A simulated racetrack-memory in-memory CNN accelerator using write-to-shift conversion and shift-based multipliers achieves up to 83.5x energy-efficiency improvement and 1.68x performance gain over an unoptimized 8-bit fixed-point baseline.
Flinkcl: An opencl- based in-memory computing architecture on heterogeneous cpu-gpu clusters for big data
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.ET 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
background 1representative citing papers
citing papers explorer
-
Hardware-software co-exploration with racetrack memory based in-memory computing for CNN inference in embedded systems
A simulated racetrack-memory in-memory CNN accelerator using write-to-shift conversion and shift-based multipliers achieves up to 83.5x energy-efficiency improvement and 1.68x performance gain over an unoptimized 8-bit fixed-point baseline.