libmesh-users

[Libmesh-users] make check reduced_basis_ex7 flags

[<ed...@op...>]

Hi.

I want to know if libMesh++ is actually using mumps and superlu, and if 
it is not how to fix it. I just recompiled libMesh++, and I am getting 
the following message with the reduced_basis_ex7 (this is actually an 
improvement, because it used to crash at that point). Thanks!

#+begin_EXAMPLE

   ... snip 8<  ...

   ,***************************************************************
   ,* Running Example reduced_basis_ex7:
   ,*   ./example-opt -online_mode 1 -pc_factor_mat_solver_package 
superlu
   ,***************************************************************

   LibMesh was configured with PETSc >= 3.9, but command-line options use 
deprecated syntax. Skipping now.
   WARNING! There are options you set that were not used!
   WARNING! could be spelling mistake, etc!
   There are 2 unused database options. They are:
   Option left: name:-online_mode value: 1
   Option left: name:-pc_factor_mat_solver_package value: superlu

   ,***************************************************************
   ,* Done Running Example reduced_basis_ex7:
   ,*   ./example-opt -online_mode 1 -pc_factor_mat_solver_package 
superlu
   ,***************************************************************
#+end_EXAMPLE

I get a similar message above with mumps. I am attaching my config.log 
as a tar ball

-------------------------------------------------
This free account was provided by VFEmail.net - report spam to ab...@vf...
 
ONLY AT VFEmail! - Use our Metadata Mitigator to keep your email out of the NSA's hands!
$24.95 ONETIME Lifetime accounts with Privacy Features!  
15GB disk! No bandwidth quotas!
Commercial and Bulk Mail Options!  

Re: [Libmesh-users] make check reduced_basis_ex7 flags

[David K. <dav...@ak...>]

I think PETSc changed the option name from "-pc_factor_mat_solver_package"
to "-pc_factor_mat_solver_type", which I guess is the issue here.

Also, you can use "-ksp_view" on the command line to get a printout of the
solver options that were used by PETSc, so you can see if it actually used
MUMPS or superlu.

Best,
David


On Sun, May 10, 2020 at 11:08 PM <ed...@op...> wrote:

> Hi.
>
> I want to know if libMesh++ is actually using mumps and superlu, and if
> it is not how to fix it. I just recompiled libMesh++, and I am getting
> the following message with the reduced_basis_ex7 (this is actually an
> improvement, because it used to crash at that point). Thanks!
>
> #+begin_EXAMPLE
>
>    ... snip 8<  ...
>
>    ,***************************************************************
>    ,* Running Example reduced_basis_ex7:
>    ,*   ./example-opt -online_mode 1 -pc_factor_mat_solver_package
> superlu
>    ,***************************************************************
>
>    LibMesh was configured with PETSc >= 3.9, but command-line options use
> deprecated syntax. Skipping now.
>    WARNING! There are options you set that were not used!
>    WARNING! could be spelling mistake, etc!
>    There are 2 unused database options. They are:
>    Option left: name:-online_mode value: 1
>    Option left: name:-pc_factor_mat_solver_package value: superlu
>
>    ,***************************************************************
>    ,* Done Running Example reduced_basis_ex7:
>    ,*   ./example-opt -online_mode 1 -pc_factor_mat_solver_package
> superlu
>    ,***************************************************************
> #+end_EXAMPLE
>
> I get a similar message above with mumps. I am attaching my config.log
> as a tar ball
>
> -------------------------------------------------
> This free account was provided by VFEmail.net - report spam to
> ab...@vf...
>
> ONLY AT VFEmail! - Use our Metadata Mitigator to keep your email out of
> the NSA's hands!
> $24.95 ONETIME Lifetime accounts with Privacy Features!
> 15GB disk! No bandwidth quotas!
> Commercial and Bulk Mail Options!
> _______________________________________________
> Libmesh-users mailing list
> Lib...@li...
> https://lists.sourceforge.net/lists/listinfo/libmesh-users
>

Re: [Libmesh-users] make check reduced_basis_ex7 flags

[John P. <jwp...@gm...>]

On Sun, May 10, 2020 at 10:08 PM <ed...@op...> wrote:

> Hi.
>
> I want to know if libMesh++ is actually using mumps and superlu, and if
> it is not how to fix it. I just recompiled libMesh++, and I am getting
> the following message with the reduced_basis_ex7 (this is actually an
> improvement, because it used to crash at that point). Thanks!
>
> #+begin_EXAMPLE
>
>    ... snip 8<  ...
>
>    ,***************************************************************
>    ,* Running Example reduced_basis_ex7:
>    ,*   ./example-opt -online_mode 1 -pc_factor_mat_solver_package
> superlu
>    ,***************************************************************
>
>
This message is from libMesh (BTW what is libMesh++?):


>    LibMesh was configured with PETSc >= 3.9, but command-line options use
> deprecated syntax. Skipping now.
>

And these messages are from PETSc:

   WARNING! There are options you set that were not used!
>    WARNING! could be spelling mistake, etc!
>    There are 2 unused database options. They are:
>    Option left: name:-online_mode value: 1
>    Option left: name:-pc_factor_mat_solver_package value: superlu
>

I think this is all working as intended, since if you look at run.sh for
reduced_basis_ex7, you can see that we run all four combinations of
"type"/"package" and "mumps"/superlu" from the script, relying on the logic
within the example itself to skip those cases which don't apply. If you
want to test whether it's working you can run the example manually with the
different cases and inspect the output of each one rather than running the
full "make check". The examples must be run from the build directory, so
for example to test mumps with PETSC >= 3.9 (assuming $LIBMESH_SRC/build is
the location of your build directory):

cd $LIBMESH_SRC/build/examples/reduced_basis/reduced_basis_ex7
./example-opt -online_mode 0 -ksp_type preonly -pc_type lu
-pc_factor_mat_solver_type mumps

Finally, there is one last check you can do, which is to look at (again,
from the build directory) the include/libmesh_config.h file and inspect the
values of the

LIBMESH_PETSC_HAVE_MUMPS
LIBMESH_PETSC_HAVE_SUPERLU_DIST

preprocessor defines.



