Pensieve: Microarchitectural Modeling for Security Evaluation

This is the artifact of our paper "Pensieve: Microarchitectural Modeling for Security Evaluation".

Our artifact is the implementation of the Pensieve framework (Figure 5) for the security evaluation of microarchitecture defenses against speculative execution attacks. The artifact includes μarch models for a variety of defense mechanisms encoded in a solver-based programming language (i.e., Rosette), and scripts for performing bounded model checking on these uarch models for counterexample detection. This artifact release can be used to reproduce the two attack examples as shown in Figure 7 and Figure 9(a), as well as the performance evaluation results for the checking time as shown in the plots in Figure 10.