Browse Prior Art Database

Fault Tolerant Branch on Value Instruction

IP.com Disclosure Number: IPCOM000088199D
Original Publication Date: 1977-May-01
Included in the Prior Art Database: 2005-Mar-04
Document File: 1 page(s) / 12K

Publishing Venue

IBM

Related People

Lang, DJ: AUTHOR [+4]

Abstract

This article relates to the adverse consequences of executing a specific type of "structured program statement" on a Von Neumann-type stack or other driven instruction sequencer. In this regard, structured programs are taken to comprise a series of instructions termed "modules" each module having precisely one entry and exit point. Among the basic types of statements which can be contained in a module are IF THEN ELSE and CASE BRANCH. The IF THEN ELSE statement connotes that if a predicate p is true, then perform the function H1 otherwise (else) perform function H2. The predicate node is limited to the testing of only a BOOLEAN value. In this respect, the CASE statement substitutes the testing of a number i lying within a range i < or = N instead of with N (conceptual) processing paths being provided.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 69% of the total text.

Page 1 of 1

Fault Tolerant Branch on Value Instruction

This article relates to the adverse consequences of executing a specific type of "structured program statement" on a Von Neumann-type stack or other driven instruction sequencer. In this regard, structured programs are taken to comprise a series of instructions termed "modules" each module having precisely one entry and exit point. Among the basic types of statements which can be contained in a module are IF THEN ELSE and CASE BRANCH. The IF THEN ELSE statement connotes that if a predicate p is true, then perform the function H1 otherwise (else) perform function H2. The predicate node is limited to the testing of only a BOOLEAN value. In this respect, the CASE statement substitutes the testing of a number i lying within a range i < or = N instead of with N (conceptual) processing paths being provided.

Hawk and Robinson [*] describe the execution of a CASE statement on a stack type micro-instruction sequencer. At page 71 they do not advert to the problem of specifying the contents of an instruction counter when the predicate value "i" being tested in a CASE statement lies outside the range, i.e., i > N, of interest. That is, they do not treat the problem of "wild branching". They do, however, point out that the functions (steps) to be performed in any selected path within the range must be equal so as to permit the statement following CASE to be executed. The prior art also teaches that "wild branching" may be avoided in ...