t****n
发帖数: 39 | 1 Any one has experience of solving a linear equation system with
block tridiagonal matrix? I really need a fast solver. Currently
I'm using Gauss elimination method. |
c*******e
发帖数: 8624 | 2 if spd, cg ;
otherwise, gmres.
I heard there is a new method found recently good for general linear systems
which is better than gmres, no idea about details.
【在 t****n 的大作中提到】
: Any one has experience of solving a linear equation system with
: block tridiagonal matrix? I really need a fast solver. Currently
: I'm using Gauss elimination method.
|
b*****y
发帖数: 163 | 3 For direct methods, platform optimized BLAS and LAPACK are
very fast.
【在 t****n 的大作中提到】
: Any one has experience of solving a linear equation system with
: block tridiagonal matrix? I really need a fast solver. Currently
: I'm using Gauss elimination method.
|
t****n
发帖数: 39 | 4 Well, I wrote a code myself, with a house-cooked matrix manipulation
subroutine. Recently I tried the Intel Math Kernel's BLAS and LAPACK
souboutine. Unfortunately, no gain.
【在 b*****y 的大作中提到】
: For direct methods, platform optimized BLAS and LAPACK are
: very fast.
|
t******t
发帖数: 13 | 5 If the dimension is really really big, you might try
some sparse solver. I use the SparsKit here:
http://www-users.cs.umn.edu/~saad/
in my research and it suits my needs.
【在 t****n 的大作中提到】
: Any one has experience of solving a linear equation system with
: block tridiagonal matrix? I really need a fast solver. Currently
: I'm using Gauss elimination method.
|
