mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   Marin's Mersenne-aries (https://www.mersenneforum.org/forumdisplay.php?f=30)
-   -   Strategic Double Clicking (https://www.mersenneforum.org/showthread.php?t=20372)

chalsall 2015-10-15 15:46

[QUOTE=Madpoo;412715]Still, I can give that a shot. All it takes is doing a datepart, more or less, on the received date for the result. What the hey, I'll try that later.[/QUOTE]

Feel free to throw a few experimental tests my way. I don't have your firepower, but I can do 10 or so a day in the lower ranges.

endless mike 2015-10-16 01:06

Earlier today I finished the last of the double checks that I had most recently claimed in this thread. All but one matched with the first time test.
50398199 Matched first test
52040731 Did not match first test
52405957 Matched first test
52939417 Matched first test
53562269 Matched first test
54170447 Matched first test
56868607 Matched first test

dragonbud20 2015-10-16 05:42

M47921893 needs a triple check

Madpoo 2015-10-16 16:26

[QUOTE=dragonbud20;412833]M47921893 needs a triple check[/QUOTE]

That looks like one of the weird ones where the previous tester had their result checked in twice by mistake:
[URL="http://www.mersenne.org/report_exponent/?exp_lo=47921893&full=1"]http://www.mersenne.org/report_exponent/?exp_lo=47921893&full=1[/URL]

I've been testing as many of those as I can. Usually I match, but I've had two now where my result was different, and yours makes a third one I'm aware of.

For me, I've just been assuming my result is correct. If your machine is stable enough then we can pretend the same for you until the double-checkers finally get up to the 47M range (could be a while). :smile:

I think I'd mentioned before that currently there are something like 4800-5000 exponents below 58M that have been DC'd without a match. *Most* of those are because one of them was suspect during it's first run, so the mismatch isn't terribly surprising.

For the other 940 where neither result was suspect (like this one), I might start taking on some of those, because it seems weirder when a machine returns a bad result without it being marked suspect. If we can identify those machines then we may have a good idea that other tests of theirs are wonky.

chalsall 2015-10-16 16:28

[QUOTE=dragonbud20;412833]M47921893 needs a triple check[/QUOTE]

I'll run it. Should take ~48 hours or so.

cuBerBruce 2015-10-17 01:22

I have confirmed dragonbud20's residue for M46102687.

Madpoo 2015-10-17 06:33

[QUOTE=Prime95;412672]Instead of program version, why not use the year the LL result was submitted. MS-SQL should be able to do a GROUP BY on that fairly efficiently. This has the advantage of catching machines that did not update the prime95 application.

I know you mentioned something like this earlier. I would give it a try before giving up on it as too burdensome for the server.[/QUOTE]

Well, I tried it out, as far as generating some reports... next step would be to actually do some LL tests on exponents I've found using the breakdown-by-year method.

And now, here's an example of a specific CPU. In aggregate this machine looks pretty good:
[CODE]Bad Good Sus Unk Solo Mis
12 74 4 41 39 6[/CODE]

Only 12 bad out of 74 good, well hey, that's decent. I wouldn't have bothered looking at it any more...odds of finding more bad in those 39 solo-checked stuff would seem to be low.

But break their results down by year and we see an interesting pattern:
[CODE]Year Bad Good Sus Unk Solo Mis
2008 0 9 0 2 2 0
2009 0 52 0 8 8 0
2010 1 12 0 26 24 2
2011 11 1 4 5 5 4[/CODE]

Seems like 2011 was a bad year for this system... 11 of the 12 known bad were from this year, and it only has 5 solo-checked exponents, so maybe it wouldn't be too bad to check those out.

In fact that's what I plan to do. I picked up all 5 of those (47M-49M range) so I'll know in a day.

If anyone else is interested in trying out a couple smaller ones using this approach, here's a few:
[CODE]37952297 19 3 3 0 3 0 DoubleCheck=37952297,71,1
38155841 12 3 3 0 1 2 DoubleCheck=38155841,71,1
41720983 3 0 3 0 3 0 DoubleCheck=41720983,72,1
42676009 3 0 4 1 4 1 DoubleCheck=42676009,72,1
42846421 5 1 6 1 6 1 DoubleCheck=42846421,72,1
43100209 3 1 6 0 5 1 DoubleCheck=43100209,72,1
43268627 4 1 2 0 2 0 DoubleCheck=43268627,72,1[/CODE]

LaurV 2015-10-17 07:01

[QUOTE=Madpoo;412911]Seems like 2011 was a bad year for this system... 11 of the 12 known bad were from this year, and it only has 5 solo-checked exponents, so maybe it wouldn't be too bad to check those out.
[/QUOTE]
Very nice report! This only reinforces again and again the fact that P95 is an extremely good "stress tester". I would recommend for all sysadmins to run it at least few hours per day in their networks, doing DC work :razz:. This guy's computer ran for years without much a problem, then it started producing junk, and at the end, got replaced. If he would pay attention earlier to the output, he had a very early indicator that his system went in the weeds, and we would have less headache now, too...

dragonbud20 2015-10-17 08:11

I've reserved the first three M37952297, M38155841, M41720983 should have them done over the next couple of days to a week.

cuBerBruce 2015-10-17 12:28

[QUOTE=Madpoo;412911]If anyone else is interested in trying out a couple smaller ones using this approach, here's a few:
[/QUOTE]

I took:
[CODE]42676009 3 0 4 1 4 1 DoubleCheck=42676009,72,1
[/CODE]

chalsall 2015-10-17 14:06

[QUOTE=Madpoo;412911]If anyone else is interested in trying out a couple smaller ones using this approach, here's a few:[/QUOTE]

I've taken the last three.

frmky 2015-10-17 20:25

[QUOTE=Madpoo;408790]Okay, I think I got it running okay. It'll spit out a list of exponents that need a triple check, and neither of the two machines that already submitted a result are suspected of being bad or good. By focusing on these we can start to get results on both by marking more bad and more good in a single test, which should improve the overall strategy of rooting out bad systems.

To start I just pulled exponents < 35.5M to keep the list smallish... there's still 128 entries though.[/QUOTE]
I just reserved the last of this TC list. Got more? :smile:

chalsall 2015-10-17 20:51

[QUOTE=LaurV;412915]I would recommend for all sysadmins to run it at least few hours per day in their networks, doing DC work :razz:. This guy's computer ran for years without much a problem, then it started producing junk, and at the end, got replaced. If he would pay attention earlier to the output, he had a very early indicator that his system went in the weeds, and we would have less headache now, too...[/QUOTE]

Seconded.

I can't actually put a dollar value on how much GIMPS has saved me in downtime.

But, as a guess, lots!

When your (possibly mission critical) machine can't do a DC "well", you must give it the evil eye. And schedule a replacement. Like, really soon now....

Madpoo 2015-10-18 03:26

[QUOTE=frmky;412954]I just reserved the last of this TC list. Got more? :smile:[/QUOTE]

Okay... I had to refresh my memory on what this was. :smile:

So, for everyone else... it's a list of exponents that need a *triple check*, and the two machines that ran it the first time don't really have a track record.

To be specific, this is a list of exponents that were checked twice already with no match, and both machines have *zero* bad results and 6 or less good results.

I was trying to think of a good way to show the list of exponents... I mean, I can do that and it'll have *two* lines for each one, showing the CPU stats for each one, but then you'd have to remember to only take one of those entries and set up a client to do the DC. That's how I was doing it when I picked up a handful of just these very things earlier today...

For now I'll just have a list of exponents and the worktodo entry... if you'd like to see more stats on each one, let me know.

