Academic Journal
Graph-based clustering cost model for performance optimization in deep learning compilers.
| Τίτλος: | Graph-based clustering cost model for performance optimization in deep learning compilers. |
|---|---|
| Συγγραφείς: | Shen, Chaoyao1 (AUTHOR) shen_chaoyao@seu.edu.cn, Shen, Yixian2 (AUTHOR) y.shen@uva.nl, Xu, Tao1 (AUTHOR) xu_tao@seu.edu.cn, Guo, Bohan1 (AUTHOR) guobohan@seu.edu.cn, Jiang, Linfeng1 (AUTHOR) linfengjiang@seu.edu.cn, Zhang, Meng1 (AUTHOR) zmeng@seu.edu.cn, Pathania, Anuj2 (AUTHOR) a.pathania@uva.nl, Pimentel, Andy D.2 (AUTHOR) a.d.pimentel@uva.nl |
| Πηγή: | Journal of Systems Architecture. Sep2026, Vol. 178, pN.PAG-N.PAG. 1p. |
| Θεματικοί όροι: | *Mathematical optimization, Clustering algorithms, Compilers (Computer programs), Graph neural networks, Cost functions |
| Περίληψη: | Deep learning (DL) compilers enable automated deployment of deep neural networks (DNNs) across diverse hardware platforms by leveraging cost models for efficient tensor program optimization. The effectiveness of these systems depends on the prediction accuracy and cross-model generalization capability of offline-trained cost models. Existing approaches mainly focus on feature engineering and cost model architecture design, while largely overlooking structural similarity at the computational subgraph level, which limits further improvements in prediction performance. To address this limitation, we propose an unsupervised subgraph clustering framework based on graph autoencoders to enhance offline-trained cost models. Specifically, DL models are partitioned into computational subgraphs, and node features are extracted. A graph autoencoder learns high-dimensional structural embeddings of these subgraphs, followed by K-means clustering to group structurally similar patterns. Dedicated cost models are trained for different subgraph clusters to predict tensor program performance. This fine-grained, cluster-aware training strategy improves both prediction accuracy and generalization capability, thereby enhancing auto-tuning efficiency. Experimental results across multiple DL models demonstrate significant improvements in search efficiency and latency optimization on both CPU and GPU platforms. Compared with Tenset, our method achieves average search speedups of 9.46 × and 5.15 × , while achieving inference latency improvements of 1.18 × and 1.07 ×. [ABSTRACT FROM AUTHOR] |
| Copyright of Journal of Systems Architecture is the property of Elsevier B.V. and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.) | |
| Βάση Δεδομένων: | Business Source Index |
καταχωρήστε σχόλιο πρώτοι!