I can't see what branch anything is happening in, so I am confused by the
commit messages.
Matt
--
What most experimenters take for granted before they begin their
experiments is infinitely more interesting than any results to which their
experiments lead.
-- Norbert Wiener
--
On Wed, Mar 13, 2013 at 2:56 PM, Richard Tran Mills rtm at eecs.utk.eduwrote:
Hi Folks,
I occasionally see messages like this from configure.py:
Error running make on HDF5: Could not execute cd
/home/rmills/proj/petsc-dev/**externalpackages/hdf5-1.8.6 make clean
make make install:
I think the model of long-running development, which is also the model
being used by the integrator (Jed) to master, is not suitable for a
community nor scalable.
I am developing in knepley/pylith, which is the branch I tell PyLith people
to look at. I know I need to keep up-to-date with other
On Sat, Mar 16, 2013 at 9:16 AM, Jed Brown jedbrown at mcs.anl.gov wrote:
Matthew Knepley knepley at gmail.com writes:
I think the model of long-running development, which is also the model
being used by the integrator (Jed) to master, is not suitable for a
community nor scalable.
Any
On Sun, Mar 17, 2013 at 9:01 PM, Jed Brown jedbrown at mcs.anl.gov wrote:
Surely it would be sufficient to check and set these only once?
I think that would just introduce fragility. We can call the test multiple
times with
different configurations if we want without worrying about caching.
On Mon, Mar 18, 2013 at 9:03 PM, Jed Brown jedbrown at mcs.anl.gov wrote:
Ever since at least here, configure has ignored compiler output by default.
https://bitbucket.org/petsc/petsc/commits/08bdf6bce1142b6c3d4d318bdb8e3320d9f5808d#Lconfig/framework.pyT287
This means that all the checks in
On Mon, Mar 18, 2013 at 9:22 PM, Jed Brown jedbrown at mcs.anl.gov wrote:
On Mon, Mar 18, 2013 at 8:10 PM, Matthew Knepley knepley at gmail.comwrote:
It was before that commit, as you can see that I just use the 1 that was
already there.
Yeah, I see you just declared the default
On Mon, Mar 18, 2013 at 9:27 PM, Jed Brown jedbrown at mcs.anl.gov wrote:
On Mon, Mar 18, 2013 at 8:24 PM, Matthew Knepley knepley at gmail.comwrote:
You can always turn that switch off for your test. What we were finding
is that some crappy compilers
were killing everything by putting
On Mon, May 20, 2013 at 8:15 PM, Barry Smith bsm...@mcs.anl.gov wrote:
Begin forwarded message:
From: Jian Du d...@math.utah.edu
Subject: Re: [petsc-maint] compile PETSc with HYPRE
Date: May 20, 2013 3:26:18 PM CDT
To: Jed Brown jedbr...@mcs.anl.gov
Cc: petsc-ma...@mcs.anl.gov,
On Tue, Mar 13, 2012 at 9:31 AM, Jed Brown jedbrown at mcs.anl.gov wrote:
On Tue, Mar 13, 2012 at 09:25, Matthew Knepley knepley at gmail.com wrote:
I don't use the makefiles.
The makefiles are the canonical statement of what is part of the build and
what isn't. You should parse them
.
Matt
Barry
On Mar 18, 2012, at 3:07 PM, Matthew Knepley wrote:
On Sun, Mar 18, 2012 at 2:49 PM, Barry Smith bsmith at mcs.anl.gov wrote:
On Mar 18, 2012, at 12:09 PM, Matthew Knepley wrote:
On Sun, Mar 18, 2012 at 11:52 AM, Barry Smith bsmith at mcs.anl.gov
wrote
On Sun, Mar 18, 2012 at 6:37 PM, Jed Brown jedbrown at mcs.anl.gov wrote:
On Sun, Mar 18, 2012 at 18:24, Matthew Knepley knepley at gmail.com wrote:
In the alternate model, you call MatAssemblyBegin/End() when the matrix
is marked as unassembled. Its easy to check this flag, so the user
On Sun, Mar 18, 2012 at 6:38 PM, Barry Smith bsmith at mcs.anl.gov wrote:
On Mar 18, 2012, at 6:24 PM, Matthew Knepley wrote:
On Sun, Mar 18, 2012 at 5:47 PM, Barry Smith bsmith at mcs.anl.gov wrote:
Let me clarify. I did not find your completely undocumented feature
that has always
On Sun, Mar 18, 2012 at 6:52 PM, Jed Brown jedbrown at mcs.anl.gov wrote:
On Sun, Mar 18, 2012 at 18:45, Matthew Knepley knepley at gmail.com wrote:
No, I don't want the nonzeros removed, I want the values set to zero,
exactly like I said. This is for robustly checking
that a matrix created
On Sun, Mar 18, 2012 at 6:46 PM, Jed Brown jedbrown at mcs.anl.gov wrote:
On Sun, Mar 18, 2012 at 18:44, Matthew Knepley knepley at gmail.com wrote:
Of course I fucking tried that first. It does not work.