52 of these below 38M. End result will be that, whatever the outcome, at least one of these previously "spotless" CPU's will have a bad result added to their account, which will hopefully feed into our strategic double checking:
[CODE]exponent worktodo
35699393 DoubleCheck=35699393,71,1
35795849 DoubleCheck=35795849,71,1
36056633 DoubleCheck=36056633,71,1
36178937 DoubleCheck=36178937,71,1
36239101 DoubleCheck=36239101,71,1
36264091 DoubleCheck=36264091,71,1
36302951 DoubleCheck=36302951,71,1
36306343 DoubleCheck=36306343,71,1
36380651 DoubleCheck=36380651,71,1
36502049 DoubleCheck=36502049,71,1
36506081 DoubleCheck=36506081,71,1
36524963 DoubleCheck=36524963,71,1
36681691 DoubleCheck=36681691,71,1
36685543 DoubleCheck=36685543,71,1
36777289 DoubleCheck=36777289,71,1
36849583 DoubleCheck=36849583,71,1
36863647 DoubleCheck=36863647,71,1
36868357 DoubleCheck=36868357,71,1
36880643 DoubleCheck=36880643,71,1
36899831 DoubleCheck=36899831,71,1
36950449 DoubleCheck=36950449,71,1
36996521 DoubleCheck=36996521,71,1
37055581 DoubleCheck=37055581,71,1
37069369 DoubleCheck=37069369,71,1
37096859 DoubleCheck=37096859,71,1
37097167 DoubleCheck=37097167,71,1
37214531 DoubleCheck=37214531,71,1
37235789 DoubleCheck=37235789,71,1
37236013 DoubleCheck=37236013,72,1
37347641 DoubleCheck=37347641,71,1
37392473 DoubleCheck=37392473,71,1
37440577 DoubleCheck=37440577,71,1
37452281 DoubleCheck=37452281,71,1
37455169 DoubleCheck=37455169,71,1
37456337 DoubleCheck=37456337,71,1
37471513 DoubleCheck=37471513,71,1
37476293 DoubleCheck=37476293,71,1
37481531 DoubleCheck=37481531,71,1
37487017 DoubleCheck=37487017,71,1
37529941 DoubleCheck=37529941,71,1
37532197 DoubleCheck=37532197,71,1
37608313 DoubleCheck=37608313,71,1
37625381 DoubleCheck=37625381,71,1
37698851 DoubleCheck=37698851,71,1
37761833 DoubleCheck=37761833,71,1
37771337 DoubleCheck=37771337,71,1
37795843 DoubleCheck=37795843,71,1
37809337 DoubleCheck=37809337,71,1
37849457 DoubleCheck=37849457,71,1
37854829 DoubleCheck=37854829,71,1
37899497 DoubleCheck=37899497,71,1
37960003 DoubleCheck=37960003,71,1[/CODE]

kladner 2015-10-18 03:32

35699393 DoubleCheck=35699393,71,1
Mine.

cuBerBruce 2015-10-18 03:56

[QUOTE=Madpoo;412980]
[CODE]35795849 DoubleCheck=35795849,71,1
36056633 DoubleCheck=36056633,71,1
[/CODE][/QUOTE]

I took these two - the next two. Will be doing these serially after my 42M finishes.

I thought I would take a couple before the lowest ones are all taken.

endless mike 2015-10-18 14:13

[QUOTE=Madpoo;412980]
[CODE]exponent worktodo
36178937 DoubleCheck=36178937,71,1
36239101 DoubleCheck=36239101,71,1
36264091 DoubleCheck=36264091,71,1
36302951 DoubleCheck=36302951,71,1
[/CODE][/QUOTE]

I grabbed these four

dragonbud20 2015-10-18 17:11

I'm grabbing these four

[QUOTE=Madpoo;412980]
[CODE]exponent worktodo
36306343 DoubleCheck=36306343,71,1
36380651 DoubleCheck=36380651,71,1
36502049 DoubleCheck=36502049,71,1
36506081 DoubleCheck=36506081,71,1
[/CODE][/QUOTE]

Madpoo 2015-10-18 21:13

[QUOTE=Madpoo;412911]...
But break their results down by year and we see an interesting pattern:
[CODE]Year Bad Good Sus Unk Solo Mis
2008 0 9 0 2 2 0
2009 0 52 0 8 8 0
2010 1 12 0 26 24 2
2011 11 1 4 5 5 4[/CODE]

Seems like 2011 was a bad year for this system... 11 of the 12 known bad were from this year, and it only has 5 solo-checked exponents, so maybe it wouldn't be too bad to check those out.

In fact that's what I plan to do. I picked up all 5 of those (47M-49M range) so I'll know in a day....[/QUOTE]

Well, that was fun. All 5 of those exponents I checked came up with different residues. :smile: If I hadn't broken this systems results down by year, I never would have bothered with doing any of them, much less been able to pick out those 5 in particular.

I can't say they'll all be like this, but it does give me a fun new way of picking out work that was probably done wrong the first time. I'll try and work up some new lists of exponents to do.

chalsall 2015-10-18 21:14

[QUOTE=dragonbud20;412833]M47921893 needs a triple check[/QUOTE]

We matched.

Madpoo 2015-10-18 21:27

[QUOTE=Madpoo;413034]I can't say they'll all be like this, but it does give me a fun new way of picking out work that was probably done wrong the first time. I'll try and work up some new lists of exponents to do.[/QUOTE]

In fact, here's some now. That system with the 8 bad, zero good... it was a bad year (2010) for it. The next year (2011) it did slightly better with 5 bad, 8 good. In total that machine did have more bad than good, but it wasn't quite bad enough that I was looking at it yet (13 bad, 8 good).

Similar story with that system with 15 bad and 2 good. That was 2013. Every other year seemed better.

Anyway, have at 'em:
[CODE]exponent Bad Good Unk Sus Solo Mis worktodo
46251281 4 0 1 1 1 1 DoubleCheck=46251281,72,1
48137273 5 1 6 1 6 1 DoubleCheck=48137273,71,1
48222241 8 0 4 3 4 3 DoubleCheck=48222241,72,1
48223573 8 0 4 3 4 3 DoubleCheck=48223573,72,1
48533047 5 1 6 1 6 1 DoubleCheck=48533047,71,1
49578379 4 1 2 0 2 0 DoubleCheck=49578379,71,1
49585297 8 0 4 3 4 3 DoubleCheck=49585297,72,1
49773601 5 1 6 1 6 1 DoubleCheck=49773601,70,1
52219579 15 2 1 2 1 2 DoubleCheck=52219579,73,1
54083921 5 0 1 3 1 3 DoubleCheck=54083921,73,1[/CODE]

chalsall 2015-10-18 22:16

[QUOTE=Madpoo;413036]Anyway, have at 'em:[/QUOTE]

I took them. I'll also TF them "appropriately" in parallel, as needed.

UBR47K 2015-10-19 12:22

Requesting a triple check to a exponent done by a new machine.
[code]Doublecheck=41516807,72,1[/code]

chalsall 2015-10-19 13:33

[QUOTE=UBR47K;413081]Requesting a triple check to a exponent done by a new machine.[/QUOTE]

I've put it on one of my highly reliable machines. ~48 hours.

endless mike 2015-10-19 14:01

[QUOTE=Madpoo;412980]
[CODE]exponent worktodo
36524963 DoubleCheck=36524963,71,1
36681691 DoubleCheck=36681691,71,1[/CODE][/QUOTE]

Grabbed these two

[QUOTE=Madpoo;412980]
[CODE]exponent worktodo
36685543 DoubleCheck=36685543,71,1
36777289 DoubleCheck=36777289,71,1
36849583 DoubleCheck=36849583,71,1
36863647 DoubleCheck=36863647,71,1
36868357 DoubleCheck=36868357,71,1
36880643 DoubleCheck=36880643,71,1
36899831 DoubleCheck=36899831,71,1
36950449 DoubleCheck=36950449,71,1
36996521 DoubleCheck=36996521,71,1
37055581 DoubleCheck=37055581,71,1[/CODE][/QUOTE]
And these also. Just adding to multiple machines worktodo files

endless mike 2015-10-19 15:06

And these also. Just adding to multiple machines worktodo files

[QUOTE=Madpoo;412980]
[CODE]exponent worktodo
37069369 DoubleCheck=37069369,71,1
37096859 DoubleCheck=37096859,71,1[/CODE][/QUOTE]

I think I have enough for now

chalsall 2015-10-19 15:24

[QUOTE=Madpoo;412911]If anyone else is interested in trying out a couple smaller ones using this approach, here's a few:[/QUOTE]

OK, the part of this batch (the last three) that I took are now complete.

