Browse Prior Art Database

C.mmp: THE CMU MULTIMINIPROCESSOR COMPUTER Requirements and Overview of the Initial Design

IP.com Disclosure Number: IPCOM000148025D
Original Publication Date: 1971-Aug-24
Included in the Prior Art Database: 2007-Mar-28
Document File: 63 page(s) / 2M

Publishing Venue

Software Patent Institute

Related People

Bell, C.G.: AUTHOR [+8]

Abstract

C.mp : THE CMU MULTIMINIPROCESSOR COMPUTER Requirement8 and Overview of the Initial Design C. G. Bell, W. Broadley, W, Wulf, A. Newell and C. Pierson R. Reddy S. Rege Department of Computer Science Carnegie-Mellon University Pittsburgh, Pennsylvania August 24, 1971 This work was supported by the Advanced Research Projects Agency of the Office of the Secretary of Defense (F44620-70-C-0107) and is monitored by the Air Force Office of Scientific Research. ABSTRACT This document deacribee a proposed CMU multiprocessor system to be constructed around a set of PDP-11 computers connected through a cross- point switch to a large sharable primary memory. The present design con- stitutes a solution to a specific set of needs existing in our environment. The syetem has reeearch consequences that reach well beyond the particular demands it was designed to satisfy. For although multiprocessors havebeen much talked about and advocated, there are remarkably few operational systems more complex than dual-processor systems, and even fewer documented ecientific investigations into their performance and operating structure.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 5% of the total text.

Page 1 of 63

  C.mp : THE CMU MULTIMINIPROCESSOR COMPUTER
Requirement8 and Overview of the Initial Design
C. G. Bell, W. Broadley, W, Wulf, A. Newell
and
C. Pierson
R. Reddy
S. Rege

Department of Computer Science
Carnegie-Mellon University

Pittsburgh, Pennsylvania
August 24, 1971

This work was supported by the Advanced Research Projects Agency
of the Office of the Secretary of Defense (F44620-70-C-0107) and
is monitored by the Air Force Office of Scientific Research.

[This page contains 1 picture or other non-text object]

Page 2 of 63

ABSTRACT

   This document deacribee a proposed CMU multiprocessor system to be
constructed around a set of PDP-11 computers connected through a cross-
point switch to a large sharable primary memory. The present design con-
stitutes a solution to a specific set of needs existing in our environment.
The syetem has reeearch consequences that reach well beyond the particular
demands it was designed to satisfy. For although multiprocessors have
been much talked about and advocated, there are remarkably few operational
systems more complex than dual-processor systems, and even fewer documented
ecientific investigations into their performance and operating structure.

    This document is limited to a presentation and analysis of the (hard-
ware) system. It gives enough description of the usage requirements, soft-
ware, and the research potentials and problems to make clear why we believe
the effort to'be a sound one, It does not attempt a systematic discussion
of the field of multiproceaaor research, nor of alternative systems that
might be of interest, either to meet our computing demands or as research
directions.

    Section I1 discusses the requirements and research potential. Section
I11 lists the design constraints adopted. Section IV laye out the PMS structure of the aystem. Section V describes the main specifications of
the operating system. Section VI provides some details on a performance
analysis.

[This page contains 1 picture or other non-text object]

Page 3 of 63

                  I. INTRODUCTION
This document describes a proposed CMU multiprocessor system to be con-
structed around a set of PDP-11 computers connected through a croaspoint
switch to a large sharable primary memory. The system is to operate as a
third node in the existing Computer Science computer system, along with two
existing PDP-lo's.

    Tha present design conetitutes a solution to a apecific set of needs
existing in our environment. It replaces a prior planned solution, which
consisted of a collection of individual stand-alone PDP-11 systems and PDP-11
I/O processor Bysterna situated in front of the PDP-10's. It is not the only
system that eatiefies our requirements, as the existence of the prior plan
indicates. However, it provides a highly effective solution and does so with-
in the cost framework with which we have been working.

    The system has research consequences that reach we...