It doesn't work because you want to structurally remove small values or it
doesn't
On Mon, Mar 19, 2012 at 9:56 AM, Satish Balay balay at mcs.anl.gov wrote:
On Sat, 17 Mar 2012, Matthew Knepley wrote:
On Sat, Mar 17, 2012 at 4:08 PM, Jed Brown jedbrown at mcs.anl.gov wrote:
On Sat, Mar 17, 2012 at 16:04, Matthew Knepley knepley at gmail.com
wrote:
On Sat, Mar 17
On Tue, Mar 20, 2012 at 1:40 PM, Barry Smith bsmith at mcs.anl.gov wrote:
It will make Matt barf but I think is way cool, anyone heard of this
http://jpetsctao.zwoggel.net/
A single tear rolls down my cheek as I contemplate the life wasted to
develop this.
Matt
Barry
--
What
On Tue, Mar 20, 2012 at 3:49 PM, Sean Farley sean at mcs.anl.gov wrote:
You have not yet modified the script that prepares the manual pages to
automatically put the date of generation in the upper right corner. As
well as other generated pages.
Could you please add that ASAP or speak to me
On Wed, Mar 21, 2012 at 7:16 AM, Aron Ahmadia aron.ahmadia at
kaust.edu.sawrote:
or am I just blind? Where is this documented outside of the developer's
list?
It is perhaps not enough, but 3.2 Changes say that VecDestroy() now zeros
the pointer so that it cannot be reused.
Matt
A
--
On Wed, Mar 21, 2012 at 8:22 AM, Barry Smith bsmith at mcs.anl.gov wrote:
Matt,
One thing that has always pissed me off about BuildSystem but for some
reason I never had the energy to bitch about is when a compile goes wrong,
BuildSystem logs the source code that wouldn't compiler BUT IT
Fixed.
Matt
On Wed, Mar 21, 2012 at 9:32 AM, Aron Ahmadia aron.ahmadia at
kaust.edu.sawrote:
It is perhaps not enough, but 3.2 Changes say that VecDestroy() now zeros
the pointer so that it cannot be reused.
It is not enough, as you know this change hit every Destroy() function in
+1
Matt
On Wed, Mar 21, 2012 at 11:09 AM, Sean Farley sean at mcs.anl.gov wrote:
Fixed.
I like how it's easy to tell what changed in the webpages now :-)
http://petsc.cs.iit.edu/petsc/petsc-dev/rev/d058bd1e3700
--
What most experimenters take for granted before they begin their
On Wed, Mar 21, 2012 at 10:24 AM, Barry Smith bsmith at mcs.anl.gov wrote:
On Mar 21, 2012, at 8:38 AM, Matthew Knepley wrote:
On Wed, Mar 21, 2012 at 8:22 AM, Barry Smith bsmith at mcs.anl.gov wrote:
Matt,
One thing that has always pissed me off about BuildSystem but for some
On Wed, Mar 21, 2012 at 11:57 AM, Jed Brown jedbrown at mcs.anl.gov wrote:
On Wed, Mar 21, 2012 at 11:46, Matthew Knepley knepley at gmail.com wrote:
The compile routine does not decide whether to crash, so I had
configure.py dump these headers if there was an exception. I
can test it later
On Wed, Mar 21, 2012 at 12:06 PM, Jed Brown jedbrown at mcs.anl.gov wrote:
On Wed, Mar 21, 2012 at 12:01, Matthew Knepley knepley at gmail.com wrote:
I am not sure what you mean wrap the whole test case up.
