pith. sign in

arxiv: 2604.03591 · v3 · submitted 2026-04-04 · 💻 cs.DC · cs.PF

Minos: Systematically Classifying Performance and Power Characteristics of GPU Workloads on HPC Clusters

classification 💻 cs.DC cs.PF
keywords powerworkloadsclustersminosperformanceprofilingapplicationsincreasingly
0
0 comments X
read the original abstract

As large-scale HPC compute clusters increasingly adopt accelerators such as GPUs to meet the voracious demands of modern workloads, these clusters are increasingly becoming power constrained. Unfortunately, modern applications can often temporarily exceed the power ratings of the accelerators ("power spikes"). Thus, current and future HPC systems must optimize for both power and performance together. However, this is made difficult by increasingly diverse applications, which often require bespoke optimizations to run efficiently on each cluster. Traditionally researchers overcome this problem by profiling applications on specific clusters and optimizing, but the scale, algorithmic diversity, and lack of effective tools make this challenging. To overcome these inefficiencies, we propose Minos, a systematic classification mechanism that identifies similar application characteristics via low-cost profiling for power and performance. This allows us to group similarly behaving workloads into a finite number of distinct classes and reduce the overhead of extensively profiling new workloads. For example, when predicting frequency capping behavior for a previously unseen application, Minos reduces profiling time by 89%. Moreover, across 18 popular graph analytics, HPC, HPC+ML, and ML workloads, Minos achieves a mean error of 4% for power predictions and 3% for performance predictions, significantly improving predictions over state-of-the-art approaches by 10%.

This paper has not been read by Pith yet.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Fine-Grained Power and Energy Attribution on AMD GPU/APU-Based Exascale Nodes

    cs.DC 2026-04 unverdicted novelty 5.0

    A sensor characterization and reconstruction method enables fine-grained power attribution on Frontier and Portage exascale nodes, separating runtime and power effects in mixed-precision benchmarks.