Name
Title | ||
|---|---|---|
Low Precision Floating Point Arithmetic for High Performance FPGA-based CNN Acceleration. |
Abstract | ||
|---|---|---|
Low precision data representation is important to reduce storage size and memory access for convolutional neural networks (CNNs). Yet, existing methods have two major limitations: (1) requiring re-training to maintain accuracy for deep CNNs, and (2) needing 16-bit floating point or 8-bit fixed point for a good accuracy.
In this paper, we propose a low precision (8-bit) floating point (LPFP) quantization method for FPGA-based acceleration to overcome the above limitations. Without any re-training, LPFP finds an optimal 8-bit data representation with negligible top-1/top-5 accuracy loss (within 0.5%/0.3% in our experiments, respectively, and significantly better than existing methods for deep CNNs). Furthermore, we implement one 8-bit LPFP multiplication by one 4-bit multiply-adder (MAC) and one 3-bit adder, and therefore implement four 8-bit LPFP multiplications using one DSP slice of Xilinx Kintex-7 family (KC705 in this paper) while one DSP can implement only two 8-bit fixed point multiplications. Experiments on six typical CNNs for inference show that on average, we improve throughput by 64.5× over Intel i9 CPU and by 1.5× over existing FPGA accelerators. Particularly for VGG16 and YOLO, compared to six recent FPGA accelerators, we improve average throughput by 3.5× and 27.5× and improve average throughput per DSP by 4.1× and 5×, respectively. To the best of our knowledge, this is the first in-depth study to simplify one multiplication for CNN inference to one 4-bit MAC and implement four multiplications within one DSP while maintaining comparable accuracy without any re-training.
|
Year | DOI | Venue |
|---|---|---|
2020 | 10.1145/3373087.3375361 | FPGA |
Field | DocType | ISBN |
Computer science,Floating point,Parallel computing,Field-programmable gate array,Acceleration | Conference | 978-1-4503-7099-8 |
Citations | PageRank | References |
2 | 0.42 | 0 |
Authors | ||
5 | ||
Name | Order | Citations | PageRank |
|---|---|---|---|
| Chen Wu | 1 | 69 | 6.20 |
| Mingyu Wang | 2 | 135 | 24.90 |
| Xinyuan Chu | 3 | 2 | 0.76 |
| Kun Wang | 4 | 425 | 56.96 |
| Lei He | 5 | 1015 | 86.74 |