Campaign Diagrams: Visualizing the March Through the Phases of a Workload
Provides a new visualization tool for performance analysts to identify optimization opportunities in modern workloads, addressing the trade-off between aggregate and fine-grained analysis.
Campaign diagrams visualize phase-level resource utilization and bottlenecks in workloads, revealing optimization opportunities like reducing operational intensity improving performance in low-rank GEMM and fusion/pipelining opportunities in Mamba, which are hard to identify with rooflines or profilers alone.
We present campaign diagrams, a visualization technique for phase-level analysis of resource utilization and bottlenecks in modern workloads. Existing tools have a trade-off: rooflines aggregate a workload into a single point and lose all notion of time, while profilers and traces expose fine-grained events but obscure what bounds performance. Instead, a campaign diagram depicts compute throughput and memory bandwidth utilization, compute and memory traffic volume, and latency in a single figure. Since they can be generated from analytical models, simulations, or profiling data, campaign diagrams capture both ideal bounds and a kernel's achieved performance. We demonstrate them on two case studies: a low-rank GEMM, where they reveal the counterintuitive result that reducing operational intensity can improve end-to-end performance, and Mamba, where they expose fusion and pipelining opportunities across phases. In both cases, our visualization technique reveals optimization opportunities that are difficult to identify with rooflines or profilers alone.