08-17 10:56 137ben PrimeNet 6 2012-03-13 04:01 Uncwilly Puzzles 8 2006-07-03 16:02

All times are UTC. The time now is 14:44.

Fri Oct 22 14:44:53 UTC 2021 up 91 days, 9:13, 1 user, load averages: 1.10, 1.13, 1.21