Bundle the environment that was in scope (via your push/pop) with the
exception so
to the new scalable MatMatMult() that is done in PETSc? Is it possible to
even compare
the flop rates sensibly?
http://arxiv.org/abs/1203.1692
This is what Jeff is planning on using in his SciDAC.
Matt
--
What most experimenters take for granted before they begin their
experiments is
On Sun, Mar 25, 2012 at 4:26 PM, Jed Brown jedbrown at mcs.anl.gov wrote:
On Sun, Mar 25, 2012 at 16:14, Satish Balay balay at mcs.anl.gov wrote:
Note: there is no such package for cusp/thrust to base a default - so
the upstream default thats currently implemented is fine [we are
supporting
On Sun, Mar 25, 2012 at 4:53 PM, Satish Balay balay at mcs.anl.gov wrote:
On Sun, 25 Mar 2012, Jed Brown wrote:
On Sun, Mar 25, 2012 at 16:32, Matthew Knepley knepley at gmail.com
wrote:
Isn't the final point here that we are trying to let the user do the
dumbest, easiest thing
I update my GPU machine today by
- Pulling BuildSystem, petsc-dev, thrust, cusp
- Reconfiguring
- Recompiling
and none of the Thrust or Cusp things will build because VecType is
again messed up. I look in the headers and the #undef VecType seems
to be missing. Where did it go? and aren't
On Sun, Jan 1, 2012 at 5:37 PM, Jed Brown jedbrown at mcs.anl.gov wrote:
This code is so fragile. Isn't there a way to say if the user requested
ParMetis in *any* way instead of enumerating the list of ways that it
could be asked for. Should we also check whether the user asked for
ParMetis
*Subject:* Re: [petsc-dev] boomerAmg scalability
** **
** **
On Dec 15, 2011, at 1:56 PM, Matthew Knepley wrote:
On Thu, Dec 15, 2011 at 10:23 AM, Ravi Kannan rxk at cfdrc.com wrote:
Dear All,
This is Ravi Kannan from CFD Research Corporation. Recently
On Thu, Jan 19, 2012 at 8:06 AM, Barry Smith bsmith at mcs.anl.gov wrote:
hdf5.py is messed up, it starts with crap from a patch??? Please fix
HDF5 in their infinite wisdom run tests in the same make target as the
build. One of the Fortran
tests breaks on the Mac. The only solution that
On Thu, Jan 19, 2012 at 10:03 AM, Satish Balay balay at mcs.anl.gov wrote:
On Thu, 19 Jan 2012, Sean Farley wrote:
HDF5 in their infinite wisdom run tests in the same make target as the
build. One of the Fortran
tests breaks on the Mac. The only solution that worked for me was to
On Thu, Jan 19, 2012 at 10:11 AM, Satish Balay balay at mcs.anl.gov wrote:
On Thu, 19 Jan 2012, Matthew Knepley wrote:
On Thu, Jan 19, 2012 at 10:03 AM, Satish Balay balay at mcs.anl.gov
wrote:
On Thu, 19 Jan 2012, Sean Farley wrote:
HDF5 in their infinite wisdom run tests
On Thu, Jan 19, 2012 at 10:35 AM, Jed Brown jedbrown at mcs.anl.gov wrote:
On Thu, Jan 19, 2012 at 10:22, Matthew Knepley knepley at gmail.com wrote:
Notice that this patch is versioned in a repo.
Well, not in a form that you can edit or re-apply to new versions from
upstream.
Not sure
http://petsc.cs.iit.edu/petsc/petsc-dev/rev/2a4f352daf49
What is going on here, and why can't it be done more succintly?
Matt
--
What most experimenters take for granted before they begin their
experiments is infinitely more interesting than any results to which their
experiments lead.
