First of all, I would like to thank Robert for continuing working with DIVA-GIS, and for making it available for 64 bit operating systems. I thought DIVA-GIS was completely migrated to Java by CIP and that the Delphi versions were just "left in the past". But, here we are.
I would like to request a simple thing (or what I think would be a simple thing). It seems that with large datasets (more than 10k species), the program fails in the Reserve selection (Rebelo complementarity analysis) as it uses too much memory in the calculation. I have tried to analyze a dataset containing more/less 500,000 occurrences of 16,000 species and havent been able to do so since everytime I try to perform the analysis, the program just bails out and shows a message saying ("Out of memory").
I was wondering if there's a way to optimize the calculation in order to better use the available RAM memory? or if there is another software with which I could perform such calculation? or perhaps a method to divide the dataset and then comprise the final results?
And a last question, are these new versions of DIVA being updated in sourceforge?
reserve selection
I have looked and yes, it would certainly be possible to make this more memory efficient and faster too, but it would take some effort. I can't do it right now. One thing you might try is to use short codes (numbers - or better yet letter combinations a .. z, aa, ab, ....). That could save some memory. I did not try it myself, no guaranties. The other thing would be to test this for a coarse grid (not too many cells) first; and then see how far you can push it, but perhaps that does not affect it as much, I do not know.
2: no DIVA is not on sourceforge, that is the CIP branch.
Robert
Many thanks for your helpful
Many thanks for your helpful reply Robert. I was able to perform the calculation with a coarse grid. However, I look forward to have a memory-efficient Reserve selection feature with which large datasets can be calculated at higher resolutions. For the moment, I got what I wanted.