42846421 matched, 43100209 and 43268627 didn't.

If I may "blow some sunshine" your way, thanks for doing this Aaron. A very creative and thoughtful idea, and tangibly beneficial to GIMPS, and for those who use Prime95/mprime for stress testing.

I had asked before, and I'll now ask again: Are any of the machines which are less than perfect still being assigned LL candidates? If so, might they be assigned DC Cat 2 for a while?

Separately, as a policy, should perhaps all machines requesting LL work be assigned a small percentage of DC Cat 1 work, just in order to ensure their ongoing sanity?

cuBerBruce 2015-10-19 16:51

[QUOTE=chalsall;413090]42846421 matched, 43100209 and 43268627 didn't.[/QUOTE]
And my check of 42676009 just finished with a mismatch.

chalsall 2015-10-19 22:18

[QUOTE=chalsall;413090]Separately, as a policy, should perhaps all machines requesting LL work be assigned a small percentage of DC Cat 1 work, just in order to ensure their ongoing sanity?[/QUOTE]

Whoops... Perhaps I pushed the envelope too far in another thread here.

Sorry about that. It's just my training (and my nature)....

frmky 2015-10-19 22:39

[QUOTE=Madpoo;412980]52 of these below 38M. [/QUOTE]
I reserved those available up to and including 37452281.

richs 2015-10-19 23:23

My DC of 35939129 didn't match. Anyone like to TC it?

dragonbud20 2015-10-20 07:46

could someone quadruple check M37952297 it seems that prime95 double submitted it and allowed to me verify my own result.

sdbardwick 2015-10-20 09:46

[QUOTE=dragonbud20;413140]could someone quadruple check M37952297 it seems that prime95 double submitted it and allowed to me verify my own result.[/QUOTE]On it. Give it a couple of days.

cuBerBruce 2015-10-20 15:44

[QUOTE=richs;413128]My DC of 35939129 didn't match. Anyone like to TC it?[/QUOTE]

It's been assigned to NoPolarBearsHere. Possibly a normal PrimeNet assignment.

kladner 2015-10-20 15:51

1 Attachment(s)
[QUOTE=kladner;412981]35699393 DoubleCheck=35699393,71,1
Mine.[/QUOTE]
CuLu matched one of two previous results.

endless mike 2015-10-20 17:12

[QUOTE=Madpoo;412980][CODE]exponent worktodo
37455169 DoubleCheck=37455169,71,1
37456337 DoubleCheck=37456337,71,1
37471513 DoubleCheck=37471513,71,1
37476293 DoubleCheck=37476293,71,1
37481531 DoubleCheck=37481531,71,1
37487017 DoubleCheck=37487017,71,1
[/CODE][/QUOTE]

[QUOTE=endless mike;413089]And these also.


[STRIKE]I think I have enough for now[/STRIKE][/QUOTE]

Maybe a few more

Madpoo 2015-10-20 17:54

[QUOTE=dragonbud20;413140]could someone quadruple check M37952297 it seems that prime95 double submitted it and allowed to me verify my own result.[/QUOTE]

Looks like you ran it with two different runs (different shift counts).

I would have noticed the self-verification when I run my check for those (I still see some pop up from time to time and I do a triple-check) :smile:

Madpoo 2015-10-20 19:57

[QUOTE=endless mike;413175]Maybe a few more[/QUOTE]

Here's more. From machines with at least a 3:1 bad:good ratio, using the "by calendar year" method:
[CODE]exponent Bad Good Unk Sus Solo Mis worktodo
36763357 10 3 4 0 4 0 DoubleCheck=36763357,71,1
36980029 14 4 5 0 5 0 DoubleCheck=36980029,71,1
37312157 14 4 5 0 5 0 DoubleCheck=37312157,71,1
42538723 3 1 8 0 8 0 DoubleCheck=42538723,72,1
43352347 3 1 8 0 8 0 DoubleCheck=43352347,72,1
44073317 3 1 2 0 2 0 DoubleCheck=44073317,72,1
44110853 3 1 2 0 2 0 DoubleCheck=44110853,72,1
48077509 17 5 5 0 5 0 DoubleCheck=48077509,72,1
49091557 3 0 3 0 3 0 DoubleCheck=49091557,72,1
49134083 17 5 5 0 5 0 DoubleCheck=49134083,72,1
49504331 17 5 5 0 5 0 DoubleCheck=49504331,72,1
49668797 7 2 1 0 1 0 DoubleCheck=49668797,72,1
50669273 17 5 5 0 5 0 DoubleCheck=50669273,73,1
50703623 3 1 2 0 2 0 DoubleCheck=50703623,73,1
50812987 7 2 11 8 11 8 DoubleCheck=50812987,73,1
51460391 7 2 11 8 11 8 DoubleCheck=51460391,73,1
51981047 7 2 11 8 11 8 DoubleCheck=51981047,73,1
52210813 3 1 2 0 2 0 DoubleCheck=52210813,73,1
55443329 7 2 11 8 11 8 DoubleCheck=55443329,73,1
55894291 7 2 11 8 11 8 DoubleCheck=55894291,73,1
56542063 7 2 11 8 11 8 DoubleCheck=56542063,73,1
56699557 7 2 11 8 11 8 DoubleCheck=56699557,73,1
57740621 7 2 11 8 11 8 DoubleCheck=57740621,73,1[/CODE]

LookAS 2015-10-21 06:41

I'll take these:

DoubleCheck=36763357,71,1
DoubleCheck=36980029,71,1
DoubleCheck=37312157,71,1
DoubleCheck=42538723,72,1
DoubleCheck=43352347,72,1
DoubleCheck=44073317,72,1
DoubleCheck=44110853,72,1
DoubleCheck=48077509,72,1
DoubleCheck=49091557,72,1
DoubleCheck=49134083,72,1
DoubleCheck=49504331,72,1
DoubleCheck=49668797,72,1

chalsall 2015-10-21 13:32

[QUOTE=UBR47K;413081]Requesting a triple check to a exponent done by a new machine.[/QUOTE]

We matched.

sdbardwick 2015-10-21 13:43

[QUOTE=dragonbud20;413140]could someone quadruple check M37952297 it seems that prime95 double submitted it and allowed to me verify my own result.[/QUOTE][URL="http://www.mersenne.org/report_exponent/?exp_lo=37952297&full=1"]We matched.[/URL] -As expected, given the different shifts noted by Madpoo.

Madpoo 2015-10-21 15:16

[QUOTE=sdbardwick;413237][URL="http://www.mersenne.org/report_exponent/?exp_lo=37952297&full=1"]We matched.[/URL] -As expected, given the different shifts noted by Madpoo.[/QUOTE]

And thanks for doing that... saved me from doing the triple-check myself. :smile:

0PolarBearsHere 2015-10-22 05:39

[QUOTE=richs;413128]My DC of 35939129 didn't match. Anyone like to TC it?[/QUOTE]

Matched your one.

cuBerBruce 2015-10-22 17:25

[QUOTE=Madpoo;412980]52 of these below 38M. End result will be that, whatever the outcome, at least one of these previously "spotless" CPU's will have a bad result added to their account, which will hopefully feed into our strategic double checking:
[CODE]exponent worktodo
...
35795849 DoubleCheck=35795849,71,1
36056633 DoubleCheck=36056633,71,1
...
[/CODE][/QUOTE]

I got a match with one of the prior results with each of these two exponents.

