> Eric
>
>
> Le 19-04-19 à 19:10, Stefano Zampini via petsc-dev a écrit :
>
> Can you please send a MWE to reproduce the errors so that I can provide a
> fix?
>
> Il Ven 19 Apr 2019, 17:00 Matthew Knepley via petsc-dev <
> petsc-dev@mcs.anl.gov> ha scri
gt; Yuck, bad bug. I pushed a one-line fix to your branch. Can you let me
>> know if there are any outstanding issues? If it solves the problem for
>> you, I'll cherry-pick it to 'maint'. Thanks.
>>
>> Stefano Zampini writes:
>>
>> > In the l
> On Feb 9, 2019, at 10:19 PM, Smith, Barry F. wrote:
>
>
>
>> On Feb 9, 2019, at 11:48 AM, Stefano Zampini
>> wrote:
>>
>>
>>
>>> On Feb 9, 2019, at 11:17 AM, Smith, Barry F. wrote:
>>>
>>>
>>> Hmm, w
> On Feb 8, 2019, at 2:37 PM, Matthew Knepley wrote:
>
> On Fri, Feb 8, 2019 at 6:30 AM Stefano Zampini via petsc-dev
> mailto:petsc-dev@mcs.anl.gov>> wrote:
> I don’t understand the need of these checks (tacky and incomplete…, as per
> the comment)
>
> htt
Pierre,
The details of the inner solvers for PCFIELDSPLIT are not known if you don’t
call PCSetUp, so the cascade approach should be preferred.
After PCSetUp, you can access the inner solves, see e.g.
https://bitbucket.org/petsc/petsc/src/4ddd364d454d7bce974e92e0b5bd4f4516dbff02/src/ksp/pc/i
If I compile PETSc with complex support, all of the TAO methods are not
registered.
Fare enough? not really, I can imagine at least three valid reasons to use
those
1) I want to check that my code works for complex builds, even if I have
pure real function evaluations. I may need complex support i
* WARNING: MAKEFLAGS (set to ) found in
> environment variables - ignoring
> >use ./configure MAKEFLAGS=$MAKEFLAGS if you
> really want to use that value **
> ===========
>
> >
&g
Pierre,
Thanks to report this. I think this should be fixed in
MatHermitianTransposeAdd. Would you like to open a PR?
Thanks
Stefani
Il Sab 20 Ott 2018, 08:16 Pierre Jolivet ha
scritto:
> Hello,
> I’m using a MatNest with some blocks made of matrices created with
> MatCreateHermitianTranspose.
2018 alle ore 21:03 Stefano Zampini <
stefano.zamp...@gmail.com> ha scritto:
> mat/examples/tests/ex1.c
>
> Yes, it raises an error for the matrix types aij and mffd, but the -h
> output is still relevant. I'll open a PR tomorrow
>
>
>
> Il Dom 9 Set 2018
mat/examples/tests/ex1.c
Yes, it raises an error for the matrix types aij and mffd, but the -h
output is still relevant. I'll open a PR tomorrow
Il Dom 9 Set 2018, 21:00 Smith, Barry F. ha scritto:
>
>
> > On Sep 9, 2018, at 4:54 AM, Stefano Zampini
> wrote:
> >
&g
I just noticed a strange behaviour in master. Take mat/examples/ex1.c
./ex1 -h -mat_type mpiaij
does not print the help for the specific MPIAIJ options, only the MPIAIJ
options header
Matrix (Mat) options -
-mat_block_size <-1>: Set the bloc
Anyone already tried/succeeded in getting an AppVeyor CI for PETSc?
https://www.appveyor.com/
--
Stefano
Lawrence,
I wrote the periodic GMSH reader and I can answer for it.
With DMPlex, when we have periodicity, we currently have two ways of
encoding it
1) through the maxCell concept (or, as Lisandro's like to call it, the
"maxCell crap")
2) through the coordinatesSection
Since periodicity with ge
f the log view. If
> the same event would be nested at several levels, the obtained total time
> wouldn't say anything useful.
>
> Thanks
>
> Vaclav
>
>
>
> 2. 7. 2018 v 17:20, Stefano Zampini :
>
>
>
>
>> I don't see why this should lead to
> I don't see why this should lead to deadlock? With current class-wise
> events you can already have many simultaneous instances of the same event
> at once.
