EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Cross-Feature Transfer Learning for Efficient Tensor Program Generation

Gaurav Verma, Siddhisanket Raskar, Murali Emani, Barbara Chapman

Applied Sciences · 06.01.2024 · 14 (2) · 513

DOI: 10.3390/app14020513 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Applied Sciences, online erschienen am 06.01.2024. Verfasst von Gaurav Verma und 3 weiteren Autor:innen. Die Arbeit wurde bisher 8-mal zitiert (Crossref) und nennt 50 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Stony Brook Research Computing and Cyberinfrastructure, Argonne Leadership Computing Facility (ALCF), Exascale Computing Project.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Tuning tensor program generation involves navigating a vast search space to find optimal program transformations and measurements for a program on the target hardware. The complexity of this process is further amplified by the exponential combinations of transformations, especially in heterogeneous environments. This research addresses these challenges by introducing a novel approach that learns the joint neural network and hardware features space, facilitating knowledge transfer to new, unseen target hardware. A comprehensive analysis is conducted on the existing state-of-the-art dataset, TenSet, including a thorough examination of test split strategies and the proposal of methodologies for dataset pruning. Leveraging an attention-inspired technique, we tailor the tuning of tensor programs to embed both neural network and hardware-specific features. Notably, our approach substantially reduces the dataset size by up to 53% compared to the baseline without compromising Pairwise Comparison Accuracy (PCA). Furthermore, our proposed methodology demonstrates competitive or improved mean inference times with only 25–40% of the baseline tuning time across various networks and target hardware. The attention-based tuner can effectively utilize schedules learned from previous hardware program measurements to optimize tensor program tuning on previously unseen hardware, achieving a top-5 accuracy exceeding 90%. This research introduces a significant advancement in autotuning tensor program generation, addressing the complexities associated with heterogeneous environments and showcasing promising results regarding efficiency and accuracy.

Förderung

  • Stony Brook Research Computing and Cyberinfrastructure
  • Argonne Leadership Computing Facility (ALCF)
  • Exascale Computing Project
  • National Science Foundation

Zitieren

APA

Verma, G., Raskar, S., Emani, M., & Chapman, B. (2024). Cross-Feature Transfer Learning for Efficient Tensor Program Generation. Applied Sciences, 14(2), 513. https://doi.org/10.3390/app14020513

Vancouver

Verma G, Raskar S, Emani M, Chapman B. Cross-Feature Transfer Learning for Efficient Tensor Program Generation. Applied Sciences. 2024;14(2):513. doi:10.3390/app14020513

BibTeX
@article{Verma2024,
  author = {Verma, Gaurav and Raskar, Siddhisanket and Emani, Murali and Chapman, Barbara},
  title = {Cross-Feature Transfer Learning for Efficient Tensor Program Generation},
  journal = {Applied Sciences},
  year = {2024},
  volume = {14},
  number = {2},
  pages = {513},
  doi = {10.3390/app14020513}
}
Kostenlos starten →