Asymmetric multiprocessing

From Citizendium
Revision as of 12:29, 28 July 2010 by imported>Howard C. Berkowitz
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search
This article is developing and not approved.
Main Article
Discussion
Related Articles  [?]
Bibliography  [?]
External Links  [?]
Citable Version  [?]
 
This editable Main Article is under development and subject to a disclaimer.

Asymmetric multiprocessing tightly couples heterogeneous computer processors, typically with specialized functions, and not necessarily having a full operating system on each processor.

History

It originated in in the days of wooden computers and iron men, IBM was struggling to get its OS/360 computers working. Engineers working on the then-current 7090 and 7094 mainframes, optimized for scientific computing attached a separate 7040/44 processor, optimized for commercial processing to the 709x, and used the 704x to manage the printers and card readers. This was IBM's Attached Support Processor (ASP). While ASP worked quite well, the management of NASA's Manned Spaceflight Center, in Houston, Texas, took a dim view of needing twice the number of large computers in their machine world.

An IBM team of six support engineers in Houston, not official software developers, wrote a software equivalent of the separate spooling software that ran on the 709x: Houston Automatic Spooling Priority, or HASP. HASP very quickly became unsupported but popular, and then, in 1973, IBM fully supported it, and operating system spooling became mainstream. HASP kept being upgraded; as much as IBM management tried to kill it, until Job Entry System 2 (JES2) replaced it on System/370 mainframes in 1978.

Modern use

Had small support computers been available in the sixties, ASP might have become routine. Today, we see this architecture on single multicore chips, which, for example, combine general-purpose processors with digital signal processors.