Supercomputers/Systems/Isambard 3
United Kingdom · operational · hpc
Isambard 3
Also known as GW4 Isambard 3
Operated by University of Bristol at National Composites Centre, Bristol .
Phase 1 seed dataset, compiled by hand. These rows were built from public operator, laboratory and vendor sources. A mechanical second-reader pass has since fetched every cited source: 1092 of 1153 systems have a readable citation that names them, and 40 are genuinely weakly sourced. Every claim carries its source and a confidence tier. Treat anything below verified as a lead, not a citation.
Measured performance
- Rmax
- -
- Rpeak
- 2.70 PFlop/s
- Rmax ÷ Rpeak
- -
- Cores
- 55,296
- Power
- -
- Per watt
- - GF/W
Figures are as last publicly reported for the configuration described below, not a live measurement. Where a system was upgraded in place, the post-upgrade configuration is shown and the earlier one appears in the timeline.
What this machine is made of
One row per supplier relationship. “Supplier at build” is the company that shipped the part at the time; where that company has since been acquired, the parent it rolls up to today is shown beside it. That distinction is what makes ticker-level aggregation possible across a thirty-year dataset.
| Role | Part | Supplier at build | Quantity | Confidence | Source |
|---|---|---|---|---|---|
| Integrator | - HPE Cray XD224 air-cooled servers | Hewlett Packard Enterprise | - | Reported | cug.orgnvidianews.nvidia.com |
| CPU | NVIDIA Grace CPUs Armv9 (Neoverse V2) · 72C · 3.1 GHz · TSMC 4N 384 NVIDIA Grace CPU Superchips (two 72-core Grace CPUs each, 144 cores per node) | NVIDIA | 384 reported | Reported | cug.orgnvidianews.nvidia.com |
| Interconnect | HPE Slingshot-11 Interconnect Dragonfly · 200 Gb/s per port · Around 2 microseconds HPE Slingshot 11 at 200 Gbps | Hewlett Packard Enterprise | - | Reported | cug.org |
The read
The Bristol Centre for Supercomputing's CPU-only Arm service for UKRI, funded with GBP 10 million and shared by the GW4 universities of Bristol, Bath, Cardiff and Exeter. The CUG 2025 evaluation paper describes 384 HPE Cray XD224 nodes, each an NVIDIA Grace CPU Superchip with 144 Arm Neoverse V2 cores, 55,296 cores in six air-cooled racks of 64 servers, joined by 200 Gbps Slingshot 11 and hosted in the same modular data centre as Isambard-AI phase 1; the service moved into operation from January 2025. The 2.7 PFlop/s figure is the FP64 peak NVIDIA gave when announcing the system in 2023, not a measured HPL result, and NVIDIA's 'less than 270 kW' was a design target, so no power figure is recorded. No separate TOP500 entry for Isambard 3 was found.
Site & facility
National Composites Centre, Bristol
- Location
- Bristol, United Kingdom
- Cooling
- Direct liquid cooling
- Commissioned
- 2024
Change history
- 2026-10-02 Added Isambard 3, Bristol's 384-node Grace CPU Superchip service
Source check
We fetched this system's own citations and recorded whether each page actually mentions it. This is a corroboration signal, not a fact check, and it is published so you can see how well the sourcing holds up rather than take it on trust.
| Cited source | Result | Found on the page |
|---|---|---|
| cug.org | does not name this system | - |
| nvidianews.nvidia.com | names system, part and count | NVIDIA Grace · counts: 384 |
Checked 2026-10-09 by pnpm verify. Re-run it and the table changes with the web.
Further reading
Operating a supercomputer
Acceptance testing, why "installed" and "in production" are six months apart, and the machine lifecycle.
Inside a node
Latency-optimised versus throughput-optimised processors, NUMA, and the end of the host-to-device copy.
The interconnect
Topologies, why latency and tail behaviour matter more than bandwidth, and how the fabric market consolidated.