PLAM: a posit logarithm-approximate multiplier

The Posit™ Number System was introduced in 2017 as a replacement for floating-point numbers. Since then, the community has explored its application in several areas, such as deep learning, and produced some unit designs which are still far from being competitive with their floating-point counterpart...

Descripción completa

Detalles Bibliográficos
Autores: Murillo Montero, Raúl, Del Barrio García, Alberto Antonio, Botella Juan, Guillermo, Kim, Min Soo, Kim, HyunJin, Bagherzadeh, Nader
Tipo de recurso: artículo
Fecha de publicación:2021
País:España
Institución:Universidad Complutense de Madrid (UCM)
Repositorio:Docta Complutense
Idioma:inglés
OAI Identifier:oai:docta.ucm.es:20.500.14352/132729
Acceso en línea:https://hdl.handle.net/20.500.14352/132729
Access Level:acceso abierto
Palabra clave:004.3
621.38
Posit arithmetic
Arithmetic and logic structures
Low-power design
Machine learning
Computer vision
Informática (Informática)
Hardware
1203 Ciencia de Los Ordenadores
3307.03 Diseño de Circuitos
Descripción
Sumario:The Posit™ Number System was introduced in 2017 as a replacement for floating-point numbers. Since then, the community has explored its application in several areas, such as deep learning, and produced some unit designs which are still far from being competitive with their floating-point counterparts. This article proposes a Posit Logarithm-Approximate Multiplication (PLAM) scheme to significantly reduce the complexity of posit multipliers, one of the most power-hungry arithmetic units. The impact of this approach is evaluated in deep neural network inference, where there are no significant accuracy drops. Compared with state-of-the-art posit multipliers, experiments show that the proposed technique reduces the area, power, and delay of 32-bit hardware multipliers up to 72.86%, 81.79%, and 17.01%, respectively.