On Dec 29, 2011, at 8:24 PM, Dave Nystrom wrote:
> Mark F. Adams writes:
>> On Dec 29, 2011, at 12:51 PM, Dave Nystrom wrote:
>>
>>> Jed Brown writes:
On Thu, Dec 29, 2011 at 11:15, Dave Nystrom
wrote:
> I have tried out -ksp_type cg -pc_type gamg -pc_gamg_type sa on my pro
Mark F. Adams writes:
> On Dec 29, 2011, at 12:51 PM, Dave Nystrom wrote:
>
> > Jed Brown writes:
> >> On Thu, Dec 29, 2011 at 11:15, Dave Nystrom
> >> wrote:
> >>
> >>> I have tried out -ksp_type cg -pc_type gamg -pc_gamg_type sa on my
> >>> problem
> >>> and am encouraged enough wit
On Dec 29, 2011, at 12:51 PM, Dave Nystrom wrote:
> Jed Brown writes:
>> On Thu, Dec 29, 2011 at 11:15, Dave Nystrom wrote:
>>
>>> I have tried out -ksp_type cg -pc_type gamg -pc_gamg_type sa on my problem
>>> and am encouraged enough with the results that I would like to try taking
>>> the
>>>
First you need to set the block size for the matrix as I said before.
Then if this is a vector Laplacian like operator then the code can construct
the null space with the node/cell coordinates, which is required for the method
to be complete for a vector Lapalcian. Here is a 3D example:
ie
On Thu, Dec 29, 2011 at 11:51, Dave Nystrom
wrote:
> Both although the most challenging problem right now is one for a Hall
> matrix
> system that is block banded with 2x2 blocks.
>
What is the null space of this operator.
> Right now, I am just inputting
> the matrix into a petsc seqaij matri
On Thu, Dec 29, 2011 at 11:15, Dave Nystrom wrote:
> I have tried out -ksp_type cg -pc_type gamg -pc_gamg_type sa on my problem
> and am encouraged enough with the results that I would like to try taking
> the
> next step with using gamg. Could you provide some advice on how to do
> that?
> I'm
Jed Brown writes:
> On Thu, Dec 29, 2011 at 11:15, Dave Nystrom wrote:
>
> > I have tried out -ksp_type cg -pc_type gamg -pc_gamg_type sa on my problem
> > and am encouraged enough with the results that I would like to try taking
> > the
> > next step with using gamg. Could you provide som
Hi Mark,
I have tried out -ksp_type cg -pc_type gamg -pc_gamg_type sa on my problem
and am encouraged enough with the results that I would like to try taking the
next step with using gamg. Could you provide some advice on how to do that?
I'm not sure how to provide the null space info you say is
On Dec 23, 2011, at 8:36 PM, Dave Nystrom wrote:
> Mark F. Adams writes:
>> It sounds like you have a symmetric positive definite systems like du/dt -
>> div(alpha(x) grad)u. The du/dt term makes the systems easier to solve.
>> I'm guessing your hard system does not have this mass term and so is
Mark F. Adams writes:
> It sounds like you have a symmetric positive definite systems like du/dt -
> div(alpha(x) grad)u. The du/dt term makes the systems easier to solve.
> I'm guessing your hard system does not have this mass term and so is
> purely elliptic. Multigrid is well suited for th
On Dec 20, 2011, at 10:33 AM, Dave Nystrom wrote:
> Hi Mark,
>
> I would like to try GAMG on some of my linear solves. Could you suggest how
> to get started? Is it more complicated than something like:
>
> -ksp_type cg -pc_type gamg
This is a good start. for scalar SPD problems '-pc_gamg_t
On Tue, Dec 20, 2011 at 09:33, Dave Nystrom wrote:
> Hi Mark,
>
> I would like to try GAMG on some of my linear solves. Could you suggest
> how
> to get started? Is it more complicated than something like:
>
> -ksp_type cg -pc_type gamg
>
Sure, start with this.
>
> I'm guessing I should firs
Hi Mark,
I would like to try GAMG on some of my linear solves. Could you suggest how
to get started? Is it more complicated than something like:
-ksp_type cg -pc_type gamg
I'm guessing I should first try it on one of my easier linear solves. I have
5 of them that would have a block size of 1.
On Dec 2, 2011, at 6:06 PM, Dave Nystrom wrote:
> Mark F. Adams writes:
>> It sounds like you have a symmetric positive definite systems like du/dt -
>> div(alpha(x) grad)u. The du/dt term makes the systems easier to solve.
>> I'm guessing your hard system does not have this mass term and so is
On Fri, Dec 2, 2011 at 5:46 PM, Dave Nystrom
wrote:
> Matthew Knepley writes:
> > On Wed, Nov 30, 2011 at 12:41 AM, Dave Nystrom >wrote:
> >
> > > I have a linear system in a code that I have interfaced to petsc that
> is
> > > taking about 80 percent of the run time per timestep. This linea
Hi Barry,
Barry Smith writes:
> Dave,
>
> Does this come from a structured 2d grid? If so, then in addition to
> algebraic multigrid you could consider geometric multigrid which could
> work very well.
Yes, this problem comes from spatial discretization on a uniform structured
2d grid. So
Mark F. Adams writes:
> It sounds like you have a symmetric positive definite systems like du/dt -
> div(alpha(x) grad)u. The du/dt term makes the systems easier to solve.
> I'm guessing your hard system does not have this mass term and so is
> purely elliptic. Multigrid is well suited for th
Matthew Knepley writes:
> On Wed, Nov 30, 2011 at 12:41 AM, Dave Nystrom comcast.net>wrote:
>
> > I have a linear system in a code that I have interfaced to petsc that is
> > taking about 80 percent of the run time per timestep. This linear system
> > is a symmetric block banded matrix wher
Dave,
Does this come from a structured 2d grid? If so, then in addition to
algebraic multigrid you could consider geometric multigrid which could work
very well.
Do you want to run in parallel, how are you generating the matrix and
managing the mesh, if you use the DMDA object it m
It sounds like you have a symmetric positive definite systems like du/dt -
div(alpha(x) grad)u. The du/dt term makes the systems easier to solve. I'm
guessing your hard system does not have this mass term and so is purely
elliptic. Multigrid is well suited for this type of problem, but the ve
On Wed, Nov 30, 2011 at 12:41 AM, Dave Nystrom wrote:
> I have a linear system in a code that I have interfaced to petsc that is
> taking about 80 percent of the run time per timestep. This linear system
> is
> a symmetric block banded matrix where the blocks are 2x2. The matrix looks
> as follo
I have a linear system in a code that I have interfaced to petsc that is
taking about 80 percent of the run time per timestep. This linear system is
a symmetric block banded matrix where the blocks are 2x2. The matrix looks
as follows:
1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0
1X X
22 matches
Mail list logo