CONV-SRAM: An Energy-Efficient SRAM With In-Memory Dot-Product Computation for Low-Power Convolutional Neural Networks.
Avishek BiswasAnantha P. ChandrakasanPublished in: IEEE J. Solid State Circuits (2019)
Keyphrases
- low power
- dot product
- power consumption
- convolutional neural networks
- energy efficiency
- low cost
- high speed
- energy saving
- power reduction
- kernel function
- similarity function
- gaussian kernels
- power dissipation
- scalar product
- single chip
- vlsi circuits
- image sensor
- low power consumption
- digital signal processing
- feature space
- random access memory
- wireless sensor networks
- cmos technology
- mac protocol
- real time
- similarity measure
- gate array
- power saving
- main memory
- pairwise
- multiscale
- machine learning