This session discusses high-performance computing (HPC) scientific applications that use tile low-rank matrix computations. It revisits tile algorithms that use low-rank matrix approximations by exploiting the data sparsity of the dense operator, which comes from computational astronomy, seismic imaging, and climate and weather prediction applications.

The Hierarchical Computations on Manycore Architectures (HiCMA) software library provides sequential numerical kernels and a oneAPI runtime system for orchestrating the resulting computational tasks onto parallel systems. HiCMA demonstrates performance superiority against state-of-the-art numerical libraries with high productivity in mind.

Hatem Ltaief is the principal research scientist at the Extreme Computing Research Center at King Abdullah University of Science and Technology (KAUST). His research interests include parallel numerical algorithms, parallel programming models, HPC, and performance optimizations. He has been collaborating with domain scientists on using their applications to meet the challenges at exascale.