>
>
Attached a MVE to better explain myself. If you run with p processes, you
will see what I mean
--
Stefano
#include "petsc.h"
int ma
2018-07-02 11:48 GMT+03:00 Vaclav Hapla :
> Barry wrote:
>
> This could get ugly real fast, for example, for vector operations, there
> may be dozens of named vectors and each one gets its own logging? You'd
> have to make sure that only the objects you care about get named, is that
> possible?
Mark,
Why don't you use the "-mkl" flag of the Intel compiler?
Il Dom 1 Lug 2018, 21:59 Smith, Barry F. ha scritto:
>
>Mark,
>
> You need to send the full configure.log (it if doesn't fit for
> petsc-dev, better to send it to petsc-maint that doesn't mind big
> attachments.
>
>Bar
Toby,
I believe snes_tutorials-ex12_p4est_exact_q2_conformal_parallel is broken
(deadlock)
2018-07-01 15:37 GMT+03:00 Tobin Isaac :
> I think it's a good idea to include p4est in at least one test arch.
>
> On June 30, 2018 6:07:13 PM EDT, Stefano Zampini <
> stefano.zamp
Satish,
when you have time can you add p4est to some of the archs we tests nightly?
Matt and/or Toby, is there any reason why p4est is not under the testsuite?
Thanks,
Stefano
> On Jun 29, 2018, at 5:33 PM, Vaclav Hapla wrote:
>
>
>
>> 22. 6. 2018 v 17:47, Smith, Barry F. :
>>
>>
>>
>>> On Jun 22, 2018, at 5:43 AM, Pierre Jolivet
>>> wrote:
>>>
>>> Hello,
>>> I’m solving a system using a MATSHELL and PCGAMG.
>>> The MPIAIJ Mat I’m giving to GAMG has a specific
Vec and DM classes should not be visible from Sys. This is why they are
PetscObject.
If they were visible, builds with --with-single-library=0 will be broken.
2018-06-29 17:06 GMT+03:00 Patrick Sanan :
> I'm looking at the VTK viewer implementation and I notice that
> PetscViewerVTKAddField() [1
It is a simple huge file compiled. No way to do a parallel build
> On Jun 6, 2018, at 4:19 PM, Smith, Barry F. wrote:
>
>
> Petsc4py takes a long time to build on my laptop (longer than PETSc
> itself?) It appears to not use a parallel build; any chance it can be changed
> to use a parallel
> On May 11, 2018, at 6:20 PM, Smith, Barry F. wrote:
>
>
>
>> On May 11, 2018, at 8:03 AM, Stefano Zampini
>> wrote:
>>
>> I don’t think changing the current TS API is best approach.
>>
>> Obtaining separate Jacobians is a need for adjoint
I don’t think changing the current TS API is best approach.
Obtaining separate Jacobians is a need for adjoints and tangent linear models
only.
This is how I implemented it in stefano_zampini/feature-continuousadjoint
https://bitbucket.org/petsc/petsc/src/7203629c61cbeced536ed8ba1dd2ef85ffb89e8f/
gonal of the matrix
> printed earlier.
>
>> On Apr 29, 2018, at 10:02 AM, Stefano Zampini
>> wrote:
>>
>> By looking at src/mat/examples/tests/output/ex88_1.out, it seems that the
>> output of MatGetDiagonal is wrong for MATSHELL.
>>
>> it should
By looking at src/mat/examples/tests/output/ex88_1.out, it seems that the
output of MatGetDiagonal is wrong for MATSHELL.
it should be
2.96782e+08
1.41731e+09
and not (as computed)
1.10918e+08
2.06459e+08
--
Stefano
Maybe this is silly, but why the default dimenson of a DM is 0; why not -1?
A DM of dimension 0 is a point, while a DM of dimension -1 is a DM in a not
yet complete setup state.
--
Stefano
The gradient norm is the one induced by the mass matrix of the DM associated
with the control.
In principle, TaoGradientNorm() can be replaced by DMCreateMassMatrix() + solve
with the mass matrix.
For PDE constrained optimization, the “gradient norm” is crucial, since we
consider optimization p
Karl,
thanks. I have deleted the "not working properly commit" and merged to next
for testing the last commit.
Ready the go...