Edit: I'll also note that richs's residue for M35939129 was verified.

richs 2015-10-23 03:31

Thanks for pointing that out. That was a quick turnaround.....

0PolarBearsHere 2015-10-23 08:12

[QUOTE=richs;413419]Thanks for pointing that out. That was a quick turnaround.....[/QUOTE]

I put yours at the top of my work queue. My first time 67M and 72M exponents weren't going to go anywhere.

cuBerBruce 2015-10-23 15:44

[QUOTE=Madpoo;413184]Here's more. From machines with at least a 3:1 bad:good ratio, using the "by calendar year" method:
[CODE]exponent Bad Good Unk Sus Solo Mis worktodo
...
51981047 7 2 11 8 11 8 DoubleCheck=51981047,73,1
...
[/CODE][/QUOTE]

I took this one.

bloodIce 2015-10-23 18:45

[QUOTE=cuBerBruce;412282]My residue did not match for M42003697.[/QUOTE]

[QUOTE=Madpoo;412352]As weird as this sounds, that's exactly what I like to hear. :smile: Strange to be rooting for *mis* matched residues with these.[/QUOTE]

The exponent M42003697 has a factor: 64458145045944198395681. I still wonder why we do not factor more, especially in cases like this. One big iron would have found the factor in less than a day...

cuBerBruce 2015-10-23 20:29

[QUOTE=bloodIce;413487]The exponent M42003697 has a factor: 64458145045944198395681. I still wonder why we do not factor more, especially in cases like this. One big iron would have found the factor in less than a day...[/QUOTE]

Well, my LL test only took a little over 2 days (using 4 cores). I don't think in general it's worth factoring such small exponents up to 76 bits. Well, thanks anyway for making my Results page a little less yellowy. :smile:

Mark Rose 2015-10-23 20:39

[QUOTE=bloodIce;413487]The exponent M42003697 has a factor: 64458145045944198395681. I still wonder why we do not factor more, especially in cases like this. One big iron would have found the factor in less than a day...[/QUOTE]

Because the same GPU ought to be able to LL it faster than doing the higher levels of TF...

Prime95 2015-10-23 20:44

I took the remaining exponents in madpoo's last post:

[CODE]
DoubleCheck=52210813,73,1
DoubleCheck=55443329,73,1
DoubleCheck=55894291,73,1
DoubleCheck=56542063,73,1
DoubleCheck=56699557,73,1
DoubleCheck=57740621,73,1[/CODE]

VBCurtis 2015-10-23 21:02

[QUOTE=bloodIce;413487]The exponent M42003697 has a factor: 64458145045944198395681. I still wonder why we do not factor more, especially in cases like this. One big iron would have found the factor in less than a day...[/QUOTE]

It takes, on average, n factoring runs from n-1 bits to n bits to find a factor. So, rather than ask how long it would take to find *this* factor, estimate how long it would take to factor 75 numbers from 75 bits to 76 bits to expect to find one factor. Expectation may be higher than 75 runs, since P-1 factoring finds some of the factors of that size before TF is run; if we guess 10% of these factors are found by P-1, you would need 83 or so runs from 75 to 76 bits to expect to find one new factor.

Compare that time to 2 LL tests, and you should no longer wonder.

bloodIce 2015-10-24 06:23

[QUOTE=Mark Rose;413503]Because the same GPU ought to be able to LL it faster than doing the higher levels of TF...[/QUOTE]
Nope, the same GPU needs to do one LL test in couple of months, may be 1.

[QUOTE]estimate how long it would take to factor 75 numbers from 75 bits to 76 bits to expect to find one factor.[/QUOTE]
I know that it is impossible to factor all for any practical time. However, I am advocating at least couple of levels in situations where you have mismatching residues or any other problematic LL test (remembering examples of 10 LL tests done, but some mismatched...). In some cases there will be a factor and the case is closed.

Madpoo 2015-10-24 16:32

[QUOTE=bloodIce;413573]I know that it is impossible to factor all for any practical time. However, I am advocating at least couple of levels in situations where you have mismatching residues or any other problematic LL test (remembering examples of 10 LL tests done, but some mismatched...). In some cases there will be a factor and the case is closed.[/QUOTE]

I suppose what I should do is look again at whether any of these strategic double-check candidates have less than the GPU72 factor level. Then, at the very least, we can get those up to those levels ahead of time since we're usually doing double-checks far ahead of when it would be done normally.

I'll check again and see where we're at with those, and also with any new ones that need triple checking.

chalsall 2015-10-24 17:08

[QUOTE=Madpoo;413639]I suppose what I should do is look again at whether any of these strategic double-check candidates have less than the GPU72 factor level.[/QUOTE]

Yes. Where GPU72 nominally takes the TF'ing level to for DC'ing is optimal. Going farther doesn't make sense from a cost / benefit perspective.

[QUOTE=Madpoo;413639]I'll check again and see where we're at with those, and also with any new ones that need triple checking.[/QUOTE]

Just generate a list like before, and I'll add them to the "hopper" for prioritization.

And, this actually brings up something... If you include any that the "GPU Factoring" account holds as well (rather than limiting the list to those candidates not currently "owned" by anyone), I can set the "Priority" flag for those as well.

Madpoo 2015-10-24 18:24

[QUOTE=chalsall;413643]Just generate a list like before, and I'll add them to the "hopper" for prioritization.

And, this actually brings up something... If you include any that the "GPU Factoring" account holds as well (rather than limiting the list to those candidates not currently "owned" by anyone), I can set the "Priority" flag for those as well.[/QUOTE]

Okay, here's a list of exponents where the system had more bad than good but need add'l TF work to get up to GPU72 goals.

