Charge Based Mixed-Signal Multiply-Accumulate Circuit for Energy Efficient In-Memory Computing

Autor:	Sebastian Kelz, Felix Wiewel, Raphael Nagele, Markus Grozing, Manfred Berroth, Bin Yang, Manuel Wittlinger
Rok vydání:	2021
Předmět:	Physics business.industry Deep learning Resolution (electron density) Charge (physics) Biological neuron model Mixed-signal integrated circuit 4-bit In-Memory Processing Electronic engineering Artificial intelligence Hardware_ARITHMETICANDLOGICSTRUCTURES business Efficient energy use
Zdroj:	2021 Kleinheubach Conference.
DOI:	10.23919/ieeeconf54431.2021.9598440
Popis:	A feasibility study of an energy efficient analog mixed-signal multiply-accumulate circuit with multi-bit resolution is presented. With a derived hardware-specific neuron model a ResNetV2 with 4 bit quantized weights is trained on a CIFAR10 dataset achieving 90% accuracy. The simulated energy efficiency of the multiply-accumulate circuit is 2.840 fJ per multiply-accumulate operation.
Databáze:	OpenAIRE
Externí odkaz:	https://explore.openaire.eu/search/publication?articleId=doi_________::96cf98c29115053c7035015040e7e506 https://doi.org/10.23919/ieeeconf54431.2021.9598440 Zobrazit plný text záznamu