Based on linear systolic array for convolutional neural network’s calculation optimization and performance analysis
Concerning the issue that the convolutional neural network (CNN) accelerator design on most FPGA ends fails to effectively use the sparsity and considering both bandwidth and energy consumption,two improved CNN calculation optimization strategies based on linear systolic array architecture are propo...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
POSTS&TELECOM PRESS Co., LTD
2018-12-01
|
| Series: | 网络与信息安全学报 |
| Subjects: | |
| Online Access: | http://www.cjnis.com.cn/thesisDetails#10.11959/j.issn.2096-109x.2018100 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1841530022772342784 |
|---|---|
| author | Qinrang LIU Chongyang LIU Jun ZHOU Xiaolong WANG |
| author_facet | Qinrang LIU Chongyang LIU Jun ZHOU Xiaolong WANG |
| author_sort | Qinrang LIU |
| collection | DOAJ |
| description | Concerning the issue that the convolutional neural network (CNN) accelerator design on most FPGA ends fails to effectively use the sparsity and considering both bandwidth and energy consumption,two improved CNN calculation optimization strategies based on linear systolic array architecture are proposed.Firstly,convolution is transformed into matrix multiplication to take advantage of sparsity.Secondly,in order to solve the problem of large I/O demand in traditional parallel matrix multiplier,linear systolic array is used to improve the design.Finally,a CNN acceleration comparative analysis of the advantages and disadvantages between parallel matrix multiplier and two improved linear systolic arrays is presented.Theoretical proof and analysis show that compared with the parallel matrix multiplier,the two improved linear systolic arrays make full use of sparsity,and have the advantages of less energy consumption and less I/O bandwidth occupation. |
| format | Article |
| id | doaj-art-00af2e41223f42d69c7104df248cd32a |
| institution | Kabale University |
| issn | 2096-109X |
| language | English |
| publishDate | 2018-12-01 |
| publisher | POSTS&TELECOM PRESS Co., LTD |
| record_format | Article |
| series | 网络与信息安全学报 |
| spelling | doaj-art-00af2e41223f42d69c7104df248cd32a2025-01-15T03:13:12ZengPOSTS&TELECOM PRESS Co., LTD网络与信息安全学报2096-109X2018-12-014162459554964Based on linear systolic array for convolutional neural network’s calculation optimization and performance analysisQinrang LIUChongyang LIUJun ZHOUXiaolong WANGConcerning the issue that the convolutional neural network (CNN) accelerator design on most FPGA ends fails to effectively use the sparsity and considering both bandwidth and energy consumption,two improved CNN calculation optimization strategies based on linear systolic array architecture are proposed.Firstly,convolution is transformed into matrix multiplication to take advantage of sparsity.Secondly,in order to solve the problem of large I/O demand in traditional parallel matrix multiplier,linear systolic array is used to improve the design.Finally,a CNN acceleration comparative analysis of the advantages and disadvantages between parallel matrix multiplier and two improved linear systolic arrays is presented.Theoretical proof and analysis show that compared with the parallel matrix multiplier,the two improved linear systolic arrays make full use of sparsity,and have the advantages of less energy consumption and less I/O bandwidth occupation.http://www.cjnis.com.cn/thesisDetails#10.11959/j.issn.2096-109x.2018100linear systolic arrayconvolutional neural networksparsityI/O bandwidthperformance analysis |
| spellingShingle | Qinrang LIU Chongyang LIU Jun ZHOU Xiaolong WANG Based on linear systolic array for convolutional neural network’s calculation optimization and performance analysis 网络与信息安全学报 linear systolic array convolutional neural network sparsity I/O bandwidth performance analysis |
| title | Based on linear systolic array for convolutional neural network’s calculation optimization and performance analysis |
| title_full | Based on linear systolic array for convolutional neural network’s calculation optimization and performance analysis |
| title_fullStr | Based on linear systolic array for convolutional neural network’s calculation optimization and performance analysis |
| title_full_unstemmed | Based on linear systolic array for convolutional neural network’s calculation optimization and performance analysis |
| title_short | Based on linear systolic array for convolutional neural network’s calculation optimization and performance analysis |
| title_sort | based on linear systolic array for convolutional neural network s calculation optimization and performance analysis |
| topic | linear systolic array convolutional neural network sparsity I/O bandwidth performance analysis |
| url | http://www.cjnis.com.cn/thesisDetails#10.11959/j.issn.2096-109x.2018100 |
| work_keys_str_mv | AT qinrangliu basedonlinearsystolicarrayforconvolutionalneuralnetworkscalculationoptimizationandperformanceanalysis AT chongyangliu basedonlinearsystolicarrayforconvolutionalneuralnetworkscalculationoptimizationandperformanceanalysis AT junzhou basedonlinearsystolicarrayforconvolutionalneuralnetworkscalculationoptimizationandperformanceanalysis AT xiaolongwang basedonlinearsystolicarrayforconvolutionalneuralnetworkscalculationoptimizationandperformanceanalysis |