Stefano
2018-03-26 10:31 GMT+02:00 Karl Rupp :
> Hi Stefano,
>
>
> Next is quite messy at the moment. How are we going to handle merging to
>> master? I have a couple o
Next is quite messy at the moment. How are we going to handle merging to
master? I have a couple of branches that are sitting in next for a while,
and that can be merged safely : stefano_zampini/fix-matis-dmda-l2g
and stefano_zampini/add-pod-test
Il 23 Mar 2018 8:21 PM, "Jed Brown" ha scritto:
I
it seems we can no longer print the available configure options in master
[szampini@localhost petsc]$ ./configure --help
===
Configuring PETSc to compile on your system
===
it.
>
> --Richard
>
> On Mon, Feb 12, 2018 at 10:04 AM, Richard Tran Mills
> wrote:
>
>> On Mon, Feb 12, 2018 at 8:47 AM, Smith, Barry F.
>> wrote:
>>
>>>
>>>
>>> > On Feb 12, 2018, at 10:25 AM, Stefano Zampini <
>>> stefano.z
I would also add --with-clanguage=C++
2018-03-01 19:03 GMT+03:00 Smith, Barry F. :
>
>
> > On Mar 1, 2018, at 8:36 AM, Patrick Sanan
> wrote:
> >
> > Proposed addition to https://bitbucket.org/petsc/
> petsc/wiki/Home#markdown-header-contributing-to-petsc :
> >
> > """
> > Suggestions for a sm
-mP2OPT_hpo_vec_remainder=F;
2018-02-12 21:04 GMT+03:00 Richard Tran Mills :
> On Mon, Feb 12, 2018 at 8:47 AM, Smith, Barry F.
> wrote:
>
>>
>>
>> > On Feb 12, 2018, at 10:25 AM, Stefano Zampini <
>> stefano.zamp...@gmail.com> wrote:
>> >
>> &g
This is the branch I'm currently working on
https://bitbucket.org/petsc/petsc/branch/stefano_zampini/feature-factor-sell
So far, it provides support for MUMPS and SUPERLU
2018-02-12 19:25 GMT+03:00 Stefano Zampini :
> Barry,
>
> for sure Amat,Pmat is the right approach; however, wi
)
create two matrices (and maybe modify the code in many other parts)
else
ok with the previous code
Just my two cents.
2018-02-12 19:10 GMT+03:00 Smith, Barry F. :
>
>
> > On Feb 12, 2018, at 9:59 AM, Stefano Zampini
> wrote:
> >
> > FYI, I just checked and Ma
FYI, I just checked and MatSOR_*SELL does not use any vectorized
instruction.
Why just not converting to SeqAIJ, factor and then use the AIJ
implementation for MatSolve for the moment?
2018-02-12 18:06 GMT+03:00 Stefano Zampini :
>
>
> 2018-02-12 17:36 GMT+03:00 Jed Brown :
>
2018-02-12 17:36 GMT+03:00 Jed Brown :
> Karl Rupp writes:
>
> > Hi Stefano,
> >
> >> Is there any plan to write code for native ILU/ICC etc for SeqSELL, at
> least to have BJACOBI in parallel?
> >
> > (imho) ILU/ICC is a pain to do with SeqSELL. Point-Jacobi should be
> > possible, yes. SELL is
Is there any plan to write code for native ILU/ICC etc for SeqSELL, at least to
have BJACOBI in parallel?
;destroy routine.
>>>
>>> Shall we allow user to pick a single matrix out of submatrices array,
>>> and use/destroy it as a regular single matrix?
>>>
>>>
>>> I accept the fact that the array of submatrices obtained by
>>> MatCreateSubMatr
ubMatrices shall be destroyed by MatDestroySubMatrices.
However, not being able to take reference on one of the submats is not properly
PETSc philosophy :-), as the Mats returned do not behave like proper Mat objects
How difficult is to allow this? and what is the common data structure the
su
Cc'ing petsc-dev
-- Messaggio inoltrato --
Da: "Stefano Zampini"
Data: 12 Gen 2018 9:32 PM
Oggetto: Re: MatDestroySubMatrices_SeqAIJ and reference counting
A: "Hong"
Cc:
Sorry, I do not understand your question. Where comes matrix B?
