簡易檢索 / 詳目顯示
| 研究生： |
林柏頲 Lin, Po-Ting |
|---|---|
| 論文名稱： |
離子通道模擬計算的GPU平行預處理技巧 GPU-based parallel preconditioning techniques for ion channel simulations |
| 指導教授： |
陳人豪
Chen, Jen-Hao |
| 口試委員: |
劉晉良
Liu, Jinn-Liang 陳仁純 Chen, Ren-Chuen |
| 學位類別： |
碩士 Master |
| 系所名稱： |
理學院 - 計算與建模科學研究所 Institute of Computational and Modeling Science |
| 論文出版年： | 2019 |
| 畢業學年度： | 107 |
| 語文別： | 英文 |
| 論文頁數： | 20 |
| 中文關鍵詞： | 平行化 、CUSPARSE 、不完整LU分解 、超鬆弛迭代法 、預處理 、圖形處理器 |
| 外文關鍵詞： | parallelization, CUSPARSE, ILU, SSOR, precondition, GPU |
| 相關次數： | 點閱：3 下載：0 |
| 分享至: |
| 查詢本校圖書館目錄 查詢臺灣博碩士論文知識加值系統 勘誤回報 |
本篇論文開發了預處理 BiCGstab 算法的 GPU 平行化，該算法能夠減 少離子通道模擬運算中矩陣的條件數。 兩個使用的預處理是不完整 的 LU 分解(ILU)和對稱的連續過度鬆弛(SSOR)。 我們使用 CUSPARSE 中的子程序來精確求解 ILU 和 SSOR 預處理的下三角和上 三角線性系統。 此外，我們還開發了 GPU 內核來執行不精確的 SSOR 預處理。結果表明，使用預處理可以減少迭代次數，不精確的 SSOR 預處理在這些方法中具有最佳性能。
This thesis develops the GPU parallelization for the preconditioned BiCGstab algorithm which is able to reduce the condition numbers of matrices in ion channel simulations. Two used preconditioners are the incomplete LU factorizations(ILU) and symmetric successive over-relaxation(SSOR). We use the subroutines in CUS- PARSE to exactly solve the lower and upper triangular linear systems for ILU and SSOR preconditioner. Moreover,we also develop a GPU kernel to perform the inexact SSOR preconditioner.The results show that the use of preconditioner can reduce the number of iteration, and the inexact SSOR preconditioner has the best performance among these methods.
[1] NVIDIA (2018) NVIDIA CUDA C programming guide 9.1.
[2] NVIDIA (2018) NVIDIA CUSPARSE LIBRARY user guide 9.1.
[3] Jen-Hao Chen, Ren-Chuen Chen, and Jinn-Liang Liu. A gpu poisson- fermi solver for ion channel simulations. Computer Physics Communications, 229(2):99–105, 2017.
[4] Burden Faires. Numerical methods, volume 4. Cengage Learning, 2012.
[5] Liu J.-L. Numerical methods for the poisson–fermi equation in electrolytes.
J. Comput. Phys., 247, 2013.
[6] Ruipeng Li and Yousef Saad. Gpu-accelerated preconditioned iterative linear
solvers. The Journal of Supercomputing, 63(2):443–466, 2013.
[7] Youcef Saad and Martin H Schultz. Gmres: A generalized minimal resid- ual algorithm for solving nonsymmetric linear systems. SIAM Journal on scientific and statistical computing, 7(3):856–869, 1986.
[8] Yousef Saad. Iterative methods for sparse linear systems, volume 2. SIAM, 2000.
[9] Hartwig Anzt Stanimire Tomov, William Sawyer Piotr Luszczek, and Jack Dongarra. Acceleration of gpu-based krylov solvers via data transfer reduc- tion. The International Journal of High Performance Computing Applica- tions, 29(3):366–383, 2015.
[10] Henk A Van der Vorst. Bi-CGSTAB: A fast and smoothly converging variant of Bi-CG for the solution of nonsymmetric linear systems. SIAM Journal on scientific and Statistical Computing, 13(2):631–644, 1992.
19
[11] Xuhong Tian Yan Chen, Zhangxin Chen Hui Liu, Wenyuan Liao Bo Yang, Ruijian He Peng Zhang, and Min Yang. Parallel ilu preconditioners in gpu computation. Soft Computing, 22(24):8187–8205, 2018.