HNPU: An Adaptive DNN Training Processor Utilizing Stochastic Dynamic Fixed-Point and Active Bit-Precision Searching

Cited 27 time in webofscience Cited 0 time in scopus
  • Hit : 731
  • Download : 42
This article presents HNPU, which is an energy-efficient deep neural network (DNN) training processor by adopting algorithm-hardware co-design. The HNPU supports stochastic dynamic fixed-point representation and layer-wise adaptive precision searching unit for low-bit-precision training. It additionally utilizes slice-level reconfigurability and sparsity to maximize its efficiency both in DNN inference and training. Adaptive bandwidth reconfigurable accumulation network enables reconfigurable DNN allocation and maintains its high core utilization even in various bit-precision conditions. Fabricated in a 28-nm process, the HNPU accomplished at least 5.9x higher energy efficiency and 2.5x higher area efficiency in actual DNN training compared with the previous state-of-the-art on-chip learning processors.
Publisher
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
Issue Date
2021-09
Language
English
Article Type
Article
Citation

IEEE JOURNAL OF SOLID-STATE CIRCUITS, v.56, no.9, pp.2858 - 2869

ISSN
0018-9200
DOI
10.1109/JSSC.2021.3066400
URI
http://hdl.handle.net/10203/287867
Appears in Collection
EE-Journal Papers(저널논문)
Files in This Item
This item is cited by other documents in WoS
⊙ Detail Information in WoSⓡ Click to see webofscience_button
⊙ Cited 27 items in WoS Click to see citing articles in records_button

qr_code

  • mendeley

    citeulike


rss_1.0 rss_2.0 atom_1.0