This paper describes two recent innovations related to the restarted Lanczo
s method for eigenvalue problems, namely the thick-restart technique and dy
namic restarting schemes. Previous restarted versions of the Lanczos method
use considerably more iterations than the non-restarted versions, largely
because too much information is discarded during restarting. The thick-rest
art technique provides a mechanism to preserve a large portion of the exist
ing basis and dynamic restarting schemes decide exactly how many vectors to
save. Combining these two new techniques we are able to implement an effic
ient eigenvalue problem solver. This paper will demonstrate its effectivene
ss on one particular class of problems for which this method is well suited
: linear eigenvalue problems generated from non-selfconsistent electronic s
tructure calculations. (C) 1999 Academic Press.