Mellanox EDR 100Gb Chosen for HPC CORAL

Mellanox Technologies, Ltd. announced its EDR 100Gb solutions have been selected as key components of new supercomputers at two of the national labs – Oak Ridge National Laboratory (ORNL) and Lawrence Livermore National Laboratory (LLNL). 

The hybrid supercomputer design will interconnect thousands of compute nodes containing both IBM POWER CPUs and NVIDIA GPUs via Mellanox’s EDR 100Gb IB-based solutions, providing one of the most advanced architectures of its kind for HPC applications.

CORAL (Collaboration of Oak Ridge, Argonne and Lawrence Livermore National Labs) is a project that was launched in 2013 to develop the technology and meet the Department of Energy’s 2017-2018 leadership computing needs with supercomputers. The collaboration between Mellanox, IBM and NVIDIA was selected by the CORAL project team after an evaluation of future technologies from a variety of vendors. The development of these supercomputers is underway with installation expected in 2017.

“The new embedded capabilities will enhance the ability of applications to effectively communicate latency-sensitive data,” said Jim Rogers, director of operations, National Center for Computational Sciences at ORNL. “The net result will be a well-balanced system that can dramatically improve performance for DOE’s mission-critical applications.”

“Simulations of the safety and reliability of elements of the American nuclear deterrent require arguably the most complex and sophisticated 3D applications on the planet,” said Michel McCoy, head of Livermore’s Advanced Simulation and Computing program. “These require the full capability of machines of this magnitude, and for these systems to effectively run these codes, extremely low latency, high bandwidth communication across the platform is essential.”

“Organizations and research facilities are required to process and analyze more information than ever before and to do it in less time,” said Michael Kagan, CTO, Mellanox. “Mellanox interconnect solutions deliver the highest performance and scalability, and provide the most advanced roadmap that paves the road to exascale computing. We are excited to collaborate with IBM, NVIDIA, ORNL and LLNL to build the most advanced supercomputers in the world.”