By GREGORY ZELLER //
The home of the Seawolves is running with a whole new pack, with the arrival of a new supercomputer that paces the world’s fastest.
Dubbed “Ookami” (loosely, the Japanese word for “wolf”), the new computer system – which runs on the same processor as Japan’s Fugaku supercomputer, recognized as the planet’s most powerful – has been installed at Stony Brook University’s Institute for Advanced Computational Science.
Operated by the IACS in partnership with the University at Buffalo’s Center for Computational Research, Ookami packs a Fujitsu A64FX processor. The lingo is thick with petaflops and ARM protocols; suffice it to say, the Fujitsu “system-on-chip” processor – the heart of what is literally Earth’s fastest supercomputer – breaks new digital ground.
And now it’s ready to break ground across a plethora of scientific pursuits, according to IACS Director Robert Harrison, who trumpeted Ookami’s “potential to transform computational research in many areas of science, engineering and industry.”
“This system has a nearly magical combination of programmability, performance and efficiency,” said Harrison, a Stony Brook professor of applied mathematics and statistics. “Ookami is a resource for researchers in academia or industry nationwide.”
The “system” Harrison referenced combines various components: primarily, the A64FX, Hewlett Packard’s HPE Apollo 80 (a high-performance computing platform) and the Cray ClusterStor E1000 storage system, another Hewlett Packard Enterprise product that tailors its software and hardware to fluctuating data-storage requirements of any size – a big deal for a supercomputer that figures to be very busy, very quickly.
“Its use opens the door for researchers to explore new computing technologies that could greatly impact the future of high-performance applications,” Harrison noted.
Ookami, which is supported by a $5 million National Science Foundation grant issued in 2019, also employs an old friend of the IACS: Bright Cluster Manager, a suite by California/Netherlands-based software specialist Bright Computing designed to manage high-performance Linux clusters (a connected array of computers or nodes working on the Linux operating system).
Not only is SBU a longtime user of Bright Cluster Manager, which will streamline the software’s deployment across new Ookami applications, but Bright Computing goes way back with Wisconsin-based Cray Computer Co., the recent HP acquisition behind the ClusterStor E1000.
“Bright Computing has a long history of working with Cray on state-of-the-art supercomputers,” Bright Computing Vice President Lee Carter said in a statement. “Our engineers have worked closely with Fujitsu’s A64FX-based development team to ensure Bright Cluster Manager seamlessly manages A64FX-based systems like the HPE Apollo 80.
“We are proud to play our part in bringing all these technologies together for our valued customer Stony Brook (University),” Carter added.