Mira: A Program-Behavior-Guided Far Memory System

Far memory, where memory accesses are non-local, has become more popular in recent years as a solution to expand memory size and avoid memory stranding. Prior far memory systems have taken two approaches: transparently swap memory pages between local and far memory, and utilizing new programming models to explicitly move fine-grained data between local and far memory. The former requires no program changes but comes with performance penalty. The latter has potentially better performance but requires significant program changes. We propose a new far-memory approach by automatically inferring program behavior and efficiently utilizing it to improve application performance. With this idea, we build Mira. Mira utilizes program analysis results, profiled execution information, and system environments together to guide code compilation and system configurations for far memory. Our evaluation shows that Mira outperforms prior swap-based and programming-model-based systems by up to 18 times.