--
On Fri, Jan 20, 2012 at 4:59 PM, Barry Smith bsmith at mcs.anl.gov wrote:
On Jan 20, 2012, at 4:51 PM, Matthew Knepley wrote:
On Fri, Jan 20, 2012 at 4:45 PM, Paul Mullowney paulm at txcorp.com
wrote:
The new code using CUSPARSE gives you triangular solves that are
supported by Nvidia
On Sat, Jan 28, 2012 at 10:11 AM, Stefano Zampini stefano.zampini at gmail.com
wrote:
Hi Jed,
you can use any code which calls the following sequence
ierr = MatCreate(PETSC_COMM_SELF,mat);CHKERRQ(ierr);
ierr = MatSetSize(mat,n,m,n,m);CHKERRQ(ierr);
ierr =
On Tue, Jan 31, 2012 at 5:12 AM, Stefano Zampini
stefano.zampini at gmail.comwrote:
2012/1/31 Jed Brown jedbrown at mcs.anl.gov
On Thu, Jan 26, 2012 at 04:04, Stefano Zampini stefano.zampini at gmail.com
wrote:
I'm partially refactoring BDDC code I contributed to petsc-dev, trying
to
This diagnosis is total crap (I think), as I tried to explain. We would
never get the same result (or the right result),
and partitioning makes no sense. Something else is going on. Can't we run on
a 2 GPU system at ANL?
Matt
On Sat, Oct 1, 2011 at 9:30 PM, Barry Smith bsmith at mcs.anl.gov
On Sat, Oct 1, 2011 at 11:26 PM, Dave Nystrom Dave.Nystrom at tachyonlogic.com
wrote:
Barry Smith writes:
On Oct 1, 2011, at 9:22 PM, Dave Nystrom wrote:
Hi Barry,
I've sent a couple more emails on this topic. What I am trying to do
at the
moment is to figure out how
was not actually copied to
the gpu.
I'm not sure yet how to check for that. Hope this might be useful.
Thanks,
Dave
ex2f_3200_3200_cuda_yes_cuda_3.2.logex2f_3200_3200_cuda_yes_cuda_4.0.log
Dave Nystrom writes:
Matthew Knepley writes:
On Sat, Oct 1, 2011 at 11:26 PM, Dave
On Mon, Oct 17, 2011 at 10:07 AM, Daniel Lowell redratio1 at gmail.com wrote:
Ah, okay thanks.
Reconfiguring, or 'make allfortranstubs', or 'python2.7 ./config/builder2.py
stubs fortran'
Matt
On Mon, Oct 17, 2011 at 10:06 AM, Jed Brown jedbrown at mcs.anl.gov wrote:
On Mon, Oct 17,
On Tue, Oct 18, 2011 at 5:23 PM, Ethan Coon ecoon at lanl.gov wrote:
Ok, but that's not in petscvariables. Maybe I should be clearer -- I'm
trying to make the python scripts for binary IO pull the default
configuration from the PETSC_DIR/PETSC_ARCH, so that binary files can be
in the correct
On Tue, Oct 18, 2011 at 5:50 PM, Ethan Coon ecoon at lanl.gov wrote:
On Tue, 2011-10-18 at 17:27 -0500, Matthew Knepley wrote:
If you are in Python, just load up the configuration directly. Take a
look at
bin/configVars.py
I don't think I have RDict or script in an installed
On Wed, Oct 19, 2011 at 7:39 PM, Jed Brown jedbrown at mcs.anl.gov wrote:
#define CHKERRXX(n)do {if (PetscUnlikely(n))
{PetscError(PETSC_COMM_SELF,__LINE__,PETSC_FUNCTION_NAME,__FILE__,__SDIR__,n,PETSC_ERROR_IN_CXX,0);}}
while(0)
Inside PetscError():
if (p ==
On Wed, Oct 19, 2011 at 7:49 PM, Barry Smith bsmith at mcs.anl.gov wrote:
On Oct 19, 2011, at 2:45 PM, Matthew Knepley wrote:
On Wed, Oct 19, 2011 at 7:39 PM, Jed Brown jedbrown at mcs.anl.gov wrote:
#define CHKERRXX(n)do {if (PetscUnlikely(n))
{PetscError(PETSC_COMM_SELF
On Wed, Oct 19, 2011 at 8:02 PM, Jed Brown jedbrown at mcs.anl.gov wrote:
On Wed, Oct 19, 2011 at 14:57, Matthew Knepley knepley at gmail.com wrote:
The properly manage the clean up of partially created objects etc don't
they, which our C approach cannot handle. Not that I'm advocating
On Thu, Oct 20, 2011 at 6:25 PM, Sean Farley sean at mcs.anl.gov wrote:
I put in a patch for HDF5 since it broke when installing with gfortran on
the Mac.
