A complete strategy for testing an on-chip multiprocessor architecture

The article proposes an approach that divides testing into three phases: router testing, RAM block testing, and distributed processor testing. This test strategy was implemented for the on-chip multiprocessor architecture of a fine-grain, massively parallel machine developed in 1995 at the National Polytechnic Institute of Grenoble. The hierarchical strategy minimizes the entire architecture's test cost by avoiding unnecessary testing. For example, testing a processor that is inaccessible because its router is faulty or that has a faulty local RAM is useless. Furthermore, a fault-free RAM cannot be used if the corresponding node router is faulty.

[1]  Kewal K. Saluja,et al.  A built-in self-test algorithm for row/column pattern sensitive faults in RAMs , 1990 .

[2]  GERNOT METZE,et al.  On the Connection Assignment Problem of Diagnosable Systems , 1967, IEEE Trans. Electron. Comput..

[3]  Peter M. Kogge,et al.  EXECUBE-A New Architecture for Scaleable MPPs , 1994, 1994 International Conference on Parallel Processing Vol. 1.

[4]  Chantal Robach,et al.  Memory testing in a massively parallel machine , 1993, Microprocess. Microprogramming.

[5]  Chantal Robach,et al.  An Implementation Approach of the IEEE 1149.1 for the Routing Test of a VLSI Massively Parallel Architecture , 1998, J. Electron. Test..

[6]  Colin M. Maunder,et al.  An introduction to the boundary scan standard: ANSI/IEEE Std 1149.1 , 1991, J. Electron. Test..

[7]  Chantal Robach,et al.  Functional testing and reconfiguration of MIMD machines , 1993, Proceedings of 1993 IEEE International Workshop on Defect and Fault Tolerance in VLSI Systems.