COMET: Neural Cost Model Explanation Framework
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2023
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866917642851319808 |
|---|---|
| author | Chaudhary, Isha Renda, Alex Mendis, Charith Singh, Gagandeep |
| author_facet | Chaudhary, Isha Renda, Alex Mendis, Charith Singh, Gagandeep |
| contents | Cost models predict the cost of executing given assembly code basic blocks on a specific microarchitecture. Recently, neural cost models have been shown to be fairly accurate and easy to construct. They can replace heavily engineered analytical cost models used in mainstream compiler workflows. However, their black-box nature discourages their adoption. In this work, we develop the first framework, COMET, for generating faithful, generalizable, and intuitive explanations for neural cost models. We generate and compare COMET's explanations for the popular neural cost model, Ithemal against those for an accurate CPU simulation-based cost model, uiCA. Our empirical findings show an inverse correlation between the prediction errors of Ithemal and uiCA and the granularity of basic block features in COMET's explanations for them, thus indicating potential reasons for the higher error of Ithemal with respect to uiCA. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2302_06836 |
| institution | arXiv |
| publishDate | 2023 |
| record_format | arxiv |
| spellingShingle | COMET: Neural Cost Model Explanation Framework Chaudhary, Isha Renda, Alex Mendis, Charith Singh, Gagandeep Performance Artificial Intelligence Hardware Architecture Distributed, Parallel, and Cluster Computing Cost models predict the cost of executing given assembly code basic blocks on a specific microarchitecture. Recently, neural cost models have been shown to be fairly accurate and easy to construct. They can replace heavily engineered analytical cost models used in mainstream compiler workflows. However, their black-box nature discourages their adoption. In this work, we develop the first framework, COMET, for generating faithful, generalizable, and intuitive explanations for neural cost models. We generate and compare COMET's explanations for the popular neural cost model, Ithemal against those for an accurate CPU simulation-based cost model, uiCA. Our empirical findings show an inverse correlation between the prediction errors of Ithemal and uiCA and the granularity of basic block features in COMET's explanations for them, thus indicating potential reasons for the higher error of Ithemal with respect to uiCA. |
| title | COMET: Neural Cost Model Explanation Framework |
| topic | Performance Artificial Intelligence Hardware Architecture Distributed, Parallel, and Cluster Computing |
| url | https://arxiv.org/abs/2302.06836 |