HPC in Asia:
(A08) Poster from Saudi Arabia: HiCMA: Hierarchical Computations on Manycore Architectures
TimeWednesday, June 27th10am - 11am
DescriptionThe Hierarchical Computations on Manycore Architectures HiCMA library aims at redesigning and leveraging existing dense linear algebra libraries to exploit the data sparsity of the matrix operator using low-rank matrix approximations. Numerical low-rank approximations have demonstrated attractive theoretical bounds, both in memory footprint and arithmetic complexity. In fact, they have even become numerical methods of choice when designing high performance big data applications, especially when looking at the forthcoming exascale era, where systems with billions of threads will be routine resources at hand.
Senior Research Scientist Applied Mathematics and Computational Science Director