*On Thu, Nov 4, 2010 at 10:47, Carl Ponder wrote:*
Error! missing file:
/home/cponder/CUDA/PFlotran/petsc-dev/src/ksp/pc/impls/gasm/makefile
*On 11/04/2010 11:35 AM, Jed Brown wrote:*
> Fixed (I think), but I don't know why configure is looking into
> that directory, I co
Did you guys have a PCgasm?
Matt
On Thu, Nov 4, 2010 at 11:35 AM, Jed Brown wrote:
> On Thu, Nov 4, 2010 at 10:47, Carl Ponder wrote:
>
>> Error! missing file:
>> /home/cponder/CUDA/PFlotran/petsc-dev/src/ksp/pc/impls/gasm/makefile
>
>
> Fixed (I think), but I don't know why configure is lo
On Thu, 4 Nov 2010, Jed Brown wrote:
> On Thu, Nov 4, 2010 at 10:47, Carl Ponder wrote:
>
> > Error! missing file:
> > /home/cponder/CUDA/PFlotran/petsc-dev/src/ksp/pc/impls/gasm/makefile
>
>
> Fixed (I think), but I don't know why configure is looking into that
> directory, I couldn't reprodu
On Thu, 4 Nov 2010, Carl Ponder wrote:
>
> Satish -- here is another error from cloning just now.
> The commands
you shouldn't have to make a fresh clone each time. Once the clone is
done, you can get updates with :
hg pull -u [in both petsc-dev and BuildSystem]
And if you need a 'clean' clone
If I do it again I'll make sure I send you the logs..
On Nov 4, 2010, at 10:56 AM, Satish Balay wrote:
> On Thu, 4 Nov 2010, Bobby Philip wrote:
>
>> Satish:
>>
>>
>> On Nov 4, 2010, at 12:11 AM, Satish Balay wrote:
>>
>>> On Wed, 3 Nov 2010, Bobby Philip wrote:
>>>
As far as I can tell
Thanks guys...the pleasures of working with multiple architectures :-)
On Nov 4, 2010, at 10:56 AM, Jed Brown wrote:
> On Thu, Nov 4, 2010 at 09:47, Bobby Philip wrote:
> Is this just an extern C type issue?
>
> No, it had to do with Sun's C compiler defining __func__, but their C++
> compiler
On Thu, Nov 4, 2010 at 10:47, Carl Ponder wrote:
> Error! missing file:
> /home/cponder/CUDA/PFlotran/petsc-dev/src/ksp/pc/impls/gasm/makefile
Fixed (I think), but I don't know why configure is looking into that
directory, I couldn't reproduce it locally, and the build system definitely
doesn't
Satish:
On Nov 4, 2010, at 12:11 AM, Satish Balay wrote:
> On Wed, 3 Nov 2010, Bobby Philip wrote:
>
>> As far as I can tell SAMRAI uses petsc assuming the C style calls
>> for petsc.
>
> Yeah - the issue comes up when building PETSc with a c compiler - and
> then compiling user code [in this
Satish -- here is another error from cloning just now.
The commands
hg clone http://petsc.cs.iit.edu/petsc/petsc-dev
cd petsc-dev/config
hg clone http://petsc.cs.iit.edu/petsc/BuildSystem
cd ..
export PETSC_ARCH=arch-linux-gnu-c-debug
export PETSC_DIR=$PWD
./configure
On Thu, Nov 4, 2010 at 09:47, Bobby Philip wrote:
> Is this just an extern C type issue?
No, it had to do with Sun's C compiler defining __func__, but their C++
compiler does not unless you give a flag that, when set, caused an internal
compiler error when building Prometheus. So I made config
On Thu, 4 Nov 2010, Bobby Philip wrote:
> Satish:
>
>
> On Nov 4, 2010, at 12:11 AM, Satish Balay wrote:
>
> > On Wed, 3 Nov 2010, Bobby Philip wrote:
> >
> >> As far as I can tell SAMRAI uses petsc assuming the C style calls
> >> for petsc.
