Sparse Periodic Systolic Dataflow for Lowering Latency and Power Dissipation of Convolutional Neural Network Accelerators.
Jung Hwan HeoArash FayyaziAmirhossein EsmailiMassoud PedramPublished in: CoRR (2022)
Keyphrases
- convolutional neural network
- power dissipation
- power consumption
- face detection
- design methodology
- low power
- digital signal processing
- data flow
- power reduction
- clock frequency
- cmos technology
- neural network
- nm technology
- parallel computing
- logic circuits
- computer vision
- finite state machines
- computing systems
- energy saving
- data transfer
- energy efficiency
- power saving
- input output
- video sequences
- case study
- real time