Profiling a warehouse-scale computer

By Svilen Kanev, Juan Pablo Darago, Kim M. Hazelwood, Parthasarathy Ranganathan, Tipp Moseley, Gu-Yeon Wei, David M. Brooks. International Symposium on Computer Architecture, 2015

513 citations recorded by the source corpus.

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 datacenter jobs, measured on more than 20,000 Google machines over a three year period, and comprising thousands of different applications. We first find that WSC workloads are extremely diverse, breeding the need for architectures that can tolerate application variability without performance loss. However, some patterns emerge, offering opportunities for co-optimization of hardware and software. For example, we identify common 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, especially in the cache hierarchy. Typical workloads place significant stress on instruction caches and prefer memory latency over bandwidth. They also stall cores often, but compute heavily in bursts. These observations motivate several interesting directions for future warehouse-scale computers.

Subfields: Computer Networks and Communications, Hardware and Architecture, Information Systems

View Paper: https://www.semanticscholar.org/paper/110c050c6c992d2b956f7b47d717810ac5c91bdc

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Networks and Communications

In Hardware and Architecture

In Information Systems

Citations in the corpus

An Open-Source Benchmark Suite for Microservices and Their Hardware-Software Implications for Cloud & Edge Systems (2019, result): “This is consistent with studies on traditional cloud applications [37, 50], although to a lesser extent for microservices than for monolithic services (memcached, mongodb), given their smaller code footprint.”
Accelerometer: Understanding Acceleration Opportunities for Data Center Overheads at Hyperscale (2020, background): “While numerous prior works focus on optimizing allocations [58, 63, 70], very few recognize that optimizing free() can result in signifcant performance wins.”
DAWG: A Defense Against Cache Timing Attacks in Speculative Execution Processors (2018, background): “Commercial studies of code cache sensitivity of production server workloads [25], [27], [41] show large instruction miss rates in L2, but even the largest code working sets fit within 1–2 L3 ways.”
The Architectural Implications of Facebook's DNN-Based Personalized Recommendation (2019, background): “However, data center scale productionenvironments are concerned with not only the average case but also tail performance [22, 38].”
Processing Data Where It Makes Sense: Enabling In-Memory Computation (2019, background): “All these steps consume substantial time and energy in order to bring the data into the CPU chip [4, 7, 151, 152].”
SoftSKU: Optimizing Server Architectures for Microservice Diversity @Scale (2019, methodology): “Relative to alternative benchmarks, our microservices exhibit (1) a greater IPC diversity than Google’s services [1] and (2) a lower IPC than most widely-studied SPEC CPU2006 benchmarks.”