I have been involved in developing the
PARSEC package .
The most recent algorithm which significantly speeds up the
expensive diagonalization in DFT calculations is documented in
this paper.
Note that this paper only discussed the sequential cases. The parallel
code is finished recently
(Thanks to Jim
for sending Murilo
from UT Austin to help me finalize the
parallelization. We did the job in
three days Oct. 16--18, working 12+ hrs per day).
A recent public release of the PARSEC package is
here . Note that our most recent (and most efficient) solvers
together with recent publications have not been put on this page yet. We
will make public release of the new solvers at a suitable time.
Numerical Linear Algebra:
Eigenvalue problems, linear systems, matrix equations
(especially in the large scale setting).
Scientific computing and software development
Interdisciplinary researches: Especially modeling and novel
algorithmic development.
Reduced basis methods for model reduction of
large-scale dynamical systems.
Optimization. Data mining.
Preconditioning and regularization methods for ill-conditioned problems:
Especially ill-conditioning related to large scale problems
from eigenvalue problems, linear systems, matrix equations,
and optimization.
Last modified:
The views and opinions expressed in this page are strictly those of the page author. The contents of this page have not been reviewed or approved by the University of Minnesota.