VALIDATION INCONCLUSIVE

Message boards : Number crunching : VALIDATION INCONCLUSIVE
Message board moderation

To post messages, you must log in.

Previous · 1 · 2 · 3 · 4 · 5 · Next

AuthorMessage
Profile Uli
Volunteer tester
Avatar

Send message
Joined: 6 Feb 00
Posts: 10923
Credit: 5,996,015
RAC: 1
Germany
Message 1297808 - Posted: 22 Oct 2012, 6:22:24 UTC

Set and forget was pre Boinc. The rules have changed. I am very happy to see you all trying to alert Wingman. Brad woke up Eric L, the starter of this thread. If you are not willing to help, so be it, but please don't diss the rest of us.
Pluto will always be a planet to me.

Seti Ambassador
Not to late to order an Anni Shirt
ID: 1297808 · Report as offensive
Profile Tron

Send message
Joined: 16 Aug 09
Posts: 180
Credit: 2,250,468
RAC: 0
United States
Message 1297812 - Posted: 22 Oct 2012, 7:02:47 UTC

who left the catnip out! ..... see what happens ... (jk mark we love you)

In my opinion, Your right and wrong , those computers running day and night pumping out invalids, seem to be folks who are interested in improving the project by donating their computers spare cycles ,if those cycles are bogus I'd have to think they would want to know and a heads up to be welcome.

No one 'likes' to waste, especially now days with the 'green' movement in full swing.

So where is this thread of offending pm's ? I have one to add , maybe more if I get bored. The inclusiveness is piling up in my history and 90% are gtx 560 both ti and non versions.


My pm went something like
Hello
Your computer participating in the Seti@home Project #010101010 <link>
has been producing more invalid work than valid. Please visit <link> number crunching forum <link> if you need assistance in resolving this problem
Thank You and Regards
Tron
ID: 1297812 · Report as offensive
Profile Uli
Volunteer tester
Avatar

Send message
Joined: 6 Feb 00
Posts: 10923
Credit: 5,996,015
RAC: 1
Germany
Message 1297814 - Posted: 22 Oct 2012, 7:15:24 UTC

Thank you Tron.
Pluto will always be a planet to me.

Seti Ambassador
Not to late to order an Anni Shirt
ID: 1297814 · Report as offensive
David S
Volunteer tester
Avatar

Send message
Joined: 4 Oct 99
Posts: 18352
Credit: 27,761,924
RAC: 12
United States
Message 1297875 - Posted: 22 Oct 2012, 13:42:58 UTC

I checked the current stats for my i7. It has 24 inconclusives, which are split evenly 12-12 between CPU and GPU. I looked at one of each at random (okay, not at random, the first of each in the list). Nobody found 30 spikes in either of them. In this one, I found 6 spikes and my partner found 5. In this one, I found 8 spikes and my partner 7, plus we both found 1 triplet.

Considering that I have 256 pending (not (yet) inconclusive) and 129 valid, I'm not overly concerned over 24 inconclusives.

Someone remind me, though: when those eventually get their 3rd opinions in, if they agree with the other guy and not me, will I still get credit?

David
Sitting on my butt while others boldly go,
Waiting for a message from a small furry creature from Alpha Centauri.

ID: 1297875 · Report as offensive
Profile skildude
Avatar

Send message
Joined: 4 Oct 00
Posts: 9541
Credit: 50,759,529
RAC: 60
Yemen
Message 1297893 - Posted: 22 Oct 2012, 15:22:34 UTC - in response to Message 1297875.  

no


In a rich man's house there is no place to spit but his face.
Diogenes Of Sinope
ID: 1297893 · Report as offensive
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 14650
Credit: 200,643,578
RAC: 874
United Kingdom
Message 1297899 - Posted: 22 Oct 2012, 15:48:55 UTC - in response to Message 1297893.  

no

But a fairly common outcome is for all three to be close enough, and all three to get cedit.
ID: 1297899 · Report as offensive
Josef W. Segur
Volunteer developer
Volunteer tester

Send message
Joined: 30 Oct 99
Posts: 4504
Credit: 1,414,761
RAC: 0
United States
Message 1297902 - Posted: 22 Oct 2012, 15:53:09 UTC - in response to Message 1297875.  
Last modified: 22 Oct 2012, 15:59:33 UTC

I checked the current stats for my i7. It has 24 inconclusives, which are split evenly 12-12 between CPU and GPU. I looked at one of each at random (okay, not at random, the first of each in the list). Nobody found 30 spikes in either of them. In this one, I found 6 spikes and my partner found 5. In this one, I found 8 spikes and my partner 7, plus we both found 1 triplet.

Considering that I have 256 pending (not (yet) inconclusive) and 129 valid, I'm not overly concerned over 24 inconclusives.

Someone remind me, though: when those eventually get their 3rd opinions in, if they agree with the other guy and not me, will I still get credit?

With the existing validators, credit is granted if at least half the signals match. You'll probably get credit, though the counts of course don't guarantee that any of your spikes match the wingmate's. For this kind of situation your result may strongly match the third one, in which case yours will become the canonical result and only the original wingmate will be in danger of not getting credit.

