HPC: Reasoning and Purpose
About the Speaker
Linux Engineer at Oak Ridge National Laboratory
With over 23 years as Linux engineer in High Performance Computing, Sean Taylor has worked in many HPC firsts-in-the-field — Parallel File Systems, Infiniband, and GPU computational systems. Early in his career, he built one of the first Beowulf clusters for Air Forces Phillips Laboratory (now ARFL), and later worked on the first Intel Pentium petaflop cluster at Sandia National Laboratory, which laid the foundation for Infiniband and the SGI Taurus Mesh network. He was lead engineer on the Thunderbird Cluster, a 4500 node, Infiniband 10Gb networked cluster, that ranked No. 3 on the Top 500 supercomputer list for over a year. Over the years, Sean has contributed to R&D for both academic and commercial research — SGI-based modelling for NOAA at Princeton University, and as senior engineer on a 13,500 node cluster for Concoo Phillips. In his work for Roche and Pfizer, he supported research in Next Generation Sequencing (NGS) computational analytics systems, and he was an integral contributor to multiple drug and diagnostic discoveries in cancer research. He considers this latter work his most significant, as lives were saved with each new discovery. His current work at Oak Ridge National Laboratory is especially important to him as he now directly supports research for the U.S. Department of Veterans Affairs programs for health, cancer, suicide, and VA modernization projects.
Presentation Resources
These three little letters — HPC — we've all heard them, and we're pretty sure they stand for High Performance Computing, but what does that mean? What kind of work and technical considerations does this entail? In this talk we'll explore this other field of computing, including the history and reasoning behind it, how it's different from standard, enterprise IT, and its present hurdles and future direction. If yours is an inquiring mind eager to explore other realms of computing, then this is a talk you won't want to miss.