Leveraging Subgraph Extraction for Performance Portable Programming Frameworks on DL Accelerators - Network and Parallel Computing
Conference Papers Year : 2018

Leveraging Subgraph Extraction for Performance Portable Programming Frameworks on DL Accelerators

Abstract

Deep learning framework plays an important role in connecting hardware platform and algorithm. In recent years, some domain-specific deep learning accelerators with better performance and energy efficiency were proposed by researchers. However, current frameworks lack enough considerations about how to better support the possible new features brought by accelerators. In this paper, we propose to build a performance portable programming framework with subgraph extraction. The intuition is that increasing ratio of optimizations are taken from the top-level framework to the low-level software stack of accelerator. In response to this development trend, framework needs to pay more attention to the splitting strategy of computation graph for the heterogeneous computation.
Fichier principal
Vignette du fichier
477597_1_En_21_Chapter.pdf (241.52 Ko) Télécharger le fichier
Origin Files produced by the author(s)
Loading...

Dates and versions

hal-02279540 , version 1 (05-09-2019)

Licence

Identifiers

Cite

Xiao Zhang, Huiying Lan, Tian Zhi. Leveraging Subgraph Extraction for Performance Portable Programming Frameworks on DL Accelerators. 15th IFIP International Conference on Network and Parallel Computing (NPC), Nov 2018, Muroran, Japan. pp.179-184, ⟨10.1007/978-3-030-05677-3_21⟩. ⟨hal-02279540⟩
105 View
76 Download

Altmetric

Share

More