54 of the 209 are already checked out to GPU72's account.
[CODE]exponent TF WorkToDo
42381011 71 Factor=42381011,71,72
42381881 71 Factor=42381881,71,72
42392789 71 Factor=42392789,71,72
42394721 71 Factor=42394721,71,72
43144741 71 Factor=43144741,71,72
43689301 71 Factor=43689301,71,72
44906783 70 Factor=44906783,70,72
44913173 70 Factor=44913173,70,72
44946037 70 Factor=44946037,70,72
44971193 70 Factor=44971193,70,72
45046607 70 Factor=45046607,70,72
45095273 70 Factor=45095273,70,72
45135451 70 Factor=45135451,70,72
45172189 71 Factor=45172189,71,72
45253927 71 Factor=45253927,71,72
45333721 70 Factor=45333721,70,72
45381379 70 Factor=45381379,70,72
45521537 70 Factor=45521537,70,72
45790939 70 Factor=45790939,70,72
45819593 70 Factor=45819593,70,72
46018891 70 Factor=46018891,70,72
46047383 71 Factor=46047383,71,72
46048973 70 Factor=46048973,70,72
46134581 70 Factor=46134581,70,72
46140331 70 Factor=46140331,70,72
46233977 70 Factor=46233977,70,72
46289773 70 Factor=46289773,70,72
46296641 70 Factor=46296641,70,72
46300987 70 Factor=46300987,70,72
46514591 70 Factor=46514591,70,72
46564261 70 Factor=46564261,70,72
46582073 71 Factor=46582073,71,72
46727689 70 Factor=46727689,70,72
46849357 70 Factor=46849357,70,72
46976291 71 Factor=46976291,71,72
47002633 70 Factor=47002633,70,72
47029163 71 Factor=47029163,71,72
47092189 71 Factor=47092189,71,72
47124751 71 Factor=47124751,71,72
47127167 71 Factor=47127167,71,72
47139277 70 Factor=47139277,70,72
47164339 71 Factor=47164339,71,72
47207833 70 Factor=47207833,70,72
47224109 71 Factor=47224109,71,72
47283289 70 Factor=47283289,70,72
47374501 71 Factor=47374501,71,72
47425219 71 Factor=47425219,71,72
47454943 70 Factor=47454943,70,72
47473697 70 Factor=47473697,70,72
47523409 70 Factor=47523409,70,72
47550073 71 Factor=47550073,71,72
47556083 71 Factor=47556083,71,72
47561929 70 Factor=47561929,70,72
47568511 70 Factor=47568511,70,72
47601637 70 Factor=47601637,70,72
47611999 70 Factor=47611999,70,72
47615479 70 Factor=47615479,70,72
47617711 70 Factor=47617711,70,72
47645839 70 Factor=47645839,70,72
47665771 70 Factor=47665771,70,72
47679571 70 Factor=47679571,70,72
47771783 70 Factor=47771783,70,72
47784553 71 Factor=47784553,71,72
47811013 71 Factor=47811013,71,72
47908891 70 Factor=47908891,70,72
48093373 70 Factor=48093373,70,72
48124261 70 Factor=48124261,70,72
48614947 71 Factor=48614947,71,72
48762059 71 Factor=48762059,71,72
48796651 70 Factor=48796651,70,72
48959681 70 Factor=48959681,70,72
48961069 71 Factor=48961069,71,72
48962813 70 Factor=48962813,70,72
48973627 71 Factor=48973627,71,72
48985051 71 Factor=48985051,71,72
49011299 70 Factor=49011299,70,72
49432261 70 Factor=49432261,70,72
49482271 71 Factor=49482271,71,72
49653937 70 Factor=49653937,70,72
49798871 70 Factor=49798871,70,72
49972387 71 Factor=49972387,71,72
50010461 71 Factor=50010461,71,73
50015437 71 Factor=50015437,71,73
50071361 72 Factor=50071361,72,73
50134391 71 Factor=50134391,71,73
50183843 70 Factor=50183843,70,73
50231387 72 Factor=50231387,72,73
50251009 71 Factor=50251009,71,73
50440301 71 Factor=50440301,71,73
50454631 70 Factor=50454631,70,73
50478761 70 Factor=50478761,70,73
50485727 72 Factor=50485727,72,73
50495573 72 Factor=50495573,72,73
50501257 72 Factor=50501257,72,73
50629741 72 Factor=50629741,72,73
50630971 72 Factor=50630971,72,73
50644907 72 Factor=50644907,72,73
50673587 72 Factor=50673587,72,73
50681039 72 Factor=50681039,72,73
50736013 71 Factor=50736013,71,73
50736601 71 Factor=50736601,71,73
50794489 71 Factor=50794489,71,73
50803771 71 Factor=50803771,71,73
50815243 71 Factor=50815243,71,73
50836231 71 Factor=50836231,71,73
50856193 72 Factor=50856193,72,73
50980277 71 Factor=50980277,71,73
51070039 72 Factor=51070039,72,73
51082337 71 Factor=51082337,71,73
51113597 72 Factor=51113597,72,73
51157151 71 Factor=51157151,71,73
51221507 72 Factor=51221507,72,73
51262039 71 Factor=51262039,71,73
51279307 71 Factor=51279307,71,73
51310661 70 Factor=51310661,70,73
51350627 70 Factor=51350627,70,73
51398227 71 Factor=51398227,71,73
51608143 71 Factor=51608143,71,73
51728381 71 Factor=51728381,71,73
51734329 71 Factor=51734329,71,73
51767627 71 Factor=51767627,71,73
51971957 71 Factor=51971957,71,73
52036157 71 Factor=52036157,71,73
52040693 71 Factor=52040693,71,73
52046131 72 Factor=52046131,72,73
52208287 72 Factor=52208287,72,73
52271633 70 Factor=52271633,70,73
52432669 70 Factor=52432669,70,73
52451699 70 Factor=52451699,70,73
52502101 72 Factor=52502101,72,73
52573063 71 Factor=52573063,71,73
52578707 70 Factor=52578707,70,73
52706201 71 Factor=52706201,71,73
52708763 72 Factor=52708763,72,73
52765847 71 Factor=52765847,71,73
52795207 72 Factor=52795207,72,73
52840859 71 Factor=52840859,71,73
52842527 72 Factor=52842527,72,73
52941061 71 Factor=52941061,71,73
52946071 71 Factor=52946071,71,73
53024641 71 Factor=53024641,71,73
53080561 71 Factor=53080561,71,73
53229229 71 Factor=53229229,71,73
53246203 71 Factor=53246203,71,73
53251339 72 Factor=53251339,72,73
53280347 71 Factor=53280347,71,73
53311441 71 Factor=53311441,71,73
53326681 71 Factor=53326681,71,73
53487659 71 Factor=53487659,71,73
53584603 71 Factor=53584603,71,73
53614793 71 Factor=53614793,71,73
53702083 72 Factor=53702083,72,73
53731529 72 Factor=53731529,72,73
53792723 72 Factor=53792723,72,73
54025357 71 Factor=54025357,71,73
54025849 71 Factor=54025849,71,73
54033829 71 Factor=54033829,71,73
54110261 72 Factor=54110261,72,73
54181007 72 Factor=54181007,72,73
54182419 72 Factor=54182419,72,73
54190127 72 Factor=54190127,72,73
54264709 71 Factor=54264709,71,73
54296309 72 Factor=54296309,72,73
54316393 72 Factor=54316393,72,73
54363629 71 Factor=54363629,71,73
54371573 72 Factor=54371573,72,73
54375719 72 Factor=54375719,72,73
54379657 72 Factor=54379657,72,73
54387589 72 Factor=54387589,72,73
54395981 72 Factor=54395981,72,73
54399437 72 Factor=54399437,72,73
54412313 72 Factor=54412313,72,73
54412541 72 Factor=54412541,72,73
54447823 72 Factor=54447823,72,73
54468251 72 Factor=54468251,72,73
54474733 71 Factor=54474733,71,73
54485917 71 Factor=54485917,71,73
54590273 72 Factor=54590273,72,73
54669541 71 Factor=54669541,71,73
54920227 72 Factor=54920227,72,73
54930289 72 Factor=54930289,72,73
54954721 71 Factor=54954721,71,73
55079753 71 Factor=55079753,71,73
55135919 71 Factor=55135919,71,73
55254491 71 Factor=55254491,71,73
55280521 71 Factor=55280521,71,73
55347857 72 Factor=55347857,72,73
55421701 72 Factor=55421701,72,73
55675793 72 Factor=55675793,72,73
55696219 72 Factor=55696219,72,73
55932521 72 Factor=55932521,72,73
56061541 72 Factor=56061541,72,73
56127031 71 Factor=56127031,71,73
56153387 71 Factor=56153387,71,73
56343121 72 Factor=56343121,72,73
56382929 71 Factor=56382929,71,73
56437903 72 Factor=56437903,72,73
56510917 71 Factor=56510917,71,73
56525813 72 Factor=56525813,72,73
56537249 72 Factor=56537249,72,73
56540269 72 Factor=56540269,72,73
56608621 72 Factor=56608621,72,73
56676901 71 Factor=56676901,71,73
56680709 71 Factor=56680709,71,73
56680747 71 Factor=56680747,71,73
56707517 72 Factor=56707517,72,73
56874541 71 Factor=56874541,71,73
57857561 72 Factor=57857561,72,73
57936677 72 Factor=57936677,72,73[/CODE]

