RESEARCH SYNOPSIS
My research interests fall into two areas:
- Structured Parallelism. I
am interested in the automatic scheduling of high-level parallel
constructs (algorithmic skeletons) in heterogeneous systems. My methodology has
employed the structural information of skeletons, especially pipelines and task farms,
to dynamically improve the overall performance of parallel programs in
non-dedicated heterogenous distributed systems. By introducing
resource-awareness to task-parallel constructs, it has led to
substantive contributions in
scheduling of divisible workloads.
- Computational Modelling of
Biomedical Applications. I have cooperated in the development of
Macaco, a
Markov-based simulator of macroscopic calcium currents which aggregates
individual calcium channel contributions for a given channel
population. Additionally, I have been involved in the specification of
the distributed agent architecture for the decision support system in HealthAgents, a
project to produce an agent-based aid for the diagnosis of brain
tumours.
|