# Edmond Chow : Parallel Computing Issues in Computational Chemistry

In computational mathematics and science, it is now essential to consider computer hardware issues if a new algorithm is to be deployed. One such issue is the prevalence of parallelism in almost all levels of computer hardware. We discuss some of the challenges presented by computer hardware and some potential solutions in the context on quantum chemistry algorithms. Important considerations include reducing data movement, load balance across processors, and use of SIMD (single instruction, multiple data) features in modern processors. Specific results we have obtained include efficient computations using Hartree--Fock approximations on more than 1.5 million processor cores, and a new library for computing electron repulsion integrals that is designed for SIMD operation. These results are joint work with Ben Pritchard, Xing Liu, and the Intel Parallel Computing Lab.

**Category**: Applied Math and Analysis**Duration**: 01:34:48**Date**: February 15, 2016 at 4:25 PM**Views**: 113-
**Tags:**seminar, Applied Math And Analysis Seminar

## 0 Comments