Enhancing Linux Scheduler Scalability

This paper examines the scalability of the Linux 2.4.x scheduler as the load and number of CPUs increases. We show that the current scheduler de-sign involving a single runqueue and lock can suffer from lock contention problems which limits its scal-ability. We present alternate designs using multiple runqueues and priority levels that can reduce lock contention while maintaining the same functional behavior as the current scheduler. These implemen-tations demonstrate better overall scheduling per-formance over a wide spectrum of loads and system configurations.