Yahoo Labs

With the increasing prevalence of warehouse-scale (WSC) and cloud computing, understanding the interactions of server applications with the underlying microarchitecture becomes ever more important in order to extract maximum performance out of server hardware. To aid such understanding, this paper presents a detailed microarchitectural analysis of live data-center jobs, measured on more than 20,000 Google machines over a three year period, and comprising thousands of differ-ent applications. We first find that WSC workloads are extremely diverse, breeding the need for architectures that can tolerate appli-cation variability without performance loss. However, some patterns emerge, offering opportunities for co-optimization of hardware and software. For example, we identify com-mon building blocks in the lower levels of the software stack. This “datacenter tax ” can comprise nearly 30 % of cycles across jobs running in the fleet, which makes its constituents prime candidates for hardware specialization in future server systems-on-chips. We also uncover opportunities for classic microarchitectural optimizations for server processors, espe-cially in the cache hierarchy. Typical workloads place signifi-cant stress on instruction caches and prefer memory latency over bandwidth. They also stall cores often, but compute heav-ily in bursts. These observations motivate several interesting directions for future warehouse-scale computers. 1