Persistent Memory functions as a hybrid of traditional storage systems with main memory. Yet, persistent memory systems impose additional memory accesses to maintain data persistence in memory, by employing logging or copy-on-write mechanisms inherited from databases and file systems. These techniques are manually handled in software by the programmer or libraries, and supported by new instructions in state-of-art architectures. These instructions increase hardware complexity and rely on the programmer to explicitly call for persistence without full knowledge of the system, leading to inefficiencies in implementation and a decrease in performance. The goal of my work is to achieve high-performance and energy-efficient hardware-driven logging in persistent memory systems. I propose an efficient, persistence-aware logging mechanism, which transfers the responsibility of logging and persisting data members to the fabric of the microarchitecture itself. Such mechanisms prove to be more cognizant by leveraging direct knowledge of the system to handle the overheads of memory persistence.