Bad CPU

Message boards : Number crunching : Bad CPU
Message board moderation

To post messages, you must log in.

AuthorMessage
Luigi Naruszewicz
Avatar

Send message
Joined: 19 Nov 99
Posts: 620
Credit: 23,910,372
RAC: 14
United Kingdom
Message 142501 - Posted: 25 Jul 2005, 12:06:31 UTC

Does BOINC/SETI have any system of informing people like this guy http://setiweb.ssl.berkeley.edu/show_host_detail.php?hostid=22724 that their computer is not working. His early WUs seem alright but then they goe haywire.
.


A person who makes no mistakes, creates nothing.
ID: 142501 · Report as offensive
Profile Jim Baize
Volunteer tester

Send message
Joined: 6 May 00
Posts: 758
Credit: 149,536
RAC: 0
United States
Message 142508 - Posted: 25 Jul 2005, 12:47:53 UTC - in response to Message 142501.  

Other than the user checking his results, no. They probably should implement something like this. I would think it would help correct some problems and maybe even lower the work load on the servers a bit. Maybe on the projects page list a percentage of WU valid. Anything under a certain percent would be signs of problems.

Jim

Does BOINC/SETI have any system of informing people like this guy http://setiweb.ssl.berkeley.edu/show_host_detail.php?hostid=22724 that their computer is not working. His early WUs seem alright but then they goe haywire.


ID: 142508 · Report as offensive
Profile Saenger
Volunteer tester
Avatar

Send message
Joined: 3 Apr 99
Posts: 2452
Credit: 33,281
RAC: 0
Germany
Message 142509 - Posted: 25 Jul 2005, 12:52:50 UTC - in response to Message 142508.  

Other than the user checking his results, no. They probably should implement something like this. I would think it would help correct some problems and maybe even lower the work load on the servers a bit. Maybe on the projects page list a percentage of WU valid. Anything under a certain percent would be signs of problems.


If you send back only crap, your max. daily WU-rate goes down from 100 to 1. So one indication would be a cold puter ;)

So in the long run, no much harm will be done.


Gruesse vom Saenger

For questions about Boinc look in the BOINC-Wiki
ID: 142509 · Report as offensive
Profile Jim Baize
Volunteer tester

Send message
Joined: 6 May 00
Posts: 758
Credit: 149,536
RAC: 0
United States
Message 142512 - Posted: 25 Jul 2005, 13:00:17 UTC - in response to Message 142509.  

You have a valid point in that respect. I still think it would be a good idea for users to be notified if one of their machines starts producing erroneous results consistantly. Perhaps via email?

Jim

Other than the user checking his results, no. They probably should implement something like this. I would think it would help correct some problems and maybe even lower the work load on the servers a bit. Maybe on the projects page list a percentage of WU valid. Anything under a certain percent would be signs of problems.


If you send back only crap, your max. daily WU-rate goes down from 100 to 1. So one indication would be a cold puter ;)

So in the long run, no much harm will be done.



ID: 142512 · Report as offensive
Luigi Naruszewicz
Avatar

Send message
Joined: 19 Nov 99
Posts: 620
Credit: 23,910,372
RAC: 14
United Kingdom
Message 142515 - Posted: 25 Jul 2005, 13:04:08 UTC

I only wondered as he seems to be wasting his and everybodie elses time. In this there is no real harm done, but I have had another instance where a computer as claiming 0 or small credit and being awarded a large credit, which I do think is unfair.

.


A person who makes no mistakes, creates nothing.
ID: 142515 · Report as offensive
Bill Barto

Send message
Joined: 28 Jun 99
Posts: 864
Credit: 58,712,313
RAC: 91
United States
Message 142522 - Posted: 25 Jul 2005, 13:38:35 UTC - in response to Message 142509.  

Other than the user checking his results, no. They probably should implement something like this. I would think it would help correct some problems and maybe even lower the work load on the servers a bit. Maybe on the projects page list a percentage of WU valid. Anything under a certain percent would be signs of problems.


