Invalid Host Messaging

Message boards : Number crunching : Invalid Host Messaging
Message board moderation

To post messages, you must log in.

Previous · 1 · 2 · 3 · 4 · 5 · 6 · 7 · 8 . . . 22 · Next

AuthorMessage
Josef W. Segur
Volunteer developer
Volunteer tester

Send message
Joined: 30 Oct 99
Posts: 4504
Credit: 1,414,761
RAC: 0
United States
Message 1413588 - Posted: 10 Sep 2013, 4:05:00 UTC - in response to Message 1413559.  
Last modified: 10 Sep 2013, 4:06:10 UTC

PM'ed owner of host 6359587 which is producing a lot of invalids, and whose pendings will most all be going to the invalids list soon.

One weird thing about his Application Details that I don't understand:

SETI@home v7 (anonymous platform, ATI GPU)
Number of tasks completed 1
Max tasks per day 33
Number of tasks today 274
Consecutive valid tasks 0
Average processing rate 4980.8622641511
Average turnaround time 0.17 days

How can his number of tasks today be so high, when his max tasks per day is only 33?

For GPUs, max tasks per day is multiplied by the project's gpu_multiplier setting which is 8 here. So the base quota is 264 per GPU. And the quota check is done early, you either get the message saying quota has been exceeded or the servers try to fulfill the work request. That often overshoots the quota (particularly on GPUs) because the project's max_wus_to_send is also multiplied by gpu_multiplier. See http://boinc.berkeley.edu/trac/wiki/ProjectOptions#Joblimits

Wiggo has suggested reducing the base quota, and it could probably be trimmed to 25 or so without making recovery from a temporary problem too slow. But perhaps it would make sense to reduce the gpu_multiplier setting too.
                                                                  Joe
ID: 1413588 · Report as offensive     Reply Quote
Profile Ageless
Volunteer tester
Avatar

Send message
Joined: 9 Jun 99
Posts: 14512
Credit: 3,677,250
RAC: 591
Netherlands
Message 1413654 - Posted: 10 Sep 2013, 8:23:22 UTC - in response to Message 1413588.  

Thanks Joe.
ID: 1413654 · Report as offensive     Reply Quote
Profile tullio Project Donor
Volunteer moderator
Volunteer tester

Send message
Joined: 9 Apr 04
Posts: 6833
Credit: 1,997,736
RAC: 422
Italy
Message 1413668 - Posted: 10 Sep 2013, 9:44:12 UTC - in response to Message 1413561.  
Last modified: 10 Sep 2013, 9:45:32 UTC

PM'ed the owner of this old machine who is trying to run v7.00 work with a 5.28 application. Not going to happen.

He is using SunOS 5.10 while I am using 5.11 on my Solaris Virtual Machine. But I only have a SETI@home enhanced 6.03 app by a developer called Dotsch and I am not getting any 7.01 tasks.
Tullio
ID: 1413668 · Report as offensive     Reply Quote
Profile Ageless
Volunteer tester
Avatar

Send message
Joined: 9 Jun 99
Posts: 14512
Credit: 3,677,250
RAC: 591
Netherlands
Message 1413756 - Posted: 10 Sep 2013, 15:29:27 UTC - in response to Message 1413668.  
Last modified: 10 Sep 2013, 15:29:33 UTC

He is using SunOS 5.10 while I am using 5.11 on my Solaris Virtual Machine. But I only have a SETI@home enhanced 6.03 app by a developer called Dotsch and I am not getting any 7.01 tasks.

Yes, but if you instruct your app_info.xml file to run v7.00 tasks with the 5.28 application, or in your case the 6.03 application, you'd get work for it. I already emailed Eric about that, but also know that when the anonymous platform is in use, that there's no actual check whether or not the application (or machine) can actually do it.
ID: 1413756 · Report as offensive     Reply Quote
Profile William
Volunteer tester
Avatar

Send message
Joined: 14 Feb 13
Posts: 2037
Credit: 17,689,662
RAC: 0
Message 1413773 - Posted: 10 Sep 2013, 15:48:53 UTC - in response to Message 1413756.  