>    ,***************************************************************
>    ,* Done Running Example reduced_basis_ex7:
>    ,*   ./example-opt -online_mode 1 -pc_factor_mat_solver_package
> superlu
>    ,***************************************************************
> #+end_EXAMPLE
>
> I get a similar message above with mumps. I am attaching my config.log
> as a tar ball
>

This mailing list does not allow attachments.

-- 
John

Re: [Libmesh-users] make check reduced_basis_ex7 flags

[<ed...@op...>]

On 2020-05-11 11:06, David Knezevic wrote:
> I think PETSc changed the option name from 
> "-pc_factor_mat_solver_package"
> to "-pc_factor_mat_solver_type", which I guess is the issue here.

Yes, thank you.

> 
> Also, you can use "-ksp_view" on the command line to get a printout of 
> the
> solver options that were used by PETSc, so you can see if it actually 
> used
> MUMPS or superlu.

Yes, thank you again.

> 
> Best,
> David


On 2020-05-11 13:47, John Peterson wrote:
> On Sun, May 10, 2020 at 10:08 PM <ed...@op...> wrote:
> 
> This message is from libMesh (BTW what is libMesh++?):

libMesh++ is a hole in my brain from Elmer's documentation, sorry (not 
blaming it on them).

>> 
> 
> And these messages are from PETSc:
> 
>> 
> 
> I think this is all working as intended, since if you look at run.sh 
> for
> reduced_basis_ex7, you can see that we run all four combinations of
> "type"/"package" and "mumps"/superlu" from the script, relying on the 
> logic
> within the example itself to skip those cases which don't apply.

That is very ingenious. Noob question: Is there still a need to test for 
options that don't exist? I know there must be a reason behind it :P .

> The examples must be run from the build directory, so
> for example to test mumps with PETSC >= 3.9 (assuming 
> $LIBMESH_SRC/build is
> the location of your build directory):
> 
> cd $LIBMESH_SRC/build/examples/reduced_basis/reduced_basis_ex7
> ./example-opt -online_mode 0 -ksp_type preonly -pc_type lu
> -pc_factor_mat_solver_type mumps

Thank you very much for the very clear instructions.

> 
> Finally, there is one last check you can do, which is to look at 
> (again,
> from the build directory) the include/libmesh_config.h file and inspect 
> the
> values of the
> 
> LIBMESH_PETSC_HAVE_MUMPS
> LIBMESH_PETSC_HAVE_SUPERLU_DIST
> 
> preprocessor defines.

I really got it with this. It seems that SuperLU is not being used, but 
MUMPS is. I will check my PETSc compilation before asking further 
questions.

All of this was very useful. Thanks, dear David and John.

-------------------------------------------------
This free account was provided by VFEmail.net - report spam to ab...@vf...
 
ONLY AT VFEmail! - Use our Metadata Mitigator to keep your email out of the NSA's hands!
$24.95 ONETIME Lifetime accounts with Privacy Features!  
15GB disk! No bandwidth quotas!
Commercial and Bulk Mail Options!  

Re: [Libmesh-users] make check reduced_basis_ex7 flags

[John P. <jwp...@gm...>]

On Mon, May 11, 2020 at 11:50 PM <ed...@op...> wrote:

>
> That is very ingenious. Noob question: Is there still a need to test for
> options that don't exist? I know there must be a reason behind it :P .
>

The options still exist for people using older versions of PETSc, so yes,
it makes sense to still test for them.

-- 
John

Re: [Libmesh-users] make check reduced_basis_ex7 flags

[<ed...@op...>]

On 2020-05-12 14:18, John Peterson wrote:
> The options still exist for people using older versions of PETSc, so 
> yes,
> it makes sense to still test for them.

Of course! Thanks! By the way, I am maintaining an AUR package 
(aur.archlinux.org) of libmesh right now.

-------------------------------------------------
This free account was provided by VFEmail.net - report spam to ab...@vf...
 
ONLY AT VFEmail! - Use our Metadata Mitigator to keep your email out of the NSA's hands!
$24.95 ONETIME Lifetime accounts with Privacy Features!  
15GB disk! No bandwidth quotas!
Commercial and Bulk Mail Options!  

Re: [Libmesh-users] make check reduced_basis_ex7 flags

[John P. <jwp...@gm...>]

On Tue, May 12, 2020 at 10:10 AM <ed...@op...> wrote:

> On 2020-05-12 14:18, John Peterson wrote:
> > The options still exist for people using older versions of PETSc, so
> > yes,
> > it makes sense to still test for them.
>
> Of course! Thanks! By the way, I am maintaining an AUR package
> (aur.archlinux.org) of libmesh right now.
>

OK, I don't know much about Arch but please let us know if we can help by
having some packaging metadata files in the repository or something along
those lines.

-- 
John