Design and Implementation of DSP-Based Configurable Heterogeneous Embedded Platform
碩士 === 國立東華大學 === 電機工程學系 === 93 === Since the reason that the demand of the embedded equipment with real time digital signal processing is increasing day by day, and various kinds of complication and requirements of specific algorithms has doubled thereupon too, it is almost impossible for us to ach...
Main Authors: | , |
---|---|
Other Authors: | |
Format: | Others |
Language: | zh-TW |
Published: |
2005
|
Online Access: | http://ndltd.ncl.edu.tw/handle/98581208810397418935 |
id |
ndltd-TW-093NDHU5442031 |
---|---|
record_format |
oai_dc |
spelling |
ndltd-TW-093NDHU54420312016-06-06T04:11:19Z http://ndltd.ncl.edu.tw/handle/98581208810397418935 Design and Implementation of DSP-Based Configurable Heterogeneous Embedded Platform DSP為基礎之可組態異質性嵌入式平台設計與實作 Cheng-Tsung Lin 林正宗 碩士 國立東華大學 電機工程學系 93 Since the reason that the demand of the embedded equipment with real time digital signal processing is increasing day by day, and various kinds of complication and requirements of specific algorithms has doubled thereupon too, it is almost impossible for us to achieve the aim of real time merely by manipulating the data processing, information transmission as well as the realization of the algorithm of the whole system with the single processor. One of the present main solutions is to allocate the complicated number value operation and system control to each individual distributed system by utilizing the multi-processor structure in order to achieve the parallel computing purpose, that is, to improve system performance. However, when carrying on the system integration with the multi-processor structure we might encounter some comparatively complicated problems, for instance: Distribution of workload, arbitration between each other, and data-synchronizations. In this thesis we have first utilized dual-DSP key structure to integrate heterogeneous distributed system, so as to realize the idea of Parallel Computing, and then explain how each component cooperates, communicates and transmits information between each other under this kind of structure. The thesis is in fact based on the application of TMS320C6000 DSK produced by Texas Instruments. We design and implement a programmable multi-functional add-on module integrating heterogeneous units such as DSPs, FPGA and Dual Port RAM. We employ TMS320F2812 DSP, on the one hand, as an instrument to control peripheral I/O device in order to gather the needed information quickly. On the other hand, TMS320C6000 DSP is responsible for the management of the data and signals so that we can have system control and data processing work efficiently at the same time. As for Dual Port RAM, it is applied as a communicator between the dual DSPs, providing the function of data buffer and temporary storage to make the anticipated smooth collaboration between two operation cores possible. At last, due to its characteristics of being programmable we use FPGA to perform high-speed DSP algorithms, for example: Fast Fourier Transform (FFT), Discrete Cosine Transform (DCT) etc.. It is the visualized Unified Modeling Language (UML) that I employ as an appliance of system analysis and design. Through the application of UML, we can easily accomplish the work such as demand analysis, system structure, the function of each component and programming, and sketch the contour of the blueprint of operation of the whole system. The carefully designed programmable multi-functional add-on module in this thesis has combined the advantages of each heterogeneous unit. Meanwhile, using parallel computing mechanism of Dual DSPs offers the systematic developer the access to rapid development and to verify various types of specific DSP application. Tsung-Ying Sun 孫宗瀛 2005 學位論文 ; thesis 114 zh-TW |
collection |
NDLTD |
language |
zh-TW |
format |
Others
|
sources |
NDLTD |
description |
碩士 === 國立東華大學 === 電機工程學系 === 93 === Since the reason that the demand of the embedded equipment with real time digital signal processing is increasing day by day, and various kinds of complication and requirements of specific algorithms has doubled thereupon too, it is almost impossible for us to achieve the aim of real time merely by manipulating the data processing, information transmission as well as the realization of the algorithm of the whole system with the single processor. One of the present main solutions is to allocate the complicated number value operation and system control to each individual distributed system by utilizing the multi-processor structure in order to achieve the parallel computing purpose, that is, to improve system performance. However, when carrying on the system integration with the multi-processor structure we might encounter some comparatively complicated problems, for instance: Distribution of workload, arbitration between each other, and data-synchronizations.
In this thesis we have first utilized dual-DSP key structure to integrate heterogeneous distributed system, so as to realize the idea of Parallel Computing, and then explain how each component cooperates, communicates and transmits information between each other under this kind of structure. The thesis is in fact based on the application of TMS320C6000 DSK produced by Texas Instruments. We design and implement a programmable multi-functional add-on module integrating heterogeneous units such as DSPs, FPGA and Dual Port RAM. We employ TMS320F2812 DSP, on the one hand, as an instrument to control peripheral I/O device in order to gather the needed information quickly. On the other hand, TMS320C6000 DSP is responsible for the management of the data and signals so that we can have system control and data processing work efficiently at the same time. As for Dual Port RAM, it is applied as a communicator between the dual DSPs, providing the function of data buffer and temporary storage to make the anticipated smooth collaboration between two operation cores possible. At last, due to its characteristics of being programmable we use FPGA to perform high-speed DSP algorithms, for example: Fast Fourier Transform (FFT), Discrete Cosine Transform (DCT) etc..
It is the visualized Unified Modeling Language (UML) that I employ as an appliance of system analysis and design. Through the application of UML, we can easily accomplish the work such as demand analysis, system structure, the function of each component and programming, and sketch the contour of the blueprint of operation of the whole system.
The carefully designed programmable multi-functional add-on module in this thesis has combined the advantages of each heterogeneous unit. Meanwhile, using parallel computing mechanism of Dual DSPs offers the systematic developer the access to rapid development and to verify various types of specific DSP application.
|
author2 |
Tsung-Ying Sun |
author_facet |
Tsung-Ying Sun Cheng-Tsung Lin 林正宗 |
author |
Cheng-Tsung Lin 林正宗 |
spellingShingle |
Cheng-Tsung Lin 林正宗 Design and Implementation of DSP-Based Configurable Heterogeneous Embedded Platform |
author_sort |
Cheng-Tsung Lin |
title |
Design and Implementation of DSP-Based Configurable Heterogeneous Embedded Platform |
title_short |
Design and Implementation of DSP-Based Configurable Heterogeneous Embedded Platform |
title_full |
Design and Implementation of DSP-Based Configurable Heterogeneous Embedded Platform |
title_fullStr |
Design and Implementation of DSP-Based Configurable Heterogeneous Embedded Platform |
title_full_unstemmed |
Design and Implementation of DSP-Based Configurable Heterogeneous Embedded Platform |
title_sort |
design and implementation of dsp-based configurable heterogeneous embedded platform |
publishDate |
2005 |
url |
http://ndltd.ncl.edu.tw/handle/98581208810397418935 |
work_keys_str_mv |
AT chengtsunglin designandimplementationofdspbasedconfigurableheterogeneousembeddedplatform AT línzhèngzōng designandimplementationofdspbasedconfigurableheterogeneousembeddedplatform AT chengtsunglin dspwèijīchǔzhīkězǔtàiyìzhìxìngqiànrùshìpíngtáishèjìyǔshízuò AT línzhèngzōng dspwèijīchǔzhīkězǔtàiyìzhìxìngqiànrùshìpíngtáishèjìyǔshízuò |
_version_ |
1718295963487961088 |