Author: |
陳政宏 Jeng-Hung Chen |
---|---|
Thesis Title: |
ARM Cortex-A8 NEON指令集與架構研究 ARM Cortex-A8 NEON Instruction Set and Architecture Study |
Advisor: |
許孟超
Mon-Chau Shie |
Committee: |
阮聖彰
none 吳晉賢 none 林昌鴻 none |
Degree: |
碩士 Master |
Department: |
電資學院 - 電子工程系 Department of Electronic and Computer Engineering |
Thesis Publication Year: | 2011 |
Graduation Academic Year: | 99 |
Language: | 中文 |
Pages: | 74 |
Keywords (in Chinese): | ARM 、DSP 、Cortex A 、SIMD 、NEON 、Tool Chain 、Boot-loader 、OS kernel |
Keywords (in other languages): | ARM, DSP, Cortex A, SIMD, NEON, Tool Chain, Boot-loader, OS kernel |
Reference times: | Clicks: 999 Downloads: 0 |
Share: |
School Collection Retrieve National Library Collection Retrieve Error Report |
ARM處理器在市場上快速發展,毫無疑問地已經成為嵌入式系統最普及以及最重要的處理器。晶片設計廠商相繼以ARM處理器為核心,再據以各自的需求增加設計硬體加速器、DSP等週邊進行差異化;以ARM為核心的SOC已成為市場的主流。
ARM Cortex是新一代V7的ARM處理器,別於之前V4/V5/V6的ARM7、 ARM9、ARM11,具有較高效能、低耗電、提供不同應用需求的全系列處理器。
多媒體廣泛應用於嵌入式系統,然而執行多媒體運算的複雜度對處理器而言仍是沉重的負擔;如何達到高效能及低耗電是重要的課題。
本論文研究ARM Cortex A8核心中的新增技術SIMD(NEON), Single Instruction Multiple Data;並介紹如何進行效能最佳化分析及應用SIMD(NEON)來增進多媒體執行效能且節省耗電。
ARM processors advance steadily on the market. Undoubtedly it has become the most popular and most important processor in embedded systems. More and more chip makers embrace ARM processor core and integrate hardware accelerator, DSP, and other peripherals according to individual needs to differentiate their products. ARM core SOC has become mainstream standard for SOC.
ARM Cortex is the newest generation ARM to replace previous ARM7, ARM9, and ARM11 of V4/V5/V6.It comes with high efficiency and low power processors such as A, R, M profiles. They provide various application needs of the whole series for all-kind systems. Multimedia has widely used in embedded environment. But implementation of multimedia computing complexity for the processor is still a heavy burden. How to achieve high performance and low power consumption is an important topic. The ARM A profile since A8, A9 ... have now provide SIMD NEON instruction set to support the portable and low power multimedia software.
In this thesis we research the new technology supported by ARM Cortex-A8 core – SIMD (NEON), Single Instruction Multiple Data. We find way to achieve performance via compiler optimization option and program scheme. The analysis shows that our method can boost multimedia program speed-up around 2-4 folds by applying SIMD (NEON).
[1] Steve Furber, “ARM system-on-chip architecture”, Addison-Wesley Professional; 2 edition (August 14, 2000)
[2] Sloss, Symes & Wright – “ARM System Developer's Guide”, Morgan Kaufmann; 1 edition (April 8, 2004)
[3] Hord; R. Michael Reston, VA,Parallel Supercomputing in SIMD Architectures, CRC Press; 1 edition (April 30, 1990)
[4] Cypher, Robert, Sanz, Jorge L.C., The SIMD Model of Parallel Computation, Springer; 1 edition (February 18, 1998)
[5] Graham R Wilson, Embedded Systems and Computer Architecture, Newnes; 1st edition (January 22, 2002)
[6] David Seal, ARM Architecture Reference Manual, Addison-Wesley Professional; 2 edition (January 6, 2001)
[7] Johan Dams, Optimizing Embedded Software – A Look at the NEON SIMD unit in the ARM Cortex Family of Processors, PAPER, Vaasa University of Applied Sciences, Vaasa - Finland
[8] Dmitry Melnik1, Andrey Belevantsev1, Dmitry Plotnikov1, and Semun Lee2, A case study: optimizing GCC on ARM for performance of libevas rasterization library, PAPER, Institute for System Programming, Russian Academy of Sciences
[9] DDI0457A_ARMv7_Performance_Monitors_v2, [Online]. Available: http://infocenter.arm.com/help/index.jsp?topic=/com.arm.doc.set.devlib/index.html
[10] DUI0471D_developing_for_arm_processors.pdf, [Online]. Available: http://infocenter.arm.com/help/index.jsp?topic=/com.arm.doc.set.devlib/index.html
[11] DUI0472D_using_the_arm_compiler.pdf, [Online]. Available: http://infocenter.arm.com/help/index.jsp?topic=/com.arm.doc.set.devlib/index.html
[12] DUI0473D_using_the_arm_assembler.pdf, [Online]. Available: http://infocenter.arm.com/help/index.jsp?topic=/com.arm.doc.set.devlib/index.html
[13] DUI0474D_using_the_arm_linker.pdf, [Online]. Available: http://infocenter.arm.com/help/index.jsp?topic=/com.arm.doc.set.devlib/index.html
[14] DUI0489D_arm_assembler_reference.pdf, [Online]. Available: http://infocenter.arm.com/help/index.jsp?topic=/com.arm.doc.set.devlib/index.html
[15] DUI0491D_arm_compiler_reference.pdf, [Online]. Available: http://infocenter.arm.com/help/index.jsp?topic=/com.arm.doc.set.devlib/index.html
[16] DDI0409F_cortex_a9_neon_mpe_r2p2_trm.pdf, [Online]. Available: http://infocenter.arm.com/help/index.jsp?topic=/com.arm.doc.set.devlib/index.html
[17] DHT0002A_introducing_neon.pdf, [Online]. Available: http://infocenter.arm.com/help/index.jsp?topic=/com.arm.doc.set.devlib/index.html
[18] DHT0004A_neon_support_in_compilation_tools.pdf, [Online]. Available: http://infocenter.arm.com/help/index.jsp?topic=/com.arm.doc.set.devlib/index.html