Automating the generation of programs maximizing the sustained switching activity in microprocessor units via evolutionary techniques

During device testing, an important parameter to be considered by the test engineers is the switching activity (SWA) of the circuit under test (CUT). It is well known that the SWA must be kept to a minimum in order to avoid catastrophic scenarios on the CUTs, e.g., unacceptable peak power consumption or over-stressing that can lead to an artificial yield loss. However, there are scenarios, where the inverse, namely the switching activity maximization, can be proven beneficial. For example, this happens during Burn-In testing, when we aim at exercising the circuit under extreme operating conditions in terms of temperature and temperature gradients to accelerate aging phenomena in a controlled manner. In some cases, this is achieved by applying to the CUT stimuli in a functional manner, based on the Software-Based Self-Test (SBST) paradigm. Though, the generation of such appropriate programs for this task represents a challenge for the test engineers. In this paper we consider the scenario where the modules to be stressed are the units of a pipelined processor. We present a method, based on an evolutionary approach, that is able to automatically generate maximum stress programs, i.e., sequences of instructions achieving a high switching activity in the target module. With respect to other approaches, the generated sequences are built from the ground-up, are short and repeatable, thus guaranteeing their easy usability to stress a module (and increase its temperature). The processor we used for our experiments is the Open RISC 1200. Results demonstrate that the generated stimuli from the proposed method are effective in achieving a high value of sustained switching activity within the considered processor modules while also achieves a better results when comparing with other methods (e.g., stuck-at oriented test programs).