MEMS Seminar: Hardware and Software Technology for Computational Atomistic Simulation
Wednesday, February 1, 2017 - 1:30pm to 2:30pm
Dr. Jeff Hammond
Computational atomistic simulation represents a significant fraction of high-performance computing, with applications ranging from molecular dynamics to quantum many-body methods. I will discuss software development for high-performance computing systems based upon my experience with the well-known computational chemistry code NWChem and highlight important new developments for tensor contractions, which are the foundation of most quantum many-body methods. An overview of the latest Xeon Phi processor (Knights Landing) will be covered, along with how NWChem is being optimized for it using the vendor-agnostic OpenMP programming model.
Jeff Hammond is a System Architect at Intel in the exascale co-design group, where he studies the behavior of computational science applications and develops hardware and software technology for next-generation supercomputers. Jeff contributes to the development of open standards for parallel programming, including MPI, OpenMP and OpenSHMEM, and is an enthusiastic developer of open-source software. Prior to joining Intel, Jeff was a computational scientist at Argonne National Laboratory. He received his PhD in chemistry from the University of Chicago and undergraduate degrees in chemistry and mathematics from the University of Washington. See https://urldefense.proofpoint.com/v2/url?u=https-3A__jeffhammond.github.io_&d=CwIGaQ&c=imBPVzF25OnBgGmVOlcsiEgHoG1i6YHLR0Sj_gZ4adc&r=oTPmdDPp1PM5aGFibB1JnetIxv0zbRu2DoUTNsaW9Ms&m=Mii5u7eOrHn_M5q56Z3PJpkLrGEQoBOvBZcsOZvBl_Q&s=afJlgYMVeayVY9mo-vYDBxSDX3GVnWsZea88amMjx2w&e= for details.
Lunch will be served from 1:00 – 1:30 pm