Supercomputers/Systems/NCSA CM-5/512
USA · operational · hpc
NCSA CM-5/512
Also known as NCSA Connection Machine 5
Operated by NCSA .
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
- 30.4 GFlop/s
- Rpeak
- 65.5 GFlop/s
- Rmax ÷ Rpeak
- 46.4%
- Cores
- 512
- 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 NCSA CM-5/512, and those in USA. Computed from our tables; a metric is ranked only when at least eight peers report it.
| Metric | Value | Against its era | Against its country |
|---|---|---|---|
| Rmax | 30.4 GFlop/s | #8 of 20 · top 37% | #328 of 332 · bottom 2% |
| Rmax ÷ Rpeak | 46.4 % | #17 of 19 · bottom 12% | #320 of 331 · bottom 4% |
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 | - Thinking Machines CM-5/512 | Thinking Machines | - | Reported | top500.orgchessprogramming.org |
| CPU | Thinking Machines CM-5 vector node CPUs SPARC + custom vector · 1C · 0.032 GHz Thinking Machines CM-5 nodes (SuperSPARC 32 MHz with vector units), 512 nodes | Thinking Machines | 512 reported | Reported | top500.orgchessprogramming.org |
| Interconnect | - Hypercube, tree | Thinking Machines | - | Reported | top500.org |
| Operating system | Thinking Machines CMost Operating systems Unix (TMC-proprietary variant) · Proprietary CMost | Thinking Machines | - | Reported | top500.org |
The read
Thinking Machines CM-5 with 512 SuperSPARC 32 MHz nodes (hypercube and tree networks, CMost) at NCSA, Rmax 30.4 and Rpeak 65.54 GFlop/s in every appearance, listed from June 1993 (rank 3, behind only the Los Alamos CM-5/1024 and the Minnesota CM-5/544) to November 1996 (rank 52) (https://www.top500.org/system/167057/). The Chess Programming Wiki, quoting Chris Joerg's PhD thesis on the *Socrates program, says the program ran in June 1994 on a 512-node CM-5 at NCSA at the University of Illinois, and that the MIT StarTech program played the 1993 ACM computer chess championship on NCSA's 512-processor CM-5; this confirms the size and the early-1990s use but not the installation date, so the 1993 TOP500 installation year is kept. Its siblings are the Los Alamos CM-5/1024 (cm-5), NCAR's Littlebear (ncar-cm5-littlebear) and the NSA machine (nsa-frostburg-cm5). NCSA moved to scalable Unix systems afterwards and the retirement date is unknown.
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
- 2026-10-18 Added NCSA CM-5/512 (ncsa)
Nearby in scale
The machines closest to NCSA CM-5/512 by Rmax, smaller to larger.
- FROSTBURG · USA, operational · 0 PFlop/s
- UTokyo HITAC S-3800/480 · Japan, decommissioned · 0 PFlop/s
- TU Delft Cray T3E · Netherlands, operational · 0 PFlop/s
- Toyota CRDL SX-4/20 · Japan, operational · 0 PFlop/s
- CM-5/1024 · USA, decommissioned · 0 PFlop/s
- HWW Stuttgart NEC SX-4/32 · Germany, operational · 0 PFlop/s
Related systems
Same operator
- NCSA Titan Itanium Cluster · USA, operational, 2001
- Tungsten · USA, decommissioned, 2003
- Lincoln (NCSA Windows) · USA, operational, 2006
- Abe · USA, decommissioned, 2007
- Blue Waters · USA, decommissioned, 2013
- Delta · USA, operational, 2022
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.