AMD to power World’s Fastest Super Computer – Frontier

U.S. Department of Energy, Oak Ridge National Laboratory, and Cray Inc to build the World’s Fastest Super Computer – Frontier, in collaboration with AMD

It’s been a few months or so that AMD has announced the fastest Server processors i.e. the EPYC Rome CPUs based on their latest Zen 2 architecture build on the 7nm process. This gives the chips immense power to solve heavy tasks and accommodate more cores and threads on to a single unit.

Already Amazon had declared their partnership with AMD in their Cloud Serving platform. And now this is a big achievement by the Red Team to collaborate with U.S. Department of Energy, Oak Ridge National Laboratory, and Cray Inc. to build the World’s Fastest Super Computer.

Named as the Frontier, this Super Computer will have a horsepower up to 1.5 exaflops. This is the highest peak processing power any Super Computer has ever achieved in the history of computing.

AMD innovations to be used in the Frontier system include:

• Future-generation High-Performance Computing (HPC) and Artificial Intelligence (AI) optimized, custom AMD EPYC CPU, and Radeon Instinct GPU processors supported by High Bandwidth Memory (HBM) and extensive mixed precision ops for optimum deep learning performance;
• A custom high-bandwidth, low-latency coherent Infinity Fabric, connecting four AMD Radeon Instinct GPUs to one AMD EPYC CPU per node;
• An enhanced version of the open source ROCm programming environment, developed with Cray to tap into the combined performance of AMD CPUs and GPUs.

Using the latest EPYC Rome CPUs that has highest up to 64 cores and 128 threads on a single CPU, you can get immense power and can be extended on a dual socket motherboard which will increase the amount to 128 cores and 256 threads.

EPYC™ processors are driving a new era in the data centres, with exceptional scalability, energy efficiency, memory bandwidth, and I/O capability.

This makes the EPYC processor an excellent choice for HPC solutions. The Frontier system will be utilizing a custom EPYC CPU optimized for HPC and AI computation.

To empower good computing, a great GPU is also required with a high performing chip as well. So here AMD has used the Radeon Instinct GPUs based on their 7mm Vega graphics architecture.

These Radeon Instinct GPUs have high bandwidth memory, and scalable fabric and communications technologies. To support the computational challenges of the Frontier system, Radeon Instinct™ accelerators will be configured in a tightly integrated, highly efficient 4:1 GPU to CPU ratio.

In order to scale performance of these EPYC processors to exascale levels, AMD has improved their original Infinity Fabric Technology. This provides unified, coherent connectivity between CPU, GPU, and memory.

The Infinity Fabric streamlines data movement, enabling faster execution of complex calculations or analysis of large scientific data sets.

The ROCm open software ecosystem is used in this Super Computer. This is an open-source platform for HPC-class heterogeneous compute and world-class datacenter system designs.

The ROCm open ecosystem incorporates optimized drivers, compilers, libraries, support for OpenMP, and the tools needed for large HPC systems.

“We are excited to work with the team at AMD to deliver the Frontier system to Oak Ridge National Laboratory.
The combination of Cray and AMD technology in the Frontier system will dramatically enhance performance at scale for AI, analytics, and simulation, enabling DOE to further push the boundaries of scientific discovery.” ”

Steve Scott, senior vice president and CTO at Cray

This is a great achievement for AMD and this can pave new directions for the Red Team in the Super Computing sector. This new age Super Computer Frontier is scheduled to be delivered to ORNL in 2021. Sources: AMD.

Do check out:

😎 TechnoSports-stay UPDATED 😎