Opportunities for Cache Friendly Process Scheduling

Operating system process scheduling has been an active area of research for many years. Process scheduling decisions can have a dramatic impact on capacity and conflict misses in on-chip caches, since processes that do not share memory will compete for entries in the cache, while processes that do share memory can effectively pre-fetch blocks or warm the cache in a symbiotic fashion. In this paper we provide a detailed characterization of context switch misses and quantify its impact. We also investigate the potential of intelligent process scheduling that minimizes cache misses across context-switch boundaries. We have developed several greedy heuristics that enable us to perform a trace driven limit study on the benefits of cache friendly process scheduling. We show that up to 37% improvements in cache miss rates are achievable in some transactional workloads. We also propose some mechanisms to leverage this potential.