Client Scheduling

Message boards : Number crunching : Client Scheduling
Message board moderation

To post messages, you must log in.

AuthorMessage
Profile Daykay
Avatar

Send message
Joined: 18 Dec 00
Posts: 647
Credit: 739,559
RAC: 0
Australia
Message 110109 - Posted: 11 May 2005, 14:18:11 UTC

From the BOINC pages, a full run down on client scheduling and work fetch policies.

Read more here

Kolch - Crunching for the BOINC@Australia team since July 2004.
Search for your own intelligence...
ID: 110109 · Report as offensive
Astro
Volunteer tester
Avatar

Send message
Joined: 16 Apr 02
Posts: 8026
Credit: 600,015
RAC: 0
Message 110116 - Posted: 11 May 2005, 14:34:54 UTC

Thanks, the page is helpful. I think they could improve it by including the formulas used. I.E 2 * (connect to) triggers "panic mode".

I'd also like to see if "Resource share" is honored when WU are regularly crunched in Panic mode. I.E I have my "connect to" set at 4 days, and it continually stays in panic mode to crunch 7 day deadlined projects. It then downloads more of the same, and those are crunched in "panic mode" as well. I believe projects with 2 week deadlines are getting robbed.

tony

ID: 110116 · Report as offensive
Profile Daykay
Avatar

Send message
Joined: 18 Dec 00
Posts: 647
Credit: 739,559
RAC: 0
Australia
Message 110118 - Posted: 11 May 2005, 14:40:59 UTC - in response to Message 110116.  

I'm seeing similar behavior on my host at the moment, however I am led to believe that the long term debts for the other projects are rising as the long term debts for my short deadline project head into negative territory.

Maybe then the other projects will play catch up and get to their 66.6% share of the CPU.

Note I recently changed my Connect to setting back to 2 days from 3 as I have broadband and it was originally increased to cope with outages, this is no longer an issue.

Kolch - Crunching for the BOINC@Australia team since July 2004.
Search for your own intelligence...
ID: 110118 · Report as offensive
Astro
Volunteer tester
Avatar

Send message
Joined: 16 Apr 02
Posts: 8026
Credit: 600,015
RAC: 0
Message 110122 - Posted: 11 May 2005, 14:46:07 UTC - in response to Message 110118.  

<blockquote>Note I recently changed my Connect to setting back to 2 days from 3 </blockquote>

I did as well, now it's only in "Panic Mode" every few days, instead of 1 day after the d/l of PPAH or Einsten. I don't know if I like having that small of a buffer for bad times. 2 days isn't really 2 days, it's more like a day and a half. That's a pretty slim margin.

tony
ID: 110122 · Report as offensive
Profile Daykay
Avatar

Send message
Joined: 18 Dec 00
Posts: 647
Credit: 739,559
RAC: 0
Australia
Message 110128 - Posted: 11 May 2005, 15:11:04 UTC - in response to Message 110122.  

<blockquote>I don't know if I like having that small of a buffer for bad times. 2 days isn't really 2 days, it's more like a day and a half. That's a pretty slim margin.

tony</blockquote>

Yeah but thanks to a known bug i now have 2 CPDN WU's anyhow so it's no biggie if my cache is depleted for a day or two.

Kolch - Crunching for the BOINC@Australia team since July 2004.
Search for your own intelligence...
ID: 110128 · Report as offensive

Message boards : Number crunching : Client Scheduling


 
©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.