What was the error that you were patching? I have HDF5 installed on my Mac
but haven't run into anything (granted, I'm not an advanced
I don't need it.
Matt
On Fri, Oct 21, 2011 at 1:35 PM, Barry Smith bsmith at mcs.anl.gov wrote:
Is there really a need for PetscTraceBackErrorHandlerCxx() that uses C++
streams for output instead of C IO but otherwise looks very similar to
PetscTraceBackErrorHandler()? It seems like a
On Fri, Oct 21, 2011 at 8:40 PM, Barry Smith bsmith at mcs.anl.gov wrote:
News to me, anyone is welcome to join me.
I think I can make that.
Matt
Barry
Begin forwarded message:
From: Lori O'Connor oconnor at mcs.anl.gov
Date: October 21, 2011 2:22:39 PM CDT
To: Paul
On Sun, Oct 23, 2011 at 8:31 PM, Barry Smith bsmith at mcs.anl.gov wrote:
On Oct 23, 2011, at 3:26 PM, Jed Brown wrote:
On Sun, Oct 23, 2011 at 15:18, Barry Smith bsmith at mcs.anl.gov wrote:
I wasn't thinking of anything so complicated. Just having a separate
routine TSGetCurrentTime()
On Sun, Oct 23, 2011 at 9:01 PM, Jed Brown jedbrown at mcs.anl.gov wrote:
On Sun, Oct 23, 2011 at 15:58, Matthew Knepley knepley at gmail.com wrote:
I am fine if you say that the interface is at the continuum level, but
then the user should not be getting back
output that depends
On Sun, Oct 23, 2011 at 9:20 PM, Jed Brown jedbrown at mcs.anl.gov wrote:
On Sun, Oct 23, 2011 at 16:15, Matthew Knepley knepley at gmail.com wrote:
TSSetExactFinalTime() ?
This I think is probably alright. I assume it means give me the closest
time to this one that I can
get without
On Tue, Oct 25, 2011 at 10:06 PM, Jed Brown jedbrown at mcs.anl.gov wrote:
On Tue, Oct 25, 2011 at 16:39, Barry Smith bsmith at mcs.anl.gov wrote:
Chebychev?
Once I learned how to configure it, I quite like Chebychev for symmetric
problems (and it is a natural fit for newer hardware).
I
On Wed, Oct 26, 2011 at 4:21 PM, Mark F. Adams mark.adams at
columbia.eduwrote:
On Oct 26, 2011, at 11:29 AM, Barry Smith wrote:
On Oct 26, 2011, at 10:19 AM, Mark F. Adams wrote:
Two things:
1) Computing C(A^{T} A) is very similar to what I need in GAMG and is
currently one of
On Mon, Oct 31, 2011 at 7:37 PM, Barry Smith bsmith at mcs.anl.gov wrote:
PETSc-folks,
Let me know if you plan to attend.
No, and there is no registration form there.
Matt
Barry
Begin forwarded message:
*From: *Snir, Marc snir at illinois.edu
*Date: *October 31, 2011
On Sat, Jul 2, 2011 at 9:39 PM, Jed Brown jedbrown at mcs.anl.gov wrote:
In refactoring TS, I want these implementations to go away (they are
special cases of TSTHETA and it's easy for TSTHETA to be as efficient for
the special case). The question is whether the names should be preserved,
On Sat, Jul 23, 2011 at 3:03 AM, Barry Smith bsmith at mcs.anl.gov wrote:
Matt,
When a user provides --download-packagename for any package that does not
support --download- configure silently ignores their request leading the
user to think that it actually did download the package.
On Mon, Jul 25, 2011 at 4:56 PM, Sean Farley sean at mcs.anl.gov wrote:
I disagree with this completely. There's a tendency to think of everything
python as a package -- this is NOT a package. It has no __init__.py, it has
no __all__, it has no submodules, and it doesn't need any of that.
On Wed, Feb 9, 2011 at 3:48 AM, Jed Brown jed at 59a2.org wrote:
On Wed, Feb 9, 2011 at 10:26, Matthew Knepley knepley at gmail.com wrote:
1) This is about fooling with the preprocessor, or I would not need any of
these goofy names. How can you deny this?
