This article includes a list ofgeneral references, butit lacks sufficient correspondinginline citations. Please help toimprove this article byintroducing more precise citations.(April 2015) (Learn how and when to remove this message)

Connection Machine
A Connection Machine CM-2 (1987) and accompanyingDataVault on display at theMimms Museum of Technology and Art in Roswell, Georgia. The CM-2 used the same casing as the CM-1.
Design
Manufacturer	Thinking Machines Corporation
Release date	1986 (CM-1) 1987 (CM-2) 1991 (CM-5)
Units sold	At least 70[1][2]
Casing
Dimensions	≈6 × 6 × 6 ft (1.8 × 1.8 × 1.8 m) (CM-1 and CM-2)
Weight	1,268 lb (575 kg) (CM-2)[1]
System
CPU	Up to 65,536 1-bit processors (CM-1 and CM-2)
Memory	512 MB (CM-2) 10 GB (FROSTBURG CM-5)
Storage	Up to 80 GB with eightDataVaults (CM-1 and CM-2) ≈ 2TB (FROSTBURG CM-5)
FLOPS	2.5 gigaFLOPS (CM-2) 65.5 gigaFLOPS (FROSTBURG CM-5)
v t e

TheConnection Machine (CM) is a member of a series ofmassively parallelsupercomputers sold byThinking Machines Corporation. The idea for the Connection Machine grew out of doctoral research on alternatives to the traditionalvon Neumann architecture of computers byDanny Hillis atMassachusetts Institute of Technology (MIT) in the early 1980s. Starting with CM-1, the machines were intended originally for applications inartificial intelligence (AI) and symbolic processing, but later versions found greater success in the field ofcomputational science.

Danny Hillis andSheryl Handler foundedThinking Machines Corporation (TMC) inWaltham, Massachusetts, in 1983, moving in 1984 to Cambridge, MA. At TMC, Hillis assembled a team to develop what would become the CM-1 Connection Machine, a design for a massively parallelhypercube-based arrangement of thousands ofmicroprocessors, springing from his PhD thesis work at MIT in Electrical Engineering and Computer Science (1985).^[3] The dissertation won the ACM Distinguished Dissertation prize in 1985,^[4] and was presented as a monograph that overviewed the philosophy, architecture, and software for the first Connection Machine, including information on its data routing betweencentral processing unit (CPU) nodes, its memory handling, and the programming languageLisp applied in the parallel machine.^[3][5] Very early concepts contemplated just over a million processors, each connected in a 20-dimensional hypercube,^[6] which was later scaled down.

Each CM-1 microprocessor has its own 4 kilobits ofrandom-access memory (RAM), and thehypercube-based array of them was designed to perform the same operation on multiple data points simultaneously, i.e., to execute tasks in single instruction, multiple data (SIMD) fashion. The CM-1, depending on the configuration, has as many as 65,536 individual processors, each extremely simple, processingone bit at a time. CM-1 and its successorCM-2 take the form of acube 1.5 meters on a side, divided equally into eight smaller cubes. Each subcube contains 16printed circuit boards and a main processor called a sequencer. Each circuit board contains 32 chips. Each chip contains arouter, 16 processors, and 16 RAMs. The CM-1 as a whole has a 12-dimensionalhypercube-basedrouting network (connecting the 2¹² chips), a main RAM, and aninput-output processor (a channel controller). Each router contains five buffers to store the data being transmitted when a clear channel is not available. The engineers had originally calculated that seven buffers per chip would be needed, but this made the chip slightly too large to build.Nobel Prize-winning physicistRichard Feynman had previously calculated that five buffers would be enough, using a differential equation involving the average number of 1 bits in an address. They resubmitted the design of the chip with only five buffers, and when they put the machine together, it worked fine. Each chip is connected to a switching device called a nexus. The CM-1 usesFeynman's algorithm for computing logarithms that he had developed atLos Alamos National Laboratory for theManhattan Project. It is well suited to the CM-1, using as it did, only shifting and adding, with a small table shared by all the processors. Feynman also discovered that the CM-1 would compute the Feynman diagrams forquantum chromodynamics (QCD) calculations faster than an expensive special-purpose machine developed at Caltech.^[7][8]

To improve its commercial viability, TMC launched the CM-2 in 1987, addingWeitek 3132floating-point numericcoprocessors and more RAM to the system. Thirty-two of the original one-bit processors shared each numeric processor. The CM-2 can be configured with up to 512 MB of RAM, and a redundant array of independent disks (RAID)hard disk system, called aDataVault, of up to 25 GB. Two later variants of the CM-2 were also produced, the smallerCM-2a with either 4096 or 8192 single-bit processors, and the fasterCM-200.