> >
> > Yeah - the issue comes up when building PET
On Wed, 3 Nov 2010, Bobby Philip wrote:
> As far as I can tell SAMRAI uses petsc assuming the C style calls
> for petsc.
Yeah - the issue comes up when building PETSc with a c compiler - and
then compiling user code [in this case SAMRAI] that includes petsc.h
with c++ compiler.
> In cases where
Guys:
As far as I can tell SAMRAI uses petsc assuming the C style calls for petsc. In
cases where
I have enabled C++ in PETSc it has broken SAMRAI.
Bobby
On Nov 3, 2010, at 7:50 PM, Satish Balay wrote:
> I think I understand the issue. I'm pushing something that should fix it.
>
> [so here -
I think I understand the issue. I'm pushing something that should fix it.
[so here - you have to do 'hg pull -u' to get this fix and rerun configure]
Jed,
The issue is: PETSc is built without a c++ compiler, but SAMRAI - is
using it with c++ - hence the error [with missing
PETSC_FUNCTION_NAME_CX
On Wed, 3 Nov 2010, Carl Ponder wrote:
> *On 11/03/2010 06:31 PM, Satish Balay wrote:*
>
> > Perhaps we are misinterpriting the errors from the snippets sent
> > earlier. Please send the complete logs [configure.log and
> > make.log] to petsc-maint so that we can diagnose the issue pr
*On 11/03/2010 06:31 PM, Satish Balay wrote:*
> Perhaps we are misinterpriting the errors from the snippets sent
> earlier. Please send the complete logs [configure.log and
> make.log] to petsc-maint so that we can diagnose the issue properly.
Satish -- I'll do that -- but could you
*On Wed, Nov 3, 2010 at 18:00, Carl Ponder mailto:CPonder at nvidia.com>> wrote:*
./$PETSC_ARCH/conf/reconfigure-$PETSC_ARCH.py
*On 11/03/2010 06:10 PM, Jed Brown wrote:*
>
> In principle, you should always rerun this. Sometimes you can get
> away with not doing it, but it is ex
On Wed, 3 Nov 2010, Carl Ponder wrote:
> *On 11/03/2010 06:07 PM, Satish Balay wrote:*
> >
> > You have some contradictory statements here:
> >
> > " Note that while I pulled BuildSystem,"
> > " although I'm skipping the steps"
> > " hg pull -u"
> >
> > How did you pull Buil
On Wed, Nov 3, 2010 at 6:00 PM, Carl Ponder wrote:
> *On 11/03/2010 05:43 PM, Satish Balay wrote:*
>
> Perhaps you pulled petsc-dev - but not BuildSystem - so they are out-of-sync?
>
>
> Satish -- here are the commands I used:
>
> hg clone http://petsc.cs.iit.edu/petsc/petsc-dev
> cd petsc-dev/
*On 11/03/2010 06:07 PM, Satish Balay wrote:*
>
> You have some contradictory statements here:
>
> " Note that while I pulled BuildSystem,"
> " although I'm skipping the steps"
> " hg pull -u"
>
> How did you pull BuildSystem - if not with 'hg pull -u'? [same
> with petsc-de
On Wed, Nov 3, 2010 at 18:00, Carl Ponder wrote:
> ./$PETSC_ARCH/conf/reconfigure-$PETSC_ARCH.py
In principle, you should always rerun this. Sometimes you can get away with
not doing it, but it is exactly what is needed to fix your current problem.
Jed
-- next part --
You have some contradictory statements here:
" Note that while I pulled BuildSystem,"
" although I'm skipping the steps"
" hg pull -u"
How did you pull BuildSystem - if not with 'hg pull -u'? [same with petsc-dev]
[I'm assuming you've done 'hg clone' a few days ago - not today]
Just to be su
*On 11/03/2010 05:43 PM, Satish Balay wrote:*
> Perhaps you pulled petsc-dev - but not BuildSystem - so they are
> out-of-sync?
