A C++ monitoring framework for past-time LTL and MTL in discrete and dense time, combining a shared computation graph with a data-oriented execution engine.
1.81–4.53× median single-property speedups; 6.4–11.7× end-to-end for 30 properties, compared with Reelay.
First-author paper with Doğan Ulus. LoomRV combines shared past-time LTL/MTL monitoring with a linear execution schedule and preallocated, double-buffered temporal state, in discrete and dense time. Compared with Reelay, the paper reports 1.81–4.53× median single-property speedups and 6.4–11.7× end-to-end speedups for 30 properties.