Supercomputers/Systems/Mahuika
New Zealand · operational · mixed
Mahuika
Also known as NeSI Mahuika HPC3
Operated by NeSI at NIWA HPC Facility (Greta Point) .
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
- -
- Rmax ÷ Rpeak
- -
- Cores
- -
- 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 supplied the AMD Milan and NVIDIA HGX A100 extension | Hewlett Packard Enterprise | - | Reported | scientific-computing.com |
| Accelerator | NVIDIA A100 SXM4 80GB Accelerators NVIDIA HGX 80GB A100, 24 in Milan and Genoa nodes | NVIDIA | - | Reported | scientific-computing.comdocs.nesi.org.nz |
| Accelerator | - NVIDIA H100 (8) in Genoa nodes | NVIDIA | - | Reported | docs.nesi.org.nz |
| CPU | AMD EPYC 7713 (Milan) CPUs x86-64 · 64C · 2 GHz · 7 nm · 225 W AMD EPYC Milan 7713, 64 dual-socket nodes | AMD | - | Reported | docs.nesi.org.nz |
| CPU | AMD EPYC 9004 (Genoa) CPUs x86-64 · 96C · 2.4 GHz · TSMC N5 · 360 W AMD EPYC Genoa 9634, 64 dual-socket nodes of 168 cores | AMD | - | Reported | docs.nesi.org.nz |
| Storage | WEKA Data Platform Storage Parallel filesystem, software-defined WEKA filesystems on HPC3 | WEKA | - | Reported | docs.nesi.org.nz |
The read
NeSI's capacity-class platform for New Zealand researchers, hosted at NIWA's Wellington facility alongside the Maui system. It went online in 2018 as a Cray CS400 with Broadwell nodes and P100 GPUs; the current HPC3 generation, per NeSI's documentation, has 64 dual AMD EPYC Milan 7713 nodes (128 cores each, inherited from the HPE-supplied Mahuika extension that carried NVIDIA HGX 80GB A100 GPUs) and 64 dual AMD EPYC Genoa 9634 nodes (168 cores each, 16 of them with 1.5 TB), with 8 A100 in Genoa nodes, 16 A100 in Milan nodes, 8 H100 and 16 L4, WEKA filesystems and Rocky Linux 9. The P100s have been retired. Totals are not published, so cores are left blank rather than summed from per-node figures that NeSI describes as in transition. No FLOPS figure is published.
Timeline
- 2018 First operational source
Site & facility
NIWA HPC Facility (Greta Point)
- Location
- Wellington, New Zealand
- Commissioned
- 2018
Change history
- 2026-10-02 Added Mahuika, NeSI's New Zealand HPC platform
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 |
|---|---|---|
| docs.nesi.org.nz | names system and part | NVIDIA A100 SXM4 80GB, WEKA Data Platform |
| scientific-computing.com | names system and part | NVIDIA A100 SXM4 80GB |
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.
Storage and I/O
Parallel file systems, why checkpointing dominates the write load, and the metadata failure mode nobody expects.