Madpoo 2015-10-24 18:26

[QUOTE=Madpoo;413651]Okay, here's a list of exponents where the system had more bad than good but need add'l TF work to get up to GPU72 goals.[/QUOTE]

Followed by 14 exponents that need a triple check but could use extra TF work as well. Half of them are already assigned to GPU72.
[CODE]exponent TF WorkToDo
43253059 71 Factor=43253059,71,72
43254349 71 Factor=43254349,71,72
45608513 70 Factor=45608513,70,72
46513769 70 Factor=46513769,70,72
46856921 70 Factor=46856921,70,72
47523583 70 Factor=47523583,70,72
47584919 70 Factor=47584919,70,72
47643977 70 Factor=47643977,70,72
47717653 70 Factor=47717653,70,72
48760711 70 Factor=48760711,70,72
49632463 70 Factor=49632463,70,72
49789249 71 Factor=49789249,71,72
49899973 71 Factor=49899973,71,72
51513167 71 Factor=51513167,71,73[/CODE]

chalsall 2015-10-24 19:34

[QUOTE=Madpoo;413651]Okay, here's a list of exponents where the system had more bad than good but need add'l TF work to get up to GPU72 goals.[/QUOTE]

OK, thanks. These are now all in the hopper and prioritized; 28 of those already assigned to GPU72 were already assigned for TF'ing (but not all 28 to the end goal).

Please feel free to bring forward such lists for additional priority TF'ing at any time, as your algorithms reveal additional suspect machines / candidates.

BTW, my machines have finished all of its SDC work (except for one, a high candidate on a slow machine). Please sir, may I have some more? :smile:

frmky 2015-10-25 07:29

[QUOTE=Madpoo;412980]52 of these below 38M. End result will be that, whatever the outcome, at least one of these previously "spotless" CPU's will have a bad result added to their account, which will hopefully feed into our strategic double checking:
[/QUOTE]
I just reserved the last of this TC list.

firejuggler 2015-10-25 16:27

Happy to report that
UID: firejuggler, M36417593 is not prime. Res64: 104739A81360C4__. We4: 968974F8,20897412,00000000, XXXXXXXXXX

and it doublecheck!

Mark Rose 2015-10-25 19:39

[QUOTE=chalsall;413655]OK, thanks. These are now all in the hopper and prioritized; 28 of those already assigned to GPU72 were already assigned for TF'ing (but not all 28 to the end goal).[/QUOTE]

A handful of those are currently assigned to me. I've prioritized them and they should be done in an hour or so.

ric 2015-10-26 15:15

Heads up
 
A brief heads-up, after the second batch of "anemic" <g> candidates:[LIST][*][URL="http://mersenneforum.org/showthread.php?t=20372&page=22"]#236 & #237[/URL] 6 candidates: 2 matching, 4 mismatches (of which 1 subsequently TC'd w/success);[*][URL="http://mersenneforum.org/showpost.php?p=411892&postcount=357"]#357[/URL] 8 candidates: 2 matching (had to be, they were both from [STRIKE]Chuck Norris[/STRIKE] ahem... ), 6 mismatches.[/LIST]Next round of other 6 cands to be started in a few days, ETC around Nov 10th. Will update as they complete.

chalsall 2015-10-26 18:35

[QUOTE=chalsall;413090]I had asked before, and I'll now ask again: Are any of the machines which are less than perfect still being assigned LL candidates? If so, might they be assigned DC Cat 2 for a while?

Separately, as a policy, should perhaps all machines requesting LL work be assigned a small percentage of DC Cat 1 work, just in order to ensure their ongoing sanity?[/QUOTE]

:hello:

Prime95 2015-10-26 19:28

Any machines that are set as "do what makes the most sense" gets a small percentage of DC work -- I'd have to look it up. A default install of prime95 will have this work preference. As an aside, I think we should up the percentage of DC work as DC is falling behind the LL wavefront by more than I like.

I'm reluctant to override settings for those that have explicitly selected first-time LL tests.


Edit: Percentage was 10%, now it is 20%.

LaurV 2015-10-27 01:24

[QUOTE=Prime95;413846]Edit: Percentage was 10%, now it is 20%.[/QUOTE]
:tu: We salute that!

cuBerBruce 2015-10-27 13:57

[QUOTE=Madpoo;413184]Here's more. From machines with at least a 3:1 bad:good ratio, using the "by calendar year" method:
[CODE]exponent Bad Good Unk Sus Solo Mis worktodo
...
51981047 7 2 11 8 11 8 DoubleCheck=51981047,73,1
...[/CODE][/QUOTE]

My residue matched on this one.

Madpoo 2015-10-28 15:59

Small set of work
 
Here's a fun and weird group of exponents up for grabs.

These are from machines with zero good, one or more bad, and only one solo-checked exponent left.

Kind of a crapshoot here... not enough data to say one way or another if their lone solo-checked thing will be good or bad except for the fact that they haven't had a good one yet. But as you can see some of them only have one bad one, so who knows. :smile:

There are a couple in there that would have been swept up with my 3:1 bad/good ratio query as well so I expect we'll have a better clue as to how those turn out.

[CODE]exponent Bad Good Unk Sus Solo Mis worktodo
34902599 1 0 1 0 1 0 DoubleCheck=34902599,71,1
34942693 1 0 1 0 1 0 DoubleCheck=34942693,71,1
35664833 1 0 2 0 1 1 DoubleCheck=35664833,71,1
35790331 1 0 1 0 1 0 DoubleCheck=35790331,71,1
37702097 4 0 2 0 1 1 DoubleCheck=37702097,71,1
38128627 3 0 1 1 1 1 DoubleCheck=38128627,71,1
42270829 1 0 1 0 1 0 DoubleCheck=42270829,72,1
43728611 1 0 1 0 1 0 DoubleCheck=43728611,72,1
44130049 1 0 1 0 1 0 DoubleCheck=44130049,72,1
44253581 1 0 1 0 1 0 DoubleCheck=44253581,72,1
46233977 1 0 1 0 1 0 DoubleCheck=46233977,72,1
46289773 1 0 1 0 1 0 DoubleCheck=46289773,72,1
47139277 2 0 1 6 1 6 DoubleCheck=47139277,72,1
50582549 1 0 1 0 1 0 DoubleCheck=50582549,71,1
52628813 1 0 1 0 1 0 DoubleCheck=52628813,72,1[/CODE]

chalsall 2015-10-28 16:10

[QUOTE=Madpoo;414089]Here's a fun and weird group of exponents up for grabs.[/QUOTE]

I'll take them.

cuBerBruce 2015-10-28 16:10

[QUOTE=Madpoo;414089]Here's a fun and weird group of exponents up for grabs.
...
[CODE]exponent Bad Good Unk Sus Solo Mis worktodo
34902599 1 0 1 0 1 0 DoubleCheck=34902599,71,1
...[/CODE][/QUOTE]

I took this first one.

chalsall 2015-10-28 16:29

[QUOTE=cuBerBruce;414091]I took this first one.[/QUOTE]

We cross posted...

I've taken all but the first; the first is yours.

cuBerBruce 2015-10-28 16:33

[QUOTE=chalsall;414093]We cross posted...

