Page 1 of 1

cluster analysis

Posted: May 31st, 2019, 4:48 am
by christianlueme
Hi,
I am running hysplit cluster analysis using 48216 backward trajectories using Linux and Cluster. It does not RUN.

48216 are for 10 years and I break them is seasonal trajectories. No need to run per year to avoid bias in the interpretation of data.

Hysplit is using only one CPU and cannot be spread in other CPUs.

What can I do?

Re: cluster analysis

Posted: May 31st, 2019, 7:19 am
by barbara.stunder
There is not enough computer memory to handle that many trajectories. To use this software you need to scale your set of trajectories down. Please see viewtopic.php?t=14

Re: cluster analysis

Posted: May 31st, 2019, 8:15 am
by christianlueme
Is using trajstat alternative to solve the issue of memory limit for big data (trajectories) in hysplit?

I need another option than cutting the data down, such as using another software, tool, program if it is not possible with Hysplit.

But another software that can use hysplit trajectories I have already.

Can you help

Re: cluster analysis

Posted: May 31st, 2019, 8:17 am
by christianlueme
Is using trajstat alternative to solve the issue of memory limit for big data (trajectories) in hysplit?

I need another option than cutting the data down, such as using another software, tool, program if it is not possible with Hysplit.

But another software that can use hysplit trajectories I have already.

Can you help