Speaker

Xin Wu

Xin Wu did his doctoral research on the GPU-accelerated semiempirical quantum chemistry on high-performance computers at Max-Planck-Institut für Kohlenforschung (MPI KoFo) and received his PhD degree from Heinrich Heine University Düsseldorf in 2014. To Harness the computational power of GPUs all hotspots (> 99% walltime) of the MNDO program were ported to the hybrid CPU-GPU architecture and the total runtime for energy evaluation and geometry optimization of large molecules were reduced by one order of magnitude. Then, he continued working as a research associate in an ERC Advanced Grant project at MPI KoFo and focused on the Big Data analytics of ab initio electron integrals. At Paderborn Center for Parallel Computing his research interest is the code optimization, especially FPGA-accelerated kernels for quantum chemistry. In addition, he is actively involved in the HPC training courses, user support, and the North Rhine-Westphalia Competence Network for High Performance Computing.

Talks at this conference:

 Overview