US Army’s Excalibur System Solves 10 Billion Equations in 3 Minutes
Stanford computer scientists working on the US Army’s Army Resource Laboratory (ARL) in Maryland conducted a unique experiment to test the computing power of the army’s newly acquired Excalibur Cray XC40 before the system is placed into regular service.
The Excalibur system comes with 101,184 processors, and the Stanford team had access to 22,00 of them. The team was working on a new scalability algorithm that would provide more efficient use of resources in high-performance environments. According to team leader Charbel Farhat, “We solved over 10 billion equations in a little over three minutes.”
According to the press release, “The entire Excalibur probably won’t be available for a repeat of this performance, but Farhat said he believes this new scalability algorithm will be tremendously useful in smaller computing systems.