PETSc still uses some deprecated MPI functions for
backward-compatibility with MPI-1, but now implementations are getting
louder about them. I assume the status quo of not requiring MPI-2 is
still in effect [1], so we should have configure test for the new
version and define them in terms of the o
will fix
On Mar 14, 2014, at 5:42 PM, Jed Brown wrote:
> Compilation error --with-parms, but other places missed as well.
>
> $ git grep 'Operators([^,()]*,[^,()]*,[^,()]*,[^,()]*)' | grep -v ,ierr
> bin/matlab/classes/PetscKSP.m:function err = SetOperators(obj,A,B,pattern)
> bin/matlab/
Compilation error --with-parms, but other places missed as well.
$ git grep 'Operators([^,()]*,[^,()]*,[^,()]*,[^,()]*)' | grep -v ,ierr
bin/matlab/classes/PetscKSP.m:function err = SetOperators(obj,A,B,pattern)
bin/matlab/classes/PetscPC.m:function err = SetOperators(obj,A,B,pattern)
bin/
mpicc -o ex56.o -c -fPIC -Wall -Wwrite-strings -Wno-strict-aliasing
-Wno-unknown-pragmas -g3 -O0 -I/sandbox/petsc/petsc.clone-3/include
-I/sandbox/petsc/petsc.clone-3/arch-linux-dbg-quad/include
-I/usr/include/mpich2`pwd`/ex56.c
In file included from /sandbox/petsc/petsc.clone-3/include/p
Barry:
>
>Beats me. It is worth keeping only if the "slight performance advantage"
> is of practical use. Like while, for example, used in a multigrid smoother.
> If it just makes a bad algorithm you won't use anyways slightly better than
> it is not work keeping
It does not improve mg smoo
CC
arch-linux-pkgs-dbg-ftn-interfaces/obj/src/ksp/pc/impls/bjacobi/bjacobi.o
CC
arch-linux-pkgs-dbg-ftn-interfaces/obj/src/ksp/pc/impls/bjacobi/ftn-auto/bjacobif.o
/sandbox/petsc/petsc.clone/include/finclude/ftn-auto/petscdm.h90:400.7:
Included at /sandbox/petsc/petsc.clo
/sandbox/petsc/petsc.clone-2/src/dm/impls/plex/plexcreate.c:315:3: warning:
initializer element is not computable at load time [enabled by default]
/sandbox/petsc/petsc.clone-2/src/dm/impls/plex/plexcreate.c:315:3: warning:
initializer element is not computable at load time [enabled by default]
Barry,
The hybrid Chebyshev was implemented for experimentation, but had not
been found useful, so the default is set as FALSE.
I can add a functional interface though.
Hong
On Tue, Mar 11, 2014 at 9:15 PM, Smith, Barry F. wrote:
>
> There should be a functional interface for turning this on a
Beats me. It is worth keeping only if the “slight performance advantage” is
of practical use. Like while, for example, used in a multigrid smoother. If it
just makes a bad algorithm you won’t use anyways slightly better than it is not
work keeping
Barry
On Mar 14, 2014, at 2:20 PM, Hong
Barry,
Currently, we have following options for hybrid Chebyshev:
-ksp_chebyshev_hybrid: Use hybrid Chebyshev ()
-ksp_chebyshev_hybrid_chebysteps <2>: Number of Chebyshev steps
in hybrid Chebyshev ()
-ksp_chebyshev_hybrid_eststeps <10>: Number of adaptive/est steps in
hybrid Chebyshev ()
I’d remove it.
Thanks
Barry
On Mar 13, 2014, at 9:49 PM, Hong Zhang wrote:
> Barry,
> I sent this email out, but apparently it failed to reach you.
>
> On Tue, Mar 11, 2014 at 10:46 PM, Hong Zhang wrote:
>> Barry,
>> The hybrid Chebyshev was implemented for experimentation, but had no
Barry,
I sent this email out, but apparently it failed to reach you.
On Tue, Mar 11, 2014 at 10:46 PM, Hong Zhang wrote:
> Barry,
> The hybrid Chebyshev was implemented for experimentation, but had not
> been found useful, so the default is set as FALSE.
> I can add a functional interface though.
On Thu, Mar 13, 2014 at 10:26 PM, Barry Smith wrote:
>
> /Users/barrysmith/Src/PETSc/src/dm/impls/plex/plexfem.c:1203:41: warning:
> variable 'cpdim' is uninitialized when used here [-Wuninitialized]
> if (fieldI != fieldJ) {offsetJ += cpdim; continue;}
>
13 matches
Mail list logo