Strategic Double Clicking - Page 131

ATH · 2017-04-18, 16:26

Quote:

Originally Posted by Madpoo

That was a count of all systems that have returned at least one LL result, but so far none of their results have been verified (or proven bad).

But as GP2 pointed out, those systems that only turned in 1 LL result ever will not be very interesting to double check, at least it will not be very "strategic" double checking. How many of those are there in the 13,683 machines?

Madpoo · 2017-04-23, 00:22

Quote:

Originally Posted by ATH

But as GP2 pointed out, those systems that only turned in 1 LL result ever will not be very interesting to double check, at least it will not be very "strategic" double checking. How many of those are there in the 13,683 machines?

As of right now:

Code:

13,651 total don't have any DC'd result
10,088 of those have 2+ results
 4,518 have 5+ results
 2,267 have 10+ results
   782 have 20+ results
   137 have 50+ results
    36 have 100+ results
     6 have 150+ results

Of those 6 with 150+ results (and not a DC'd result among them) here are 4 exponents that represent the lowest available. I excluded any CPU where someone's already doing a DC of any of their work.

Code:

Exponent	Bad	Good	Unk	Sus	Solo	Mis	worktodo
50162537	0	0	167	0	167	0	DoubleCheck=50162537,73,1
50190247	0	0	150	0	150	0	DoubleCheck=50190247,73,1
51464701	0	0	155	0	155	0	DoubleCheck=51464701,73,1
57849343	0	0	159	0	159	0	DoubleCheck=57849343,73,1

3 of those 4 are Curtis machines which explains the large amount of work.

GP2 · 2017-04-23, 10:48

Quote:

Originally Posted by Madpoo

Of those 6 with 150+ results (and not a DC'd result among them) here are 4 exponents that represent the lowest available. I excluded any CPU where someone's already doing a DC of any of their work.

Fascinating. Any machine that's prolific enough to turn in 150+ results should have done one of the new automatic once-yearly double checks on its own by now, unless that setting was deliberately turned off or unless the machine is no longer active.

This does look like a promising vein to pursue, although the fact that the fourth-lowest exponent jumps all the way to 57M suggests that most of these exponents will be in the fairly large ranges.

Anyways, I grabbed these four.

Mark Rose · 2017-04-23, 13:43

All the exponents from the thread are already queued/done. If we're out of normal SDCs, that would be a good list to post (perhaps from the 50+ machines).

Madpoo · 2017-04-23, 21:57

Quote:

Originally Posted by GP2

Fascinating. Any machine that's prolific enough to turn in 150+ results should have done one of the new automatic once-yearly double checks on its own by now, unless that setting was deliberately turned off or unless the machine is no longer active.

This does look like a promising vein to pursue, although the fact that the fourth-lowest exponent jumps all the way to 57M suggests that most of these exponents will be in the fairly large ranges.

Anyways, I grabbed these four.

Many of those CPUs are probably older and don't participate anymore. Otherwise I imagine they would have picked up the occasional DC by now. It's also telling that none of these have ever turned in a "suspect" result either since that would have triggered a DC of their result much earlier.

That's not to say that a system with no suspect results is always good... in fact there are quite a few bad results out there where the system thought everything was fine. Some of them turned in all bad results and not a single one was suspect.

LaurV · 2017-04-25, 03:57

Quote:

Originally Posted by Madpoo

<snip>

The accountant in me: somehow the numbers in that table do not add up... :razz:
Is the first row (13k) machines with a [U]single[/U] LL done (as it seems from the numbers), or the total (as I understand from the surrounding text)? (the issue may be with my English)

Edit: Scrap that, I just realized that the categories are not exclusive. The 2+ includes 5+, of course....

If there are 5 or more, then there are 2 or more too... Need more coffee...

Mark Rose · 2017-05-01, 02:11

Quote:

Originally Posted by Madpoo

Cool. Yeah, I figured that large one is in no immediate need of a strategic double check anyway, but figured I'd include it just to be thorough.

Turns out 102141577 needs a triple check.

Madpoo · 2017-05-01, 04:03

Quote:

Originally Posted by Mark Rose

Turns out 102141577 needs a triple check.

Yeah, saw that and picked it up.

Here's a new list of the top 100 exponents ordered by that cpu's total bad/good ratio.

