3 ms·
I always thought it would be cool to have a top500 list that does not restrict the algorithm, just accuracy and precision of fft. It would be interesting to see
by daveguy 10y ago
I always thought it would be cool to have a top500 list that does not restrict the algorithm, just accuracy and precision of fft. It would be interesting to see how algorithms would get us from 10 petaflops to "1 exaflop equivalent" -- the flops equivalent would be time speedup to get the same answer in less time. As it is they have very strict algorithm requirements (which makes sense if you are regularizing on flops). Just curious, but of course execution time on the top supercomputers isn't cheap.
- privong 10y agoWhat about doing something like the inverse? Top500 for algorithms on a fixed hardware configuration. You have an input dataset and an answer (perhaps something with an analytic solution, for sake of comparison). That would let you compare the algorithm more directly.