Systems developers from IBM and NVIDIA will work with applications scientists to develop the tools, technologies and applications that will perform best on the supercomputers.
When the newest supercomputers come online at the Department of Energy national labs, they’ll be working at peak efficiency, thanks to two new Centers of Excellence.
Working with IBM and NVIDIA, the Lawrence Livermore National Laboratory and the Oak Ridge National Laboratory have each established mission-specific Centers of Excellence to support the advanced, large-scale scientific and engineering applications that will run on two new supercomputers to be built by IBM with NVIDIA processors.
The Summit and Sierra supercomputers use IBM’s OpenPOWER architecture and POWER processors, NVIDIA Tesla GPU accelerators and the NVIDIA NVLink high-speed processor interconnect, which will minimize data in motion, optimize problem solving and time to solution while driving down energy consumption and computing cost, IBM said.
Computational scientists from IBM and NVIDIA will work with applications scientists from each lab to create the tools, technologies, programming models, algorithms and applications that will perform best on the supercomputers, which will be delivered in 2017 and operational in 2018.
This communication between the hardware and software specialists with the lab’s application developers will ensure the ongoing system design is compatible with user applications.
Specifically, the Center for Excellence will support development of at least 13 applications for Oak Ridge’s Summit supercomputer, and the program at Lawrence Livermore will support the lab’s efforts to evolve national security applications in areas of safety, reliability and security of the nuclear supply, as well as national security computational science measures like biosecurity, energy security and global warming.
“Along with NVIDIA, our scientists are ensuring Oak Ridge and Lawrence Livermore are able to get the most out of these revolutionary supercomputers to reach the next level of scientific discovery,” said Dave Turek, IBM vice president of HPC market engagement.
The wide collaboration will also guarantee that developed applications are capable of running on all of the DOE supercomputers.
NEXT STORY: Comatose servers draining data center efficiency