COMPUTE ATLAS Supercomputer supply-chain graph
1581 systems 558 sites Sourced data

Supercomputers/Systems/ZIB Cray T3D

Germany · operational · hpc

ZIB Cray T3D

RSS for this system Your watchlist

Also known as Berlin Landeshoechstleistungsrechner T3D

Operated by Zuse Institute Berlin .

Researched with AI assistance, checked by scripts. These rows were researched by AI agents reading public operator, laboratory and vendor sources, appended by scripts and checked by automated validation; no subject-matter expert has reviewed them (how this site is made). A mechanical second-reader pass has since fetched every cited source: 1350 of 1581 systems have a readable citation that names them, and 64 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
25.3 GFlop/s
HPL measured
Rpeak
38.4 GFlop/s
Rmax ÷ Rpeak
65.9%
Cores
256
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.

How it compares

Ranked against the other machines in this dataset that report the same figure: those that first came into service within two years of ZIB Cray T3D, and those in Germany. Computed from our tables; a metric is ranked only when at least eight peers report it.

MetricValueAgainst its eraAgainst its country
Rmax25.3 GFlop/s#21 of 29 · bottom 30%#94 of 95 · bottom 2%
Rmax ÷ Rpeak65.9 %#17 of 28 · bottom 42%#64 of 94 · bottom 33%

Every metric is ranked largest-first, so a high power-draw rank means a big power draw, not a good one. Rmax ÷ Rpeak uses HPL-measured results only. Rankings move as machines are added; they describe this dataset, not the world.

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 -
Cray T3D SC256-8/464
 Cray
HPENYSE rolls up to Hewlett Packard Enterprise
- Reported top500.orgnetlib.org
CPU -
DEC Alpha EV4 149 MHz, 256 processors
 Digital Equipment Corporation
HPENYSE rolls up to Hewlett Packard Enterprise
256
reported
 Reported top500.org
Interconnect -
Cray T3D 3-D torus
 Cray
HPENYSE rolls up to Hewlett Packard Enterprise
- Reported top500.org
Operating system -
UNICOS
 Cray
HPENYSE rolls up to Hewlett Packard Enterprise
- Reported top500.org

The read

Cray T3D with 256 DEC Alpha EV4 149 MHz processors on a 3-D torus running UNICOS at the Konrad Zuse Zentrum (ZIB) in Berlin, Rmax 25.3 and Rpeak 38.4 GFlop/s, listed from November 1995 (rank 38) to June 1998 (rank 209) (https://www.top500.org/system/170975/). The TOP500 report on the 1994 German market (https://netlib.org/benchmark/top500/reports/report94/Europe/node5.html) says a 192-processor T3D had been installed at ZIB in December and was not yet in the 1994 list, with an extension to 256 processors planned for later that year; read in context (the report follows the November 1994 list) the installation month is December 1994, which is the start date used. TOP500 lists the 192-processor stage as a separate June 1995 entry at rank 39 (Rmax 19.05 GFlop/s, https://www.top500.org/system/170966/); this record covers the final 256-processor configuration. The report does not name the exact build-out date of the 256-processor stage; TOP500 gives 1995. ZIB's current systems are lise and konrad-hlrn-iii.

Timeline

  • 1994-12 First operational source
  • 1995-11 List debut at #38 (November 1995) source

Public list appearances

Pointers only: rank and edition, linking to the canonical entry. We do not reproduce list tables. See sourcing policy.

Change history

Nearby in scale

The machines closest to ZIB Cray T3D by Rmax, smaller to larger.

Same operator

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.