A 12nm 137 TOPS/W Digital Compute-In-Memory using Foundry 8T SRAM Bitcell supporting 16 Kernel Weight Sets for AI Edge Applications.
Gajanan JedheChetan DeshpandeSushil KumarCheng-Xin XueZijie GuoRitesh GargKim Soon JwayEn-Jui ChangJenwei LiangZhe WanZhenhao PanPublished in: VLSI Technology and Circuits (2023)
Keyphrases
- random access memory
- artificial intelligence
- dynamic random access memory
- power consumption
- embedded dram
- case based reasoning
- kernel methods
- edge information
- linear separability
- memory usage
- machine learning
- ai systems
- data transmission
- edge detector
- kernel function
- feature space
- memory requirements
- cmos technology
- metal oxide semiconductor
- support vector
- low power
- intelligent systems
- edge weights
- edge detection
- multiscale
- expert systems
- leakage current