It’s not impor
Hong,
can you explain what’s the rationale behind calling explicitly the Layout and
header destroy for the submit case in the loop here
https://bitbucket.org/petsc/petsc/src/ac3af1b492556bac9856a6aee1c73992bd0b1779/src/mat/impls/aij/seq/aij.c?at=master&fileviewer=file-view-default#aij.c-2531
<
MatSetType + MatSetLocalToGlobalMapping
is equivalent to MatCreateIS. Look at the code for MatCreateIS.
Il 13 Dic 2017 7:32 PM, "Franck Houssen" ha
scritto:
> Thanks. Working now. I missed that. I believed "MatSetType +
> MatSetLocalToGlobalMapping" <=> "MatCreateIS(..., rmap, cmap, ...)" : I go
ally, the test language was not written to
>>> generate
>>>scripts that have an `rm` command in them, so perhaps
>>> explaining why
>>> an extension would be a good idea.]
>>> I think what Stefano means is the foll
think what Stefano means is the following;
>>I write a test for the HDF5 Viewer. It creates a file called
>>"vector.dat". I want it removed at the end of the test.
>>I think the answer is, the test directory is blown away, so as
>&
How can I remove a file created by the executable I'm testing with the new
test harness?
--
Stefano
Pierre,
If you have access to the indices used to create the "unblocked" vecscatter,
you can then call ISCreateBlock and obtain the vecscatter you want.
To obtain the garray, MatMPIAIJGetSeqAIJ
> On Nov 25, 2017, at 9:26 PM, Pierre Jolivet
> wrote:
>
> OK, so I guess I need to tinker aroun
> On Oct 18, 2017, at 7:31 PM, Barry Smith wrote:
>
>
>> On Oct 18, 2017, at 4:47 AM, Stefano Zampini
>> wrote:
>>
>> I don't have a real API, just a sketch of a possible general framework (with
>> holes)
>>
>> First,
I don't have a real API, just a sketch of a possible general framework
(with holes)
First, I prefer having TSCreateAdjointTS(ts,&ats) (and possibly
SNESGetAdjointSNES for optimization with nonlinear time-independent
problems) that allows users to replace their Jacobian evaluation routines.
Th
attached the patch for ts ex2.c to reproduce the problem.
It runs 30 times the same TSSolve with a TS that has a RHSJacobian and
BEULER as a solver.
This are the times for each solve on my workstation
[szampini@localhost tutorials]$ ./ex2
0.164751
0.1406
0.148703
0.158375
0.169456
0.172364
0.17992
> >> have you thought about how to fix this?
> >>
> >> https://lists.mcs.anl.gov/pipermail/petsc-dev/2017-October/021387.html
>
> I feel like I'm missing context. I know I failed to follow up in a
> thread at some time in the past, but I don't really understand the
> problem or what is being attem
>
>Yes, we are waiting for Hong to provide his proposed API.
>
>BTW: Stefano has not really submitted a proposed API as I requested,
> just some minor explanation of his current API and minor changes.
>
>
My proposal would be based on what we are currently discussing, and it will
build on t
>
> Sure, but changing the way parameters are passed isn't an essential part
> of the adjoint interface. This particular debate spawned from Stefano's
> claim that passing the parameter vector to one of his functions was a
> feature of his interface relative to Hong's. I replied that it was an
>
> I want to say again how much I dislike ad hoc memory layouts through
> contexts
> and the like. We have a perfectly good layout descriptor (DM) that should
> be used
> to describe data layout.
>
>
Why do you say ad-hoc memory layout? the design vector is an input, and
users can decide to layout m
>What happens if you call TSCreateAdjointsTS() on a TS obtained with
> TSCreateAdjointsTS()? Is the resulting TS useful?
>
>
I don't think so. Theoretically, it should be a tangent linear model with a
different forcing function
> Barry
>
>
> > On Oct 17, 20
>
>
> I don't follow why multi-objective optimization would be any different.
> Are you supposing that the different objective functions would be
> implemented by different software packages?
>
>
It was related on how I was thinking to support multi-objective
optimization (and thus multi-gradient)
> >
> >
> >> In case of multiple objectives, there may be a performance reason to
> >> amortize evaluation of several at once, though the list interface is
> >> convenient. Consider common objectives being quantities like lift and
> >> drag on different surfaces of a fluids simulation or stress/st
Jed,
have you thought about how to fix this?
https://lists.mcs.anl.gov/pipermail/petsc-dev/2017-October/021387.html
2017-10-16 11:26 GMT+03:00 Stefano Zampini :
>
>
>>
>> Yeah, but they are sparse quadratures. That's sort of how Stefano has
>> implemented it,
>
>
> Yeah, but they are sparse quadratures. That's sort of how Stefano has
> implemented it, though the concept should be generalized and
> fixtime=PETSC_MIN_REAL is not a nice way to denote an integral.
