This code puts the error in x:
call VecAXPY(x,neg_one,u,ierr)
I suspect you printed these numbers after this statement.
On Mon, Jan 14, 2019 at 8:10 PM Maahi Talukder via petsc-users <
petsc-users@mcs.anl.gov> wrote:
> Hello all,
>
> I complied and run the example *ex2f.F90* located in
>
The memory requested is an insane number. You may need to use 64 bit
integers.
On Mon, Jan 14, 2019 at 8:06 AM Sal Am via petsc-users <
petsc-users@mcs.anl.gov> wrote:
> I ran it by: mpiexec -n 8 valgrind --tool=memcheck -q --num-callers=20
> --log-file=valgrind.log-osa.%p ./solveCSys -malloc
>
>
>
> > The Riemann solver computes the flux and the Jacobian is assembled in
> PETSc.
>
> If you run this example with an implicit integrator, it uses FD coloring
> to compute the Jacobian. Manav wants an analytic Jacobian. I was
> hoping Matt could answer this; I haven't used Plex in that
On Sat, Jan 12, 2019 at 6:59 PM Manav Bhatia via petsc-users <
petsc-users@mcs.anl.gov> wrote:
> I have been studying the source code and the manual entries in the
> documentation and have a better understanding of setting up a high-order DG
> analysis.
>
> I have been able to identify the
On Tue, Dec 25, 2018 at 12:10 AM Jed Brown wrote:
> Mark Adams writes:
>
> > On Mon, Dec 24, 2018 at 4:56 PM Jed Brown wrote:
> >
> >> Mark Adams via petsc-users writes:
> >>
> >> > Anyway, my data for this is in my SC 2004 paper (MakeNextMa
On Mon, Dec 24, 2018 at 4:56 PM Jed Brown wrote:
> Mark Adams via petsc-users writes:
>
> > Anyway, my data for this is in my SC 2004 paper (MakeNextMat_private in
> > attached, NB, this is code that I wrote in grad school). It is memory
> > efficient and simple, just f
Wow, this is an old thread.
Sorry if I sound like an old fart talking about the good old days but I
originally did RAP. in Prometheus, in a non work optimal way that might be
of interest. Not hard to implement. I bring this up because we continue to
struggle with this damn thing. I think this
OK, so this thread has drifted, see title :)
On Fri, Dec 21, 2018 at 10:01 PM Fande Kong wrote:
> Sorry, hit the wrong button.
>
>
>
> On Fri, Dec 21, 2018 at 7:56 PM Fande Kong wrote:
>
>>
>>
>> On Fri, Dec 21, 2018 at 9:44 AM Mark Adams wrote:
>>
>>> Also, you mentioned that you are using
Also, you mentioned that you are using 10 levels. This is very strange with
GAMG. You can run with -info and grep on GAMG to see the sizes and the
number of non-zeros per level. You should coarsen at a rate of about 2^D to
3^D with GAMG (with 10 levels this would imply a very large fine grid
Woops petsc-users, mistake, sorry about that.
On Mon, Dec 17, 2018 at 11:29 AM Mark Adams wrote:
> Dan and Sherry,
>
> I am taking over blind postdoc call for Andy Nonaka. Do you have any
> particular priorities that you would like me to refer good candidates for?
>
> Mark
>
Dan and Sherry,
I am taking over blind postdoc call for Andy Nonaka. Do you have any
particular priorities that you would like me to refer good candidates for?
Mark
On Wed, Dec 5, 2018 at 1:27 PM Sajid Ali via petsc-users <
petsc-users@mcs.anl.gov> wrote:
> I have created a file as per the specification as shown below
>
> [sajid@xrm temp]$ cat vector.dat
> 00010010001101001110
> 00010100
>
There is a reason PETSc does not have variable block matrices. It is messy
and rarely a big win. It is doable, ML/Aztec does it, but it is messy.
On Sat, Dec 1, 2018 at 6:20 PM Smith, Barry F. via petsc-users <
petsc-users@mcs.anl.gov> wrote:
>
>Well, you need to start somewhere. It is going
On Wed, Nov 28, 2018 at 10:27 AM Sal Am via petsc-users <
petsc-users@mcs.anl.gov> wrote:
> Thank you indeed --download-mpich and using PETSC_ARCH/bin/mpiexec seems
> to work.
