I need to extract the lowest, e.g., 30 eigenvalues and their eigenvectors from a real symmetric sparse double-precision matrix which has a size about 200k x 200k. I know it's huge (about 310 GB if they are 8-byte doubles) but I have enough RAM for it. A full diagonalization is insane. Would you please point me to any packages/libraries that does a decent eigenvalue/eigenvector extraction? Thank you for your time.
Asked
Active
Viewed 77 times
0
-
Seeking for library recommendations is off-topic on SO. Maybe [math.se] can give you pointers on what kind of algorithm to use. – chtz May 13 '17 at 09:20
-
1I would suggest [softwarerecs.se] – May 13 '17 at 12:45
-
1http://slepc.upv.es – ewcz May 14 '17 at 00:15