However, Eric Korpela is at least considering tightening the "weakly similar" criteria, see his post 43994 at SETI Beta.
                                                                  Joe
ID: 1297902 · Report as offensive
Profile David Anderson (not *that* DA) Project Donor
Avatar

Send message
Joined: 5 Dec 09
Posts: 215
Credit: 74,008,558
RAC: 74
United States
Message 1297919 - Posted: 22 Oct 2012, 17:08:13 UTC

I was alarmed at my 100 Inconclusive (1800+ pending) and
did find a couple wingmen with rather bad records, meaning
10 times as many Invalid as Valid. So I PM'd them today.
Hopefully it gets them to ask for help, and soon.

ID: 1297919 · Report as offensive
kittyman Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 9 Jul 00
Posts: 51468
Credit: 1,018,363,574
RAC: 1,004
United States
Message 1297920 - Posted: 22 Oct 2012, 17:14:01 UTC
Last modified: 22 Oct 2012, 17:30:18 UTC

This morning, after the kitty farm crunched hard all weekend, I have 7,193 pending, 1,241 validation inconclusive, 7 actual invalids, and 168 errors. Oh, I forgot the 5,708 valid results. Almost all of the errors were those 5 minutes 'till deadline resends that are impossible to return in time.
"Freedom is just Chaos, with better lighting." Alan Dean Foster

ID: 1297920 · Report as offensive
musicplayer

Send message
Joined: 17 May 10
Posts: 2430
Credit: 926,046
RAC: 0
Message 1298029 - Posted: 23 Oct 2012, 0:30:03 UTC
Last modified: 23 Oct 2012, 0:31:06 UTC

I had a couple of inconclusives myself so I checked my most recent tasks against my wingmen before finishing off yesterday.

Maybe a little tempting perhaps, but for some of these tasks I may get a sense of what is the more correct results based on my earlier results (at least for the CUDA-tasks where I had 30 spikes several times).

If my wingman possibly had 30 spikes, my guess is that my result may have been slightly better if I found at least one pulse or triplet (and possibly gaussian) in addition in such a task. As usual, the spike count could be anything better than 0 as well.
ID: 1298029 · Report as offensive
Profile Gary Charpentier Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 25 Dec 00
Posts: 30608
Credit: 53,134,872
RAC: 32
United States
Message 1298068 - Posted: 23 Oct 2012, 3:13:54 UTC

Just checked myself, not a big cruncher but
740 tasks, 307 in progress, 346 Pending, 24 inconclusive, 63 valid, 0 invalid, 0 error. 20 of those inconclusive are the same wing man.

I'm thinking that Seti needs to "punish" bad hosts more severely than it presently does and restrict their ability to get work. This would cut down on waste bandwidth and wasted disk space. It would even be green if some of these hosts went into idle and drew less power.

As to BOINC, I'm thinking it needs a new feature that is, if say more than 25% of the work units come up invalid and error, but not timed out, it automatically puts a notice up on the notices tab to that effect and suggests there may be a handware/software issue the user should look into. Not just for Seti but all projects.

ID: 1298068 · Report as offensive
Profile Tron

Send message
Joined: 16 Aug 09
Posts: 180
Credit: 2,250,468
RAC: 0
United States
Message 1298070 - Posted: 23 Oct 2012, 3:23:06 UTC - in response to Message 1298068.  

Gary Charpenter wrote:

I'm thinking that Seti needs to "punish" bad hosts more severely than it presently does


*me Prepares the whipping post ;P
ID: 1298070 · Report as offensive
David S
Volunteer tester
Avatar

Send message
Joined: 4 Oct 99
Posts: 18352
Credit: 27,761,924
RAC: 12
United States
Message 1298211 - Posted: 23 Oct 2012, 14:30:13 UTC

Okay, you guys got me curious enough to check into my wingmates on my inconclusives. My number has dropped by 1 to 23 (I guess this means the 24th unit's 3rd user validated one or both of us; I show no invalids and no new errors). It turns out I do have several with the same partners.

I share 3 inconclusives each with:
5567915 (8600GTS, 1906 inc., 365 inv., 207 error)
6598140 (560 Ti, 3908 inc., 35 valid, 592 inv., 194 error)
6611681 (550 Ti, driver 280.26, 43 inc., 0 inv., 3 error)

2 with:
6700211 (560 Ti, driver 285.62, didn't note other stats)

6 with:
5564597 (Radeon HD 79x0 series, 108 inc., 0 inv., 37 error)

and 1 each with 6 other hosts; notable among these are 2 560 Ti GPUs, a GTX 570, and a GT 220 and a GT 440 both running driver 296.10. For the record, I also have a GT 440, but a newer driver.

I note that, among the hosts whose stats I wrote down, there are 2 really egregious ones in terms of invalids and errors. One of these is a 560 Ti, one is not. I also note that even the ones with high numbers of inconclusives mostly do not have high numbers of invalids and errors, so I tend to think there may be something wonky in the work itself lately. I think it's a strong enough possibility that at least Eric should look into it.

