Supercomputers/Systems/Helios
Poland · operational · mixed
Helios
Operated by Cyfronet AGH at Cyfronet AGH, Krakow .
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
- 19.14 PFlop/s
- Rpeak
- 30.44 PFlop/s
- Rmax ÷ Rpeak
- 62.9%
- Cores
- 89,760
- Power
- 316.88 kW
- Per watt
- 60.4 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 EX4000 platform | Hewlett Packard Enterprise | - | Reported | cyfronet.pltop500.org |
| Accelerator | NVIDIA GH200 Grace Hopper Superchip Accelerators Arm CPU + CUDA GPU · TSMC 4N GPU partition: 440 NVIDIA GH200 Grace Hopper superchips | NVIDIA | 440 reported | Reported | cyfronet.pltop500.org |
| Accelerator | - Interactive partition: 24 NVIDIA H100 accelerators with NVMe storage | NVIDIA | 24 reported | Reported | cyfronet.pl |
| CPU | AMD EPYC 9004 (Genoa) CPUs x86-64 · 96C · 2.4 GHz · TSMC N5 · 360 W CPU partition: 98,304 AMD Zen 4 cores (4th Gen EPYC) | AMD | - | Reported | cyfronet.pl |
| Interconnect | HPE Slingshot-11 Interconnect Dragonfly · 200 Gb/s per port · Around 2 microseconds Slingshot network, 200 Gb/s | Hewlett Packard Enterprise | - | Reported | cyfronet.pltop500.org |
| Cooling | - Direct liquid cooling of the CPU and GPU partitions | Hewlett Packard Enterprise | - | Reported | cyfronet.pl |
AI datacenter metrics
Figures beyond an accelerator count, each with who stated it. What can and cannot be compared across AI datacenters.
| Metric | Value | As of | Basis | Note | Source |
|---|---|---|---|---|---|
| Accelerators | 440 accelerators | 2026-10 | Operator stated | GH200 superchips in the GPU partition (the page also lists 24 H100 in an interactive partition). Cyfronet cites a 'very low PUE' but gives no number. | cyfronet.pl |
The read
Cyfronet AGH's fifth-generation flagship, an HPE Cray EX4000 in Krakow built as three partitions: a CPU partition, a GPU partition of 440 NVIDIA GH200 superchips and an interactive partition of 24 NVIDIA H100 accelerators, all on 200 Gb/s Slingshot with direct liquid cooling of the CPU and GPU partitions. Cyfronet's current page states 37 PFlop/s theoretical peak, more than 131 thousand cores and 1.8 EFlop/s of AI performance, and gives the CPU partition as 98,304 AMD Zen 4 cores; earlier launch material quoted 75,264 cores, so the CPU core count is recorded only in the component edge. The 19.14 PFlop/s Rmax, 30.44 PFlop/s Rpeak, 89,760 cores and 316.88 kW are TOP500's figures for the GPU partition alone, which ranked 96th on the November 2025 list.
Timeline
Public list appearances
Pointers only: rank and edition, linking to the canonical entry. We do not reproduce list tables. See sourcing policy.
Site & facility
Change history
- 2026-10-04 Added AI datacenter metrics for helios
- 2026-10-01 Added Helios (Cyfronet AGH)
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 |
|---|---|---|
| top500.org | names system and part | NVIDIA GH200 Grace Hopper Superchip, HPE Slingshot-11 |
| cyfronet.pl | names system, part and count | AMD EPYC 9004 (Genoa), NVIDIA GH200 Grace Hopper Superchip · counts: 440 |
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.