Improving the Configuration of the Predictable ACDC Data Cache for Real-Time Systems

In real-time systems, analyzing the worst-case execution time (WCET) of a task in presence data caches is hard. The ACDC cache that provides predictability, facilitating WCET analysis. It works by granting replacement permission to specific load/store instructions. Nonetheless, knowing how select these instructions minimize WCET, i.e., configuring ACDC, not trivial. this paper, we propose four new methods configure and compare them with existing methods. Unlike those previous studies, our proposed provide configurations for different phases given task, instead single configuration per task. We evaluate bounds obtained when using on TACLeBench benchmark suite. Our results show most complex benchmarks work better multiple-content configurations, which indicates realistic tasks may also benefit from kind configuration. study improve more than 60% cases, an average improvement nearly 5% up 50% some cases.