I've taken all but the first; the first is yours.[/QUOTE]

Yes, I noticed that. I grabbed, then posted. I guess you posted, then tried to grab.

chalsall 2015-10-28 16:41

[QUOTE=cuBerBruce;414094]Yes, I noticed that. I grabbed, then posted. I guess you posted, then tried to grab.[/QUOTE]

Yes. It actually worked out perfectly... I have 14 available high-speed(ish) CPUs, but there were 15 in the list. I tried to put the lowest one on a slower machine, but noticed it wasn't available.

lycorn 2015-10-29 08:26

@Madpoo,
I´ve noticed a fair amount of 2M exponents that have been very recently triple checked by you. As the previous runs were matching, and they appear legit, I´m curious about the criteria used to select those ones for TC.

cuBerBruce 2015-10-30 06:28

[QUOTE=Madpoo;414089]Here's a fun and weird group of exponents up for grabs.
...
[CODE]exponent Bad Good Unk Sus Solo Mis worktodo
34902599 1 0 1 0 1 0 DoubleCheck=34902599,71,1
...
[/CODE][/QUOTE]

The residue I got for this one did not match.

sdbardwick 2015-10-30 07:07

[QUOTE=cuBerBruce;414277]The residue I got for this one did not match.[/QUOTE]
Ok, I'll run the TC on M34902599.

LookAS 2015-10-30 09:13

I'd welcome TC these exponents:

M44110853, M48077509, M49134083, M49504331

chalsall 2015-10-30 16:25

[QUOTE=cuBerBruce;414277]The residue I got for this one did not match.[/QUOTE]

The lowest six of my runs from this batch have now completed; three matched, three didn't.

chalsall 2015-10-30 16:57

[QUOTE=LookAS;414284]I'd welcome TC these exponents:[/QUOTE]

I took them. Check them in ~60 hours.

cuBerBruce 2015-10-30 19:07

[QUOTE=UBR47K;412432]Requesting DC on:
...
Doublecheck=N/A,58475341,75,1
Doublecheck=N/A,58491773,75,1
...
[/QUOTE]
I'm doing these two.

Madpoo 2015-10-31 01:25

[QUOTE=lycorn;414177]@Madpoo,
I´ve noticed a fair amount of 2M exponents that have been very recently triple checked by you. As the previous runs were matching, and they appear legit, I´m curious about the criteria used to select those ones for TC.[/QUOTE]

Just the fact that they hadn't been triple-checked already. :smile:

Similar to how I did triple checks of everything below 2M. No real purpose to it except just to confirm that earlier versions and newer versions of the software agree.

In this case, I don't really plan on doing all of the 2M-3M range, but I picked out the first 112 so I could do a test with my 28-core box. I set it up with 28 workers on one core each with 4 assignments, and let 'er rip.

Sad news there was that having 14 cores on the same CPU, even the small ones (which is why I picked these little 2M guys) was pretty lousy. When running a single worker it did something like 0.35 ms per iteration, but if all the cores were going it jumped to around 0.45 ms / iter.

I experimented with varying amounts of workers and cores-per-worker and just settled on 2 cores per worker which had per iteration times of something like 0.22 ms. Essentially no different throughput than having them all going with one core each, but oh well.

I let them finish out the 112 assignments and called it good. :smile:

LaurV 2015-10-31 16:14

One of the three 59M's which I took, just ended with a [URL="http://www.mersenne.org/report_exponent/?exp_lo=59662607&full=1"]mismatch[/URL]. Free to be taken, I won't triple check it.
The other two still going for a couple of hours.
I was curious if doing this activity increases my number of "bad results". In which case I won't report the other if they mismatches. It seems like not, the exponent appears in my "unverified" list, and not in the "bat LL" list, which is ok. Point scored for the server this time... :smile:
(I looked into that because in the past there was a bug where the mismatched result was recorded as bad, and if someone TC and proves your result good and the older bad, then your "bad LL" list wasn't amended, retaining the "wrong" "bad result" for you - it seems like that bug is solved now).

lycorn 2015-10-31 16:53

[QUOTE=Madpoo;414374]No real purpose to it except just to confirm that earlier versions and newer versions of the software agree.

[/QUOTE]

Fair enough. Thx for answering.

chalsall 2015-10-31 17:37

[QUOTE=Madpoo;413651]Okay, here's a list of exponents where the system had more bad than good but need add'l TF work to get up to GPU72 goals.[/QUOTE]

OK, so you know Aaron, most of these suspect candidates which needed additional TF'ing have now completed. There are still [URL="https://www.gpu72.com/reports/available/sdc/"]40 to go[/URL] at the moment -- these should trickle in in the next few days -- but this gives you a couple of hundred to offer for SDC'ing.

Tangentially... might it make sense to have a page on Primenet which your algorithm(s) list as desirous of a SDC? Not necessarily linked anywhere from Primenet itself, but posted here. This would allow those of us interested in helping out to request such candidates (by placing them in our worktodo files and ensuring we get a valid AID before proceeding) without bothering you to post additional candidates to be worked.

Just a thought.

sdbardwick 2015-10-31 22:34

[QUOTE=cuBerBruce;414277]The residue I got for this one did not match.[/QUOTE]
[URL="http://www.mersenne.org/report_exponent/?exp_lo=34902599&full=1"]We match[/URL].

Madpoo 2015-11-01 01:37

[QUOTE=LaurV;414419]... Free to be taken, I won't triple check it. [/QUOTE]
Good... I'm already doing triple-checks on 3 more of your self-verified work. :smile:

You may have noticed that I picked up a bunch of your older results and did a double-check for you, free of charge. LOL

[QUOTE=LaurV;414419]I was curious if doing this activity increases my number of "bad results". In which case I won't report the other if they mismatches. It seems like not, the exponent appears in my "unverified" list, and not in the "bat LL" list, which is ok. Point scored for the server this time... :smile:[/QUOTE]

Not sure how it used to work, but right now a mismatch will keep both of them set as "unverified" just like it was a first time check.

The only other thing that *could* happen is being marked "suspect" if it had certain errors along the way, but even then it's not marked as bad or anything until it's either verified for better or worse.

In a looser sense, I mentioned before that when I'm looking for potentially bad systems, in cases like this where there's a mismatch, I look at the total history of both systems and if one of them has been consistently good, I assume the other one is bad (just for purposes of my analysis of course, not actually marking the other result bad in the DB or anything).

In this particular instance for exponent M59662607, your system (for 2015) has a track record of 14 good, zero bad. Thumbs up. The other one has 1 good and 5 bad... 2 suspect results and 3 total mismatches. Ouch... thumbs down.

I still use my own rule of requiring 40 good and zero bad before I assume a machine that contributed one result of a mismatched set is the correct one. I should probably loosen that up to 10 good, zero bad and see what kind of newly "bad" machines I might find. And now that I only count a machine's results for a single calendar year, it's a little harder for some to meet that 40 result threshold... hmm... Yeah, I should figure out what constitutes a "reliable" machine, or maybe include *all* of the years for that system, not just one year.

Madpoo 2015-11-01 02:04

[QUOTE=sdbardwick;414459][URL="http://www.mersenne.org/report_exponent/?exp_lo=34902599&full=1"]We match[/URL].[/QUOTE]

Ah, the famous "Phil Frakes" account was the bad one there. When I first started this project out on my own, looking for bad systems, that was the very first account I identified as having an unusually high amount of bad stuff.

Since then I've broken it down more so instead of just by the user's account, I look at the CPU and now also by year. By those breakdowns his account isn't as bad as I originally thought, but of the 211 total records (account/cpu/year) for him, only 27 of those had more bad than good. Total, he has 721 good, 173 bad, and another 261 unknown.

Not the worst track record out there in terms of the hit/miss ratio.

So, not to single him out or anything, but it did trigger my interest in the idea and helped me refine the analysis.

LaurV 2015-11-01 05:42

[QUOTE=Madpoo;414469]Good... I'm already doing triple-checks on 3 more of your self-verified work. :smile:
You may have noticed that I picked up a bunch of your older results and did a double-check for you, free of charge. LOL
[/QUOTE]
You were very welcome to do that for "self-doublechecked" part of my old work, especially for v4 work (where I was contributed with over 50 computers, but gave up when EFF money were awarded - some of those computers [U]did[/U] produce bad results, and DC/TC made sense).

