臺灣博碩士論文加值系統

利用CUDA (Compute Unified Device Architecture)來讓GPU (Graphic Processor Unit)執行平行運算是這近幾年才興起的一項技術。早在過去，GPU就可以用於平行運算，但是礙於程式編寫上的不易，所以一直沒有被廣泛的應用。CUDA是C語言的延伸，改善了程式編寫上的障礙，再加上GPU核心技術的改良，IEEE浮點數的支援，和相對於超級電腦低成本的效應，其應用已逐漸拓展到不同領域。磁振頻譜是以非侵入性的特色，探測人體組織中代謝物質的濃度分布，可以輔助醫生在臨床上的診斷分析。而磁振頻譜影像則是將單一的磁振頻譜，變成多維度的磁振頻譜影像，所包含的磁振頻譜更多，所能提供的訊息也更多。CUDA在磁振影像處理被廣泛應用在影像的重建加速和影像品質的提升，但是在磁振頻譜這方面卻少有相關的應用。在本研究，主要是把CUDA應用在磁振頻譜這領域上，做磁振頻譜影像的前處理動作，加速磁振頻譜影像空間位置的解析。
在這份研究中，首先我們先使用一維、二維、三維，這三種不同維度，去亂數取得複數資料，並利用CUDA的快速傅立葉轉換去觀察他們的加速情形。最後我們將其應用對應使用到二維和三維的磁振頻譜影像上，去看看利用CUDA加速後的情形。在結果中，我們可以得到利用CUDA在一維到三維亂數取得的複數資料中，當資料量增加時對CUDA的加速會很有幫助。然後應用在磁振頻譜影像上，可以發現在使用CUDA去做磁振頻譜影像RAW檔案的產生時，應該避免資料搬移的次數，並且可知在資料量太小的情況下，對CUDA的運算是沒有幫助的。因此，要如何解決磁振頻譜影像的格式和CUDA的傅立葉轉換的對應關係，還有要如何減少資料搬移的次數，都會在本研究中做個討論。

Using Graphic Processor Unit (GPU) to process the parallel operation via Compute Unified Device Architecture (CUDA) is a new technology in recent years. In the past, the GPU has been used in parallel operation but it was not easy for programming so that it couldn’t be widely used in applications. CUDA is the newly-developed environment based on C language mainly for improving the complexity in programming with CUDA. The applications of GPU with CUDA has been expending to various fields gradually due to support of IEEE floating point as well as its lower cost in hardware while comparing to the super computers. Magnetic Resonance Spectroscopy (MRS) has the feature of non-invasive to probe the concentration distributed of metabolites in vivo. It can assist doctor in clinical diagnosis. The Magnetic Resonance Spectroscopy Imaging (MRSI) is imaging by many Signal Voxel Spectroscopy (SVS) to become multi-dimension MRS image. In MRSI, it can offer more information than SVS. CUDA are applied to MR image widely such as accelerating the image reconstruction and promoting the image quality, but in MRS it is seldom for the related application. In this paper, we using the CUDA to applied in MRS, the MRSI data pre-processing, to accelerate the spatial location in MRSI.

In this work, we firstly use random data with different dimensions: 1D (one-dimension), 2D and 3D to evaluate the performance of Fourier transformation by using CUDA. We also finally apply some GE 2D/3D MRSI data to see how the acceleration of using CUDA works. Our results show that the acceleration rate of Fastest Fourier Transform (FFT) with CUDA in 1D, 2D and 3D random data largely increases as the data size increases. In the experiment of 2D/3D MRSI data, we find that using CUDA for accelerating the MRSI RAW-file generating procedure would avoid the data moving times, and it is not good for CUDA 1D FFT with parallel architecture while too small data amount processing in kernel. Therefore, how to solve the relationship between MRSI data format with CUDA FFT library and how to decrease the data moving time will discuss in the study.

