Browse Prior Art Database

Overlapped System Clocking

IP.com Disclosure Number: IPCOM000085005D
Original Publication Date: 1976-Feb-01
Included in the Prior Art Database: 2005-Mar-02
Document File: 2 page(s) / 46K

Publishing Venue

IBM

Related People

Zurla, FA: AUTHOR

Abstract

Microcoded computers generally have two types of basic cycles: one in which all functions are completed in one cycle and one in which functions of one microword are overlapped with the second microword cycle. The advantages of the first type is that race conditions and complexity are reduced while the second type gives greater performance. The complexity of the second type is increased more when the system is one which runs variable-length clock cycles. The method shown here is designed to take advantage of the improved performance of the overlapped clocks, while retaining much of the simplicity of the nonoverlapped functions.

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 64% of the total text.

Page 1 of 2

Overlapped System Clocking

Microcoded computers generally have two types of basic cycles: one in which all functions are completed in one cycle and one in which functions of one microword are overlapped with the second microword cycle. The advantages of the first type is that race conditions and complexity are reduced while the second type gives greater performance. The complexity of the second type is increased more when the system is one which runs variable-length clock cycles. The method shown here is designed to take advantage of the improved performance of the overlapped clocks, while retaining much of the simplicity of the nonoverlapped functions.

For design simplicity, every cycle is designed to work in a non-overlapped mode. Also, every function is made to occur during one, and only one, time period of the clock. Fig. 1 shows a clock cycle with a basic 176 ns clock cycle (22 ns/T-pulse) where the source data is read out during T0T1 time, the destination and status registers are set at T4 time, and the result is destined at T6T7 time. By assigning designated time slots for functions to occur, the individual designers job is made easier.

To run an effective 88ns cycle, the second microword cycle is started at T4 time or the first microword cycle, thus overlapping the functions of the two microwords. That is, T0/T4, T1/T5, T2/T6, and T3/T7 pulses are on at the same time. (Fig. 2.) Any effective cycle length can be run by simply picking different points at wh...