Hi, I am now using "eigs" to solve an eigenvalue problem for large complex sparse matrices (approx. 20000*20000):
opts.isreal = 0;
opts.maxit = 100;
Gamma = eigs(A,-B,2,'sm',opts);
Due to the size of the matrix the program is extremely slow, sometimes takes 1.5 hrs for computing the first two smallest eigenvalues. So I plan to speedup the program by upgrading my PC's hardware.
Now I suspect that the main factor which bottlenecks the speed is the size of the memory, because when the program is running, the memory of my PC is often used up (32 GB RAM, 64-bit OS, Intel Xeon CPU E3-1276 v3 @3.6GHz 3.6GHz):
My question is, is the memory size indeed the bottleneck? If yes, How much improvement can the program achieve if I expand the RAM, say, from 32 GB to 64 GB? Is it worth upgrading
Thanks for the suggestions!