Supercomputers/Systems/CRIEPI SGI ICE X
Japan · operational · hpc
CRIEPI SGI ICE X
Also known as CRIEPI ICE X (2012); Central Research Institute of Electric Power Industry SGI ICE X
Operated by CRIEPI .
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: 1263 of 1345 systems have a readable citation that names them, and 56 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
- 582.10 TFlop/s
- Rpeak
- 670.92 TFlop/s
- Rmax ÷ Rpeak
- 86.8%
- Cores
- 32,256
- Power
- 720 kW
- Per watt
- 0.8 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 | - SGI ICE X blade cluster, 2,016 nodes | SGI HPENYSE
rolls up to Hewlett Packard Enterprise | - | Verified | storagenewsletter.comtop500.org |
| CPU | Intel Xeon E5-2670 CPUs x86-64 · 8C · 2.6 GHz · 32 nm · 115 W 4,032 Intel Xeon E5-2670 8C 2.6GHz processors, 32,256 cores | Intel | 4,032 reported | Verified | storagenewsletter.comtop500.org |
| Interconnect | Mellanox InfiniBand FDR Interconnect Fat tree · 56 Gb/s per 4x port · About 1 microsecond 4x FDR InfiniBand | NVIDIA | - | Verified | storagenewsletter.comtop500.org |
| Storage | Lustre Storage Parallel filesystem, separate metadata and object servers 2.88 PB of disk on the Lustre parallel file system, supported by SGI | SGI HPENYSE
rolls up to Hewlett Packard Enterprise | - | Reported | storagenewsletter.com |
| Cooling | - Water cooling system that avoids additional air conditioning | SGI HPENYSE
rolls up to Hewlett Packard Enterprise | - | Reported | storagenewsletter.com |
| Operating system | SUSE Linux Enterprise Server Operating systems Linux distribution · Open source, commercial support subscription SUSE Linux Enterprise Server 11 | SUSE | - | Reported | top500.org |
The read
SGI ICE X scale-out cluster of the Central Research Institute of Electric Power Industry. SGI's November 2012 announcement (reproduced by StorageNewsletter) gives 2,016 blade nodes with 4,032 Xeon E5-2670 processors (32,256 cores), 67.5 TB of memory, 4x FDR InfiniBand, 2.88 PB of disk on a Lustre file system, water cooling, a 670.9 TFlop/s theoretical peak and operation since July 2012. TOP500 independently lists 32,256 cores, 582.1 TFlop/s Rmax, 670.92 TFlop/s Rpeak and 720 kW, 39th on the November 2012 list and last seen in November 2017. CRIEPI later ran an HPE SGI 8600 (2018) and an AMD EPYC cluster (2024), which are not in this dataset. The machine room is not named on the pages fetched, so no site is recorded; status is held as of the last TOP500 appearance.
Timeline
Public list appearances
Pointers only: rank and edition, linking to the canonical entry. We do not reproduce list tables. See sourcing policy.
Change history
- 2026-10-09 Added CRIEPI SGI ICE X
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 |
|---|---|---|
| storagenewsletter.com | names system, part and count | Intel Xeon E5-2670, Lustre · counts: 4032 |
| top500.org | names system and part | Intel Xeon E5-2670, Mellanox InfiniBand FDR, SUSE Linux Enterprise Server |
Checked 2026-10-13 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.