WPC: Whole-picture Workload Characterization

TL;DR

This paper introduces WPC, a comprehensive methodology and tool for analyzing how various software and hardware components across the stack contribute to system bottlenecks, aiding co-design and optimization.

Contribution

WPC provides an iterative framework combining observation, normalization, and exploration to systematically characterize workload bottlenecks across multiple system layers.

Findings

WPC can quantify the impact of each component on pipeline efficiency.

The methodology is validated through evaluations demonstrating its effectiveness.

Open-source tool available for community use.

Abstract

This article raises an important and challenging workload characterization issue: can we uncover each critical component across the stacks contributing what percentages to any specific bottleneck? The typical critical components include languages, programming frameworks, runtime environments, instruction set architectures (ISA), operating systems (OS), and microarchitecture. Tackling this issue could help propose a systematic methodology to guide the software and hardware co-design and critical component optimizations. We propose a whole-picture workload characterization (WPC) methodology to answer the above issue. In essence, WPC is an iterative ORFE loop consisting of four steps: Observation, Reference, Fusion, and Exploration. WPC observes different level data (observation), fuses and normalizes the performance data (fusion) with respect to the well-designed standard reference workloads suite (reference), and explores the software and hardware co-design space (exploration) to investigate the impacts of critical components across the stacks. We build and open-source the WPC tool. Our evaluations confirm WPC can quantitatively reveal the contributions of the language, framework, runtime environment, ISA, OS, and microarchitecture to the primary pipeline efficiency.