Wondering why host gets Invalid


log in

Advanced search

Message boards : Number crunching : Wondering why host gets Invalid

Author Message
Profile David Anderson (not *that* DA)Project donor
Avatar
Send message
Joined: 5 Dec 09
Posts: 111
Credit: 23,389,311
RAC: 5,644
United States
Message 1385254 - Posted: 27 Jun 2013, 23:40:49 UTC

Host : http://setiathome.berkeley.edu/results.php?hostid=5940343
GPU/s : NVIDIA GeForce GTX 570 (1279MB)
State : All (1567) · In progress (112) · Validation pending (506) · Validation inconclusive (136) · Valid (113) · Invalid (699) · Error (1)
pm : Impossible. Unclear why Invalid.


task 3053496078 wu 1270505166

Idly I was wondering why a wingman with Inconclusive was getting mostly
Invalid, and looking at this wu and task left me puzzled.
I am not involved in this wu, it was just one I happened to look at
from the host Invalid list
and ... no obvious error. But marked Invalid.

Idle curiosity.

Sten-Arne
Volunteer tester
Send message
Joined: 1 Nov 08
Posts: 3752
Credit: 21,472,107
RAC: 14,692
Sweden
Message 1385256 - Posted: 27 Jun 2013, 23:48:05 UTC - in response to Message 1385254.
Last modified: 28 Jun 2013, 0:17:14 UTC

Host : http://setiathome.berkeley.edu/results.php?hostid=5940343
GPU/s : NVIDIA GeForce GTX 570 (1279MB)
State : All (1567) · In progress (112) · Validation pending (506) · Validation inconclusive (136) · Valid (113) · Invalid (699) · Error (1)
pm : Impossible. Unclear why Invalid.


task 3053496078 wu 1270505166

Idly I was wondering why a wingman with Inconclusive was getting mostly
Invalid, and looking at this wu and task left me puzzled.
I am not involved in this wu, it was just one I happened to look at
from the host Invalid list
and ... no obvious error. But marked Invalid.

Idle curiosity.


It's invalid because he's been tampering with his app_info, to run v7 with an old v6 app. Notice that his result doesn't have any Autocorr count line.

Edit: there seems to be no end to what people can come up with when it comes to cheating, to get more credits. Yes, the old v6 app can run v7 and it does it faster, because it doesn't do any autocorr. However the science becomes invalid. This was caught some time ago, so Eric made sure, those results becomes invalid immediately, instead of going into the science database and corrupting the science.

Edit 2: Looking closer to his results, it seems that he manages to sneak some of them into the science database too, despite that Eric did something to stop that from happening. He even manages to get some as canonical results. Example: http://setiathome.berkeley.edu/workunit.php?wuid=1270273091
____________

Profile betregerProject donor
Avatar
Send message
Joined: 29 Jun 99
Posts: 2661
Credit: 5,506,207
RAC: 6,695
United States
Message 1385257 - Posted: 27 Jun 2013, 23:50:24 UTC - in response to Message 1385254.

A validate error could be the cause. As I understand it, that's a result outside of the valid range, such as a 0 when the field requires a 1 or 2.
____________

Profile betregerProject donor
Avatar
Send message
Joined: 29 Jun 99
Posts: 2661
Credit: 5,506,207
RAC: 6,695
United States
Message 1385261 - Posted: 28 Jun 2013, 0:20:18 UTC - in response to Message 1385256.

Sten, good spot, and the best part is he is shooting himself in the foot.
____________

Profile BilBg
Volunteer tester
Avatar
Send message
Joined: 27 May 07
Posts: 2894
Credit: 6,626,064
RAC: 8,057
Bulgaria
Message 1385264 - Posted: 28 Jun 2013, 1:34:43 UTC - in response to Message 1385256.

Edit 2: Looking closer to his results, it seems that he manages to sneak some of them into the science database too, despite that Eric did something to stop that from happening. He even manages to get some as canonical results. Example: http://setiathome.berkeley.edu/workunit.php?wuid=1270273091

Maybe this happens only when Autocorr count: 0


____________



- ALF - "Find out what you don't do well ..... then don't do it!" :)

Sten-Arne
Volunteer tester
Send message
Joined: 1 Nov 08
Posts: 3752
Credit: 21,472,107
RAC: 14,692
Sweden
Message 1385266 - Posted: 28 Jun 2013, 1:46:46 UTC - in response to Message 1385264.