>
>
My quadrature implementation is very primitive: it is just function
sampling and trapezoi
>
> It'd have to be TSAddObjective if you're building a list. Convention is
> that Set replaces what might have been there before.
>
>
Fair enough
> In case of multiple objectives, there may be a performance reason to
> amortize evaluation of several at once, though the list interface is
> conve
>
>> Point-form functionals (i.e., objective functions that are not integrated
>> over time, but just sampled at a given time) drive the initialization of
>> the adjoint variables, and they are not supported explicitly in Hong's
>> code. S: TSSetObjective()
>>
>
> Should these just be handled by a
If anyone wants to know what this discussion is about, see
https://bitbucket.org/petsc/petsc/pull-requests/766/support-for-pde-constrained-optimization
I'll try to summarize the interfaces here. Hong's code API is labeled with
H, mine with S.
Both methods support cost functions (i.e. objectives f
Instead of reporting a leak, the below code, when run with -log_view,
triggers an error
#include
int main(int argc,char **args)
{
PetscErrorCode ierr;
PetscViewerview;
ierr = PetscInitialize(&argc,&args,(char*)0,help);CHKERRQ(ierr);
ierr = PetscViewerASCIIGetStdout(PETSC_COMM_WORLD,
s patch [it might work on one setup - but not
>> the other?] - but Matt/Jed might have better suggestions.
>>
>> cc:ing petsc-dev
>>
>> Satish
>>
>> On Tue, 10 Oct 2017, Stefano Zampini wrote:
>>
>> > Satish,
>> >
>> > today
I was wondering what was the reason behind using a PetscScalar argument for
the function that is tested for zero-crossings in TSEvent. Why can't we use
PetscReal for it?
tsevent.c has plenty of PetscRealPart calls that seems to indicate that
PetscReal should be used instead. I can open a PR if ther
mat in PETSc? I
> would like to play around, and fix any issue I encounter.
>
Take a look at src/mat/examples/tests/ex115.c
>
> Thanks,
>
> Fande,
>
>
> On Wed, Oct 4, 2017 at 2:27 AM, Stefano Zampini > wrote:
>
>> Does this error arise at the first assem
Does this error arise at the first assembly of the matrix or in a
successive assembly?
In the second case, I guess the issue is with HYPRE, or the way we are
passing the preallocation to HYPRE routines.
For some reasons that I'm not aware of, reassembling an HypreParCSR matrix
may have issues.
My
45
2017-09-25 18:37 GMT+03:00 Stefano Zampini :
> Jed,
>
> Today I fought for several hours against a strange behaviour occuring when
> reusing a TS for multiple solves. You can reproduce it by applying the
> attached patch.
>
> The issue is with DMCoarsenHookAdd called in SNES
If you know the union of the different sparsity patterns, after you
preallocate you can set all zeros to use all the entries. This way PETSc
will not complain about new nonzeros in successive assemblies.
Il 25 Set 2017 7:01 PM, "Kong, Fande" ha scritto:
> Hi Matt,
>
> Thanks for your reply.
>
>
Jed,
Today I fought for several hours against a strange behaviour occuring when
reusing a TS for multiple solves. You can reproduce it by applying the
attached patch.
The issue is with DMCoarsenHookAdd called in SNESSetUpMatrices, and git
blames you ;-).
Every time you call TSGetRHSMats_Private,
Maybe this has been answered before, but here it is. When trying to build
HTML pages with
$ make alldoc LOC=$PETSC_DIR
With the latest master, I get this strange warning
Could not open output file
/home/szampini/src/petsc/docs/manualpages/DM/MLoad_Moab*/.html
Reason: No such file or directory
a
y want a fix. I want a code that I can run
> that demonstrates a problem (then we can evaluate fixes)
>
>Barry
>
> > On Sep 8, 2017, at 11:51 AM, Stefano Zampini
> wrote:
> >
> > I encountered the problem while working on the adjoint code. Here is the
> fix
Should the references to ADIFOR listed below be removed? It seems PETSc
does no longer provide support for automatic differentiation. I never used
AD support in PETSc, but I was considering adding some reverse mode support
to compute the transpose of the Jacobian matrices in adjoint computations.
