Model taking too long to run

questions concerning analysis/theory using program MARK

Model taking too long to run

Postby gabrielcaetano » Fri Jul 13, 2012 1:31 am

Hello, I´m trying to run a full POPAN model with 144 occasions and 528 individuals, and it's taking way too long. The strange thing is that with just a few less occasions it would take only 3 days to run, now it seems it will take months, because it's been running for almost 20 hours and it's only made about 40 iterations. Does anyone know some way to speed it up? Here are my system specifications:

AMD Athlon 64 X2 Dual core processor 3800+ 2.00 GHz
RAM: 2 Gb
32 bits

Thank you,

Gabriel
gabrielcaetano
 
Posts: 5
Joined: Mon Dec 13, 2010 9:21 pm

Re: Model taking too long to run

Postby gabrielcaetano » Fri Jul 13, 2012 11:04 pm

Also, it got slower now that i updated to 6.2
gabrielcaetano
 
Posts: 5
Joined: Mon Dec 13, 2010 9:21 pm

Re: Model taking too long to run

Postby bacollier » Mon Jul 16, 2012 12:02 pm

gabrielcaetano wrote:Hello, I´m trying to run a full POPAN model with 144 occasions and 528 individuals, and it's taking way too long. The strange thing is that with just a few less occasions it would take only 3 days to run, now it seems it will take months, because it's been running for almost 20 hours and it's only made about 40 iterations. Does anyone know some way to speed it up? Here are my system specifications:

AMD Athlon 64 X2 Dual core processor 3800+ 2.00 GHz
RAM: 2 Gb
32 bits

Thank you,

Gabriel


Gabriel,
There are a few possible reasons it could be taking a while to run, most are hard to debug online without the data/model structure to tinker with. With a simple g*t model with 144 occasions if you have 2 groups and are just looking at apparent survival, recapture prob, entry, that's a bunch of parameters MARK has to chug through. You might be able to speed it up a bit by providing some initial parameter estimates, but will probably not help much. Do less complicated models fit faster?

I have had some models take a week to run, and a recent bootstrap analysis in MARK took 9 weeks on my system (64bit quad core with 16gb ram using 7 of 8 processors 24/7), so I will tell you what Gary told me during that analysis (which may be of little consolation to you): be happy, computer time is cheap...

Bret
bacollier
 
Posts: 231
Joined: Fri Nov 26, 2004 10:33 am
Location: Louisiana State University

Re: Model taking too long to run

Postby gabrielcaetano » Sun Jul 22, 2012 10:17 pm

Thanks for the answer. I tried moving it to a six-core 3.3GHz with 2.75 GB of RAM and it is still at the same speed, it's using about 90% of each of the six and only 600mb RAM. I guess I'll just have to wait and hope it doesn't give me an error message, but thanks anyway.
gabrielcaetano
 
Posts: 5
Joined: Mon Dec 13, 2010 9:21 pm


Return to analysis help

Who is online

Users browsing this forum: No registered users and 4 guests