Edit 2: Looking closer to his results, it seems that he manages to sneak some of them into the science database too, despite that Eric did something to stop that from happening. He even manages to get some as canonical results. Example: http://setiathome.berkeley.edu/workunit.php?wuid=1270273091

Maybe this happens only when Autocorr count: 0



Or when the task is an overflow anyhow. If it only happens with overflow tasks, then it isn't so serious for the science, but nevertheless, people running old apps, to gain credit, should in my opinion be cut off from the project immediately.

This is nothing that could happen accidentally, so he deliberately changes the app_info (and/or change the file name). So without doubt he's a cheater, and should be treated like that. Pull the plug on that host pronto, ban the IP-address he connects from also.
____________

Profile BilBg
Volunteer tester
Avatar
Send message
Joined: 27 May 07
Posts: 2894
Credit: 6,626,064
RAC: 8,057
Bulgaria
Message 1385269 - Posted: 28 Jun 2013, 2:11:13 UTC - in response to Message 1385266.

... ban the IP-address he connects from also.

:))) You are joking, right?


____________



- ALF - "Find out what you don't do well ..... then don't do it!" :)

Sten-Arne
Volunteer tester
Send message
Joined: 1 Nov 08
Posts: 3752
Credit: 21,472,107
RAC: 14,692
Sweden
Message 1385270 - Posted: 28 Jun 2013, 2:14:51 UTC - in response to Message 1385269.
Last modified: 28 Jun 2013, 2:18:40 UTC

... ban the IP-address he connects from also.

:))) You are joking, right?



That part was a joke yes. He's probably on a dynamic address, so it wouldn't help much.

Had it been a static address, then it should be stopped from ever again trying to be part of this project IMO. Once a cheater, always a cheater. That he doesn't care about the science, is painfully clear. He only cares about the credit. There's a reason why he is Anonymous, and that reason is pretty obvious.

I could say much more about how I feel about those who try to cheat here, to gain credit, and destroy the science, but the rules of this board, does not allow such language.
____________

Josef W. SegurProject donor
Volunteer developer
Volunteer tester
Send message
Joined: 30 Oct 99
Posts: 4347
Credit: 1,126,619
RAC: 878
United States
Message 1385271 - Posted: 28 Jun 2013, 2:23:37 UTC - in response to Message 1385256.

Sten-Arne wrote:
...
Edit 2: Looking closer to his results, it seems that he manages to sneak some of them into the science database too, despite that Eric did something to stop that from happening. He even manages to get some as canonical results. Example: http://setiathome.berkeley.edu/workunit.php?wuid=1270273091

For a result_overflow task with no reportable Autocorr signals, there's no easy way for the Validator to detect the work was by a v6 app. Also, there's no need to protect the science since v6 and v7 apps will report the same signals (assuming the apps work as designed).

In addition, of the 33 "Completed and validated" v7 tasks for the host 22 were reported June 18 and earlier and have gotten that status by a credit granting script. Most of those do not yet have a canonical result chosen, a sure sign of action outside normal BOINC processing. For those which do have a canonical result because a third task has been completed, the result from 5940343 was not chosen.

As there is no Linux CUDA SETI@home v7 application in release status yet, it's not surprising the anonymous owner tried to make the old version do the work. At least the small percentage of tasks which overflow without Autocorrs can be assimilated, that's some useful work.
Joe

Profile BilBg
Volunteer tester
Avatar
Send message
Joined: 27 May 07
Posts: 2894
Credit: 6,626,064
RAC: 8,057
Bulgaria
Message 1385272 - Posted: 28 Jun 2013, 2:30:54 UTC - in response to Message 1385270.

... ban the IP-address he connects from also.

:))) You are joking, right?



That part was a joke yes. He's probably on a dynamic address, so it wouldn't help much.

Had it been a static address, then it should be stopped from ever again trying to be part of this project IMO.

