Strategic Double Clicking - Page 74

Madpoo · 2016-02-06, 02:32

Quote:

Originally Posted by Madpoo

In my ongoing quest to weed out bad systems (well, bad tests), I'm toying with another revision to my query.

And here's a list of 3:1 or more bad/good based on this approach:

Code:

exponent	Bad	Good	Unk	Sus	Solo	Mis	worktodo
35429437	3	0	1	0	1	0	DoubleCheck=35429437,71,1
35675153	4	0	3	0	3	0	DoubleCheck=35675153,71,1
35904383	3	1	1	0	1	0	DoubleCheck=35904383,71,1
36118051	4	0	3	0	3	0	DoubleCheck=36118051,71,1
36583447	3	0	1	0	1	0	DoubleCheck=36583447,71,1
37237547	3	1	1	0	1	0	DoubleCheck=37237547,72,1
45865801	4	1	4	1	4	1	DoubleCheck=45865801,72,1
46101953	6	0	1	0	1	0	DoubleCheck=46101953,72,1
46189607	4	1	4	1	4	1	DoubleCheck=46189607,72,1
46344919	4	1	4	1	4	1	DoubleCheck=46344919,72,1
50015437	9	2	2	0	2	0	DoubleCheck=50015437,73,1
50559623	3	0	2	1	2	1	DoubleCheck=50559623,73,1
50803771	9	2	2	0	2	0	DoubleCheck=50803771,73,1
50967491	4	1	4	1	4	1	DoubleCheck=50967491,73,1
51753259	3	0	2	1	2	1	DoubleCheck=51753259,73,1
55125053	3	0	3	3	3	3	DoubleCheck=55125053,73,1
55622053	5	0	3	1	3	1	DoubleCheck=55622053,73,1
55972967	5	0	3	1	3	1	DoubleCheck=55972967,73,1
56829233	3	0	3	3	3	3	DoubleCheck=56829233,73,1
57384337	4	0	3	0	3	0	DoubleCheck=57384337,73,1
58141243	3	0	1	0	1	0	DoubleCheck=58141243,73,1
58229173	5	1	11	4	11	4	DoubleCheck=58229173,73,1
58229179	5	1	11	4	11	4	DoubleCheck=58229179,73,1
58229333	5	1	11	4	11	4	DoubleCheck=58229333,73,1
58229621	5	1	11	4	11	4	DoubleCheck=58229621,73,1
58232081	5	1	11	4	11	4	DoubleCheck=58232081,73,1
58374847	4	0	3	0	3	0	DoubleCheck=58374847,73,1
58407949	4	0	2	0	2	0	DoubleCheck=58407949,73,1
58773223	4	0	3	0	3	0	DoubleCheck=58773223,73,1
59136449	5	1	11	4	11	4	DoubleCheck=59136449,73,1
59136463	5	1	11	4	11	4	DoubleCheck=59136463,73,1
60470117	5	1	11	4	11	4	DoubleCheck=60470117,73,1
60489577	5	1	11	4	11	4	DoubleCheck=60489577,73,1

cuBerBruce · 2016-02-06, 02:49

Quote:

Originally Posted by Madpoo

And here's a list of 3:1 or more bad/good based on this approach:

Code:

exponent	Bad	Good	Unk	Sus	Solo	Mis	worktodo
35429437	3	0	1	0	1	0	DoubleCheck=35429437,71,1
35675153	4	0	3	0	3	0	DoubleCheck=35675153,71,1

I took these, the first two.

kladner · 2016-02-06, 05:48

I will queue this one:
DoubleCheck=35904383,71,1

ATH · 2016-02-06, 06:06

I queued 2:

DoubleCheck=36118051,71,1
DoubleCheck=36583447,71,1

Dubslow · 2016-02-06, 07:25

Although not directly related to strategic double checking, the user known as Amy Pond seems to have a box doing Cat 1 DC that is... unreliable. Madpoo, could you please investigate? Is there anything we can do about it?

ET_ · 2016-02-06, 12:31

Taking the following:

Code:

37237547	3	1	1	0	1	0	DoubleCheck=37237547,72,1
45865801	4	1	4	1	4	1	DoubleCheck=45865801,72,1
46101953	6	0	1	0	1	0	DoubleCheck=46101953,72,1
46189607	4	1	4	1	4	1	DoubleCheck=46189607,72,1
46344919	4	1	4	1	4	1	DoubleCheck=46344919,72,1

Luigi

ATH · 2016-02-06, 18:55

Quote:

Originally Posted by ET_

38194643 unverified.