But it makes no sense for you to continue to TC my own, [U]new[/U] self-DC work, especially [B][U]if[/U][/B] the first check and the DC came in the same time day:hour:minute [U]and[/U] at least one of them has a valid UID, [U]and[/U] they have different shifts*. Which is almost everything above 50M. You just waste your time, you can use the resources to do a more useful job. Enough people knows me here already to say I won't "cheat", I am a "credit whore" but I don't go over some borders :razz:

Moreover, I have no reason to "hide" a prime. :loco:

I was doing LL tests in parallel, in different hardware, with different shifts, and the tests were combined in such a way to take about the same time, so I can compare the residues at every check point, like 1M or 100k iterations, or so, depending on the exponent, if they are different, then BOTH tests will resume from a previous checkpoint. If they match, previous checkpoints are deleted (always only the last two are kept, this in case the last one had errors when the file was written, but never happened). This is as simple as comparing file names, because cudaLucas saves the residue in the file name (I had long arguing with people compiling/working on cudaLucas, to get this feature inside, you may not remember, you are kind of "newer" here :wink:).

And this saves me a lot of time, and [U]for the project too[/U], because when a mismatch is found, no time is wasted - contrarily to the classical work path when a test with a mistake in the beginning will still continue to the end, and be reported, to find the mismatch, for both LL and DC, a lot of wasted time.

When the tests end with a match, I report both. Because I have done both. I used two times the amount of hardware, even if the time was spent only once. Why should I not get credit for both? This is (partially) the reason we pushed to implement "shifting" in cudaLucas. The reports come in the same time.

I also proposed (in fact, sustained, it was not my proposition, but other's forum member) in the past to implement a feature on the server side to be able to mark an exponent for TC, with a low priority. If that would be implemented, you could use it to mark all my self-DC work, instead of wasting your time to TC it.

I can understand that my hardware can produce crap and some of my tests are bad, so other people will DC them and find them bad, but under no circumstance I would accept the fact that a self-DC which I did and matched could be found bad. You TC them if you like, is your time and resources.

And I will continue to do that, in spite of what you or others say, because I feel better to know that I didn't miss a prime :razz:, and I didn't report a wrong result by mistake.

----------
* all these conditions have some importance to avoid the case when someone can report fake results using my name, but I don't think anybody has any reason, and would waste his/her time to do such a stupid thing anyhow.

chalsall 2015-11-01 14:26

[QUOTE=chalsall;414324]The lowest six of my runs from this batch have now completed; three matched, three didn't.[/QUOTE]

OK, all fourteen from this batch have now completed. Five matched, nine didn't. Looks like you're onto something here Aaron.

Madpoo 2015-11-01 19:26

[QUOTE=chalsall;414493]OK, all fourteen from this batch have now completed. Five matched, nine didn't. Looks like you're onto something here Aaron.[/QUOTE]

Woot! :smile:

Here are some more. These are 3:1 bad/good and for this query, I loosened up what I consider an "awesome" system to include anything with 15 good and zero bad. Most are from a single new system that has 10 solo-checked stuff... If someone grabs all of these, I might suggest trying that smallest one first just as a trial to get an idea of how the rest might go. Still, that same system has 7 suspect/7 mismatches which means all of those were probably bad too and just awaiting a triple-check.

In fact, when the # of suspect is close to the # of mismatches, that means someone double-checked their suspect results and got a different result, so more often than not (like 90%+ of the time) those will be bad...you can mentally include those in the bad column if it helps. LOL

[CODE]exponent Bad Good Unk Sus Solo Mis worktodo
43101913 3 0 4 0 4 0 DoubleCheck=43101913,72,1
43102177 3 0 4 0 4 0 DoubleCheck=43102177,72,1
43985531 4 1 2 2 2 2 DoubleCheck=43985531,72,1
44802613 3 1 10 7 10 7 DoubleCheck=44802613,72,1
44954269 3 1 10 7 10 7 DoubleCheck=44954269,72,1
45678811 3 1 10 7 10 7 DoubleCheck=45678811,72,1
46013971 3 1 10 7 10 7 DoubleCheck=46013971,72,1
46296641 4 1 2 2 2 2 DoubleCheck=46296641,72,1
47973161 3 1 10 7 10 7 DoubleCheck=47973161,72,1
48686327 3 1 10 7 10 7 DoubleCheck=48686327,72,1
48761711 3 1 10 7 10 7 DoubleCheck=48761711,72,1
49334413 3 1 10 7 10 7 DoubleCheck=49334413,72,1
49482271 3 1 10 7 10 7 DoubleCheck=49482271,72,1[/CODE]

chalsall 2015-11-01 21:23

[QUOTE=Madpoo;414509]Here are some more.[/QUOTE]

Mine.

Madpoo 2015-11-02 02:08

[QUOTE=LaurV;414479]You were very welcome to do that for "self-doublechecked" part of my old work, especially for v4 work (where I was contributed with over 50 computers, but gave up when EFF money were awarded - some of those computers [U]did[/U] produce bad results, and DC/TC made sense).[/QUOTE]

I'm just OCD about it I guess. Nothing personal. I know you wouldn't cheat, but on general principle I think it "looks cool" if different a different person does the double-check.

And yeah, you do have the occasional bad result. In fact, if you want, you can do a triple-check of this one I just turned in a result for: :smile:

[URL="http://www.mersenne.org/M45196597"]M45196597[/URL]

LaurV 2015-11-02 05:36

[QUOTE=chalsall;414493]Looks like you're onto something here Aaron.[/QUOTE]
Agree, my [URL="http://www.mersenne.org/report_exponent/?exp_lo=59273327&full=1"]second one[/URL] is also a mismatch, up to now, two from two. Interesting they were done by the same user before. They are both still unassigned.

[QUOTE=Madpoo;414544]And yeah, you do have the occasional bad result. In fact, if you want, you can do a triple-check of this one I just turned in a result for: :smile:
[URL="http://www.mersenne.org/M45196597"]M45196597[/URL][/QUOTE]
Sure boss! Queued! Thanks! [STRIKE](it seems that was not a GPU test, which computer does it come from?)[/STRIKE] scrap that, it was a GPU test, I see from James' site. I have to go home to look into the logs to see which card sent it (in few hours. Lunch break here).

endless mike 2015-11-02 08:14

[QUOTE=endless mike;413175]Maybe a few more[/QUOTE]

I'm halfway through the 24 triple checks that I took earlier in the thread; no quadruple checks needed yet. It does mean that Madpoo should hopefully have a few more machines for his list that are no longer perfect.

henryzz 2015-11-02 09:42

@Madpoo It might make sense if you import the data into R and do some machine learning to find candidates. Logistic regression and random forests would be fairly decent for this.

firejuggler 2015-11-02 20:09

a long overdue doublecheck on M36433591

Verified 2007-09-13 Yu Peng Chen 4103C5EBC1F27D48
Verified 2015-11-02 firejuggler 4103C5EBC1F27D48

LaurV 2015-11-03 12:09

[QUOTE=Madpoo;414544][URL="http://www.mersenne.org/M45196597"]M45196597[/URL][/QUOTE]
And done, matching yours.
Note that the period is mid-of-Jan to mid-of-March 2012, when I was testing cudaLucas (switching from powers of 2 to non powers of two), as we already discussed. So this is not a hardware failure, for the peace of my heart :razz:
You may find more like that, from the same period.

Next time however, when you find a mistake in my tests, you should find a prime! Otherwise we both are wasting the time... :razz:

Mark Rose 2015-11-03 12:14

[QUOTE=LaurV;414794]Next time however, when you find a mistake in my tests, you should find a prime! Otherwise we both are wasting the time... :razz:[/QUOTE]

Always good advice.


All times are UTC. The time now is 07:17.

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2021, Jelsoft Enterprises Ltd.