I don't think I have to explain to you why this is bad or useless ;)

(because I remember your responses here (so you have to know ways around IP):
http://setiathome.berkeley.edu/forum_thread.php?id=64691&postid=1123474#1123474
)

You know that behind one IP can be many people (family/building/school/...)
even millions in cases like this:
http://www.securitykiss.com/
http://www.hotspotshield.com/

bad: you'll block all these people
useless: anyone can get other/new IP this way


____________



- ALF - "Find out what you don't do well ..... then don't do it!" :)

Profile BilBg
Volunteer tester
Avatar
Send message
Joined: 27 May 07
Posts: 2894
Credit: 6,626,064
RAC: 8,057
Bulgaria
Message 1385273 - Posted: 28 Jun 2013, 2:39:56 UTC - in response to Message 1385271.

As there is no Linux CUDA SETI@home v7 application in release status yet, it's not surprising the anonymous owner tried to make the old version do the work. At least the small percentage of tasks which overflow without Autocorrs can be assimilated, that's some useful work.
Joe

Is "Linux 64bit Cuda Multibeam (x41g)" compatible with SETI@home v7 ?
http://lunatics.kwsn.net/index.php?module=Downloads;sa=dlview;id=311


____________



- ALF - "Find out what you don't do well ..... then don't do it!" :)

Sten-Arne
Volunteer tester
Send message
Joined: 1 Nov 08
Posts: 3752
Credit: 21,472,107
RAC: 14,692
Sweden
Message 1385274 - Posted: 28 Jun 2013, 2:42:27 UTC - in response to Message 1385272.

... ban the IP-address he connects from also.

:))) You are joking, right?



That part was a joke yes. He's probably on a dynamic address, so it wouldn't help much.

Had it been a static address, then it should be stopped from ever again trying to be part of this project IMO.

I don't think I have to explain to you why this is bad or useless ;)

(because I remember your responses here (so you have to know ways around IP):
http://setiathome.berkeley.edu/forum_thread.php?id=64691&postid=1123474#1123474
)

You know that behind one IP can be many people (family/building/school/...)
even millions in cases like this:
http://www.securitykiss.com/
http://www.hotspotshield.com/

bad: you'll block all these people
useless: anyone can get other/new IP this way



I know the drawbacks, let those innocents who can't connect to SETI call the ISP then, and let them track him down, and stop him :-)

Come on, don't be so serious about everything I write. Start up your humour machine.
____________

Sten-Arne
Volunteer tester
Send message
Joined: 1 Nov 08
Posts: 3752
Credit: 21,472,107
RAC: 14,692
Sweden
Message 1385275 - Posted: 28 Jun 2013, 2:46:59 UTC - in response to Message 1385273.

As there is no Linux CUDA SETI@home v7 application in release status yet, it's not surprising the anonymous owner tried to make the old version do the work. At least the small percentage of tasks which overflow without Autocorrs can be assimilated, that's some useful work.
Joe

Is "Linux 64bit Cuda Multibeam (x41g)" compatible with SETI@home v7 ?
http://lunatics.kwsn.net/index.php?module=Downloads;sa=dlview;id=311



I think it is, since Petri33 is running it:

http://setiathome.berkeley.edu/forum_thread.php?id=71971&postid=1385236#1385236
____________

Sten-Arne
Volunteer tester
Send message
Joined: 1 Nov 08
Posts: 3752
Credit: 21,472,107
RAC: 14,692
Sweden
Message 1385276 - Posted: 28 Jun 2013, 3:05:41 UTC - in response to Message 1385271.
Last modified: 28 Jun 2013, 3:09:36 UTC

Sten-Arne wrote:
...
Edit 2: Looking closer to his results, it seems that he manages to sneak some of them into the science database too, despite that Eric did something to stop that from happening. He even manages to get some as canonical results. Example: http://setiathome.berkeley.edu/workunit.php?wuid=1270273091

For a result_overflow task with no reportable Autocorr signals, there's no easy way for the Validator to detect the work was by a v6 app. Also, there's no need to protect the science since v6 and v7 apps will report the same signals (assuming the apps work as designed).

In addition, of the 33 "Completed and validated" v7 tasks for the host 22 were reported June 18 and earlier and have gotten that status by a credit granting script. Most of those do not yet have a canonical result chosen, a sure sign of action outside normal BOINC processing. For those which do have a canonical result because a third task has been completed, the result from 5940343 was not chosen.