M
asking on the list.
2017-09-08 0:02 GMT+03:00 Barry Smith :
>
> Example that demonstrates the current code is problematic in some cases?
>
> > On Aug 28, 2017, at 11:59 AM, Stefano Zampini
> wrote:
> >
> >
> > I was wondering if the following code in TSComp
As it is now, if you create any object and use PetscObjectRegisterDestroy,
the output from log_view may be misleading, as it will show something like
this
Memory usage is given in bytes:
Object Type Creations Destructions Memory Descendants' Mem.
Reports information only for proc
hat that x and y
> are the same vector.
>
> Wim
>
>
> On Wed, Aug 30, 2017 at 2:55 PM, Stefano Zampini <
> stefano.zamp...@gmail.com> wrote:
>
>> if you don't use KSPCreateVecs, but just a sequence of VecDuplicates,
>> then you can swap pointers
>>
if you don't use KSPCreateVecs, but just a sequence of VecDuplicates, then
you can swap pointers
Vec tmp = Z_Vec[pipe_l];
for(int i = pipe_l; i>0; i--){
Z_Vec[i] = Z_Vec[i-1];
}
if you still want to use KSPCreateVecs, you need to keep track of what was
the first vector of the sequence at t
I was wondering if the following code in TSComputeRHSJacobian is correct
when the user pass the reuse flag via TSRHSJacobianReuse
if (ts->rhsjacobian.time == t && (ts->problem_type == TS_LINEAR ||
(ts->rhsjacobian.Xid == Uid && ts->rhsjacobian.Xstate == Ustate)) &&
(rhsfunction != TSComputeRHSFu
I have recently noticed a regression with using Libs.private of PETSc.pc
(used by MFEM configure). For some reason, the mumps and mkl linking lines
are scrambled up. At the bottom of the message you can see the undefined
references I'm getting, coming from having libdmumps before libscalapack,
and
017, at 2:18 PM, Matthew Knepley wrote:
>
> On Mon, Aug 7, 2017 at 6:09 AM, Stefano Zampini <mailto:stefano.zamp...@gmail.com>> wrote:
> It's the first time I notice PETSC_HAVE_CONSTRAINTS in MatMult interface
>
> https://bitbucket.org/petsc/petsc/src/5ddda37bd9
ish Balay :
> Merged now
>
> Satish
>
> On Fri, 4 Aug 2017, Stefano Zampini wrote:
>
> > pushed to next for testing
> >
> > https://bitbucket.org/petsc/petsc/commits/2999d9857170fb34ae2e13612caedb
> 3c52f4aad7 <https://bitbucket.org/petsc/petsc/commits/
> 29
thanks
> On Aug 5, 2017, at 6:07 PM, Satish Balay wrote:
>
> Merged now
>
> Satish
>
> On Fri, 4 Aug 2017, Stefano Zampini wrote:
>
>> pushed to next for testing
>>
>> https://bitbucket.org/petsc/petsc/commits/2999d9857170fb34ae2e13612caedb3c52f
> On Aug 5, 2017, at 1:07 AM, Jed Brown wrote:
>
> Stefano Zampini writes:
>> This is what I’m currently doing. Storing the last used shift and don’t
>> update the matrices if the next shift is the same.
>> However, I don’t trust 100% the implementation, as PETSc, i
> On Aug 4, 2017, at 9:53 PM, Jed Brown wrote:
>
> Stefano Zampini writes:
>
>> As it is now, we have support for xdot = A(t) x and xdot = A x via
>> TSComputeRHSFunctionLinear and TSComputeRHSJacobianConstant
>>
>> We also have support for A(t) xdot = f
As it is now, we have support for xdot = A(t) x and xdot = A x via
TSComputeRHSFunctionLinear and TSComputeRHSJacobianConstant
We also have support for A(t) xdot = f(t) via TSComputeIFunctionLinear and
TSComputeIJacobianConstant
However, we miss a class of linear time-independent PDEs A xdot +
t 2:15 PM, Stefano Zampini
> wrote:
>
> It appears that dmswarm include petsc.h.
> Attached is a patch that removes this dependency.