Due to its origins in AI research, the software for the CM-1/2/200 single-bit processor was influenced by theLisp programming language and a version ofCommon Lisp,*Lisp (spoken:Star-Lisp), was implemented on the CM-1. Other early languages includedKarl Sims' IK and Cliff Lasser's URDU. Much system utility software for the CM-1/2 was written in *Lisp. Many applications for the CM-2, however, were written inC*, a data-parallel superset ofANSI C.

With theCM-5, announced in 1991, TMC switched from the CM-2's hypercubic architecture of simple processors to a new and different multiple instruction, multiple data (MIMD) architecture based on afat tree network ofreduced instruction set computing (RISC)SPARC processors. To make programming easier, it was made to simulate aSIMD design. The laterCM-5E replaces the SPARC processors with faster SuperSPARCs. A CM-5 was the fastest computer in the world in 1993 according to theTOP500 list, running 1024 cores with Rpeak of 131.0 GFLOPS, and for several years many of the top 10 fastest computers were CM-5s.^[9]

Connection Machines were noted for their striking visual design. The CM-1 and CM-2 design teams were led byTamiko Thiel.^[10] The physical form of the CM-1, CM-2, and CM-200 chassis was a cube-of-cubes, referencing the machine's internal 12-dimensionalhypercube network, with the redlight-emitting diodes (LEDs), by default indicating the processor status, visible through the doors of each cube.

By default, when a processor is executing an instruction, its LED is on. In a SIMD program, the goal is to have as many processors as possible working the program at the same time – indicated by having all LEDs being steady on. Those unfamiliar with the use of the LEDs wanted to see the LEDs blink – or even spell out messages to visitors. The result is that finished programs often have superfluous operations to blink the LEDs.

The CM-5, in plan view, had a staircase-like shape, and also had large panels of red blinking LEDs. Prominent sculptor-architectMaya Lin contributed to the CM-5 design.^[11]

• The very first CM-1 is on permanent display in theComputer History Museum, Mountain View, California, which also has two other CM-1s and CM-5.^[12]
• There is a decommissioned CM-1 or CM-2 on display in the main building of theKarlsruhe Institute of Technology computer science department. Students have converted it into aBluetooth-controlled LED matrix display which can be used to play games or display art.^[13]
• A CM-2 with flashing red LED arrays and its accompanyingDataVault storage unit are on permanent display at theMimms Museum of Technology and Art in Roswell, Georgia.^[14]
• There is a CM-200 on permanent display atÉcole Polytechnique Fédérale de Lausanne in Switzerland, part of the exhibition space ofMusée Bolo.^[15]

• TheMuseum of Modern Art in New York City displayed a CM-2 in 2018. They continue to house the machine in their collection.^[16]
• A CM-2 is in the collection of theSwedish National Museum of Science and Technology (Tekniska Museet) in Stockholm, Sweden.^[1]
• Several parts of a CM-1 are in the collection of the Smithsonian InstitutionNational Museum of American History, though it may not be a complete example.^[17][18]
• TheLiving Computers: Museum + Labs in Seattle displayed a CM-2 with flashing LEDs prior to its closing in 2020.^{[19][better source needed]} It is possible this machine is now in private hands, though it is not listed among the objects auctioned by Christie's.^[20]
• As of June 2025, a CM-2 is held in storage by theInterim Computer Museum.^[21] It is stated on the museum's website that they would like to display the computer publicly in the future.^[22]

• As of 2007, a preserved CM-2a was owned by the Corestore, a type of online-only museum.^[23]

A CM-5 was featured in the filmJurassic Park in thecontrol room for theisland (instead of aCray X-MPsupercomputer as in the novel). Two banks, one bank of 4 Units and a single off to the right of the set could be seen in the control room.^[24]

The computer mainframes inFallout 3 were inspired heavily by the CM-5.^[25]

Cyberpunk 2077 features numerous CM-1/CM-2 style units in various portions of the game.

The b-side toClock DVA's 1989 single "The Hacker" is titled "The Connection Machine" in reference to the CM-1.

• Blinkenlights
• Brewster Kahle – lead engineer on the Connection Machine projects
• Danny Hillis – inventor of the Connection Machine
• David E. Shaw – creator of NON-VON machine, which preceded the Connection machine slightly
• FROSTBURG – a CM-5 used by theNSA
• Goodyear MPP
• ICL DAP
• MasPar
• Parallel computing

• Gallery of CM-5 images
• CM-5 ManualsArchived 2006-09-02 at theWayback Machine
• Tamiko Thiel on the visual design of the CM-1/2/200
• Feynman and the Connection Machine
• Liquid Selves, an animated short film rendered on a CM-2

• Supercomputers
• Parallel computing
• Massively parallel computers
• Thinking Machines supercomputers
• Computer-related introductions in 1984

• Articles with short description
• Short description matches Wikidata
• Articles lacking in-text citations from April 2015
• All articles lacking in-text citations
• All articles lacking reliable references
• Articles lacking reliable references from October 2024
• Webarchive template wayback links
• Use dmy dates from March 2017