The preprocessor doesn't care what
On Wed, Feb 9, 2011 at 5:49 AM, Jed Brown jed at 59a2.org wrote:
On Wed, Feb 9, 2011 at 12:33, Matthew Knepley knepley at gmail.com wrote:
If you refer to the first part of your comment above, it suggested
generating the correct code,
So every function in PETSc that currently does
On Wed, Feb 9, 2011 at 6:05 AM, Jed Brown jed at 59a2.org wrote:
On Wed, Feb 9, 2011 at 12:56, Matthew Knepley knepley at gmail.com wrote:
We used to have everything in the preprocessor, and we could happily edit
that and never have to think
about Python at all. We abandoned that approach
On Wed, Feb 9, 2011 at 6:34 AM, Jed Brown jed at 59a2.org wrote:
On Wed, Feb 9, 2011 at 13:17, Matthew Knepley knepley at gmail.com wrote:
What does this sentence mean? If petscconf.h has no logic, then isn't all
the logic on the Python side? Isn't all that logic used from C?
The result
I have changed the stuff that did not belong in #requirepackage to use
either #requirefunction
or #requiredefine. No behavior should change.
Matt
--
What most experimenters take for granted before they begin their experiments
is infinitely more interesting than any results to which their
On Wed, Feb 9, 2011 at 1:21 PM, Jose E. Roman jroman at dsic.upv.es wrote:
El 09/02/2011, a las 15:02, Matthew Knepley escribi?:
I have changed the stuff that did not belong in #requirepackage to use
either #requirefunction
or #requiredefine. No behavior should change.
Matt
On Tue, Feb 15, 2011 at 4:47 PM, Barry Smith bsmith at mcs.anl.gov wrote:
In MPI one calls MPI_Comm_free(comm) to allow the MPI implementation to
set the pointer explicitly to 0 after the object is destroyed.
In Petsc XXXDestroy() does not pass the pointer (because it seemed too
unnatural
On Tue, Feb 15, 2011 at 7:58 PM, Barry Smith bsmith at mcs.anl.gov wrote:
On Feb 15, 2011, at 5:26 PM, Matthew Knepley wrote:
On Tue, Feb 15, 2011 at 4:47 PM, Barry Smith bsmith at mcs.anl.gov wrote:
In MPI one calls MPI_Comm_free(comm) to allow the MPI implementation to
set
On Tue, Feb 15, 2011 at 8:10 PM, Barry Smith bsmith at mcs.anl.gov wrote:
On Feb 15, 2011, at 8:03 PM, Matthew Knepley wrote:
On Tue, Feb 15, 2011 at 7:58 PM, Barry Smith bsmith at mcs.anl.gov wrote:
On Feb 15, 2011, at 5:26 PM, Matthew Knepley wrote:
On Tue, Feb 15, 2011 at 4:47 PM
On Tue, Feb 15, 2011 at 8:18 PM, Barry Smith bsmith at mcs.anl.gov wrote:
On Feb 15, 2011, at 8:12 PM, Matthew Knepley wrote:
On Tue, Feb 15, 2011 at 8:10 PM, Barry Smith bsmith at mcs.anl.gov wrote:
On Feb 15, 2011, at 8:03 PM, Matthew Knepley wrote:
On Tue, Feb 15, 2011 at 7:58 PM
On Sat, Feb 19, 2011 at 10:08 PM, Barry Smith bsmith at mcs.anl.gov wrote:
What about
extern PetscErrorCode VecDestroy_(Vec);
#define VecDestroy(a) (VecDestroy_(a) || (((a) = 0),0))
Not exactly PETSc style, but allows the switch without changing the API.
Why not let VecDestroy_()
I believe Jed has suggested before that we multiplex this check over
different architectures. This is also
starting to bite me and I looked at MPI.py. I think just by changing
args.petsc to args.arch.petsc we
can achieve this. Of course, it will not work for the unfortunate poop that
use make.inc,
On Tue, Feb 22, 2011 at 11:41 AM, Matthew Knepley knepley at gmail.com wrote:
I believe Jed has suggested before that we multiplex this check over
different architectures. This is also
starting to bite me and I looked at MPI.py. I think just by changing
args.petsc to args.arch.petsc we
can
I have pushed a new version:
1) F90 modules now build correctly
2) It has been refactored so that all dependency analysis is done up
front, and
all building happens afterwards.