He is using SunOS 5.10 while I am using 5.11 on my Solaris Virtual Machine. But I only have a SETI@home enhanced 6.03 app by a developer called Dotsch and I am not getting any 7.01 tasks.

Yes, but if you instruct your app_info.xml file to run v7.00 tasks with the 5.28 application, or in your case the 6.03 application, you'd get work for it. I already emailed Eric about that, but also know that when the anonymous platform is in use, that there's no actual check whether or not the application (or machine) can actually do it.

Some people are hopeless. Maybe he'll notice that the only tasks validating are the -9 - the validator rejects the other ones for not having autocorr entries.

Gosh. Sparcs. Oh they still produce and develop them...
A person who won't read has no advantage over one who can't read. (Mark Twain)
ID: 1413773 · Report as offensive     Reply Quote
rob smith Special Project $250 donor
Volunteer tester

Send message
Joined: 7 Mar 03
Posts: 16295
Credit: 326,346,835
RAC: 231,017
United Kingdom
Message 1429735 - Posted: 17 Oct 2013, 16:17:47 UTC

This one appears to be going for the record:
http://setiathome.berkeley.edu/results.php?hostid=7104886
Bob Smith
Member of Seti PIPPS (Pluto is a Planet Protest Society)
Somewhere in the (un)known Universe?
ID: 1429735 · Report as offensive     Reply Quote
Profile Wiggo "Socialist"
Avatar

Send message
Joined: 24 Jan 00
Posts: 14679
Credit: 189,752,791
RAC: 72,104
Australia
Message 1429836 - Posted: 17 Oct 2013, 20:22:35 UTC - in response to Message 1429735.  

This one appears to be going for the record:
http://setiathome.berkeley.edu/results.php?hostid=7104886

That person has a GTX590 running with a GTX460 and the 460 doesn't like it so I can see why he's generating those bad numbers.

Cheers.
ID: 1429836 · Report as offensive     Reply Quote
Profile Wiggo "Socialist"
Avatar

Send message
Joined: 24 Jan 00
Posts: 14679
Credit: 189,752,791
RAC: 72,104
Australia
Message 1431316 - Posted: 21 Oct 2013, 4:42:37 UTC

ID: 1431316 · Report as offensive     Reply Quote
Profile Bernie Vine
Volunteer moderator
Volunteer tester
Avatar

Send message
Joined: 26 May 99
Posts: 9417
Credit: 57,868,877
RAC: 35,163
United Kingdom
Message 1432154 - Posted: 23 Oct 2013, 10:25:51 UTC

I wonder what is going on here:

ID9708545

Over 430 apparent computers in 18 months.

Looks like WU's are downloaded then the machine is removed from the project.

The added back and the process repeated.
"A liberal and proud to be"
ID: 1432154 · Report as offensive     Reply Quote
Jason

Send message
Joined: 20 May 99
Posts: 43
Credit: 8,944,900
RAC: 5,787
United Kingdom
Message 1440569 - Posted: 9 Nov 2013, 9:08:43 UTC

PM'ed owner of host 4974846 which is producing invalids
ID: 1440569 · Report as offensive     Reply Quote
Terry & Laura
Avatar

Send message
Joined: 15 May 99
Posts: 5
Credit: 6,159,255
RAC: 0
United States
Message 1440898 - Posted: 10 Nov 2013, 11:45:32 UTC - in response to Message 1440569.  

An anonymous host with serious issues, that seems to be playing havoc with more than an ordinary number of AP jobs:

http://setiathome.berkeley.edu/show_host_detail.php?hostid=6072583

State: All (4010) · In progress (200) · Validation pending (1563) · Validation inconclusive (1712) · Valid (36) · Invalid (498) · Error (1)
Application: All (4010) · AstroPulse v6 (736) · SETI@home Enhanced (0) · SETI@home v7 (3274)
ID: 1440898 · Report as offensive     Reply Quote
Jason

Send message
Joined: 20 May 99
Posts: 43
Credit: 8,944,900
RAC: 5,787
United Kingdom
Message 1441370 - Posted: 11 Nov 2013, 21:28:18 UTC - in response to Message 1440569.  

