Allan Snavely


Associate Director

San Diego Supercomputer Center
University of California, San Diego
La Jolla, CA 92093

Office phone: (858) 534-5158
Office fax: (858) 534-5117

E-mail: asnavely at ucsd dot edu
URL: http://www.sdsc.edu/~allans




Allan Snavely is Associate Director at the San Diego Supercomputer Center (SDSC), and an Adjunct Associate Professor in the Department of Computer Science and Engineering (CSE) both at the University of California, San Diego (UCSD). Besides running a research lab at SDSC he also serves on the executive committee for the inter-departmental CSME program in computational science, mathematics, and engineering, is the computer science faculty advisor to the Free Clinic Project and shares running CSE294 Largescale Systems Seminar (alternating quarters) with Professor Scott Baden of CSE.

He is a noted expert in high performance computing (HPC), has published more than 75 cited papers with more than 1,400 citations on this subject (his current h-index is 23 i.e. 23 publications have been cited 23 or more times in other journal articles or conference proceedings) , he has presented numerous invited talks including briefing U.S. congressional staff on the importance of the field to economic competitiveness, has twice been a finalist for the Gordon Bell Prize (2007 and 2008) in recognition for outstanding achievement in HPC applications, and in 2009 shared the SC09 Storage Challenge Award for the design of Dash, an innovative new supercomputer that makes extensive use of flash memory. Dash is just a prototype of a much larger system he and and his team have built: Gordon, the most powerful supercomputer in the world for accessing data with 100 GB/s of bandwidth to disk and 36 million IOPS of random data access from 300 TB of NAND flash. He is Co-principal investigator, and Co-architect of Gordon: A Data Intensive Supercomputer .
Here is a video interview with Linux Magazine about the project. Here is a Wired magazine article about Gordon.

In 2000, he established the Performance Modeling and Characterization Laboratory where he has supervised numerous graduate students (both MS and Ph.D), post-docs, visiting scholars, and senior research staff. His research interests cover a wide spectrum in the areas of high performance computing. A common thread among his research projects focuses on understanding, and improving the Von Neumann Bottleneck that limited throughput (data transfer rate) between the CPU and memory compared to the amount of memory that in turn limits the performance of computers past and modern.

Allan Snavely received his Ph.D. in Computer Science from the University of California, San Diego


Selected Publications with Impact on High Performance Computing