News
News Categories

NVIDIA pushes the Cloud and supercomputing to new heights with the new Quantum-2

By Ken Wong - on 10 Nov 2021, 5:53am

NVIDIA pushes the Cloud and supercomputing to new heights with the new Quantum-2

Welcome to the power of Quantum-2. Image source: NVIDIA.

As part of its NVIDIA GTC event, NVIDIA Chief Executive Jensen Huang unveiled an update to its InfiniBand networking platform, called Quantum-2.

Huang introduced NVIDIA Quantum-2 as “the most advanced networking platform ever built.” And with the BlueField-3 DPU, Quantum-2 offers the extreme performance, broad accessibility and strong security needed by cloud computing providers and supercomputing centres, he said.

It comes as more businesses and institutions are using supercomputers for research, simulations, stock analysis, and computational number crunching. In these instances, Quantum-2 will accelerate the most demanding workloads.

According to Huang, until Quantum-2, you had to settle for either the performance of a supercomputer or the shareability of Cloud Computing. With Quantum-2, the supercomputer is now Cloud-native and will be far better utilised. NVIDIA said that the activities of a single tenant won't affect the others through an advanced telemetry-based congestion control system that ensures sufficient throughput is available, regardless of spikes in demand or users.

 

Helping it deliver more than the original

Nvidia’s InfiniBand is a computer-networking communications standard used in high-performance computing for data flow between processors and input/output devices, known for its high throughput and very low latency.

The NVIDIA Quantum-2 is a 400Gbps InfiniBand networking platform that consists of an NVIDIA Quantum-2 switch, a ConnectX-7 network adapter, a BlueField-3 data processing unit (DPU), and all the software that supports the new architecture.

The new Quantum-2 InfiniBand switch has 57 billion transistors on 7-nanometer silicon.

The new Quantum-2 InfiniBand switch has 57 billion transistors on 7-nanometer silicon, it is slightly bigger than the NVIDIA A100 GPU with 54 billion transistors. It features 64 ports at 400Gbps or 128 ports at 200Gbps and will be offered in a variety of switch systems up to 2,048 ports at 400Gbps or 4,096 ports at 200Gbps, more than 5x the switching capability over the previous generation, Quantum-1.

The NVIDIA Quantum-2 platform provides two networking end-point options, the NVIDIA ConnectX-7 NIC and NVIDIA BlueField-3 DPU InfiniBand.

ConnectX-7 has 8 billion transistors in a 7-nanometer design.

ConnectX-7, with 8 billion transistors in a 7-nanometer design, doubles the data rate of the world’s current leading HPC networking chip, the NVIDIA ConnectX-6. It also doubles the performance of RDMA, GPUDirect Storage, GPUDirect RDMA and In-Networking Computing.

The BlueField-3 InfiniBand has 22 billion transistors in a 7-nanometer design.

BlueField-3 InfiniBand, with 22 billion transistors in a 7-nanometer design, offers sixteen 64-bit Arm CPUs to offload and isolate the data centre infrastructure stack.

 

Support from industry partners

The NVIDIA Quantum-2 switch is now available from a wide range of leading infrastructure and system vendors around the world, including Atos, DataDirect Networks (DDN), Dell Technologies, Excelero, GIGABYTE, HPE, IBM, Inspur, Lenovo, NEC, Penguin Computing, QCT, Supermicro, VAST Data and WekaIO.

 

Join HWZ's Telegram channel here and catch all the latest tech news!
Our articles may contain affiliate links. If you buy through these links, we may earn a small commission.