-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Hi all,
I have multi-field operators that have the same layout as given by the
matrices created by DMComposite and would like to be able to switch
between a MatNest representation (if I'm using fieldsplit PCs) or AIJ
(if not). This works fine if the
I modified the DMNetwork example to include the new DM with the modified
section. It has the same problems. Please find attached the code to this
email.
Thanks
On Tue, Feb 24, 2015 at 6:49 PM, Matthew Knepley knep...@gmail.com wrote:
On Tue, Feb 24, 2015 at 6:42 PM, Miguel Angel Salazar de
On Wed, Feb 25, 2015 at 9:59 AM, Abhyankar, Shrirang G. abhy...@mcs.anl.gov
wrote:
Miguel,
I'm a bit tied up today. I'll try to debug this issue tomorrow and get
back to you.
The problem is the way that DMNetwork is using the default section. When
DMCreateGlobalVec() is called,
it uses
Miguel,
I'm a bit tied up today. I'll try to debug this issue tomorrow and get back
to you.
Thanks,
Shri
From: Miguel Angel Salazar de Troya
salazardetr...@gmail.commailto:salazardetr...@gmail.com
Date: Wed, 25 Feb 2015 08:48:10 -0600
To: Matthew Knepley
Hi Dmitry,
Thanks so much, that's great! I've updated the pull request now to include
your patch:
https://github.com/dknez/libmesh/commit/247cc12f7536a5f848fd5c63765d97af307d9fa7
We can update the code later, once MatReset() is available in PETSc.
Regarding the convergence issues: I think the
The RAP is often more than 50% of PCSetUp, so this might not save much.
Hee,hee in some of my recent runs of GAMG the RAP was less than 25% of
the time. Thus skipping the other portions could really pay off.*
This percentage is very problem dependant. 3D elasticity is higher, 2D
Matt,
Thanks for debugging. I'll try to fix this tomorrow.
Shri
From: Matthew Knepley knep...@gmail.commailto:knep...@gmail.com
Date: Wed, 25 Feb 2015 10:12:43 -0600
To: Shri abhy...@mcs.anl.govmailto:abhy...@mcs.anl.gov
Cc: Miguel Angel Salazar de Troya
Now it works, thanks a lot!
Miguel
On Wed, Feb 25, 2015 at 10:12 AM, Matthew Knepley knep...@gmail.com wrote:
On Wed, Feb 25, 2015 at 9:59 AM, Abhyankar, Shrirang G.
abhy...@mcs.anl.gov wrote:
Miguel,
I'm a bit tied up today. I'll try to debug this issue tomorrow and get
back to you.
I want to do 3 steps of gauss seidel from a Mat A to another Mat B. Is there a
way to do this?
I mean, what I can think of is to get the column vectors of B by
ierr = MatGetColumnVector(B,v,col);CHKERRQ(ierr);
and apply Gauss seidel from A to v:
ierr = MatSOR(A, v, 1,
Let me try to understand what you wish to do. You have a sequential matrix A
and a matrix B and you wish to compute the (dense) matrix C where each column
of C is obtained by running three iterations of SOR (using the matrix A) with
the corresponding column of B as the right hand side for
Thanks for the report. I ran the example with those options under Linux with
valgrind and found no memory leaks, I cannot run valgrind on my Mac. I suspect
the issue is related to some internal library memory problems on the Apple and
is not in the PETSc library or that example.
Barry
Thank you, Barry. In fact, my matrix B is sparse. Is it possible to drop the
terms less than a certain threshold from C, so that C can be sparse?
Another question, there are pc_sor_its and pc_sor_lits in MatSOR. What is local
its?
Best,
Hui
From:
On Feb 25, 2015, at 3:02 PM, Sun, Hui hus...@ucsd.edu wrote:
Thank you, Barry. In fact, my matrix B is sparse. Is it possible to drop the
terms less than a certain threshold from C, so that C can be sparse?
You can do that but it becomes much more complicated and expensive,
especially
Barry Smith bsm...@mcs.anl.gov writes:
Thanks for the report. I ran the example with those options under
Linux with valgrind and found no memory leaks, I cannot run
valgrind on my Mac.
When are you going to get a real operating system? In the mean time,
several people have reported
On Wed, Feb 25, 2015 at 4:09 PM, Jed Brown j...@jedbrown.org wrote:
Barry Smith bsm...@mcs.anl.gov writes:
Thanks for the report. I ran the example with those options under
Linux with valgrind and found no memory leaks, I cannot run
valgrind on my Mac.
When are you going to get
Matthew Knepley knep...@gmail.com writes:
It works for me. I can't wait for Jed's suffering when a kid wants to
watch PBS Kids on his custom Debian build.
Touché, but in fact, PBS Kids plays great on my Linux box.
signature.asc
Description: PGP signature
Thank you Barry. It is a step in forming my PC matrix for the schur complement.
I think I will try something else.
Best,
Hui
From: Barry Smith [bsm...@mcs.anl.gov]
Sent: Wednesday, February 25, 2015 1:12 PM
To: Sun, Hui
Cc: petsc-users@mcs.anl.gov
By the way, how do you add up two matrices? Assume A and B are of same size,
both sparse and of type Matmpi. I want to perform this operation: C=A+B. What
should I do? I haven't found anything related in the user manual.
Best,
Hui
From: Sun, Hui
On Wed, Feb 25, 2015 at 6:36 PM, Sun, Hui hus...@ucsd.edu wrote:
By the way, how do you add up two matrices? Assume A and B are of same
size, both sparse and of type Matmpi. I want to perform this operation:
C=A+B. What should I do? I haven't found anything related in the user
manual.
On Feb 25, 2015, at 4:09 PM, Jed Brown j...@jedbrown.org wrote:
Barry Smith bsm...@mcs.anl.gov writes:
Thanks for the report. I ran the example with those options under
Linux with valgrind and found no memory leaks, I cannot run
valgrind on my Mac.
When are you going to get a
Thank you Matt.
Hui
From: Matthew Knepley [knep...@gmail.com]
Sent: Wednesday, February 25, 2015 5:00 PM
To: Sun, Hui
Cc: Barry Smith; petsc-users@mcs.anl.gov
Subject: Re: [petsc-users] about MATSOR
On Wed, Feb 25, 2015 at 6:36 PM, Sun, Hui
21 matches
Mail list logo