Satish -- here are the commands I used:
hg clone http://petsc.cs.iit.edu/petsc/petsc-dev
cd petsc-dev/config
hg clone http://petsc.cs.iit.edu/petsc/Bu
On Wed, Nov 3, 2010 at 5:43 PM, Satish Balay wrote:
> Perhaps you pulled petsc-dev - but not BuildSystem - so they are
> out-of-sync?
>
With petsc-dev, sometimes changes are made that require reconfiguration
(like this one)
since the include is generated. If you pull BuildSystem and rerun config
Perhaps you pulled petsc-dev - but not BuildSystem - so they are out-of-sync?
Satish
On Wed, 3 Nov 2010, Carl Ponder wrote:
>
> Satash -- the filepetsc-dev/include/petscsys.h contains a statement
>
>25 #if defined(__cplusplus)
>*26 # define PETSC_FUNCTION_NAME PETSC_FUNCTION_NAME_
Satash -- the filepetsc-dev/include/petscsys.h contains a statement
25 #if defined(__cplusplus)
*26 # define PETSC_FUNCTION_NAME PETSC_FUNCTION_NAME_CXX*
27 #else
28 # define PETSC_FUNCTION_NAME PETSC_FUNCTION_NAME_C
29 #endif
I don't see thisPETSC_FUNCTION_NAME_C
On Wed, 3 Nov 2010, Matthew Knepley wrote:
> Since you use our version control system, you will never have to worry
> about something breaking. If something does not work after a pull, just
>
> a) rollback to the previous revision, 'hg update -r '
Or if you don't remember the previous 'working
On Wed, Nov 3, 2010 at 3:12 PM, Satish Balay wrote:
> On Wed, 3 Nov 2010, Carl Ponder wrote:
>
> > *On 11/03/2010 01:33 PM, Satish Balay wrote:*
> >
> > > for http there is no '$USER:$PASSWORD@' in the URL [I guess its
> just
> > > ignored - so it works..]
> > > I pushed a fix now.. You c
On Wed, 3 Nov 2010, Carl Ponder wrote:
> *On 11/03/2010 01:33 PM, Satish Balay wrote:*
>
> > for http there is no '$USER:$PASSWORD@' in the URL [I guess its just
> > ignored - so it works..]
> > I pushed a fix now.. You can pull and retry [after 'hg revert configure'
> > - to remove your
*On 11/03/2010 01:33 PM, Satish Balay wrote:*
> for http there is no '$USER:$PASSWORD@' in the URL [I guess its just
> ignored - so it works..]
> I pushed a fix now.. You can pull and retry [after 'hg revert configure'
> - to remove your local changes]
Satish -- thanks -- it works now.
On Wed, 3 Nov 2010, Satish Balay wrote:
> >./configure --with-cc=gcc --with-fc=mpif90
> >--download-f-blas-lapack=1 --download-mpich=1 --download-hdf5=1
BTW: the above is inconsistant.
If you already have mpif90 - then you should have a corresponding
mpicc to use [ so use '--with-cc=mpic
On Wed, 3 Nov 2010, Carl Ponder wrote:
>
> I just copied thepetsc-dev subtree using the commands
>
>hg clone http://$USER:$PASSWORD at petsc.cs.iit.edu/petsc/petsc-dev
for http there is no '$USER:$PASSWORD@' in the URL [I guess its just ignored -
so it works..]
>cd petsc-dev/config
I just copied thepetsc-dev subtree using the commands
hg clone http://$USER:$PASSWORD at petsc.cs.iit.edu/petsc/petsc-dev
cd petsc-dev/config
hg clone http://$USER:$PASSWORD at petsc.cs.iit.edu/petsc/BuildSystem
cd ..
export PETSC_DIR=$PWD
./configure --with-cc=gcc --with-
33 matches
Mail list logo