PM'ed owner of host 6204067 which is producing invalids
ID: 1441370 · Report as offensive     Reply Quote
MiklaDfar

Send message
Joined: 24 Apr 12
Posts: 4
Credit: 11,021,325
RAC: 4,857
United States
Message 1443033 - Posted: 16 Nov 2013, 15:02:51 UTC

As of this weekend I will pull all 6 computers offline from the SETI project. I volunteer a great deal of compute HP that has NO/ZIP/ZERO issues with other projects. I don't have time to babysit the SETI project or come on this forum. If the project hase issues with certain results from certain computers, than find a way to kick them off and quickly disregard the data. This is a software design problem being blamed on a user... good luck with the project.
ID: 1443033 · Report as offensive     Reply Quote
Profile Fred E.
Volunteer tester

Send message
Joined: 22 Jul 99
Posts: 768
Credit: 24,140,697
RAC: 0
United States
Message 1443084 - Posted: 16 Nov 2013, 16:41:07 UTC

As of this weekend I will pull all 6 computers offline from the SETI project. I volunteer a great deal of compute HP that has NO/ZIP/ZERO issues with other projects. I don't have time to babysit the SETI project or come on this forum. If the project hase issues with certain results from certain computers, than find a way to kick them off and quickly disregard the data. This is a software design problem being blamed on a user... good luck with the project.

Suggest you wait until one of the more savvy users can look at the tasks for computer xxxx106. It has only 2 successful "good" S@H gpu tasks since S@H v7 rolled out. The tasks seem to be finding signals that aren't there and then reaching a space limitation, leading to an early exit. Those results don't match the results of the other computers that crunched the same tasks, so your results are marked as invalid. You aren't earning any credit crunching S@H v7 with that powerful gpu. The gpu is handling Astropulse for Nvida okay, so I'm puzzled and my usual suggestions don't seem appropriate.

There is indeed a software problem. The server software is supposed to limit you to one task/day for the gpu applications when you produce invalid results. The project hasn't found that problem yet.
Another Fred
Support SETI@home when you search the Web with GoodSearch or shop online with GoodShop.
ID: 1443084 · Report as offensive     Reply Quote
Profile Bernie Vine
Volunteer moderator
Volunteer tester
Avatar

Send message
Joined: 26 May 99
Posts: 9417
Credit: 57,868,877
RAC: 35,163
United Kingdom
Message 1443119 - Posted: 16 Nov 2013, 17:35:52 UTC - in response to Message 1443033.  
Last modified: 16 Nov 2013, 17:37:52 UTC

As of this weekend I will pull all 6 computers offline from the SETI project. I volunteer a great deal of compute HP that has NO/ZIP/ZERO issues with other projects. I don't have time to babysit the SETI project or come on this forum. If the project has issues with certain results from certain computers, then find a way to kick them off and quickly disregard the data. This is a software design problem being blamed on a user... good luck with the project.


Seems a strange attitude, one of my computers is having problems so I will pull the other 5 as well.
But c'est la vie.
ID: 1443119 · Report as offensive     Reply Quote
MiklaDfar

Send message
Joined: 24 Apr 12
Posts: 4
Credit: 11,021,325
RAC: 4,857
United States
Message 1443157 - Posted: 16 Nov 2013, 20:57:40 UTC - in response to Message 1443119.  

Seems a strange attitude, one of my computers is having problems so I will pull the other 5 as well.
But c'est la vie.


Very helpful post you got there...
I got this message...
"...you are more of a danger to the Seti project than a help..."
you tell me why I should not pull all my computers when I clearly stated I do not have time... Messages like this do not inspire me to continue.

