Exascale Day Audio: Tzanio Kolev

Impact Area: Discretization Algorithms

Tzanio Kolev, Lawrence Livermore National Laboratory

Transcript

My name is Tzanio Kolev and I am a computational mathematician at Lawrence Livermore National Laboratory.

I am also the director of the Center for Efficient Exascale Discretizations (CEED) in the ECP [Exascale Computing Project] where we develop cutting-edge (discretization) algorithms for large-scale simulation codes.

These algorithms, including high-order methods, are important because they can take advantage of advanced hardware to produce more accurate scientific predictions while running faster.

I am excited about exascale computing because it can open the door to such next-generation simulation algorithms which can really transform the world of HPC [high-performance computing] as we know it today.

The power of exascale can make the current hard simulations routine and the problems that are impossible to solve today, possible tomorrow.

DOE [the US Department of  has always been at the forefront of scientific computing, and in the ECP we are pushing the mathematics and the software to their limits.

For example, with the high-order methods we are developing under the ECP, exascale applications will be able to tune performance to different hardware, and fully utilize accelerators such as GPUs.

This will be an opportunity to push the boundaries of scientific computing and foster new collaborations between application scientist, algorithm developers, and hardware partners.

This is a truly exciting time in computational science because of the really challenging problems we have to solve and the amazing research teams and computing resources we are putting together to solve them.

HPC will look very different at the exascale and we, at the ECP, are excited to help build the software ecosystem and exascale applications of tomorrow.