并联超低功耗系统中的离散余弦变换硬件加速器

Alen Duspara, M. Kovač, H. Mlinaric
{"title":"并联超低功耗系统中的离散余弦变换硬件加速器","authors":"Alen Duspara, M. Kovač, H. Mlinaric","doi":"10.1109/ELMAR52657.2021.9550946","DOIUrl":null,"url":null,"abstract":"In this paper, the architecture of a fully pipelined discrete cosine transform ( DCT) hardware accelerator for a JPEG encoder is proposed. The integration of the accelerator into the parallel ultra-low power (PULP) platform is also demonstrated. The accelerator architecture is divided into two one-dimensional transform cores with one transpose buffer between them . With the designed accelerator, it is possible to calculate one 2D DCT operation in 32 cycles with a latency of 80 cycles. The JPEG DCT hardware accelerator is integrated into the PULP cluster as a separate processing element (PE) and successfully implemented on the Xilinx ZC706 evaluation board. The accelerator can achieve the performance of up to 1.78M transformations per second working on the clock frequency of 57 MHz.","PeriodicalId":410503,"journal":{"name":"2021 International Symposium ELMAR","volume":"11 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2021-09-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"2","resultStr":"{\"title\":\"Discrete Cosine Transform Hardware Accelerator in Parallel Ultra-low Power System\",\"authors\":\"Alen Duspara, M. Kovač, H. Mlinaric\",\"doi\":\"10.1109/ELMAR52657.2021.9550946\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"In this paper, the architecture of a fully pipelined discrete cosine transform ( DCT) hardware accelerator for a JPEG encoder is proposed. The integration of the accelerator into the parallel ultra-low power (PULP) platform is also demonstrated. The accelerator architecture is divided into two one-dimensional transform cores with one transpose buffer between them . With the designed accelerator, it is possible to calculate one 2D DCT operation in 32 cycles with a latency of 80 cycles. The JPEG DCT hardware accelerator is integrated into the PULP cluster as a separate processing element (PE) and successfully implemented on the Xilinx ZC706 evaluation board. The accelerator can achieve the performance of up to 1.78M transformations per second working on the clock frequency of 57 MHz.\",\"PeriodicalId\":410503,\"journal\":{\"name\":\"2021 International Symposium ELMAR\",\"volume\":\"11 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2021-09-13\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"2\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"2021 International Symposium ELMAR\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1109/ELMAR52657.2021.9550946\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"2021 International Symposium ELMAR","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ELMAR52657.2021.9550946","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 2

摘要

提出了一种用于JPEG编码器的全流水线离散余弦变换(DCT)硬件加速器的结构。并演示了将加速器集成到并联超低功耗(PULP)平台中的方法。加速器结构分为两个一维变换核心,中间有一个转置缓冲器。利用设计的加速器,可以在32个周期内计算一个2D DCT操作,延迟为80个周期。JPEG DCT硬件加速器作为单独的处理元件(PE)集成到PULP集群中,并成功地在Xilinx ZC706评估板上实现。在57mhz的时钟频率下,加速器可以达到每秒1.78M变换的性能。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
Discrete Cosine Transform Hardware Accelerator in Parallel Ultra-low Power System
In this paper, the architecture of a fully pipelined discrete cosine transform ( DCT) hardware accelerator for a JPEG encoder is proposed. The integration of the accelerator into the parallel ultra-low power (PULP) platform is also demonstrated. The accelerator architecture is divided into two one-dimensional transform cores with one transpose buffer between them . With the designed accelerator, it is possible to calculate one 2D DCT operation in 32 cycles with a latency of 80 cycles. The JPEG DCT hardware accelerator is integrated into the PULP cluster as a separate processing element (PE) and successfully implemented on the Xilinx ZC706 evaluation board. The accelerator can achieve the performance of up to 1.78M transformations per second working on the clock frequency of 57 MHz.
求助全文
通过发布文献求助,成功后即可免费获取论文全文。 去求助
来源期刊
自引率
0.00%
发文量
0
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:604180095
Book学术官方微信