Mods... per your Warning on the first page of this thread I think pulling down this comment and the comment I replied to might be appropriate.
ID: 1443157 · Report as offensive     Reply Quote
Profile betreger Project Donor
Avatar

Send message
Joined: 29 Jun 99
Posts: 7956
Credit: 18,982,463
RAC: 9,252
United States
Message 1443162 - Posted: 16 Nov 2013, 21:10:17 UTC - in response to Message 1443033.  

A quick look at your results show that it is your 560ti that is having problems.
IIRC some tome back a lot of 560ti were having the problem. It seems Seti works them pretty hard and some of them do what yours is doing. I know fixes were found and the problem went pretty much away. The info is buried in these forums. I know the fix is pretty easy but I don't remember the exact details.
ID: 1443162 · Report as offensive     Reply Quote
Profile Bernie Vine
Volunteer moderator
Volunteer tester
Avatar

Send message
Joined: 26 May 99
Posts: 9417
Credit: 57,868,877
RAC: 35,163
United Kingdom
Message 1443165 - Posted: 16 Nov 2013, 21:12:09 UTC

you are more of a danger to the Seti project than a help.

My apologies, that message is not helpful and should not have been said.

I do not condone that sort of language and you are right to be annoyed, you are certainly no "danger" to the project.
"A liberal and proud to be"
ID: 1443165 · Report as offensive     Reply Quote
rob smith Special Project $250 donor
Volunteer tester

Send message
Joined: 7 Mar 03
Posts: 16295
Credit: 326,346,835
RAC: 231,017
United Kingdom
Message 1443173 - Posted: 16 Nov 2013, 21:33:23 UTC

First, let's have a look at what is causing the problem.
There are a number of possible causes for the problems being exhibited by ONE of your computers:
Impending hardware failure - SETI pushes the GPUs far harder than most other projects, thus if your hardware is close to its limits then it is more likely to fail when processing SETI tasks than most other projects (and the 560Ti is a GPU is often said to be very sensitive to its operating temperature and individual cards can be worse than others)
Wrong drivers, there are several threads around that work through GPU/driver combinations. I have to admit that it can be difficult to find the correct combination, and that generally one that works well for SETI will work well for everything else.
Overclocking, as above, some cards will overclock very well and others are very poor.
Dust bunnies - yes these are a pain, almost everyone suffers from them every for time to time, high error rates, and high rates of invalids can be signs that it is time to clean the dust out of a computer.

One solution is to run one of the optimised applications rather than the stock one. But given your desire not to have to baby sit your computers I would hesitate to suggest that you go down that route as they can take a bit of effort to get them installed properly.


As for your request to hide some posts, the correct way to get the mods to read a request is to hit the "red-x" that appears at the bottom of every post. This has several advanatages - we get near instant notification of the request, we are pointed straight to the post, and if yo have filled in the report box we know what you think is wrong with that post.
Bob Smith
Member of Seti PIPPS (Pluto is a Planet Protest Society)
Somewhere in the (un)known Universe?
ID: 1443173 · Report as offensive     Reply Quote
Profile Wiggo "Socialist"
Avatar

Send message
Joined: 24 Jan 00
Posts: 14679
Credit: 189,752,791
RAC: 72,104
Australia
Message 1443193 - Posted: 16 Nov 2013, 22:03:43 UTC

I had a GTX560Ti until last year, but I got rid of it as it required to much baby sitting because as soon as it hit 74C it would only produce crap so I was always reducing the clock rates when needed.

These are the main 4 fixes that have helped others here out with them,

1/ Increase minimum fan speed to at least 65% or higher (as those cards run hot).

2/ Most 560Ti's are overclocked well above Nvidia spec's so return clocks back to stock speeds.

3/ Early 560Ti's also have their power settings set to low and a slight increase of that can help.

4/ 12v multi rail PSU's are generally no good for these either, replace with 12v single rail PSU with at least a 45amp rating on the 12V rail.

Cheers.
ID: 1443193 · Report as offensive     Reply Quote
Previous · 1 · 2 · 3 · 4 · 5 · 6 · 7 · 8 . . . 22 · Next

Message boards : Number crunching : Invalid Host Messaging


 
©2018 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.