Sorry, but at this point I feel I've wasted enough work time and don't feel inclined to PM any of these guys. If anyone else cares to, go ahead.

David
Sitting on my butt while others boldly go,
Waiting for a message from a small furry creature from Alpha Centauri.

ID: 1298211 · Report as offensive
David S
Volunteer tester
Avatar

Send message
Joined: 4 Oct 99
Posts: 18352
Credit: 27,761,924
RAC: 12
United States
Message 1298213 - Posted: 23 Oct 2012, 14:36:24 UTC - in response to Message 1298211.  

Sorry, but at this point I feel I've wasted enough work time and don't feel inclined to PM any of these guys. If anyone else cares to, go ahead.

I just checked the list of users already PMed and 3 of mine are on it.

David
Sitting on my butt while others boldly go,
Waiting for a message from a small furry creature from Alpha Centauri.

ID: 1298213 · Report as offensive
Profile Tron

Send message
Joined: 16 Aug 09
Posts: 180
Credit: 2,250,468
RAC: 0
United States
Message 1298246 - Posted: 23 Oct 2012, 22:39:28 UTC - in response to Message 1298213.  


I just checked the list of users already PMed and 3 of mine are on it.


PM's are being sent only to users with greater invalid than valid tasks, reported within the last 3 days.
Errors are not included, nor are inconclusive.
ID: 1298246 · Report as offensive
David S
Volunteer tester
Avatar

Send message
Joined: 4 Oct 99
Posts: 18352
Credit: 27,761,924
RAC: 12
United States
Message 1298709 - Posted: 25 Oct 2012, 19:10:58 UTC

I poked into a few more inconclusives this morning and found one where the 3rd host (mine) didn't validate either of the first two and it's now out to a fourth. http://setiathome.berkeley.edu/workunit.php?wuid=1086502146

The first host and I agree on the spikes/pulses/gaussians/triplets, while the second came back with wildly different numbers. I guess something else between the first and mine disagrees, thus it's out to a fourth.

Whoa! I just checked on these other hosts. The first one has a few more inconclusives than I do, but the second has a huge number. 252 pending, 282 inconclusive, 8 valid, 59 invalid, 1 error. The owner hasn't been PMed, but he definitely needs to be. He is NOT running a 560 Ti.

David
Sitting on my butt while others boldly go,
Waiting for a message from a small furry creature from Alpha Centauri.

ID: 1298709 · Report as offensive
Profile MusicGod
Avatar

Send message
Joined: 7 Dec 02
Posts: 97
Credit: 24,782,870
RAC: 0
United States
Message 1298732 - Posted: 25 Oct 2012, 20:18:29 UTC

I too had this problem from the last batch....it came up with 425 inconclusives and I don`t know why....never had that happen before....
ID: 1298732 · Report as offensive
Profile skildude
Avatar

Send message
Joined: 4 Oct 00
Posts: 9541
Credit: 50,759,529
RAC: 60
Yemen
Message 1298738 - Posted: 25 Oct 2012, 20:34:31 UTC
Last modified: 25 Oct 2012, 20:37:23 UTC

I noticed that I had quite a few invalids 200+. The first one I picked had http://setiathome.berkeley.edu/hosts_user.php?userid=9319681 having a majority of his ATI GPU WU's being pending or inconclusive.


In a rich man's house there is no place to spit but his face.
Diogenes Of Sinope
ID: 1298738 · Report as offensive
Profile shizaru
Volunteer tester
Avatar

Send message
Joined: 14 Jun 04
Posts: 1130
Credit: 1,967,904
RAC: 0
Greece
Message 1298740 - Posted: 25 Oct 2012, 20:41:54 UTC - in response to Message 1298709.  

He is NOT running a 560 Ti.


(This might have been mentioned and/or disproven)

I had a quick look at the thread with the PMs and it looks like there are TWO scenarios:

a) either 560 Ti

b) or 300 series driver
ID: 1298740 · Report as offensive
David S
Volunteer tester
Avatar

Send message
Joined: 4 Oct 99
Posts: 18352
Credit: 27,761,924
RAC: 12
United States
Message 1298958 - Posted: 26 Oct 2012, 14:33:15 UTC - in response to Message 1298740.  

He is NOT running a 560 Ti.


(This might have been mentioned and/or disproven)

I had a quick look at the thread with the PMs and it looks like there are TWO scenarios:

a) either 560 Ti

b) or 300 series driver

I thought it was the 290 series drivers that were bad. I have a 300 series myself and a relatively low number of inconclusives, no invalids, and only timeout errors.

I see I have one new inconclusive today, and the wingmate on it is giving huge numbers of bad returns. (He's not running a 560 Ti either; it's an ATI.)

Okay, I finally bit the bullet and sent him (her?) a PM myself instead of trying to palm it off on you guys.

David
Sitting on my butt while others boldly go,
Waiting for a message from a small furry creature from Alpha Centauri.

ID: 1298958 · Report as offensive
Previous · 1 · 2 · 3 · 4 · 5 · Next

Message boards : Number crunching : VALIDATION INCONCLUSIVE


 
©2024 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.