Advanced search

Message boards : Number crunching : frsmd_f_ga version 5.22

Author Message
Profile Krunchin-Keith [USA]
Avatar
Send message
Joined: 17 May 07
Posts: 512
Credit: 111,288,061
RAC: 0
Level
Cys
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 137 - Posted: 1 Jul 2007 | 17:14:33 UTC

I see we have a new version.

I just completed successfully a task which is uploading now, but got no more work.

So what was added to this new version ?
____________
Alpha Tester ~~ BOINCin since 10-Apr-2004 (2.28) ~~~ Join team USA

Profile GDF
Volunteer moderator
Project administrator
Project developer
Project tester
Volunteer developer
Volunteer tester
Project scientist
Send message
Joined: 14 Mar 07
Posts: 1957
Credit: 629,356
RAC: 0
Level
Gly
Scientific publications
watwatwatwatwat
Message 139 - Posted: 2 Jul 2007 | 20:35:51 UTC - in response to Message 137.

I see we have a new version.

I just completed successfully a task which is uploading now, but got no more work.

So what was added to this new version ?


Version .22 had just minimal changes. The new application we are about to distribute instead is able to chain works in a workflow, so when one result is returned another one is created restarting the job on another computer.
This allows us to run very long workunits as a sequence of 20 hours ones.

GDF

Profile Krunchin-Keith [USA]
Avatar
Send message
Joined: 17 May 07
Posts: 512
Credit: 111,288,061
RAC: 0
Level
Cys
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 140 - Posted: 2 Jul 2007 | 20:38:00 UTC - in response to Message 139.

Very interesting.

Mark Gallaher
Send message
Joined: 15 Jun 07
Posts: 3
Credit: 10,063,350
RAC: 0
Level
Pro
Scientific publications
watwatwatwatwatwatwat
Message 141 - Posted: 3 Jul 2007 | 16:46:43 UTC

Out of curiosity, how many workunits *might* you be chaining that way?

Profile GDF
Volunteer moderator
Project administrator
Project developer
Project tester
Volunteer developer
Volunteer tester
Project scientist
Send message
Joined: 14 Mar 07
Posts: 1957
Credit: 629,356
RAC: 0
Level
Gly
Scientific publications
watwatwatwatwat
Message 142 - Posted: 3 Jul 2007 | 17:41:17 UTC - in response to Message 141.
Last modified: 3 Jul 2007 | 17:42:20 UTC

Out of curiosity, how many workunits *might* you be chaining that way?


As many as we like. Now we are just testing with two. In practice a result can spawn as many workunits as needed, even more than one at a time.

Mark Gallaher
Send message
Joined: 15 Jun 07
Posts: 3
Credit: 10,063,350
RAC: 0
Level
Pro
Scientific publications
watwatwatwatwatwatwat
Message 143 - Posted: 3 Jul 2007 | 21:35:59 UTC

I think I follow what you are saying: you can have a problem that can be sent out as (xxx) 20-hour results for us to crunch.

I was curious about the size of the problem :) I don\'t understand if the problem would become 100 20-hour chunks, or 100,000 20-hour chunks ... millions?

Profile GDF
Volunteer moderator
Project administrator
Project developer
Project tester
Volunteer developer
Volunteer tester
Project scientist
Send message
Joined: 14 Mar 07
Posts: 1957
Credit: 629,356
RAC: 0
Level
Gly
Scientific publications
watwatwatwatwat
Message 145 - Posted: 4 Jul 2007 | 14:27:47 UTC - in response to Message 143.
Last modified: 4 Jul 2007 | 16:40:32 UTC

I think I follow what you are saying: you can have a problem that can be sent out as (xxx) 20-hour results for us to crunch.

I was curious about the size of the problem :) I don\'t understand if the problem would become 100 20-hour chunks, or 100,000 20-hour chunks ... millions?


Most likely just less than 10, but this allows us to use more complex thermodynamic protocols to compute the quantities we are looking at.

Post to thread

Message boards : Number crunching : frsmd_f_ga version 5.22

//