Systolic Array Based Accelerator and Algorithm Mapping for Deep Learning Algorithms - Inria - Institut national de recherche en sciences et technologies du numérique Accéder directement au contenu
Communication Dans Un Congrès Année : 2018

Systolic Array Based Accelerator and Algorithm Mapping for Deep Learning Algorithms

Résumé

As the depth of DNN increases, the need for DNN calculations for the storage and computing power of the underlying computing platform is increasing. In this work, we implement an accelerator on FPGA for deep learning algorithms (CNN and RNN). The core computing module of the accelerator is a 32 * 32 systolic array of PEs. A mapping method for variable size of CNN and RNN algorithms is proposed. The experiment result shows that the maximum power consumption of the accelerator is 7.5W@100Mhz, the peak performance is 0.2Tops/s, and the real performance is 7.6Mops@100Mhz when running the 1st layer of LeNet-5.
Fichier principal
Vignette du fichier
477597_1_En_16_Chapter.pdf (489.55 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-02279547 , version 1 (05-09-2019)

Licence

Paternité

Identifiants

Citer

Zhijie Yang, Lei Wang, Dong Ding, Xiangyu Zhang, Yu Deng, et al.. Systolic Array Based Accelerator and Algorithm Mapping for Deep Learning Algorithms. 15th IFIP International Conference on Network and Parallel Computing (NPC), Nov 2018, Muroran, Japan. pp.153-158, ⟨10.1007/978-3-030-05677-3_16⟩. ⟨hal-02279547⟩
129 Consultations
720 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More