>
> The patch also fixes some PetscViewerDraw API declared in petscdraw.h and not
> in petscviewer.h
>
> Satish, can you check if the pa
It appears that dmswarm include petsc.h.
Attached is a patch that removes this dependency.
The patch also fixes some PetscViewerDraw API declared in petscdraw.h and
not in petscviewer.h
Satish, can you check if the patch does not break anything?
I ran make alltests with a couple of configurations
017, at 9:05 AM, Jed Brown wrote:
>
> Stefano Zampini <mailto:stefano.zamp...@gmail.com>> writes:
>
>> 2017-07-24 19:28 GMT+03:00 Barry Smith :
>>
>>>
>>>
>>>>
>>>> IMHO I think that it is not worth the complicati
MatShift, MatZeroEntries…. there are for sure others that modify the entries. I
would suggest you to quickly go over the online Mat pages.
Just a curiosity: Why the inspector should care about the magnitude of the
entries?
If it’s clear in the code how to change the entry of the CSR that the ins
2017-07-24 19:28 GMT+03:00 Barry Smith :
>
>
> >
> > IMHO I think that it is not worth the complication. I value the
> generality of the current API, that already is somewhat cumbersome to
> maintain.
> > Users providing the RHS function + mass matrix already know that they
> should do M^-1 * rhsv
_steps in TSMonitor and
TSSetTrajectory? We should be really careful when modifying interface code.
2017-07-26 13:11 GMT+03:00 Stefano Zampini :
>
>
> 2017-07-25 22:45 GMT+03:00 Zhang, Hong :
>
>>
>>
>> >
>> > Actually, it is not usable: take for example TSTraj
2017-07-25 22:45 GMT+03:00 Zhang, Hong :
>
>
> >
> > Actually, it is not usable: take for example TSTrajectoryGet() with a
> given step number: then, if I want to use BASIC it will always return the
> last step recorded from the previous TSSolve (this behavior is common to
> all the TSTrajectory i
Satish
Thanks. I'll fix those tomorrow.
Stefano
Il 26 Lug 2017 12:17 AM, "Satish Balay" ha scritto:
> On Mon, 24 Jul 2017, Satish Balay wrote:
>
> > On Mon, 24 Jul 2017, Barry Smith wrote:
> >
> > > Can you please put
> > > stefano_zampini/feature-glvis-viewer-fixdmdependency2
> into next
Is TSTrajectory and its API (in the public header petscts.h) supposed to be
used from outside a TSAdjointSolve?
Actually, it is not usable: take for example TSTrajectoryGet() with a given
step number: then, if I want to use BASIC it will always return the last
step recorded from the previous TSSol
Barry, I would like to have the new GLVis based viewer for DMDAs, FEM an
FVM in.
The branch I would like to merge to next for testing and then to master is
https://bitbucket.org/petsc/petsc/branch/stefano_zampini/feature-glvis-viewer-fixdmdependency2
Thanks
Stefano
2017-07-24 11:18 GMT+03:00 Lisa
2017-07-24 11:26 GMT+03:00 Lisandro Dalcin :
> On 23 July 2017 at 21:42, Barry Smith wrote:
> >
> >Then explicit methods could automatically use a solve to apply the
> mass matrix inverse for each "right hand side" application and implicit
> methods could automatically "drop in" the mass matr
ybe I'm doing something wrong
>> in bddcschurs.c.
>>
>> Thanks,
>> Stefano
>>
>>
>> 2017-07-12 22:20 GMT+03:00 Hong :
>>
>>> MatCreateSubMatrices_xxx() calloc an array of size n+1, n=ismax and
>>> submat[n]=null as default.
Yes if you're done using it.
MAT_REUSE_MATRIX is just a flag that will skip preallocation for the
returned matrix.
2017-07-13 13:18 GMT+03:00 Franck Houssen :
> Do I need to MatDestroy the Mat returned by MatISGetMPIXAIJ ? Does the
> answer depend on reuse (MAT_INITIAL_MATRIX or MAT_REUSE_MATRIX
tures for reuse.
>
> When submat[n] != null, these data structures are destroyed.
>
> Hong
>
>
> On Wed, Jul 12, 2017 at 11:29 AM, Stefano Zampini <
> stefano.zamp...@gmail.com> wrote:
>
>> Is line number 6890 of matrix.c correct???
>>
>> https://bi
101 - 200 of 377 matches
Mail list logo