NVIDIA (CEO Jensen Huang) accelerates AI, digital twins, quantum computing, edge, and HPC.

NVIDIA announced acceleration solutions for various industries at ISC 2022 on the 7th.

“Researchers are pushing high-performance computing (HPC) to the cutting edge by using highly sensitive equipment or accelerating simulations of hybrid quantum systems,” said Ian Buck, vice president of accelerated computing at NVIDIA, in a special address at the ISC.

■ CPU acceleration up to 3x with super chips

The Venado supercomputer at Los Alamos National Laboratory delivers 10 exaflops of AI performance for work in areas such as materials science and renewable energy.

Benado said the NVIDIA Grace Hopper superchip can run workloads up to three times faster than previous GPUs and double the performance per watt of traditional CPUs in non-accelerated applications.

■ BlueField-2

Texas firstThe single computing center said it is adding BlueField-2 DPUs to Lonestar6's NVIDIA Quantum InfiniBand network.

It is expected to be a development platform for cloud-native supercomputing that safely isolates workloads and hosts multiple users and applications with bare-metal performance, he said.

■ Expanding the Exascale Computing Deployment Ecosystem

NVIDIA and SiPearl announced a collaboration to expand the developer ecosystem building exascale computing on Arm.

Nvidia's accelerated computing and networking technologies could port applications to systems using Cypher's Rhea and future Arm-based CPUs, he added.

■Project progress with heterogeneous 3D applications

Ian Buck demonstrated a virtual fusion power plant being built in NVIDIA Omniverse by researchers at the UK Atomic Energy Agency and the University of Manchester.

NVIDIA Omniverse is a 3D design collaboration and world simulation platform that allows remote researchers to collaborate on projects in real time using different 3D applications.

■ MONAI, an AI framework

Ian Buck explains how researchers created a library of 100,000 synthetic images of the human brain on NVIDIA's Cambridge-1, a supercomputer dedicated to advancing healthcare with AI.

King's College London said it has used MONAI, an AI framework for medical imaging, to generate lifelike images that could help researchers study diseases like Parkinson's.br />
■ Clara Holoscan

At Lawrence Berkeley National Laboratory, we're using NVIDIA Clara Holoscan to see in real time, down to the nanometer, what would take days on a CPU.

NVIDIA is developing HPC Holoscan, a highly scalable version of its imaging software, to bring supercomputing to the edge.

■ Accelerating Quantum Simulation

Ian Buck said the NVIDIA cuQuantum software development kit, which integrates with major quantum software frameworks, allows users to access GPU acceleration without any additional coding.

“Quantum computing has enormous potential, and simulating quantum computers on GPU supercomputers is essential to getting us closer to meaningful quantum computing,” said Ian Buck. “We’re really excited to be at the forefront of this work.”