CPU run time

Message boards : Number crunching : CPU run time

To post messages, you must log in.

AuthorMessage
Azurrio

Send message
Joined: 27 Jun 07
Posts: 12
Credit: 8,020
RAC: 0
Message 3865 - Posted: 9 Apr 2008, 13:28:20 UTC

So I was just wondering...
Would it be better for RALPH to have a bigger cpu run time per WU? Is the default 1 hour run time enough for testing or would it be better for me to raise it up by few hours?
I was just figuring that since this project is focusing on error detection etc., is it really useful to run the WU's longer than the deafault 1 hour?
ID: 3865 · Report as offensive    Reply Quote
Pepo
Avatar

Send message
Joined: 8 Sep 06
Posts: 104
Credit: 36,890
RAC: 0
Message 3868 - Posted: 9 Apr 2008, 13:55:24 UTC - in response to Message 3865.  

So I was just wondering...

...and not only you...

Would it be better for RALPH to have a bigger cpu run time per WU? Is the default 1 hour run time enough for testing or would it be better for me to raise it up by few hours?
I was just figuring that since this project is focusing on error detection etc., is it really useful to run the WU's longer than the deafault 1 hour?

We were discussing this question two months ago in the Run time defaults thread, unluckily without any agreement or an official response from devs' side.

The only known note is a one year old David Kim's comment:
We would prefer lower run times so that results are returned quicker.


Peter
ID: 3868 · Report as offensive    Reply Quote
Azurrio

Send message
Joined: 27 Jun 07
Posts: 12
Credit: 8,020
RAC: 0
Message 3874 - Posted: 10 Apr 2008, 4:21:39 UTC

Ok, thanks Pepo, I missed that topic...
Well, I guess I'm going to keep my run time at default to keep the WU flowing... Hope it detects all the errors in that time :)
I would also presume that if the staff would want the WU's to run longer they would propaply raise the default run time...
ID: 3874 · Report as offensive    Reply Quote
Profile David Emigh
Avatar

Send message
Joined: 6 Jan 08
Posts: 27
Credit: 26,482
RAC: 0
Message 4032 - Posted: 17 May 2008, 19:42:29 UTC

Bump

In light of the recent difficulties with memory leaks in Rosetta Mini 1.1x, especially those leaks that took upwards of 20 hours to cause a failure, would it be appropriate to increase the default runtime?
RALPH and Rosie sittin' in a tree,
F - O - L - D - I - N - G!
ID: 4032 · Report as offensive    Reply Quote
Pepo
Avatar

Send message
Joined: 8 Sep 06
Posts: 104
Credit: 36,890
RAC: 0
Message 4034 - Posted: 18 May 2008, 0:24:47 UTC - in response to Message 4032.  

In light of the recent difficulties with memory leaks in Rosetta Mini 1.1x, especially those leaks that took upwards of 20 hours to cause a failure, would it be appropriate to increase the default runtime?

This is possibly not the case, but IMO the events prove the necessity of rather broad variance of used runtimes - not only the short ones (for fast turnaround of tasks and apps as possibly the devs' main goal) but also the mid and long(est) ones to test the tasks' and apps' stability.

Peter
ID: 4034 · Report as offensive    Reply Quote

Message boards : Number crunching : CPU run time



©2024 University of Washington
http://www.bakerlab.org