jhatfiel@nmsu.edu (Joseph T. Hatfield) (09/09/90)
I'm trying to determine if the the architecture of the cray-1 provided any support for operating system. Some source claim that the cray-1 did not HAVE an operating system and other source mention something call the DEMOS file system and COS. Our library has a lot of information on the Cray-1 architecture and nothing on DEMOS or COS. Can anyone shed any light on this topic? Thanks -- ============================================================= Joseph T. Hatfield New Mexico State University ------------------------------------------------------------- BITNET : G4210791 at NMSUVM1 USMail : Box 4217, Las Cruces, NM 88003 ------------------------------------------------------------- Artificial Life/Virtual Worlds/Nanotech/APL/Recursion/Dead or Alive/Dead Kennedy's/Ministry/MDC/Tina Turner/Goldberg/Haydn/ Bach/Brahms/Scriabin/Glass/Delaney/Gibson/STH:The N.Y. Review of... ------------------------------------------------------------- Send me risque stories. =============================================================
eugene@eos.UUCP (Eugene Miya) (09/10/90)
In article <JHATFIEL.90Sep8142612@silicon.nmsu.edu> jhatfiel@nmsu.edu (Joseph T. Hatfield) writes: >I'm trying to determine if the the architecture of the cray-1 provided >any support for operating system. Some source claim that the cray-1 >did not HAVE an operating system and other source mention something >call the DEMOS file system and COS. Our library has a lot of >information on the Cray-1 architecture and nothing on DEMOS or COS. Support or OS? There were something like 6 OSes for Crays. Such an old machine.... There was a joke that if you brought a machine from Seymour Cray, you basically had to write your own operating system..... DEMOS was designed by F. Baskett and others, several papers published, the best known of which appeared in a SOSP (Symp. on Oper. Syst. Prin.) proceedings. It was a very controversial system. %A F. Baskett %A J. H. Howard %A J. Monague %T Task Communication in DEMOS %J Proceedings of the 6th Symposium on Operating Systems Principles %I ACM %D November 1977 %P 2-7 %X This paper is reproduced in Kuhn and Padua's (1981, IEEE) survey "Tutorial on Parallel Processing." COS: The pedigree of these machines comes thru. Bill Joy called this once "C/PM Cray." An OS derived from earlier CDC machines which trace their way back to EXEC*8 when Cray worked for a Univac company. The interactive version of this OS was preferable except to managers. Basically, you have to ACCESS (or ASSIGN [@ASG]) files explicitly. Flat file system. Hard to use tasking structure. Not really designed to be interactive. If you want to get a feel of what this OS is like either find a COS machine or a Unisys/Univac 1100 or 2200. Not as bad as explicit DD statements on an IBM, but an increasingly hard machine to find. The other operating systems include CTSS [derived from LTSS on the 7600], NLTSS, and a couple of others which existed in secure environments or have vanished into well-deserved obsurity (like NCAROS). --e.n. miya, NASA Ames Research Center, eugene@eos.arc.nasa.gov {uunet,mailrus,most gateways}!ames!eugene