s. keeling wrote:
That's almost trivial. The datasets you see in the petrochemical industry can be in the terabyte range. They're so big, they have to edit in place, not write another output file. perl handles even this well. I/O performance is pretty much hardware bound. This is binary data, btw. Seismic data.
Interesting. Never had chance to see this kind of applications. Is this about doing signal processing and pattern recognition with that data?
BTW, is that tera byte denoting one single file?