Integrating Performance Tools in Model Reasoning for GPU Kernel Optimization
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866911220634746880 |
|---|---|
| author | Nichols, Daniel Parasyris, Konstantinos Jekel, Charles Bhatele, Abhinav Menon, Harshitha |
| author_facet | Nichols, Daniel Parasyris, Konstantinos Jekel, Charles Bhatele, Abhinav Menon, Harshitha |
| contents | Language models are now prevalent in software engineering with many developers using them to automate tasks and accelerate their development. While language models have been tremendous at accomplishing complex software engineering tasks, there are still many areas where they fail to deliver desirable results, for instance code performance related tasks. Tasks like optimization depend on many complex data from the environment, hardware, etc. that are not directly represented in source code. Recent efforts have seen large improvements in general code modeling tasks using chain-of-thought style reasoning, but these models still fail to comprehend how the environment interacts with code performance. In this paper we propose a methodology to train language models that can interact with performance tools during their reasoning process. We then demonstrate how this methodology can be used to train a state-of-the-art GPU kernel optimization model. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_17158 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Integrating Performance Tools in Model Reasoning for GPU Kernel Optimization Nichols, Daniel Parasyris, Konstantinos Jekel, Charles Bhatele, Abhinav Menon, Harshitha Distributed, Parallel, and Cluster Computing Performance Software Engineering Language models are now prevalent in software engineering with many developers using them to automate tasks and accelerate their development. While language models have been tremendous at accomplishing complex software engineering tasks, there are still many areas where they fail to deliver desirable results, for instance code performance related tasks. Tasks like optimization depend on many complex data from the environment, hardware, etc. that are not directly represented in source code. Recent efforts have seen large improvements in general code modeling tasks using chain-of-thought style reasoning, but these models still fail to comprehend how the environment interacts with code performance. In this paper we propose a methodology to train language models that can interact with performance tools during their reasoning process. We then demonstrate how this methodology can be used to train a state-of-the-art GPU kernel optimization model. |
| title | Integrating Performance Tools in Model Reasoning for GPU Kernel Optimization |
| topic | Distributed, Parallel, and Cluster Computing Performance Software Engineering |
| url | https://arxiv.org/abs/2510.17158 |