Code:

exponent	Bad	Good	Unk	Sus	Solo	Mis	worktodo
78464039	8	0	1	1	1	1	DoubleCheck=78464039,75,1
40484341	7	1	1	0	1	0	DoubleCheck=40484341,72,1
75458441	5	0	5	0	5	0	DoubleCheck=75458441,75,1
76565029	5	0	5	0	5	0	DoubleCheck=76565029,75,1
76565473	5	0	5	0	5	0	DoubleCheck=76565473,75,1
76565663	5	0	5	0	5	0	DoubleCheck=76565663,75,1
76565837	5	0	5	0	5	0	DoubleCheck=76565837,75,1
78247601	5	1	13	0	13	0	DoubleCheck=78247601,75,1
78325889	5	1	13	0	13	0	DoubleCheck=78325889,75,1
78343427	5	1	13	0	13	0	DoubleCheck=78343427,75,1
78354667	5	1	13	0	13	0	DoubleCheck=78354667,75,1
78435713	5	1	13	0	13	0	DoubleCheck=78435713,75,1
78595963	5	1	13	0	13	0	DoubleCheck=78595963,75,1
79374233	5	1	13	0	13	0	DoubleCheck=79374233,75,1
79517539	5	1	13	0	13	0	DoubleCheck=79517539,75,1
79594373	5	1	13	0	13	0	DoubleCheck=79594373,75,1
79829011	5	1	13	0	13	0	DoubleCheck=79829011,75,1
80036263	5	1	13	0	13	0	DoubleCheck=80036263,75,1
80053069	5	1	13	0	13	0	DoubleCheck=80053069,75,1
80070589	5	1	13	0	13	0	DoubleCheck=80070589,75,1
75226579	4	0	1	1	1	1	DoubleCheck=75226579,75,1
54460597	133	41	24	0	22	2	DoubleCheck=54460597,73,1
55474673	133	41	24	0	22	2	DoubleCheck=55474673,73,1
55524773	133	41	24	0	22	2	DoubleCheck=55524773,73,1
55657067	133	41	24	0	22	2	DoubleCheck=55657067,73,1
55835551	133	41	24	0	22	2	DoubleCheck=55835551,73,1
55980481	133	41	24	0	22	2	DoubleCheck=55980481,73,1
56735933	133	41	24	0	22	2	DoubleCheck=56735933,73,1
56926579	133	41	24	0	22	2	DoubleCheck=56926579,73,1
58037909	133	41	24	0	22	2	DoubleCheck=58037909,73,1
59214593	133	41	24	0	22	2	DoubleCheck=59214593,73,1
59908361	133	41	24	0	22	2	DoubleCheck=59908361,73,1
60552893	133	41	24	0	22	2	DoubleCheck=60552893,74,1
61311953	133	41	24	0	22	2	DoubleCheck=61311953,74,1
61469383	133	41	24	0	22	2	DoubleCheck=61469383,74,1
75847703	6	2	4	0	4	0	DoubleCheck=75847703,75,1
75856579	6	2	4	0	4	0	DoubleCheck=75856579,75,1
76351169	6	2	4	0	4	0	DoubleCheck=76351169,75,1
77116451	6	2	4	0	4	0	DoubleCheck=77116451,75,1
78111461	3	0	0	1	1	0	DoubleCheck=78111461,75,1
79503253	3	0	0	1	1	0	DoubleCheck=79503253,75,1
80356853	3	0	0	1	1	0	DoubleCheck=80356853,75,1
75649417	3	0	1	0	1	0	DoubleCheck=75649417,75,1
79593601	3	0	1	0	1	0	DoubleCheck=79593601,75,1
74711501	3	1	4	0	4	0	DoubleCheck=74711501,76,1
76895393	3	1	4	0	4	0	DoubleCheck=76895393,75,1
78551183	3	1	4	0	4	0	DoubleCheck=78551183,75,1
78993269	3	1	4	0	4	0	DoubleCheck=78993269,75,1
70131329	3	1	22	0	22	0	DoubleCheck=70131329,75,1
70602353	3	1	22	0	22	0	DoubleCheck=70602353,75,1
72001729	3	1	22	0	22	0	DoubleCheck=72001729,75,1
73028581	3	1	22	0	22	0	DoubleCheck=73028581,75,1
73029409	3	1	22	0	22	0	DoubleCheck=73029409,75,1
74066401	3	1	22	0	22	0	DoubleCheck=74066401,75,1
74767633	3	1	22	0	22	0	DoubleCheck=74767633,75,1
75253049	3	1	22	0	22	0	DoubleCheck=75253049,75,1
75722363	3	1	22	0	22	0	DoubleCheck=75722363,75,1
75843913	3	1	22	0	22	0	DoubleCheck=75843913,75,1
75850783	3	1	22	0	22	0	DoubleCheck=75850783,75,1
77021183	3	1	22	0	22	0	DoubleCheck=77021183,75,1
77030879	3	1	22	0	22	0	DoubleCheck=77030879,75,1
78142787	3	1	22	0	22	0	DoubleCheck=78142787,75,1
78224677	3	1	22	0	22	0	DoubleCheck=78224677,75,1
78226651	3	1	22	0	22	0	DoubleCheck=78226651,75,1
78229273	3	1	22	0	22	0	DoubleCheck=78229273,75,1
79017877	3	1	22	0	22	0	DoubleCheck=79017877,75,1
79017889	3	1	22	0	22	0	DoubleCheck=79017889,75,1
79018187	3	1	22	0	22	0	DoubleCheck=79018187,75,1
79018309	3	1	22	0	22	0	DoubleCheck=79018309,75,1
77293871	42	15	1	1	2	0	DoubleCheck=77293871,75,1
77463277	27	12	8	5	12	1	DoubleCheck=77463277,75,1
77570957	27	12	8	5	12	1	DoubleCheck=77570957,75,1
77808737	27	12	8	5	12	1	DoubleCheck=77808737,75,1
78141127	27	12	8	5	12	1	DoubleCheck=78141127,75,1
78212293	27	12	8	5	12	1	DoubleCheck=78212293,75,1
78244297	27	12	8	5	12	1	DoubleCheck=78244297,75,1
78328181	27	12	8	5	12	1	DoubleCheck=78328181,75,1
78448613	27	12	8	5	12	1	DoubleCheck=78448613,75,1
78479413	27	12	8	5	12	1	DoubleCheck=78479413,75,1
78607747	27	12	8	5	12	1	DoubleCheck=78607747,75,1
78700397	27	12	8	5	12	1	DoubleCheck=78700397,75,1
78807727	27	12	8	5	12	1	DoubleCheck=78807727,75,1
78428699	8	4	4	0	4	0	DoubleCheck=78428699,75,1
77725871	6	3	1	0	1	0	DoubleCheck=77725871,75,1
61626077	4	2	18	1	19	0	DoubleCheck=61626077,74,1
62632363	4	2	18	1	19	0	DoubleCheck=62632363,74,1
62634337	4	2	18	1	19	0	DoubleCheck=62634337,74,1
64433371	4	2	18	1	19	0	DoubleCheck=64433371,74,1
65280883	4	2	18	1	19	0	DoubleCheck=65280883,74,1
65446441	4	2	18	1	19	0	DoubleCheck=65446441,74,1
66847993	4	2	18	1	19	0	DoubleCheck=66847993,74,1
67476193	4	2	18	1	19	0	DoubleCheck=67476193,74,1
67878047	4	2	18	1	19	0	DoubleCheck=67878047,74,1
67878137	4	2	18	1	19	0	DoubleCheck=67878137,74,1
67878289	4	2	18	1	19	0	DoubleCheck=67878289,74,1
67878389	4	2	18	1	19	0	DoubleCheck=67878389,74,1
68839013	4	2	18	1	19	0	DoubleCheck=68839013,74,1
68924729	4	2	18	1	19	0	DoubleCheck=68924729,74,1
69826747	4	2	18	1	19	0	DoubleCheck=69826747,74,1
69846011	4	2	18	1	19	0	DoubleCheck=69846011,74,1