If you send back only crap, your max. daily WU-rate goes down from 100 to 1. So one indication would be a cold puter ;)

So in the long run, no much harm will be done.



The daily WU quota is reduced when there are client errors. In this case it is the validation process that is failing for this host and the quota is not being reduced. This one is different from others like this I have seen in that he is not getting the -9 result overload error.
ID: 142522 · Report as offensive
Bill Barto

Send message
Joined: 28 Jun 99
Posts: 864
Credit: 58,712,313
RAC: 91
United States
Message 142523 - Posted: 25 Jul 2005, 13:42:19 UTC - in response to Message 142515.  

I only wondered as he seems to be wasting his and everybodie elses time. In this there is no real harm done, but I have had another instance where a computer as claiming 0 or small credit and being awarded a large credit, which I do think is unfair.


In the cases where hosts claim low credit but are granted normal credit it is because of a problem with the benchmarking. In fact the host has taken the normal amount of time to process the WU and the result is validated. You can tell this by looking at the amount of time it takes the host to return results. I don't have a link for one of these hosts handy or I would post it.
ID: 142523 · Report as offensive
Luigi Naruszewicz
Avatar

Send message
Joined: 19 Nov 99
Posts: 620
Credit: 23,910,372
RAC: 14
United Kingdom
Message 142526 - Posted: 25 Jul 2005, 13:56:20 UTC - in response to Message 142523.  

I only wondered as he seems to be wasting his and everybodie elses time. In this there is no real harm done, but I have had another instance where a computer as claiming 0 or small credit and being awarded a large credit, which I do think is unfair.


In the cases where hosts claim low credit but are granted normal credit it is because of a problem with the benchmarking. In fact the host has taken the normal amount of time to process the WU and the result is validated. You can tell this by looking at the amount of time it takes the host to return results. I don't have a link for one of these hosts handy or I would post it.


I tried find an example of large credit for little or no credit claimed but it was a little while back and it must have fallllen of the end of my results.

.


A person who makes no mistakes, creates nothing.
ID: 142526 · Report as offensive
Profile Saenger
Volunteer tester
Avatar

Send message
Joined: 3 Apr 99
Posts: 2452
Credit: 33,281
RAC: 0
Germany
Message 142527 - Posted: 25 Jul 2005, 13:58:45 UTC - in response to Message 142523.  

I don't have a link for one of these hosts handy or I would post it.


Here is one with a benchmark problem. He has benchmarked 0.00/0.00, and therefore claims just that: 0.00 ;)
Gruesse vom Saenger

For questions about Boinc look in the BOINC-Wiki
ID: 142527 · Report as offensive
Profile Jim Baize
Volunteer tester

Send message
Joined: 6 May 00
Posts: 758
Credit: 149,536
RAC: 0
United States
Message 142544 - Posted: 25 Jul 2005, 15:10:26 UTC - in response to Message 142527.  

His computer has ZERO CPU's. Wow... talk about super computing!

I don't have a link for one of these hosts handy or I would post it.


Here is one with a benchmark problem. He has benchmarked 0.00/0.00, and therefore claims just that: 0.00 ;)


ID: 142544 · Report as offensive
Profile tekwyzrd
Volunteer tester
Avatar

Send message
Joined: 21 Nov 01
Posts: 767
Credit: 30,009
RAC: 0
United States
Message 142581 - Posted: 25 Jul 2005, 16:11:01 UTC - in response to Message 142544.  

His computer has ZERO CPU's. Wow... talk about super computing!

I don't have a link for one of these hosts handy or I would post it.


Here is one with a benchmark problem. He has benchmarked 0.00/0.00, and therefore claims just that: 0.00 ;)



Maybe it's overclocked.
ID: 142581 · Report as offensive

Message boards : Number crunching : Bad CPU


 
©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.