mersenneforum.org  

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

Closed Thread
 
Thread Tools
Old 2015-09-11, 01:49   #243
UBR47K
 
UBR47K's Avatar
 
Aug 2015

22·17 Posts
Default

Requesting a triple check to a CUDALucas DC test
http://www.mersenne.org/report_expon...6131497&full=1
UBR47K is offline  
Old 2015-09-11, 01:49   #244
Mark Rose
 
Mark Rose's Avatar
 
"/X\(‘-‘)/X\"
Jan 2013

B7216 Posts
Default

Quote:
Originally Posted by UBR47K View Post
Requesting a triple check to a CUDALucas DC test
http://www.mersenne.org/report_expon...6131497&full=1
I'll do it.
Mark Rose is offline  
Old 2015-09-13, 17:52   #245
Madpoo
Serpentine Vermin Jar
 
Madpoo's Avatar
 
Jul 2014

3,313 Posts
Default slightly new approach

I've been toying with a little different way of trying to find bad systems where we don't quite know they're bad yet.

Previously I've looked at the history of their bad versus good results, but I realized we also have data on how many times each CPU has been double-checked without a match. In cases where the other system has a good track record, I can make an assumption, but most of the time the other CPU is also an unknown quantity... or at least not "awesome" enough to make a bet one way or the other.

But if there's a system with maybe at least as many bad as good, but also has a lot of mismatches under it's belt, I think I can count that as another indicator of badness.

I did a test the past week of taking some of those exponents where maybe the system has an equal amount of bad and good, or even examples where the bad and good are zero but those mismatches are high, and had pretty good success at showing that the first result was bad.

With all of that said, here's a list of exponents below 58M that meet this criteria, based on the system that checked it:
  • At least twice as many bad as good
  • Mismatches greater-equal than 4

You'll note a couple new columns... "Mis" which is the # of times an exponent by that same CPU has been double-checked without a match, and "Solo" which is the count of exponents done by that system that have only had a single check.

In theory, the # of "solos" and "mismatches" should total the same as the "unknown" and "suspect". "Suspect" results are almost always re-assigned out pretty quick unless it was a larger exponent, by the way.

Questions or comments? No? Good...here's the list: I took the ones in the 48M and 52M ranges for myself just for fun, if anyone wonders why there are gaps there... I had previously done the 54M and up.
Code:
exponent	Bad	Good	Unk	Sus	Solo	Mis	worktodo
35987977	23	11	3	8	2	9	DoubleCheck=35987977,71,1
37452329	13	5	12	1	8	5	DoubleCheck=37452329,71,1
37713343	13	6	2	7	1	8	DoubleCheck=37713343,71,1
42014249	13	5	12	1	8	5	DoubleCheck=42014249,72,1
42230479	10	4	5	9	5	9	DoubleCheck=42230479,72,1
42321029	2	0	7	2	5	4	DoubleCheck=42321029,72,1
42395803	10	4	5	9	5	9	DoubleCheck=42395803,72,1
42853367	2	1	5	6	4	7	DoubleCheck=42853367,72,1
42929083	13	5	12	1	8	5	DoubleCheck=42929083,72,1
43558883	22	9	16	21	16	21	DoubleCheck=43558883,72,1
43566329	3	0	6	6	6	6	DoubleCheck=43566329,72,1
43993823	22	9	16	21	16	21	DoubleCheck=43993823,72,1
44861221	3	0	6	6	6	6	DoubleCheck=44861221,72,1
45418559	3	0	6	6	6	6	DoubleCheck=45418559,72,1
45628951	22	9	16	21	16	21	DoubleCheck=45628951,72,1
45912649	3	0	6	6	6	6	DoubleCheck=45912649,72,1
46367329	3	0	6	6	6	6	DoubleCheck=46367329,72,1
46555081	22	9	16	21	16	21	DoubleCheck=46555081,72,1
46561043	22	9	16	21	16	21	DoubleCheck=46561043,72,1
47453207	22	9	16	21	16	21	DoubleCheck=47453207,72,1
47955643	3	0	6	6	6	6	DoubleCheck=47955643,72,1
47963197	18	9	7	10	7	10	DoubleCheck=47963197,72,1
49052473	18	9	7	10	7	10	DoubleCheck=49052473,72,1
49088243	18	9	7	10	7	10	DoubleCheck=49088243,72,1
49440151	18	9	7	10	7	10	DoubleCheck=49440151,72,1
50312729	22	9	16	21	16	21	DoubleCheck=50312729,73,1
50399549	22	9	16	21	16	21	DoubleCheck=50399549,73,1
50401231	22	9	16	21	16	21	DoubleCheck=50401231,73,1
53004233	22	9	16	21	16	21	DoubleCheck=53004233,73,1
53418059	22	9	16	21	16	21	DoubleCheck=53418059,73,1
53418073	22	9	16	21	16	21	DoubleCheck=53418073,73,1
53816729	22	9	16	21	16	21	DoubleCheck=53816729,73,1
53829631	22	9	16	21	16	21	DoubleCheck=53829631,73,1
Madpoo is offline  
Old 2015-09-15, 16:02   #246
Madpoo
Serpentine Vermin Jar
 