3) It could take other dependency trackers (insert your favorite)
4) It should be ready for a nightly
:/home/balay/tmp/petsc-devar t asterix64/lib/libpetsc.a
petscsysmod.o
petscvecmod.o
petscmatmod.o
petscdmmod.o
petsckspmod.o
petscsnesmod.o
petsctsmod.o
asterix:/home/balay/tmp/petsc-dev
On Wed, 23 Feb 2011, Matthew Knepley wrote:
I have pushed a new version:
1) F90 modules now build
I had the same understanding as Jed. What was the problem which prompted the
fix?
Matt
On Mon, May 3, 2010 at 5:41 AM, Jed Brown jed at 59a2.org wrote:
On Sun, 2 May 2010 19:17:51 -0700, Barry Smith bsmith at mcs.anl.gov wrote:
Reminder, don't use
#include mylocalinclude.h
On Tue, May 4, 2010 at 12:50 PM, Jed Brown jed at 59a2.org wrote:
On Tue, 4 May 2010 12:22:10 -0500, Matthew Knepley knepley at gmail.com
wrote:
It looks like tup only has a Linux daemon, so it would run the same as
make
everywhere else. That does not seem like a strong enough case to use
11:37:14 -0500, Matthew Knepley knepley at gmail.com
wrote:
I see. Yes, it currently uses the makefile organization. This is the
kind of metadata that Barry would like in a DB rather than in
makefiles.
It would be easy to convert between being spread out in the makefiles
and being held
On Tue, May 4, 2010 at 2:42 PM, Jed Brown jed at 59a2.org wrote:
PetscFunctionBegin has logic to to avoid overflowing the array, but the
stack still gets skewed because PetscFunctionReturn always pops the
stack. This is a problem because many simple examples already overflow
the stack at
On Tue, May 4, 2010 at 3:09 PM, Jed Brown jed at 59a2.org wrote:
On Tue, 4 May 2010 14:52:32 -0500, Matthew Knepley knepley at gmail.com
wrote:
I really hate variables in macros that are necessary in other
macros. It so opaque. I do not have a better solution.
I agree, let's just live
On Thu, May 6, 2010 at 10:16 PM, Barry Smith bsmith at mcs.anl.gov wrote:
I'd like to add a MPI_Comm as the first argument to PetscError() and
friends.
In this way, if the same error is known over all the communicator ranks
it can print just one nice error message and stack instead of
I believe this is an ugly workaround for the terrible behavior of
MPI_Abort()
in some early implementations. Is this still the case?
Matt
On Thu, May 13, 2010 at 11:48 AM, Jed Brown jed at 59a2.org wrote:
I'm concerned by the following which appears in
PetscAttachDebuggerErrorHandler() and
I agree with Barry.
By making command line splits permanent you mean they are not overridden
by API splits?
Matt
On Sat, May 15, 2010 at 5:20 PM, Barry Smith bsmith at mcs.anl.gov wrote:
On May 15, 2010, at 5:09 PM, Jed Brown wrote:
Is there a case for putting a field in more than one
Kick ass. Can it run on my iPhone?
Matt
On Sat, May 15, 2010 at 9:09 PM, Barry Smith bsmith at mcs.anl.gov wrote:
Jed,
I have successfully run your ex48 on the iPad :-) Gets around 15
megaflops
Barry
--
What most experimenters take for granted before they begin their
On Tue, May 18, 2010 at 4:25 PM, Barry Smith bsmith at mcs.anl.gov wrote:
Yes it could do with some love, but that love is not coming from me.
Well, I am waiting for someone else to use it. All the examples from
Spiegelman
run correctly and he uses it. Yes, I tried to make it more general
Richard must be right. I guess we only talked about it. I can put it in
easily if we need it.
Matt
On Wed, May 19, 2010 at 2:56 AM, Richard Katz
richard.katz at earth.ox.ac.ukwrote:
I use this too (in fact, I wrote the original version), and would be very
happy to see it cleaned up.
I'm
http://insidehpc.com/2010/05/19/qlogic-aims-for-the-fences-with-infiniband-fabric-suite/
Matt
--
What most experimenters take for granted before they begin their experiments
is infinitely more interesting than any results to which their experiments
lead.
