Britain’s most powerful supercomputer to use NVIDIA Superchips

The UK’s new supercomputer will be housed in containers in a car park at the National Composites Centre in Bristol
The UK’s new supercomputer will be housed in containers in a car park at the National Composites Centre in Bristol

During the AI Safety Summit in Bletchley Park, the U.K. government announced a £225 million ($273 million) investment to build one of the world’s fastest AI supercomputers, Isambard-AI.

British Prime Minister Rishi Sunak said in a speech that AI will bring a transformation as far-reaching as the industrial revolution, the coming of electricity, or the birth of the internet. He also emphasized the importance of international collaboration to ensure the safe and responsible development of AI.

The announcement of Isambard-AI is a significant step forward for the U.K.’s AI research capabilities and its commitment to leading the way in the responsible development of this transformative technology.

grace hopper superchip 1 e1670969676880
The NVIDIA GH200 Grace Hopper Superchip combines the NVIDIA Grace™ and Hopper™ architectures using NVIDIA® NVLink®-C2C to deliver a CPU+GPU coherent memory model for accelerated AI and HPC applications.

TL;DR: At the AI Safety Summit, the U.K. government announced a £225 million investment to build one of the world’s fastest AI supercomputers, called Isambard-AI, it’s the latest in a series of systems named after a legendary 19th century British engineer and it will pack 5,448 NVIDIA GH200 Grace Hopper Superchips to deliver a whopping 21 exaflops of AI performance for researchers across the country and beyond.

AI contributes nearly £4 billion to the UK economy and employs over 50,000 people, according to the nation’s Science, Innovation and Technology Secretary. The government’s investment in a new AI Research Resource in Bristol will boost scientific discovery and keep the UK at the forefront of AI development.

“Isambard-AI represents a huge leap forward for AI computational power in the U.K.,” said Simon McIntosh-Smith, a Bristol professor and director of the Isambard National Research Facility. “Today, Isambard-AI would rank within the top 10 fastest supercomputers in the world and, when in operation later in 2024, it will be one of the most powerful AI systems for open science anywhere.”

Hewlett Packard Enterprise, which is building Isambard-AI, is also collaborating with the University of Bristol on energy-efficiency plans that support net-zero carbon targets mandated by the British government.

Isambard-3 is part of a research alliance between the universities of Bath, Bristol, Cardiff, and Exeter.

Energy-Efficient HPC

A new supercomputer coming to the National Composites Centre (NCC) in 2024 will showcase the energy efficiency of Arm processors for non-accelerated high performance computing (HPC) workloads.

grace hopper overview
Figure 2. NVIDIA Grace Hopper Superchip logical overview

Isambard-3 is expected to deliver 2.7 petaflops of FP64 peak performance while consuming less than 270 kilowatts of power, making it one of the three greenest non-accelerated supercomputers in the world. The system will be powered by 384 Arm-based NVIDIA Grace CPU Superchips and will be used for medical and scientific research.

  • NVIDIA Grace CPU:
    • Up to 72x Arm Neoverse V2 cores with Armv9.0-A ISA and 4×128-bit SIMD units per core.
    • Up to 117 MB of L3 Cache.
    • Up to 512 GB of LPDDR5X memory delivering up to 546 GB/s of memory bandwidth.
    • Up to 64x PCIe Gen5 lanes.
    • NVIDIA Scalable Coherency Fabric (SCF) mesh and distributed cache with up to 3.2 TB/s memory bandwidth.
    • High developer productivity with a single CPU NUMA node.
  • NVIDIA Hopper GPU:
    • Up to 144 SMs with fourth-generation Tensor Cores, Transformer Engine, DPX, and 3x higher FP32 and FP64 throughout compared to the NVIDIA A100 GPU.
    • Up to 96 GB of HBM3 memory delivering up to 3000 GB/s.
    • 60 MB L2 Cache.
    • NVLink 4 and PCIe 5.
  • NVIDIA NVLink-C2C:
    • Hardware-coherent interconnect between the Grace CPU and Hopper GPU.
    • Up to 900 GB/s total bandwidth, 450 GB/s/dir.
    • The Extended GPU Memory feature enables the Hopper GPU to address all CPU memory as GPU memory. Each Hopper GPU can address up to 608 GB of memory within a superchip.
  • NVIDIA NVLink Switch System:
    • Connects up to 256x NVIDIA Grace Hopper Superchips using NVLink 4.
    • Each NVLink-connected Hopper GPU can address all HBM3 and LPDDR5X memory of all superchips in the network, for up to 150 TB of GPU addressable memory.
0 0 votes
Article Rating
1 Comment
Oldest
Newest Most Voted
Inline Feedbacks
View all comments

[…] Britain’s most powerful supercomputer to use NVIDIA Superchips […]