>
> Now I am wondering about the other problem namely getting the residual, is
> the residual only computed when using
Note, your FormInitialGuess does not initialize bb[0] and bb[1].
On Thu, Nov 22, 2018 at 11:23 AM barry via petsc-users <
petsc-users@mcs.anl.gov> wrote:
> Hi,
>
> I want to solve a nonlinear equation tan(x) = x, and the code i wrote
> occur some error.
>
> In the error, it say that I need to
Yes, that's right.
Also, it is best to send this to the list, like petsc-users, for a few
reasons.
It is useful for the core developers, as well as any lurkers really, to
casually see what people are doing. And the threads are archived and are a
good source of deep documentation. eg, one could
The local indices of the local mesh and local vectors, which includes ghost
vertices. There are global-to-local methods to fill in ghost values and
local-to-global methods to create global vectors that you can use for
computation.
On Mon, Nov 19, 2018 at 5:16 PM Sajid Ali
wrote:
> Bingo!
>
>
You seem to be confusing the degree of the mesh and the "degree" of the
matrix and vector.
A matrix is always N x M (2D if you like), a vector is always N (or 1 x N,
or 1D if you like).
The mesh in a DM or DA can be 1, 2 or 3D.
On Mon, Nov 19, 2018 at 4:44 PM Sajid Ali via petsc-users <
>
>
>
> Mark would have better comments on the scalability of the setup stage.
>
>
The first thing to verify is that the algorithm is scaling. If you coarsen
too slowly then the coarse grids get large, with many non-zeros per row,
and the cost of the matrix triple product can explode. You can
The manual is pretty clear on this, but your code is purely algebraic, that
is you do not use a DM, and so you will need to use algebraic multigrid
(AMG). So you want to look at AMG preconditioners (-pc_type gamg [or hypre
if you configured with --download-hypre]).
Mark
On Sun, Nov 18, 2018 at
There is a lot of load imbalance in VecMAXPY also. The partitioning could
be bad and if not its the machine.
On Thu, Nov 15, 2018 at 1:56 PM Smith, Barry F. via petsc-users <
petsc-users@mcs.anl.gov> wrote:
>
> Something is odd about your configuration. Just consider the time for
> VecMAXPY
I am not that familiar with hypre's options. AMG is complicated and I
barely keep my options straight.
OpenFOAM seems to have highly specialized solvers so being with 50% of them
is decent.
On Thu, Nov 8, 2018 at 12:03 PM Edoardo alinovi
wrote:
> Yes, it is like you are saying. This is mostly
To repeat:
You seem to be saying that OpenFOAM solves the problem in 10 seconds and
PETSc solves it in 14 seconds. Is that correct?
On Thu, Nov 8, 2018 at 3:42 AM Edoardo alinovi via petsc-users <
petsc-users@mcs.anl.gov> wrote:
> Hello Mark,
>
> Yes, there are 5 KSP calls within a time-step
>
>
> I did not configured PETSc with ParMetis support. Should I?
>
> I figured it out when I tried to use "-pc_gamg_repartition". PETSc
> complained that it was not compiled with ParMetis support.
>
You need ParMetis, or some parallel mesh partitioner, configured to use
repartitioning. I would
First I would add -gamg_est_ksp_type cg
You seem to be converging well so I assume you are setting the null space
for GAMG.
Note, you should test hypre also.
You probably want a bigger "-pc_gamg_process_eq_limit 50". 200 at least but
you test your machine with a range on the largest problem.
please respond to petsc-users.
You are doing 5 solves here in 14 seconds. You seem to be saying that the
two pressure solves are taking all of this time. I don't know why the two
solves are different.
You seem to be saying that OpenFOAM solves the problem in 10 seconds and
PETSc solves it in 14
On Wed, Nov 7, 2018 at 10:16 AM Yingjie Wu via petsc-users <
petsc-users@mcs.anl.gov> wrote:
> Dear Petsc developer:
> Hi,
> Recently, I'm solving the problems of nonlinear systems of PDEs, I
> encountered some problems about precondition and wanted to seek help.