-- Norbert Wiener
-- next
On Thu, May 20, 2010 at 1:00 AM, Barry Smith bsmith at mcs.anl.gov wrote:
We need to get rid of cygwin ASAP, this is just absurd. Everyone is
having these problems!
I totally agree. Preferably before June 20. Are there free compilers that do
not require Cygwin?
Also, do we have a
Fixed.
Matt
On Fri, May 21, 2010 at 6:39 AM, Barry Smith bsmith at mcs.anl.gov wrote:
#undef __FUNCT__
#define __FUNCT__ MatPartitioningSetPartitionWeights
/*@C
MatPartitioningSetPartitionWeights - Sets the weights for each partition.
What does this mean? For each partition WHAT
Can we make Monday code freeze? I am only back Thurs night, and I need to
get the builder stuff done.
Matt
On Tue, Mar 2, 2010 at 3:05 PM, Barry Smith bsmith at mcs.anl.gov wrote:
Can we make next Monday our release date for 3.0.1?
Thanks
Barry
--
What most experimenters take
Yep, I have already setup a meeting with Dave next week to remedy
this and told Paul I would help him use it.
Matt
On Thu, Mar 4, 2010 at 10:43 AM, Barry Smith bsmith at mcs.anl.gov wrote:
I don't see a manual page here for PCBFBT!
I will do the builder.py stuff this weekend (modulo the Fortran module
stuff).
Matt
On Fri, Mar 12, 2010 at 2:53 PM, Barry Smith bsmith at mcs.anl.gov wrote:
PETSc-dev folks,
What's left to do in preparation for the 3.1 release? Can we get it done
early next week?
Thanks
On Sun, Mar 14, 2010 at 9:52 AM, Jed Brown jed at 59a2.org wrote:
On Sat, 13 Mar 2010 21:16:22 -0600, Barry Smith bsmith at mcs.anl.gov
wrote:
Perhaps the status should be set to 0 and the error to '' in the
situation where I put in the wrong values (thinking it was failure).
I tried
On Sun, Mar 14, 2010 at 2:49 PM, Jed Brown jed at 59a2.org wrote:
On Sun, 14 Mar 2010 13:29:57 -0500, Matthew Knepley knepley at gmail.com
wrote:
Okay, whatever 'yuckiness' is here is imposed on us by the threading
package. I think what is confusing you guys is the reprocessing of
output
The C Object System: Using C as a High-Level Object-Oriented Language.
(arXiv:1003.2547v1 [cs.PL]) http://arxiv.org/abs/1003.2547
Has anyone seen this?
Matt
--
What most experimenters take for granted before they begin their experiments
is infinitely more interesting than any results to
There is a request to make a more memory efficient version of ASM for
running
very large system of equations. They are using LU on small diagonal blocks,
and
have asked that the blocks be factored, applied, and discarded, rather than
being
saved for all iterations. Does anyone think this is easy
On Tue, Mar 16, 2010 at 1:51 PM, Barry Smith bsmith at mcs.anl.gov wrote:
On Mar 16, 2010, at 11:28 AM, Matthew Knepley wrote:
There is a request to make a more memory efficient version of ASM for
running
very large system of equations. They are using LU on small diagonal
blocks
On Tue, Mar 16, 2010 at 2:35 PM, Jed Brown jed at 59a2.org wrote:
On Tue, 16 Mar 2010 14:12:03 -0400, Matthew Knepley knepley at gmail.com
wrote:
You have a preconditioner which is ASM, but with a LOT of small
blocks. You do not save very much by discarding one big block, but if
you
On Wed, Mar 17, 2010 at 12:30 PM, Barry Smith bsmith at mcs.anl.gov wrote:
Some of the PETSc examples are crashing under some circumstances (like
with complex) using PLAPACK and valgrind is showing uninitialized variables
etc in PLAPACK code. I don't want to debug PLAPACK code, so will let
On Wed, Mar 17, 2010 at 12:36 PM, Barry Smith bsmith at mcs.anl.gov wrote:
Poor Jack, he should change his name immediately.
If even Robert can see past that, it tells you have fantastic Jack is.
Matt
Barry
On Mar 17, 2010, at 11:34 AM, Matthew Knepley wrote:
On Wed, Mar 17
601 - 700 of 2694 matches
Mail list logo