Madpoo's Avatar
 
Jul 2014

3,313 Posts
Default Quad check needed if anyone is interested

Just had another case of one of my triple checks not matching either of the previous two.

M34938083

If anyone felt like doing a quad check, there it be.
Madpoo is offline  
Old 2015-09-15, 16:51   #247
Prime95
P90 years forever!
 
Prime95's Avatar
 
Aug 2002
Yeehaw, FL

19·397 Posts
Default

Quote:
Originally Posted by Madpoo View Post
Questions or comments? No? Good...here's the list: I took the ones in the 48M and 52M ranges
I took the ones below 43M
Prime95 is offline  
Old 2015-09-15, 17:21   #248
ric
 
ric's Avatar
 
Jul 2004
Milan, Ita

2·7·13 Posts
Default

Quote:
Originally Posted by Madpoo View Post
If anyone felt like doing a quad check, there it be.
Enqueued.
ric is offline  
Old 2015-09-15, 17:24   #249
Prime95
P90 years forever!
 
Prime95's Avatar
 
Aug 2002
Yeehaw, FL

754310 Posts
Default

Quote:
Originally Posted by Prime95 View Post
I took the ones below 43M
And now I took all those below 47M for another computer
Prime95 is offline  
Old 2015-09-16, 06:44   #250
frmky
 
frmky's Avatar
 
Jul 2003
So Cal

11×193 Posts
Default

Quote:
Originally Posted by Prime95 View Post
And now I took all those below 47M for another computer
And I took the rest below 50M.
frmky is online now  
Old 2015-09-16, 17:14   #251
Madpoo
Serpentine Vermin Jar
 
Madpoo's Avatar
 
Jul 2014

3,313 Posts
Default

Quote:
Originally Posted by frmky View Post
And I took the rest below 50M.
I'll snag the ones 50M and up... I have a few worktodo files running on empty.
Madpoo is offline  
Old 2015-09-17, 05:59   #252
UBR47K
 
UBR47K's Avatar
 
Aug 2015

22×17 Posts
Default

Requesting a Double (?Quad?) check on a CUDALucas DC result:
http://www.mersenne.org/report_expon...6125533&full=1
UBR47K is offline  
Old 2015-09-17, 06:44   #253
sdbardwick
 
sdbardwick's Avatar
 
Aug 2002
North San Diego County

5·137 Posts
Default

Quote:
Originally Posted by UBR47K View Post
Requesting a Double (?Quad?) check on a CUDALucas DC result:
http://www.mersenne.org/report_expon...6125533&full=1
I'll do it. ETA <15 hours.
sdbardwick is offline  
Closed Thread



Similar Threads
Thread Thread Starter Forum Replies Last Post
Double-Double Arithmetic Mysticial Software 52 2021-04-23 06:51
Clicking an exponent leads to 404 page marigonzes Information & Answers 2 2017-02-14 16:56
x.265 half the size, double the computation; so if you double again? 1/4th? jasong jasong 7 2015-08-17 10:56
What about double-checking TF/P-1? 137ben PrimeNet 6 2012-03-13 04:01
Double the area, Double the volume. Uncwilly Puzzles 8 2006-07-03 16:02

All times are UTC. The time now is 10:00.


Fri Aug 6 10:00:29 UTC 2021 up 14 days, 4:29, 1 user, load averages: 4.21, 4.41, 4.19

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.