This goal of this semester is to study the parallel numerical algorithms for 
<br>solving large linear systems.  We will focus on iterative linear solvers and its 
<br>parallelization, especially the algorithms for multi-GPU (graphics processing 
<br>unit).  Students who take this coarse can learn various numerical linear solvers 
<br>and their implementations on massively parallel environments.  
<br>
<br>Textbook: 
<br>Iterative methods for sparse linear systems (2nd edition) by Yousef Saad 2003.
<br>http://www-users.cs.umn.edu/~saad/IterMethBook_2ndEd.pdf
<br>http://ec-securehost.com/SIAM/ot82.html  
<br>
<br>Topics:
<br>1. Basic linear algebra review: LU, Cholesky, QR, Eigen-problems
<br>2. Introduction to GPU programming
<br>3. Stationary iterative linear solver
<br>4. Krylov subspace methods
<br>5. Lanczos based methods
<br>6. Pre-conditioning techniques
<br>7. Parallel preconditioner
<br>8. Algebraic multigrid method 
<br>
<br>Grading:
<br>1. 3 Programming assignments on GPU: 45%
<br>2. Class note: 25%
<br>3. Final project: 30% 
<br>