Abstract. Software-Based Self-Test (SBST) approaches have shown to be an effective solution to detect permanent faults, both at the end of the production process, and during the operational phase. However, when Very Long Instruction Word (VLIW) processors are addressed these techniques require some optimization steps in order to properly exploit the parallelism intrinsic in these architectures. In this chapter we present a new method that, starting from previously known algorithms, automatically generates an effective test program able to still reach high fault coverage on the VLIW processor under test, while minimizing the test duration and the test code size. Moreover, using this method, a set of small SBST programs can be generated aimed at the diagnosis of the VLIW processor. Experimental results gathered on a case study show the effectiveness of the proposed approach.Keywords: SBST, VLIW processor, Fault Simulation, Fault Diagnosis.
IntroductionThe continuous scaling in the semiconductor fabrication process combined with the progressive growth of the integrated circuits operation frequency pushes processor cores to face more difficult testability problems. Furthermore, several phenomena such as metal migration or aging become more likely, thus increasing the occurrence of permanent faults in the generic system, in particular during the circuit operational phase. For these reasons, in order to provide high fault coverage with acceptable costs, new test solutions are being investigated and evaluated (e.g., in terms of silicon area overhead, required test infrastructure and test time). Software-Base Self-Test (SBST) has been demonstrated to be a promising and effective approach for the test of processors and processor-based systems [1]. The SBST main idea is to generate test programs to be executed by the processor under test, able to fully stimulate the processor itself or other components belonging to the system, and to detect possible faults by looking at the produced results. The SBST technique does not require any additional hardware; therefore, the whole test cost is reduced and no performance penalty is introduced. Moreover, the SBST technique allows at-speed testing and can be easily used even for on-line test purposes. Hence,