Z-PIM: A Sparsity-Aware Processing-in-Memory Architecture With Fully Variable Weight Bit-Precision for Energy-Efficient Deep Neural Networks.
Ji-Hoon KimJuhyoung LeeJinsu LeeJaehoon HeoJoo-Young KimPublished in: IEEE J. Solid State Circuits (2021)
Keyphrases
- energy efficient
- neural network
- processing elements
- wireless sensor networks
- associative memory
- energy consumption
- memory management
- sensor networks
- multi core architecture
- real time
- network architecture
- parallel architecture
- base station
- random access
- management system
- energy efficiency
- routing protocol
- data processing
- hardware architecture
- memory access
- analog to digital converter
- multi layer
- data dissemination
- data transmission
- processing units
- data center
- power consumption
- response time
- data streams