2022 IEEE Symposium on VLSI Technology and Circuits (VLSI Technology and Circuits) · 2022 · 39 citations · 0 references
EngineeringComputer EngineeringElectrical EngineeringComputer ArchitectureMicroelectronicsSemiconductor DevicesCmos Circuit DesignEmbedded Memory BlocksLeakage Power ReductionLow-power Digital Circuit DesignMemory DevicesMemory Subsystem DesignCmos ScalingConvertersEnergy-efficient Memory DesignLeakage-aware Memory DesignLeakage-constrained Circuit DesignMemory Circuit DesignNon-volatile Memory DesignTriggers
We present a leakage and read bitline (RBL) swing aware Compute-in-Memory (CIM) design leveraging a promising high-density gain-cell embedded DRAM bitcell and the intrinsic RBL capacitors to perform CIM computations within the limited RBL swing available in a 2T1C eDRAM. The CIM D/A converters (DAC) are realized intrinsically with variable RBL precharge voltage levels. A/D converters (ADC) are realized using Schmitt Triggers (ST) as compact and reconfigurable Flash comparators. A 65nm CMOS prototype achieves energy efficiency of 7.4-236 TOPS/W, 13.1-411 GOPS/mm <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> for the CIFAR-10 dataset with ResNet-20 and improves the defined FoM by 2.3-4.3X over prior CIM designs.