Browse Prior Art Database

Block Oriented Forward Trace Through Blocks for Which the Turn-On and Turn-Off Delays Are Not Equal

IP.com Disclosure Number: IPCOM000050521D
Original Publication Date: 1982-Nov-01
Included in the Prior Art Database: 2005-Feb-10
Document File: 3 page(s) / 57K

Publishing Venue

IBM

Related People

Hitchcock, RB: AUTHOR [+2]

Abstract

This article is related to information contained in other articles on pages 2826-2830 and 2835 of this issue.

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 53% of the total text.

Page 1 of 3

Block Oriented Forward Trace Through Blocks for Which the Turn-On and Turn-Off Delays Are Not Equal

This article is related to information contained in other articles on pages 2826-2830 and 2835 of this issue.

Two techniques, the "propagation" algorithm and the "traceback" algorithm, are described which are used to calculate arrival times and identify all bad paths in a logic structure with delays. The Propagation Technique:

The propagation technique is illustrated by an example in Fig. 1. The conventions used in the example are shown at the top of the diagram. When associated with a net, the figures in parenthesis give the times when a rising/falling signal will arrive on the net. When associated with a block, the figures give the delay through the block when the block output is rising/falling. Note that some information about the block function is necessary in order to know whether a rising output will be caused by an input rising or falling or both. A wedge on the output of a block in the figure indicates that an inversion has taken place.

To start with, we are given the block delays which, in general, are a function of the number of blocks driven, the amount of wiring driven, whether the block is turning on or off, and the power or drive capability of the block. We are also given the times at which a rising (falling) signal will start. (This is for Primary Input nets and for Storage Element Output nets - the latter has actually been determined by using the clock information.) Net AA is generated by block AA. The last time a rising signal will start on net AA is the last time a rising signal is input to block AA (t = 1) plus the delay of block AA if the output is rising (delay =
2). This is the same for the falling signal.

The signal on net BA, on the other hand, is generated by block BA which is an inverting block. Thus, the last time a rising signal could start on net BA is the last time a falling signal was input to block BA (t = 4) plus the block delay when the output is rising (delay = 3). Similarly we get the time for a falling signal.

Summarizing the calculations for each block: Block AA rising: 3 = (delay = 2) + max of (0,1) the input rise times falling: 4 = (delay = 3) + max of (1,0) the input fall times Block AB rising: 2 = (delay = 2) + max of (0,0) the input fall times falling: 5 = (delay = 4) + max of (1,0) the input rise times Block BA rising: 7 = (delay = 3) + max of (3,4) the input fall times falling: 4 = (delay = 1) + max of (2,3) the input rise times Block CB rising: 8 = (delay = 3) + max of (4,5) the input fall times falling: 12 = (delay = 5) + max of (7,2) the input rise times

1

Page 2 of 3

The propagation technique just described illustrates the method of determining the last time a given pulse transition could occur at the output...