Browse Prior Art Database

Method to Generate Care Tests for Failures

IP.com Disclosure Number: IPCOM000083811D
Original Publication Date: 1975-Jul-01
Included in the Prior Art Database: 2005-Mar-01
Document File: 2 page(s) / 13K

Publishing Venue

IBM

Related People

Roth, JP: AUTHOR

Abstract

In the testing for failures of a computer in particular, of the hardware, there may be computed tests for which the purpose of the device may be indifferent. It may be a "DONTCARE". On the other hand, the test may be an "illegal" input, which could not be brought up under operating -- but only test -- conditions. In this case, the illegal condition corresponds or plays exactly the same role with respect to the test computation problem, as the DONTCAREs. In each case, it is desirable and, in a high reliable operating environment, even mandatory to assure that the tests computed to detect any of a given category of failures, be CARE or legal inputs.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 53% of the total text.

Page 1 of 2

Method to Generate Care Tests for Failures

In the testing for failures of a computer in particular, of the hardware, there may be computed tests for which the purpose of the device may be indifferent. It may be a "DONTCARE". On the other hand, the test may be an "illegal" input, which could not be brought up under operating -- but only test -- conditions. In this case, the illegal condition corresponds or plays exactly the same role with respect to the test computation problem, as the DONTCAREs. In each case, it is desirable and, in a high reliable operating environment, even mandatory to assure that the tests computed to detect any of a given category of failures, be CARE or legal inputs.

The D-algorithm is a procedure to compute tests for failures, used by AT&T, IBM, TI, ICL, Olivetti, Hughes Aircraft, GE, RCA, etc., (IBM Journal, July 1966). It assumes, however, that any input logically valid is an admissible test. All other algorithms published assume the same.

A review of the assumptions follow. First, assume a definition of the function to be implemented in either R-notation or as an R-design, wherein for each, precisely the CARE conditions are specified. Second, assume that the hardware implementation, the regular logic design (RLD), is implemented from the R- design in modular fashion. This means that it is assumed that the optimization conforms to the module structure, i.e., the R-algorithm, the R-design are structurally isomorphic to the hardware implementation.

Assume then that there is a failure which is contained within a given module. There is by this method of construction, a one-to-one correspondence between the hardware module and the R-design module, for each such module and their interconnections of the entire design.

Call the hardware design the RLD and that of the one specifying especially the CARE conditions, the R-design. It is impor...