As there is no Linux CUDA SETI@home v7 application in release status yet, it's not surprising the anonymous owner tried to make the old version do the work. At least the small percentage of tasks which overflow without Autocorrs can be assimilated, that's some useful work.
Joe


Well, as long as the science is not at risk, there's no big harm. However, I still question why someone would run an obsolete app, or even come up with the idea to try, even if there isn't any v7 app for Linux. (It seems to be a working app for Linux though, which is v7 capable, x41g, which do autocorr. That's the one he should run until there's a new v7 for Linux, and not the one he runs now)

If the day comes when there isn't any app for my system, I sure wouldn't try to run an obsolete one, by changing the file name and/or editing app_info. I would just simply stop crunching of course, until there is a proper app available.
____________

JohnDKProject donor
Volunteer tester
Avatar
Send message
Joined: 28 May 00
Posts: 899
Credit: 49,149,665
RAC: 44,047
Denmark
Message 1385405 - Posted: 28 Jun 2013, 15:27:01 UTC - in response to Message 1385275.

As there is no Linux CUDA SETI@home v7 application in release status yet, it's not surprising the anonymous owner tried to make the old version do the work. At least the small percentage of tasks which overflow without Autocorrs can be assimilated, that's some useful work.
Joe

Is "Linux 64bit Cuda Multibeam (x41g)" compatible with SETI@home v7 ?
http://lunatics.kwsn.net/index.php?module=Downloads;sa=dlview;id=311



I think it is, since Petri33 is running it:

http://setiathome.berkeley.edu/forum_thread.php?id=71971&postid=1385236#1385236

I also ran it on my linux machine before switching to the new AP GPU app, had no problem and all WUs got validated.

Profile petri33Project donor
Volunteer tester
Send message
Joined: 6 Jun 02
Posts: 409
Credit: 74,126,398
RAC: 80,745
Finland
Message 1385743 - Posted: 29 Jun 2013, 9:02:16 UTC - in response to Message 1385275.

As there is no Linux CUDA SETI@home v7 application in release status yet, it's not surprising the anonymous owner tried to make the old version do the work. At least the small percentage of tasks which overflow without Autocorrs can be assimilated, that's some useful work.
Joe

Is "Linux 64bit Cuda Multibeam (x41g)" compatible with SETI@home v7 ?
http://lunatics.kwsn.net/index.php?module=Downloads;sa=dlview;id=311



I think it is, since Petri33 is running it:

http://setiathome.berkeley.edu/forum_thread.php?id=71971&postid=1385236#1385236


The Lunatics linux Cuda Multibeam (x41g) works and is v7 compatible. When it detects a v7 workload it enables autocorr and the processing of v7 packets takes much longer compared to the v6.

Stderr output <core_client_version>6.10.58</core_client_version> <![CDATA[ <stderr_txt> setiathome_CUDA: Found 2 CUDA device(s): Device 1: GeForce GTX 560 Ti, 1279 MiB, regsPerBlock 32768 computeCap 2.0, multiProcs 14 clockRate = 1500000 Device 2: GeForce GTX 660, 2047 MiB, regsPerBlock 65536 computeCap 3.0, multiProcs 5 clockRate = 1071500 In cudaAcc_initializeDevice(): Boinc passed DevPref 1 setiathome_CUDA: CUDA Device 1 specified, checking... Device 1: GeForce GTX 560 Ti is okay SETI@home using CUDA accelerated device GeForce GTX 560 Ti Cuda Active: Plenty of total Global VRAM (>300MiB). All early cuFft plans postponed, to parallel with first chirp. ) _ _ _)_ o _ _ (__ (_( ) ) (_( (_ ( (_ ( not bad for a human... _) Multibeam x41g Preview, Cuda 3.20 Detected setiathome_enhanced_v7 task. Autocorrelations enabled, size 128k elements. Work Unit Info: ............... WU true angle range is : 4.153101 re-using dev_GaussFitResults array for dev_AutoCorrIn, 4194304 bytes re-using dev_GaussFitResults+524288x8 array for dev_AutoCorrOut, 4194304 bytes Thread call stack limit is: 1k setiathome_CUDA: Found 2 CUDA device(s): Device 1: GeForce GTX 560 Ti, 1279 MiB, regsPerBlock 32768 computeCap 2.0, multiProcs 14 clockRate = 1500000 Device 2: GeForce GTX 660, 2047 MiB, regsPerBlock 65536 computeCap 3.0, multiProcs 5 clockRate = 1071500 In cudaAcc_initializeDevice(): Boinc passed DevPref 2 setiathome_CUDA: CUDA Device 2 specified, checking... Device 2: GeForce GTX 660 is okay SETI@home using CUDA accelerated device GeForce GTX 660 Cuda Active: Plenty of total Global VRAM (>300MiB). All early cuFft plans postponed, to parallel with first chirp. Restarted at 53.06 percent, with Lunatics x41g preview, Cuda 3.20 Detected setiathome_enhanced_v7 task. Autocorrelations enabled, size 128k elements. re-using dev_GaussFitResults array for dev_AutoCorrIn, 4194304 bytes re-using dev_GaussFitResults+524288x8 array for dev_AutoCorrOut, 4194304 bytes Thread call stack limit is: 1k cudaAcc_free() called... cudaAcc_free() running... cudaAcc_free() PulseFind freed... cudaAcc_free() Gaussfit freed... cudaAcc_free() AutoCorrelation freed... cudaAcc_free() DONE. Flopcounter: 16059461390835.248047 Spike count: 4 Autocorr count: 3 Pulse count: 0 Triplet count: 3 Gaussian count: 0 10:46:53 (1868): called boinc_finish

