A Cloud Server Oriented FPGA Accelerator for LSTM Recurrent Neural Network
Long Short-Term Memory network (LSTM) is the most widely used recurrent neural network architecture. It plays an important role in a number of research areas, such as language modeling, machine translation and image captioning. However, owing to its recurrent nature, general-purpose processors like...
Main Authors: | , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
IEEE
2019-01-01
|
Series: | IEEE Access |
Subjects: | |
Online Access: | https://ieeexplore.ieee.org/document/8819996/ |
id |
doaj-3a97a08e64cb4b16b373ac8b6ec9a39e |
---|---|
record_format |
Article |
spelling |
doaj-3a97a08e64cb4b16b373ac8b6ec9a39e2021-03-29T23:24:45ZengIEEEIEEE Access2169-35362019-01-01712240812241810.1109/ACCESS.2019.29382348819996A Cloud Server Oriented FPGA Accelerator for LSTM Recurrent Neural NetworkJun Liu0https://orcid.org/0000-0003-4007-6109Jiasheng Wang1Yu Zhou2Fang Liu3Center for Data Science Beijing University of Posts and Telecommunications, Beijing, ChinaCenter for Data Science Beijing University of Posts and Telecommunications, Beijing, ChinaCenter for Data Science Beijing University of Posts and Telecommunications, Beijing, ChinaCenter for Data Science Beijing University of Posts and Telecommunications, Beijing, ChinaLong Short-Term Memory network (LSTM) is the most widely used recurrent neural network architecture. It plays an important role in a number of research areas, such as language modeling, machine translation and image captioning. However, owing to its recurrent nature, general-purpose processors like CPUs and GPGPUs can achieve limited parallelism while consuming high power energy. FPGA accelerators can outperform general-purpose processors with flexibility, energy-efficiency and more delicate optimization capabilities for the recurrent based algorithms. In this paper, we present the design and implementation of a cloud-oriented FPGA accelerator for LSTM. Different from most of previous works designed for embedded systems, our FPGA accelerator transfers data sequences from and to the host server through PCIe and performs multiple time series predictions in parallel. We optimize both the on-chip computation and the communication between the host server and the FPGA board. We perform experiments to evaluate the overall performance as well as the computation and the PCIe communication efforts. The results show that the performance of our implementation is better than the CPU-based and other hardware-based implementations.https://ieeexplore.ieee.org/document/8819996/Long short-term memoryFPGAcloud computingrecurrent neural networkOpenCL |
collection |
DOAJ |
language |
English |
format |
Article |
sources |
DOAJ |
author |
Jun Liu Jiasheng Wang Yu Zhou Fang Liu |
spellingShingle |
Jun Liu Jiasheng Wang Yu Zhou Fang Liu A Cloud Server Oriented FPGA Accelerator for LSTM Recurrent Neural Network IEEE Access Long short-term memory FPGA cloud computing recurrent neural network OpenCL |
author_facet |
Jun Liu Jiasheng Wang Yu Zhou Fang Liu |
author_sort |
Jun Liu |
title |
A Cloud Server Oriented FPGA Accelerator for LSTM Recurrent Neural Network |
title_short |
A Cloud Server Oriented FPGA Accelerator for LSTM Recurrent Neural Network |
title_full |
A Cloud Server Oriented FPGA Accelerator for LSTM Recurrent Neural Network |
title_fullStr |
A Cloud Server Oriented FPGA Accelerator for LSTM Recurrent Neural Network |
title_full_unstemmed |
A Cloud Server Oriented FPGA Accelerator for LSTM Recurrent Neural Network |
title_sort |
cloud server oriented fpga accelerator for lstm recurrent neural network |
publisher |
IEEE |
series |
IEEE Access |
issn |
2169-3536 |
publishDate |
2019-01-01 |
description |
Long Short-Term Memory network (LSTM) is the most widely used recurrent neural network architecture. It plays an important role in a number of research areas, such as language modeling, machine translation and image captioning. However, owing to its recurrent nature, general-purpose processors like CPUs and GPGPUs can achieve limited parallelism while consuming high power energy. FPGA accelerators can outperform general-purpose processors with flexibility, energy-efficiency and more delicate optimization capabilities for the recurrent based algorithms. In this paper, we present the design and implementation of a cloud-oriented FPGA accelerator for LSTM. Different from most of previous works designed for embedded systems, our FPGA accelerator transfers data sequences from and to the host server through PCIe and performs multiple time series predictions in parallel. We optimize both the on-chip computation and the communication between the host server and the FPGA board. We perform experiments to evaluate the overall performance as well as the computation and the PCIe communication efforts. The results show that the performance of our implementation is better than the CPU-based and other hardware-based implementations. |
topic |
Long short-term memory FPGA cloud computing recurrent neural network OpenCL |
url |
https://ieeexplore.ieee.org/document/8819996/ |
work_keys_str_mv |
AT junliu acloudserverorientedfpgaacceleratorforlstmrecurrentneuralnetwork AT jiashengwang acloudserverorientedfpgaacceleratorforlstmrecurrentneuralnetwork AT yuzhou acloudserverorientedfpgaacceleratorforlstmrecurrentneuralnetwork AT fangliu acloudserverorientedfpgaacceleratorforlstmrecurrentneuralnetwork AT junliu cloudserverorientedfpgaacceleratorforlstmrecurrentneuralnetwork AT jiashengwang cloudserverorientedfpgaacceleratorforlstmrecurrentneuralnetwork AT yuzhou cloudserverorientedfpgaacceleratorforlstmrecurrentneuralnetwork AT fangliu cloudserverorientedfpgaacceleratorforlstmrecurrentneuralnetwork |
_version_ |
1724189526056763392 |