Look at the above picture. It should absolutely awe you. Our computational power has exponential looking growth In a log plot!!! The types of calculations will will be able to do during our own lifetimes will just be staggering to think about!
Furthermore, supercomputers today can make calculations 1000 times faster than just ~10 years ago and a million times faster than ~20 years ago. This means the top supercomputer computing speed doubles about every year.
And Nick, this also means I will be able to solve the complex problems you are solving with Fortran in ~5 years in Python just as quickly. :) I'll just have to bask in the easy to use, batteries included programming language, running on lightning fast hardware. I doubt I will ever solve a problem more numerically complicated than those. (And this is assuming I don't convert the slow parts of the code to C and call it from Python. It would be more work, but I would only have to wait 1-2 years in that case.)
That is assuming that you can effectively use all that computing power. Right now the development of programming techniques for parallel programming is lagging far behind that super-exponential curve you plotted. And while everybody gets excited and sends out press-releases over large supercomputers, the same doesn't happen when someone comes up with better MPI tools.
ReplyDeleteYour absolutely right. I didn't see any press release over this cool MPI tool:http://tinyurl.com/kljsvn :)
ReplyDelete