This paper presents first steps towards automatic generation of distributed tests. We first define a characterization of the tests for which the property of unbias is preserved by the existence of an asynchronous environment. Then, starting from a centralized test case, we propose a method to derive automatically its corresponding distributed test case in an asynchronous environment. We prove that the generated distributed test case is not biased, it tests the same behaviors of an implementation and has the same testing power as the centralized test case.
In this paper, we present an end-to-end industrial case-study concerning the automatic generation of tests suites for the Cache Coherency Protocol of a Multiprocessor Architecture. It consists of the following stages: (1) formal specification of the architecture using Lotos language, (2) formal description of the test purposes, (3) automatic generation of abstract test suites using the prototype TGV, and (4) automatic generation and analysis of executable test suites. Through the description of each of the previous stages, this paper demonstrates that tools designed for protocol conformance testing can be efficiently used to generate executable tests for hardware concurrent systems.
We present an experiment which has demonstrated that methods and tools developed in the context of black box conformance testing of communication protocols can be efficiently used for testing the cache coherency protocol of a hardware multi-processor architecture. We have used the automatic conformance tests generator tgv developed by IN-RIA to generate abstract tests and we have developed a software in order to make them executable in the real test environment of Bull. The tgv approach has been considered by the hardware testing community as a serious alternative to usual random test generation. It overwhelms the well known debugging and coverage problems linked to this kind of technic.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.