jaclyn_brown@yale_edu
New Member
I've been running POP on a Linux cluster with 8 processors. It takes about 12 mins to run a day. I increased the processors to 32 and the run time dropped to 9 mins/day. I expected it to be a lot faster than this for 4 times as many processors. I also tried 8 processors, and it was still 12 min/day. Does any one know how to make the model run faster? I am outputing the 3 day averages as I go. I only need the equatorial Pacific - is there some way to write out less information - would this speed up the model?
Thanks,
Jaci Brown
Thanks,
Jaci Brown