致謝 i
中文摘要 ii
Abstract iii
List of Figures vii
List of Tables ix
Chapter 1 Introduction 1
1.1 Background 1
1.2 Related Work 3
1.3 Motivation 5
1.4 Outline 6
Chapter 2 Materials and Methods 7
2.1 Compute Unified Device Architecture (CUDA) 7
2.1.1 Hardware 7
2.1.2 Thread Model 15
2.1.3 Memory Hierarch 20
2.1.4 CUDA Label 28
2.1.5 CUDA Compiler 31
2.2 The Fourier Transform 33
2.3 Experiment Design 36
2.3.1 Subjects and data format 36
2.3.2 Data processing 38
2.3.3 LCModel 48
Chapter 3 Results 50
3.1 Results of random data 50
3.2 Results of 2D/3D MRSI data 54
Chapter 4 Discussions and Conclusions 62
4.1 Discussions 62
4.2 Conclusions 65
References 66

1.Nickolls, J., et al., Scalable Parallel Programming with CUDA. Queue, 2008. 6(2): p. 40-53.
2.Garland, M., et al., Parallel Computing Experiences with CUDA. IEEE Micro, 2008. 28(4): p. 13-27.
3.G.Nishimura, D., Principles of Magnetic Resonance Imaging. 1 ed. 2010.
4.David D.Stark, W.G.B., MAGNETIC RESONANCE IMAGING. Vol. 1. 1999.
5.Nouha Salibi, M.A.B., Clinical MR Spectroscopy : First Principles. 1 ed. 1998.
6.Hongyu, G., D. Jianping, and H. Yanfa. GPU Acceleration of PROPELLER MRI Using CUDA. in Bioinformatics and Biomedical Engineering , 2009. ICBBE 2009. 3rd International Conference on. 2009.
7.Stone, S.S., et al., Accelerating advanced MRI reconstructions on GPUs. Journal of Parallel and Distributed Computing, 2008. 68(10): p. 1307-1318.
8.Gregerson, A., Implementing Fast MRI Gridding on GPUs via CUDA.
9.Schiwietz, T., et al. MR image reconstruction using the GPU. 2006. San Diego, CA, USA: SPIE.
10.Provencher, S. (2010) LCModel1 & LCMgui User''s Manual.
11.John Owens, U.D., GPU Architecture Owerview. SIGGRAPH, 2007.
12.Kayvon Fatahalian, M.H., GPU a Close Look. ACM Queue, 2008.
13.Dinh, M.T.D., GPUs-Graphics Processing Units. 2008.
14.Davis, C.E., Graphics Processing Unit Computation of Neural Networks. 2005.
15.N.Glaskowsky, P., NVIDIA''s Fermi: The First Complete GPU Computing Architecture. 2009.
16.NVIDIA, NVIDIA Programming Guide Version 2.1. 2008.
17.NVIDIA, NVIDIA CUDA C Programming - Best Practices Guide CUDA Toolkit 2.3. 2009.
18.David Kirk, W.-m.H., ed. KirkHwuTextbook. 2006-2008.
19.David Luebke, G.H., How GPUs Work. IEEE, 2008.
20.Zhiyi, Y., Z. Yating, and P. Yong. Parallel Image Processing Based on CUDA. in Computer Science and Software Engineering, 2008 International Conference on. 2008.
21.Oberman, E.L.S., NVIDIA GeForce 8800 Graphic Card.
22.Breitbart, J., Case studies on GPU usage and data structure design. 2008.
23.NVIDIA (2008) The CUDA compiler driver NVCC.
24.Graaf, R.A.d., In Vivo NMR Spectroscopy Principles and Techniques 2ed.
25.Moreland, K. and E. Angel, The FFT on a GPU, in Proceedings of the ACM SIGGRAPH/EUROGRAPHICS conference on Graphics hardware. 2003, Eurographics Association: San Diego, California. p. 112-119.
26.Govindaraju, N.K., et al., High performance discrete Fourier transforms on graphics processors, in Proceedings of the 2008 ACM/IEEE conference on Supercomputing. 2008, IEEE Press: Austin, Texas. p. 1-12.
27.Haynes, P.D. and M. Cote, Parallel fast Fourier transforms for electronic structure calculations. Computer Physics Communications, 2000. 130(1-2): p. 130-136.
28.NVIDIA, CUDA CUFFT Library. 2008.
29.FFTW homepage. Available from: http://www.fftw.org/.
30.M. Alley, rdgehdr. 2006, Radiological Sciences Laboratory , Standford University.