>
> 1.I set the precondition
On Mon, Nov 5, 2018 at 4:11 PM Appel, Thibaut
wrote:
> "Local" as in serial?
>
Block Jacobi with ILU as the solver on each block. Each block corresponds
to an MPI process by default. So it is completely parallel it is just not a
true ILU. I the limit of one equation per processor it is just
On Mon, Nov 5, 2018 at 12:50 PM Thibaut Appel
wrote:
> Hi Mark,
>
> Yes it doesn't seem to be usable. Unfortunately we're aiming to do 3D so
> direct solvers are not a viable solution and PETSc' ILU is not parallel and
> we can't use HYPRE (complex arithmetic)
>
I think SuperLU has a parallel
On Mon, Nov 5, 2018 at 10:37 AM Yingjie Wu wrote:
> Thank you very much for your reply.
> My equation is a neutron diffusion equation with eigenvalues, which is why
> I use DMConposite because there is a single non-physical field variable,
> eigenvalue.
>
OK, DMComposite might be your best
DMComposite is not very mature, the last time I checked and I don't of
anyone having worked on it recently, and it is probably not what you want
anyway. FieldSplit is most likely what you want.
What are your equations and discretization? eg, Stokes with cell centered
pressure? There are probably
On Wed, Oct 31, 2018 at 8:11 PM Smith, Barry F. wrote:
>
>
> > On Oct 31, 2018, at 5:39 PM, Appel, Thibaut via petsc-users <
> petsc-users@mcs.anl.gov> wrote:
> >
> > Well yes naturally for the residual but adding -ksp_true_residual just
> gives
> >
> > 0 KSP unpreconditioned resid norm
On Wed, Oct 31, 2018 at 3:43 PM Manav Bhatia wrote:
> Here are the updates. I did not find the options to make much difference
> in the results.
>
> I noticed this message in the GAMG output for cases 2, 3: HARD stop of
> coarsening on level 3. Grid too small: 1 block nodes
>
Yea, this is
These are indefinite (bad) Helmholtz problems. Right?
On Wed, Oct 31, 2018 at 2:38 PM Matthew Knepley wrote:
> On Wed, Oct 31, 2018 at 2:13 PM Thibaut Appel
> wrote:
>
>> Hi Mark, Matthew,
>>
>> Thanks for taking the time.
>>
>> 1) You're not suggesting having -fieldsplit_X_ksp_type *f*gmres
Again, you probably want to avoid Cheby. with ‘-mg_levels_ksp_type
richardson -mg_levels_pc_type sor’ with the proper prefix.
I'm not sure about "-fieldsplit_pc_type gamg" GAMG should work on one
block, and hence be a subpc. I'm not up on fieldsplit syntax.
On Wed, Oct 31, 2018 at 9:22 AM
On Tue, Oct 30, 2018 at 5:23 PM Appel, Thibaut via petsc-users <
petsc-users@mcs.anl.gov> wrote:
> Dear users,
>
> Following a suggestion from Matthew Knepley I’ve been trying to apply
> fieldsplit/gamg for my set of PDEs but I’m still encountering issues
> despite various tests. pc_gamg simply
t;>>> rotational DOF using some plate and shell formulations.
>>>> This can explain poor convergence of a multilevel approach, which needs
>>>> to restrict and extrapolate the unknowns. In order to check this
>>>> hypothesis, you can try a test ca
for the displacement DOF but is not for the
>>> rotational DOF using some plate and shell formulations.
>>> This can explain poor convergence of a multilevel approach, which needs
>>> to restrict and extrapolate the unknowns. In order to check this
>>> hypothesis,
shell formulations.
> This can explain poor convergence of a multilevel approach, which needs to
> restrict and extrapolate the unknowns. In order to check this hypothesis,
> you can try a test case with zero rotations.
>
> Nicolas
>
> Le lun. 29 oct. 2018 à 22:13, Mark Adams via
* the two level results tell us that MG is not doing well on the coarse
grids. So the coarse grids are the problem.
* Do not worry about timing now. Get the math correct. The two level solve
is not meant to be a solution just a diagnostic so don't try to optimize it
by squaring the graph. Use
101 - 140 of 140 matches
Mail list logo