Thanks for the reply.
> "What I want to do is to exchange data (probably just MPI_Reduce)" which
confuses me, because halo exchange is a point-to-point exchange and not a
reduction. Can you clarify?
PetscSFReduceBegin/End seems to be the function that do reduction for
PetscSF object. What I
The PetscSF that is created automatically is the "point sf" (
https://petsc.org/main/docs/manualpages/DM/DMGetPointSF/): it says which
mesh points (cells, faces, edges and vertices) are duplicates of others.
In a finite volume application we typically want to assign degrees of
freedom just to
Hello PETSc users,
I have been trying to run example ex1 from
src/tao/constrained/tutorials, which requires the use of the MUMPS
package. I have downloaded MUMPS and configured PETSc with it using
./configure --download-make --download-cmake --download-mumps
--download-metis
> Yes, this is what I did.
> But I have a function that sets up the preconditioner by taking a KSP
> as an argument. Depending on the problem being solved, it can be a
> single KSP or a sub-KSP coming from a field split (which if there are
> lagrange multipliers, can be further split again). So
You can also use -log_view -log_sync to sync before timing so that you can
clearly see which operations are really imbalanced.
--Junchao Zhang
On Fri, May 20, 2022 at 12:37 PM Ernesto Prudencio via petsc-users <
petsc-users@mcs.anl.gov> wrote:
> Thank you, Barry. I will dig more on the issue
Barry Smith writes:
>> 1. Is there a way to set the mg_levels_pc_type via an API call?
>> 2. Are there any changes in efficiency expected with this new PC?
>
> This was changed mainly because PCSOR is not effective (currently in PETSc)
> for GPUs. For CPUs
> it will definitely be problem
> On May 20, 2022, at 2:39 PM, Jeremy Theler wrote:
>
> The default smoothing PC changed from sor to jacobi in 3.17.
Note that this is only for GAMG, it is not for geometric multigrid (using
PCMG directly).
> The
> Changelog says the old behavior can be recovered by using
>
The default smoothing PC changed from sor to jacobi in 3.17. The
Changelog says the old behavior can be recovered by using
-mg_levels_pc_type sor.
1. Is there a way to set the mg_levels_pc_type via an API call?
2. Are there any changes in efficiency expected with this new PC?
Regards
--
jeremy
Thank you, Barry. I will dig more on the issue with your suggestions.
Schlumberger-Private
From: Barry Smith
Sent: Friday, May 20, 2022 12:33 PM
To: Ernesto Prudencio
Cc: PETSc users list
Subject: [Ext] Re: [petsc-users] Very slow VecDot operations
Ernesto,
If you ran (or can run)
Ernesto,
If you ran (or can run) with -log_view you could see the time "ratio" in
the output that tells how much time the "fastest" rank spent on the dot product
versus the "slowest". Based on the different counts per rank you report that
ratio might be around 3. But based on the times
I am using LSQR to minimize || L x - b ||_2, where L is a sparse rectangular
matrix with 145,253,395 rows, 209,423,775 columns, and around 54 billion non
zeros.
The numbers reported below are for a run with 27 compute nodes, each compute
node with 4 MPI ranks, so a total of 108 ranks.
11 matches
Mail list logo
