Thank you Barry for your enlightenment. I'll just continue to use
BoomerAMG for the poisson eqn. I'll also check up on FFTW. Last time, I
recalled that there seemed to be some restrictions for FFT on solving
poisson eqn. It seems that the grids must be constant in at least 1
dimension. I
Sorry Barry, I just would like to confirm that as long as it's a
constant constant coefficient Poisson eqn with Neumann or Dirchelet
boundary conditions, I can use FFT. It doesn't matter if the grids are
uniform or not. Is that correct? Thanks.
Barry Smith wrote:
On Feb 5, 2008, at 8:48 PM,
Whoops, actually the grids in each direction would need to be
uniform.
Barry
On Feb 5, 2008, at 9:39 PM, Ben Tay wrote:
Sorry Barry, I just would like to confirm that as long as it's a
constant constant coefficient Poisson eqn with Neumann or Dirchelet
boundary conditions, I
On Tue, 2008-02-05 at 07:31 -0600, Matthew Knepley wrote:
On Feb 5, 2008 3:26 AM, Erlend Pedersen :.
erlend.pedersen at holberger.com wrote:
On Sun, 2008-02-03 at 19:59 -0600, Matthew Knepley wrote:
On Feb 1, 2008 5:54 AM, Erlend Pedersen :.
erlend.pedersen at holberger.com wrote:
I
Well, after taking into accout Barry's comments, you have have the
following choices.
* You can use a direct method based on LU factorization using
'-ksp_type preonly -pc_type lu' . This way, PETSc will compute the LU
factors the fist time they are needed; after that, every call to
KSPSolve will