Using FPGAs to accelerate ConvNets has attracted significant attention in recent years. However, FPGA accelerator design has not leveraged the latest progress of ConvNets. As a result, the key application characteristics such as frames-per-second (FPS) are ignored in favor of simply counting GOPs, and results on accuracy, which is critical to application success, are often not even reported. In this work, we adopt an algorithm-hardware co-design approach to develop a ConvNet accelerator called Synetgy and a novel ConvNet model called DiracDeltaNet†. Both the accelerator and ConvNet are tailored to FPGA requirements. DiracDeltaNet, as the name suggests, is a ConvNet with only 1x1 convolutions while spatial convolutions are replaced by more efficient shift operations. DiracDeltaNet achieves competitive accuracy on ImageNet (89.0% top-5), but with 48x fewer parameters and 65x fewer OPs than VGG16. We further quantize DiracDeltaNet’s weights to 1-bit and activations to 4-bits, with less than 1% accuracy loss. These quantizations exploit well the nature of FPGA hardware. In short, DiracDeltaNet’s small model size, low computational OP count, ultra-low precision and simplified operators allow us to co-design a highly customized computing unit for an FPGA. We implement the computing units for DiracDeltaNet on an Ultra96 SoC system through high-level synthesis. Our accelerator’s final top-5 accuracy of 88.2% on ImageNet, is higher than all the previously reported embedded FPGA accelerators. In addition, the accelerator reaches an inference speed of 96.5 FPS on the ImageNet classification task, surpassing prior works with similar accuracy by at least 16.9x.

Synetgy: Algorithm-hardware Co-design for ConvNet Accelerators on Embedded FPGAs / Yang, Yifan; Huang, Qijing; Wu, Bichen; Zhang, Tianjun; Ma, Liang; Gambardella, Giulio; Blott, Michaela; Lavagno, Luciano; Vissers, Kees; Wawrzynek, John; Keutzer, Kurt. - ELETTRONICO. - (2019), pp. 23-32. (Intervento presentato al convegno 27th ACM/SIGDA International Symposium on Field-Programmable Gate Arrays) [10.1145/3289602.3293902].

Synetgy: Algorithm-hardware Co-design for ConvNet Accelerators on Embedded FPGAs

Liang Ma;Giulio Gambardella;Luciano Lavagno;
2019

Abstract

Using FPGAs to accelerate ConvNets has attracted significant attention in recent years. However, FPGA accelerator design has not leveraged the latest progress of ConvNets. As a result, the key application characteristics such as frames-per-second (FPS) are ignored in favor of simply counting GOPs, and results on accuracy, which is critical to application success, are often not even reported. In this work, we adopt an algorithm-hardware co-design approach to develop a ConvNet accelerator called Synetgy and a novel ConvNet model called DiracDeltaNet†. Both the accelerator and ConvNet are tailored to FPGA requirements. DiracDeltaNet, as the name suggests, is a ConvNet with only 1x1 convolutions while spatial convolutions are replaced by more efficient shift operations. DiracDeltaNet achieves competitive accuracy on ImageNet (89.0% top-5), but with 48x fewer parameters and 65x fewer OPs than VGG16. We further quantize DiracDeltaNet’s weights to 1-bit and activations to 4-bits, with less than 1% accuracy loss. These quantizations exploit well the nature of FPGA hardware. In short, DiracDeltaNet’s small model size, low computational OP count, ultra-low precision and simplified operators allow us to co-design a highly customized computing unit for an FPGA. We implement the computing units for DiracDeltaNet on an Ultra96 SoC system through high-level synthesis. Our accelerator’s final top-5 accuracy of 88.2% on ImageNet, is higher than all the previously reported embedded FPGA accelerators. In addition, the accelerator reaches an inference speed of 96.5 FPS on the ImageNet classification task, surpassing prior works with similar accuracy by at least 16.9x.
2019
9781450361378
File in questo prodotto:
Non ci sono file associati a questo prodotto.
Pubblicazioni consigliate

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11583/2726980
 Attenzione

Attenzione! I dati visualizzati non sono stati sottoposti a validazione da parte dell'ateneo