Hi Heartwood-Wealth,
The scientific programs that we run often use large arrays of floating point numbers. The Help Defeat Cancer project (recently completed) had one section that it would regularly execute that had a working set of 540 MB (plus a fair amount of virtual memory). We started out requiring 768 MB of RAM for computers to run it, but after a number of people ran into surprises while running other programs which they did not understand, we increased the minimum to 1025 MB (> 1024!). Our smallest program is Genome Comparison. The Human Proteome Folding 2 project was expected to require a 512 MB computer but it was successfully trimmed to run on a 250 MB computer.

This is not unusual. Back when I was starting to learn programming back in the 1970s, I was shocked to realize that a simple physics problem could easily take several hundred megabytes. In those days that seemed enormous, but now it is possible to run a real scientific program on a personal computer rather than a supercomputer. Rick Alther regularly cuts out 100-300 megabytes from the project programs submitted to make them easier to run.

Lawrence