Trippple Checks - Page 36

kladner · 2016-02-18, 22:46

DoubleCheck=38746907,71,1

I'll take this one.
registered

kladner · 2016-03-02, 02:56

Quote:

Originally Posted by kladner

DoubleCheck=38746907,71,1

I'll take this one.
registered

Third mismatch. Once again I am unsure of my own result.

This exponent must be on an FFT boundary. CuLu selected FFT 2408. This produced a round off error of ~0.3. About midway through, I found the program short cycling (many lines quickly scrolling) with an error message concerning round off, ending with the statement, "Something is wrong." I backtracked through the Save files to a previous point, and changed the FFT to 2304. This reduced the error range from around 0.3 to about 0.03. It finished without reporting an error again.

I would be delighted if someone wants to do a quad check.

NOTE: To reiterate, this was run on a GPU, so a quad will have to be run on a CPU, or abacus, or something.

Code:

LL    
Status    Date    User    Residue
Unverified    2007-07-04    C. Cooper / S. Boone    962C36A4F78247__
Unverified    2016-01-31    Mark Rose               80C6A70A7F04DA__
Unverified    2016-03-02    ktony                   3099732E406766__

sdbardwick · 2016-03-02, 05:35

Quote:

Originally Posted by kladner

Third mismatch. Once again I am unsure of my own result.

This exponent must be on an FFT boundary. CuLu selected FFT 2408. This produced a round off error of ~0.3. About midway through, I found the program short cycling (many lines quickly scrolling) with an error message concerning round off, ending with the statement, "Something is wrong." I backtracked through the Save files to a previous point, and changed the FFT to 2304. This reduced the error range from around 0.3 to about 0.03. It finished without reporting an error again.

I would be delighted if someone wants to do a quad check.

NOTE: To reiterate, this was run on a GPU, so a quad will have to be run on a CPU, or abacus, or something.

I grabbed it. ETA 18 hours or less.

kladner · 2016-03-02, 12:37

Quote:

Originally Posted by sdbardwick

I grabbed it. ETA 18 hours or less.

Many Thanks!

sdbardwick · 2016-03-02, 23:28

Quote:

Originally Posted by kladner

Code:

LL    
Status    Date    User    Residue
Unverified    2007-07-04    C. Cooper / S. Boone    962C36A4F78247__
Unverified    2016-01-31    Mark Rose               80C6A70A7F04DA__
Unverified    2016-03-02    ktony                   3099732E406766__

New result for 38746907:

Code:

LL	
Status	          Date	         User	                Residue
Bad	        2007-07-04	C. Cooper / S. Boone	962C36A4F7824731
Verified	2016-01-31	Mark Rose	        80C6A70A7F04DACE
Bad	        2016-03-02	ktony	                3099732E40676654
Verified	2016-03-02	Clara Oswald	        80C6A70A7F04DACE

kladner · 2016-03-03, 05:18

Quote:

Originally Posted by sdbardwick

New result for 38746907:

Code:

LL    
Status              Date             User                    Residue
Bad            2007-07-04    C. Cooper / S. Boone    962C36A4F7824731
Verified    2016-01-31    Mark Rose            80C6A70A7F04DACE
Bad            2016-03-02    ktony                    3099732E40676654
Verified    2016-03-02    Clara Oswald            80C6A70A7F04DACE

Aw Shoot! Thanks!

I am not surprised that a very recent Mark Rose result is the correct one.

Mark Rose · 2016-03-20, 07:50

Requesting double checks:

DoubleCheck=38827301,71,1
DoubleCheck=52276529,73,1

kladner · 2016-03-20, 15:44

I will queue

Code:

DoubleCheck=38827301,71,1

on the CPU. It will take about 12 days, but I am confident of a good result, unlike the GPU

.

Mark Rose · 2016-03-20, 21:04

Quote:

Originally Posted by kladner

I will queue

Code:

DoubleCheck=38827301,71,1

on the CPU. It will take about 12 days, but I am confident of a good result, unlike the GPU

.

Thanks!

NBtarheel_33 · 2016-03-21, 14:55

Quote:

Originally Posted by kladner

NOTE: To reiterate, this was run on a GPU, so a quad will have to be run on a CPU, or abacus, or something.

I wish I could be of help, but I have been getting some bad results from my abacus lately; seems as though it has been overclocked too aggressively. I performed a stress test but that only resulted in the counting wires getting bent and several of the beads flying off and hitting me in the eye. I might try increasing the voltage just a tad to see if that helps.

endless mike · 2016-03-28, 01:13

I thought I turned off "Use Primenet" in the options. Anyone willing to quad check a self verified result?

43313779

http://www.mersenne.org/report_expon...3313779&full=1

2016-02-18, 22:46	#386
kladner "Kieren" Jul 2011 In My Own Galaxy! 2·3·1,693 Posts	DoubleCheck=38746907,71,1 I'll take this one. registered Last fiddled with by kladner on 2016-02-18 at 23:06

2016-03-20, 15:44	#393
kladner "Kieren" Jul 2011 In My Own Galaxy! 2·3·1,693 Posts	I will queue Code: DoubleCheck=38827301,71,1 on the CPU. It will take about 12 days, but I am confident of a good result, unlike the GPU .

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Double checks	casmith789	PrimeNet	7	2015-05-26 00:53
Help doing some quadrup1e+ checks	Madpoo	Data	28	2015-04-06 17:01
Double checks	Rastus	Data	1	2003-12-19 18:20
How do I get rid of the Triple Checks??	outlnder	Lounge	4	2003-04-07 18:06
Double-checks come in pairs?	BigRed	Software	1	2002-10-20 05:29

2016-03-20, 07:50	#392
Mark Rose "/X\(‘-‘)/X\" Jan 2013 B72₁₆ Posts	Requesting double checks: DoubleCheck=38827301,71,1 DoubleCheck=52276529,73,1

2016-03-28, 01:13	#396
endless mike Jan 2004 Milwaukee, WI 2×3×23 Posts	I thought I turned off "Use Primenet" in the options. Anyone willing to quad check a self verified result? 43313779 http://www.mersenne.org/report_expon...3313779&full=1