DocumentCode :
3547739
Title :
Instruction-based delay fault self-testing of pipelined processor cores
Author :
Singh, Virendra ; Inoue, Michiko ; Saluja, Kewal K. ; Fujiwara, Hideo
Author_Institution :
Nara Inst. of Sci. & Technol., Japan
fYear :
2005
fDate :
23-26 May 2005
Firstpage :
5686
Abstract :
Although nearly all modern processors use a pipelined architecture, no method has yet been proposed in the literature to model these for the purpose of test generation. The paper proposes a graph theoretic model of pipelined processors and develops a systematic approach for delay fault testing of such processor cores using the processor instruction set. Our methodology consists of using a graph model of the pipelined processor, extraction of architectural constraints, classification of paths, and generation of tests using a constrained ATPG. These tests are then converted to a test program, a sequence of instructions, for testing the processor. Thus, the tests generated by our method can be applied in a functional mode of operation and can also be used for self-test. We applied our method to two example processors, namely a 16 bit five stage VPRO pipelined processor and a 32 bit pipelined DLX processor, to demonstrate the effectiveness of our methodology.
Keywords :
automatic test pattern generation; automatic test software; delays; graph theory; instruction sets; microprocessor chips; pipeline processing; 16 bit; 32 bit; architectural constraints; constrained ATPG; graph theoretic model; instruction-based delay fault self-testing; pipelined architecture; pipelined processor cores; processor instruction set; test generation; test program; Automatic test pattern generation; Automatic testing; Built-in self-test; Costs; Delay systems; Electronic equipment testing; Energy consumption; Pipelines; Software testing; System testing;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Circuits and Systems, 2005. ISCAS 2005. IEEE International Symposium on
Print_ISBN :
0-7803-8834-8
Type :
conf
DOI :
10.1109/ISCAS.2005.1465928
Filename :
1465928
Link To Document :
بازگشت