MSDF-Based Hardware Accelerators for Energy-Efficient Neural Networks in Edge Computing Applications

27 de agosto de 2025

Edge computing applications, such as IoT and wearable devices, demand energy-efficient and high-performance hardware accelerators to support neural networks in resource-constrained environments. This paper presents a novel Multiply-Accumulate (MAC) unit based on Most Significant Digit First (MSDF) arithmetic, designed to address these challenges. The proposed MSDF MAC unit exploits digit-level parallelism and eliminates carry propagation delays to enhance throughput and reduce energy consumption. We integrate this unit into three different widely used applications: (I) Sparse Matrix-Vector Multiplication (SpMV), (II) Multi-Layer Perceptron (MLP), and (III) Denoising Autoencoder. All designs are synthesized using TSMC 45 nm CMOS technology, and the proposed designs achieve significant improvements, including 1.69× speedup and up to 60.4% energy reduction on SpMV, and at least 39% power savings on MLP with maintained accuracy for the MLP and the autoencoder. These results demonstrate the potential of MSDF-based accelerators for efficient neural network deployment on edge devices.

Autores da comunidade :

• 

  S

  Sahar Moradi Cherati

  ist1101188

  Afiliações e fontes

  IST

  Instituto Superior Técnico

• 

  L

  Leonel Augusto Pires Seabra de Sousa

  ist13348

  Afiliações e fontes

  IST > INESC-ID

  Instituto de Engenharia de Sistemas e Computadores - Investigação e Desenvolvimento em Lisboa

  IST > DEEC

  Departamento de Engenharia Electrotécnica e de Computadores

NA - Versão desconhecida

IEEE

https://ieeexplore.ieee.org/document/11130344

2025 IEEE Computer Society Annual Symposium on VLSI (ISVLSI)

1

4

1

2159-3477

10.1109/ISVLSI65124.2025.11130344

electrical-engineering-electronic-engineering-information-engineering - Engenharia Eletrotécnica, Eletrónica e Informática

Palavras-chave

• MSDF arithmetic
• hardware accelerators
• edge computing
• neural networks
• energy efficiency

eng - Inglês