rudi_m · 2017-05-01, 07:11

Quote:

Originally Posted by Madpoo

Here's a new list of the top 100 exponents ordered by that cpu's total bad/good ratio.
[/CODE]

I took all below 60M.

Mark Rose · 2017-05-05, 04:29

I took the remaining below 77M.

LaurV · 2017-05-07, 14:53

Here is one I am not happy with, M76453229. I ran both tests and the residues matched on the way, but one is clLucas and one is gpuOwl, both with null shift, in the same card (HD7970). Most probably (99.999999%, hehe) my residue is right, but a "shifted" check won't hurt.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Double-Double Arithmetic	Mysticial	Software	50	2017-10-30 19:16
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

2017-04-23, 13:43	#1434
Mark Rose "/X\(‘-‘)/X\" Jan 2013 101101110001₂ Posts	All the exponents from the thread are already queued/done. If we're out of normal SDCs, that would be a good list to post (perhaps from the 50+ machines).

2017-05-05, 04:29	#1440
Mark Rose "/X\(‘-‘)/X\" Jan 2013 2929₁₀ Posts	I took the remaining below 77M.

2017-05-07, 14:53	#1441
LaurV Romulan Interpreter Jun 2011 Thailand 2³·19·61 Posts	Here is one I am not happy with, M76453229. I ran both tests and the residues matched on the way, but one is clLucas and one is gpuOwl, both with null shift, in the same card (HD7970). Most probably (99.999999%, hehe) my residue is right, but a "shifted" check won't hurt.