It did not match unfortunately: http://mersenne.org/M38194643
Titan Black 2048K FFT. Max error: 0.25781.

If I had noticed the "high" error level I would have forced the next FFT probably.

Madpoo · 2016-02-07, 18:20

Quote:

Originally Posted by Dubslow

Although not directly related to strategic double checking, the user known as Amy Pond seems to have a box doing Cat 1 DC that is... unreliable. Madpoo, could you please investigate? Is there anything we can do about it?

Unreliable in what way?

Of the 4 DC exponents that user has right now, the CPUs working on them have okay track records (well, one of them must be newer since it has no returned results so far).

All 4 of the assignments have checked in recently (today) and were only assigned recently (last few days).

fivemack · 2016-02-07, 20:58

Taking

Code:

50015437	9	2	2	0	2	0	DoubleCheck=50015437,73,1
50559623	3	0	2	1	2	1	DoubleCheck=50559623,73,1
50803771	9	2	2	0	2	0	DoubleCheck=50803771,73,1
50967491	4	1	4	1	4	1	DoubleCheck=50967491,73,1

Dubslow · 2016-02-07, 21:39

Quote:

Originally Posted by Madpoo

Unreliable in what way?

Of the 4 DC exponents that user has right now, the CPUs working on them have okay track records (well, one of them must be newer since it has no returned results so far).

All 4 of the assignments have checked in recently (today) and were only assigned recently (last few days).

As in, almost all the recent results I could see were bad.

kladner · 2016-02-08, 00:23

I see that i5-4670-main has a poor record, and two results pending double check.
That computer seems to have had 5 matches this year.

I understand the concern, but what recourse might there be? If the user-of-record isn't on the Forum, then it seems that PrimeNet is the only other source of contact information.

But think about it. The "Microsoft Tech Support" scam is regularly derided here. If one got an email from GIMPS saying "your computer is doing bad things," how much credence might it get? At least, this is a named user, with a variety of machines running Prime95. This suggests someone a bit more knowledgeable. Such a person might respond well if approached carefully, if contact is even possible.

EDIT: The last screeny is just to say that bad spells do happen. 2014 was rotten for me, until I replaced a failing PSU. The 2015 failures came in the course of trying to find successful hardware settings for running CUDALucas

2016-02-07, 20:58	#812
fivemack (loop (#_fork)) Feb 2006 Cambridge, England 2×13²×19 Posts	Taking Code: 50015437 9 2 2 0 2 0 DoubleCheck=50015437,73,1 50559623 3 0 2 1 2 1 DoubleCheck=50559623,73,1 50803771 9 2 2 0 2 0 DoubleCheck=50803771,73,1 50967491 4 1 4 1 4 1 DoubleCheck=50967491,73,1

2016-02-08, 00:23	#814
kladner "Kieren" Jul 2011 In My Own Galaxy! 2×3×1,693 Posts	I see that i5-4670-main has a poor record, and two results pending double check. That computer seems to have had 5 matches this year. I understand the concern, but what recourse might there be? If the user-of-record isn't on the Forum, then it seems that PrimeNet is the only other source of contact information. But think about it. The "Microsoft Tech Support" scam is regularly derided here. If one got an email from GIMPS saying *"your computer is doing bad things," how much credence might it get? At least, this is a named user, with a variety of machines running Prime95. This suggests someone a bit more knowledgeable. Such a person might respond well if approached carefully, if contact is even possible. EDIT: The last screeny is just to say that bad spells do happen. 2014 was rotten for me, until I replaced a failing PSU. The 2015 failures came in the course of trying to find successful hardware settings for running CUDALucas Attached Thumbnails Last fiddled with by kladner on 2016-02-08 at 00:27*

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

2016-02-06, 05:48	#806
kladner "Kieren" Jul 2011 In My Own Galaxy! 10158₁₀ Posts	I will queue this one: DoubleCheck=35904383,71,1

2016-02-06, 06:06	#807
ATH Einyen Dec 2003 Denmark 7·11·41 Posts	I queued 2: DoubleCheck=36118051,71,1 DoubleCheck=36583447,71,1

2016-02-06, 07:25	#808
Dubslow Basketry That Evening! "Bunslow the Bold" Jun 2011 40<A<43 -89<O<-88 3×29×83 Posts	Although not directly related to strategic double checking, the user known as Amy Pond seems to have a box doing Cat 1 DC that is... unreliable. Madpoo, could you please investigate? Is there anything we can do about it?