Attention-Informed Surrogates for Navigating Power-Performance Trade-offs in HPC
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2026
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866917216754073600 |
|---|---|
| author | Ahmed, Ashna Nawar Banday, Banooqa Jones, Terry Islam, Tanzima Z. |
| author_facet | Ahmed, Ashna Nawar Banday, Banooqa Jones, Terry Islam, Tanzima Z. |
| contents | High-Performance Computing (HPC) schedulers must balance user performance with facility-wide resource constraints. The task boils down to selecting the optimal number of nodes for a given job. We present a surrogate-assisted multi-objective Bayesian optimization (MOBO) framework to automate this complex decision. Our core hypothesis is that surrogate models informed by attention-based embeddings of job telemetry can capture performance dynamics more effectively than standard regression techniques. We pair this with an intelligent sample acquisition strategy to ensure the approach is data-efficient. On two production HPC datasets, our embedding-informed method consistently identified higher-quality Pareto fronts of runtime-power trade-offs compared to baselines. Furthermore, our intelligent data sampling strategy drastically reduced training costs while improving the stability of the results. To our knowledge, this is the first work to successfully apply embedding-informed surrogates in a MOBO framework to the HPC scheduling problem, jointly optimizing for performance and power on production workloads. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_15399 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Attention-Informed Surrogates for Navigating Power-Performance Trade-offs in HPC Ahmed, Ashna Nawar Banday, Banooqa Jones, Terry Islam, Tanzima Z. Machine Learning High-Performance Computing (HPC) schedulers must balance user performance with facility-wide resource constraints. The task boils down to selecting the optimal number of nodes for a given job. We present a surrogate-assisted multi-objective Bayesian optimization (MOBO) framework to automate this complex decision. Our core hypothesis is that surrogate models informed by attention-based embeddings of job telemetry can capture performance dynamics more effectively than standard regression techniques. We pair this with an intelligent sample acquisition strategy to ensure the approach is data-efficient. On two production HPC datasets, our embedding-informed method consistently identified higher-quality Pareto fronts of runtime-power trade-offs compared to baselines. Furthermore, our intelligent data sampling strategy drastically reduced training costs while improving the stability of the results. To our knowledge, this is the first work to successfully apply embedding-informed surrogates in a MOBO framework to the HPC scheduling problem, jointly optimizing for performance and power on production workloads. |
| title | Attention-Informed Surrogates for Navigating Power-Performance Trade-offs in HPC |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2601.15399 |