Visit the official project webpage: https://www.stonybrook.edu/ookami/

Ookami

Ookami is a testbed for researchers nationwide to explore future supercomputing technologies and to advance computational and data-driven research on pressing science challenges.  

• Robert J. Harrison (IACS/SBU)
• Barbara Chapman (IACS/SBU)
• Alan Calder (IACS/SBU)
• Firat Coskun (IACS/DoIT/SBU)
• Matt Jones (CCR/UB)
• Eva Siegmann (IACS/DoIT)

Description
Ookami has been the first computer outside of Japan with the A64fx processor developed by Riken and Fujitsu for the Japanese path to exascale computing. Now in early silicon, by focusing on crucial architectural details, the ARM-based, multi-core, 512-bit SIMD-vector processor with ultrahigh-bandwidth memory promises to retain familiar and successful programming models while achieving very high performance for a wide range of applications. It supports a wide range of data types and enables both HPC and big data applications.

Unique and novel features include:

1. first access for US research and first planned install of A64fx outside of Japan;
2. first CPU worldwide with on-package high bandwidth memory (1 TB/s; 82% peak measured on STREAM triad is 8x Sky Lake socket);
3. first CPU worldwide to implement the ARM scalable vector extensions with multiple features to facilitate auto-vectorization;
4. high CPU performance (2.76dpTFLOP/s; dgemm 94% peak) balanced to match memory speed giving ~0.4 bytes/dpFLOP is unparalleled in modern CPUs (4.6x Sky Lake, 3x NVIDIA V100);
5. green technology with high power efficiency.

What does this mean for science? Compared with the best CPUs anticipated for the deployment era, A64fx offers 2-4x better performance on memory-intensive applications such as sparse-matrix solvers found in many engineering and physics codes. For nearly all other applications performance is also better or competitive. This transformational performance should be available nearly out of the box, with additional performance possible from tuning. To the scientist or engineer this means faster time to solution with significantly less programmer effort. The target applications to be enabled are memory-bandwidth intensive with 32GB/node memory, with significant gains anticipated for many other applications. Analysis of XSEDE workload shows 86% of all jobs (85% cycles) will fit within the available memory per node and that the majority of jobs are memory-bandwidth intensive.

 

Institute for Advanced Computational Science (IACS)
IACS presently has 13 faculty spanning chemistry, materials by design, condensed matter, astrophysics, atmospheric science, linguistics, nano-science, sociology, applied mathematics, and computer science. The institute began with a transformational $10 million anonymous donation plus matching funds of equal value from the Simons Foundation that enabled Stony Brook University to establish our institute. Our integrated, multidisciplinary team of faculty, students, and staff overcome the limitations at the very core of how we compute, collectively take on challenges of otherwise overwhelming complexity and scale, and individually and jointly define new frontiers and opportunities for discovery through computation. 

University at Buffalo Center for Computational Research
CCR is more than just a supercomputing center.  It has been our goal since CCR was formed in 1998 to not only enable research at UB but also support our local economy and provide educational outreach to local schools.  CCR has hosted annual workshops for high school students in a variety of topics including computer programming, bioinformatics, and computer visualization.  We support computational intensive courses offered at UB in fields such as Computer Science & Engineering, Bioinformatics, and Chemistry.  We have developed industrial partnerships with many local companies to help aid in their business development and now offer a cluster just for local industrial clients.  We also collaborate with our collegues in the high performance computing field and have been at the forefront of the development of open source tools for use by HPC centers to provide quantitative and qualitative metrics relevant to HPC. 

Support for this award (OAC-1927880) comes from the National Science Foundation:  
(https://www.nsf.gov/awardsearch/showAward?AWD_ID=1927880)