Dismiss
InnovationQ will be updated on Sunday, Oct. 22, from 10am ET - noon. You may experience brief service interruptions during that time.
Browse Prior Art Database

Multiprocessor System for Executing Concurrent Sequential Processes with Shared Variables

IP.com Disclosure Number: IPCOM000052775D
Original Publication Date: 1981-Jul-01
Included in the Prior Art Database: 2005-Feb-11
Document File: 2 page(s) / 38K

Publishing Venue

IBM

Related People

Lorie, RC: AUTHOR [+2]

Abstract

This invention relates to a method for controlling concurrent access to global variables stored in a shared memory by tasks executing simultaneously in different processors having access to the memory. Each processor executes the respective instruction streams and includes global and/or local variables therein. The method steps include (1) communicating sequences of global variables and their associated instructions to a common controller by each processor, (2) serializing and executing the instructions against a dedicated memory by the controller, and (3) placing the results in registers local to the counterpart processors.

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 58% of the total text.

Page 1 of 2

Multiprocessor System for Executing Concurrent Sequential Processes with Shared Variables

This invention relates to a method for controlling concurrent access to global variables stored in a shared memory by tasks executing simultaneously in different processors having access to the memory. Each processor executes the respective instruction streams and includes global and/or local variables therein. The method steps include (1) communicating sequences of global variables and their associated instructions to a common controller by each processor, (2) serializing and executing the instructions against a dedicated memory by the controller, and (3) placing the results in registers local to the counterpart processors.

Referring now to the figure, there is shown a multiprocessor configuration utilizing the invention.

The figure depicts one serializer processing unit (PU) and two other local PUs. Local PU-1 executes code from its local memory (MEM) until some hardware mechanism (i.e., a bit in an instruction or a special instruction) indicates that the serializer should execute the next instruction or sequence of instructions. At this time, local PU-1 sets bit i=l in a control register, indicating that the serializer should use register set i to continue the sequence of executions for this task. By way of hardware control means within, PU-1 scans its control register. If bit j=0, then PU-1 uses register set j to execute code for another ready task.

Via hardware, the ser...