Together with James N. Lyness I'm working on constructing lattice rules for
high dimensional integration. Extensive searches has been done to find
optimal rules. Current direction is to limit the search space by imposing
more structure on the rules. We then hope to be able to construct sequences
of rules of high dimension and/or degree.
Here I've worked on a number of different stuff like: data partitioning
and load balancing, programming models and parallel applications.
This used to be my main research focus, but lately one of the application
area has sucked up most of my time, leaving little time left for
specific parallel computing research. That application is..
Numerical solution of the time dependent Schrödinger equation
This work is done in close collaboration with Prof. Jan P. Hansen at the Physics department. Our current focus
is on building an efficient and flexible code for multi particle systems.
This will extremely compute intensive so we do need to have well optimized
computational kernels as well as scalable parallel implementation.