ESSCIRC 2022- IEEE 48th European Solid State Circuits Conference (ESSCIRC) · 2022 · 25 citations · 10 references
EngineeringComputer EngineeringComputer ScienceComputer ArchitectureComputer VisionMachine LearningAccelerator ArchitectureDeep LearningEnergy-efficient Embedded SystemsHardware AccelerationLow-power Digital Circuit DesignPower OptimizationDeep InferenceDeep Learning AccelerationDeep Learning AcceleratorsEnergy-efficient ComputationEnergy-efficient Hardware AccelerationFloating-point InstructionsFloating-point Unit DesignPrecision
This paper presents a digital compute-in-memory (CIM) macro for accelerating deep neural networks. The macro provides high-precision computation required for training deep neural networks and running state-of-the-art models by supporting floating-point MAC operations. Additionally, the design supports variable computation precision, enabling optimized processing for different models and tasks. The design achieves 1.644TFLOPS/W energy efficiency and 57.9GFLOPS/mm <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> computation density while supporting a wide range of floating-point data formats and computation precisions.
10