____________

S@NL - John van GorselProject donor
Volunteer tester
Avatar
Send message
Joined: 5 Jul 99
Posts: 190
Credit: 137,850,501
RAC: 5,165
Netherlands
Message 1386814 - Posted: 2 Jul 2013, 13:44:20 UTC

I have v7 cuda running under Linux on 2 pc's now using the x41g app. So far, I have 20 valid tasks, 2 pending and 1 inconclusive.

Here's the app_info.xml file:

<app_info> <app> <name>astropulse_v6</name> </app> <file_info> <name>ap_6.01r546_sse3_linux64</name> <executable/> </file_info> <app_version> <app_name>astropulse_v6</app_name> <version_num>601</version_num> <file_ref> <file_name>ap_6.01r546_sse3_linux64</file_name> <main_program/> </file_ref> </app_version> <app> <name>setiathome_v7</name> </app> <file_info> <name>setiathome_7.01_i686-pc-linux-gnu</name> <executable/> </file_info> <file_info> <name>setiathome_x41g_x86_64-pc-linux-gnu_cuda32</name> <executable/> </file_info> <file_info> <name>libcudart.so.3</name> <executable/> </file_info> <file_info> <name>libcufft.so.3</name> <executable/> </file_info> <app_version> <app_name>setiathome_v7</app_name> <version_num>701</version_num> <file_ref> <file_name>setiathome_7.01_i686-pc-linux-gnu</file_name> <main_program/> </file_ref> </app_version> <app_version> <app_name>setiathome_v7</app_name> <version_num>700</version_num> <plan_class>cuda32</plan_class> <avg_ncpus>0.040000</avg_ncpus> <max_ncpus>0.040000</max_ncpus> <coproc> <type>CUDA</type> <count>0.5</count> </coproc> <file_ref> <file_name>setiathome_x41g_x86_64-pc-linux-gnu_cuda32</file_name> <main_program/> </file_ref> <file_ref> <file_name>libcudart.so.3</file_name> </file_ref> <file_ref> <file_name>libcufft.so.3</file_name> </file_ref> </app_version> <app_info>



____________


Seti@Netherlands website

Profile David Anderson (not *that* DA)Project donor
Avatar
Send message
Joined: 5 Dec 09
Posts: 111
Credit: 23,389,311
RAC: 5,644
United States
Message 1386914 - Posted: 2 Jul 2013, 22:32:01 UTC - in response to Message 1386814.

Thanks for the app_info.xml, John.

Looks like it should work for me
assuming my 1GB 560 TI can do just 2 instances on V7. Was doing 3
instances <count>0.33</count> successfully on V6.

I too have x41g installed on Linux 64bit from when we were on V6,
but was a bit unclear if it was ok to use, so was not using it.
Extremely busy the next 2 weeks, so x41g will have to wait.

Never imagined this thread would result in app_info.xml help :-)

Message boards : Number crunching : Wondering why host gets Invalid

Copyright © 2014 University of California