Hi Ann, It's hard for me to answer your question as it greatly depends on the actual processing you're doing. However, as a general rule, if you can fit all of the data into memory, it makes processing it much simpler. But if the data is too large, you don't have that option and you'll need to write it to disk. It's also hard to define what "too large" is, as it depends on your hardware and how you've configured the JVM, but I tend to get uneasy at the 1 GB mark. Do you need to have all of the data available to do any processing, or is only a small chunk needed to compute part of the result? For example, maybe you only need one time slice loaded into memory at a time? That would make things easier. I don't expect caching to help here, unless you plan to have a lot of local NetCDF files open. And NCO might be useful, but only if part or all of your processing could be done by their operators. I expect that NCO has been coded to properly handle very large datasets as well. Cheers, Christian Ticket Details =================== Ticket ID: QQO-385265 Department: Support netCDF Java Priority: Normal Status: Closed
NOTE: All email exchanges with Unidata User Support are recorded in the Unidata inquiry tracking system and then made publicly available through the web. If you do not want to have your interactions made available in this way, you must let us know in each email you send to us.