Connection Machine

From Wikipedia, the free encyclopedia

Thinking Machines CM-2 at the Computing Museum in San Jose. One of the face plates has been partially removed to show the circuit boards inside.

The Connection Machine was a series of supercomputers that grew out of Danny Hillis's research in the early 1980s at MIT on alternatives to the traditional von Neumann architecture of computation. The CM-1, originally conceived of at MIT, was a "massively parallel" hypercubic arrangement of thousands of very simple microprocessors or CPUs, each with its own RAM, which together executed in a SIMD fashion. The Connection Machine was originally intended for applications in artificial intelligence and symbolic processing, but later found greater success in the field of computational science. The Connection Machine looks a bit like the C Cube in the novel Artemis Fowl: The Eternity Code.

Danny Hillis and Sheryl Handler founded Thinking Machines in Waltham, Massachusetts (it was later moved to Cambridge, Massachusetts) in 1983 and assembled a team to develop the CM-1 and then the CM-2, which, depending on the configuration, had as many as 65,536 processors. The individual processors were extremely simple, processing one bit at a time. The CM-2, launched in 1987, added Weitek 3132 floating-point numeric co-processors to the system, with 32 of the original simple processors sharing each numeric processor. Two later variants of the CM-2 were also produced, the smaller CM-2a with either 4096 or 8192 single-bit processors, and the faster CM-200.

The light panels of FROSTBURG, a CM-5, on display at the National Cryptologic Museum. The panels were used to check the usage of the processing nodes, and to run diagnostics.

Due to its origins in AI research, the software for the CM-1/CM-2 single-bit processor was influenced by the Lisp programming language and a version of Common Lisp, *Lisp (spoken: "Star-Lisp"), was implemented on the CM-1. Other early languages included Kark Sims' IK and Cliff Lasser's URDU. Much system utility software for the CM-1/2 was written in *Lisp.

With the CM-5, announced in 1991, Thinking Machines switched from the CM-2's hypercubic architecture of simple processors to an entirely new MIMD architecture based on a fat tree network of SPARC RISC processors. The later CM-5E replaced the SPARC processors with faster SuperSPARCs.

The Connection Machine has a length, width, and height of 1.5 metres. It is divided into 8 equally large cubic sections. Each cubic section contains 16 printed circuit boards and a main processor called a sequencer. Each printed circuit board contains 32 chips. Each chip contains a communication channel called a router, 16 processors, 16 RAMs, and many electronic components such as microchips, transistors, diodes, resistors, and capacitors. The Connection Machine as a whole has a routing network, a main RAM, and an input/output processor. It is connected to a switching device called a nexus. The Connection Machine also has an internal memory of 512 megabytes and a permanent external memory of 10 gigabytes.

Connection Machines were noted for their (intentional) striking visual design. The CM-2 was cube-shaped with red blinking LEDs visible over a large portion of the surface. The CM-5, when viewed from above, looked like a lightning bolt, and also had a large panel of red blinking LEDs. Perhaps because of its design, a CM-5 was featured in the movie Jurassic Park in the control room for the island.

Danny Hillis's original thesis paper, on which the Connection Machine was based, is The Connection Machine (MIT Press Series in Artificial Intelligence) (ISBN 0-262-08157-1) . The title is out of print as of 2005. The book provides an overview of the philosophy, architecture and software for the Connection Machine, including data routing between CPU nodes, memory handling, Lisp programming for parallel machines, etc.

[edit] See also

Transputer
FROSTBURG — a CM-5 used by the NSA
David Shaw's NON-VON machine, which preceded the Connection machine slightly.

[edit] References

Hillis, D. 1982 "New Computer Architectures and Their Relationship to Physics or Why CS is No Good", Int J. Theoretical Physics 21 (3/4) 255-262.
Lewis W. Tucker, George G. Robertson, "Architecture and Applications of the Connection Machine," Computer ,vol. 21, no. 8, pp. 26-38, August, 1988.
Arthur Trew and Greg Wilson (eds.) (1991). Past, Present, Parallel: A Survey of Available Parallel Computing Systems. New York: Springer-Verlag. ISBN 0-387-19664-1.
W. Daniel Hillis and Lewis W. Tucker. The CM-5 Connection Machine: A Scalable Supercomputer. In Communications of the ACM, Vol. 36, No. 11 (November 1993).