>> PetscMallocAlign,
>>>>>> ?FP=7fff342ca080
>>>>>> PetscTrMallocDefault,
>>>>>> ?FP=7fff342ca180
>>>>>> PetscStrallocpy,
>>>>>> FP=7fff342ca230
>>>>>> PetscFListGetPathAndFunction,
>>>>>> ?FP=7fff342cb2e0
>>>>>> PetscFListFind,
>>>>>> ?FP=7fff342cb520
>>>>>> PetscObjectQueryFunction_Petsc,
>>>>>> ?FP=7fff342cb590
>>>>>> PetscObjectQueryFunction,
>>>>>> ?FP=7fff342cb620
>>>>>> MatMumpsSetIcntl,
>>>>>> ?FP=7fff342cb720
>>>>>> MatMumpsSetIcntl,
>>>>>> ?FP=7fff342cb820
>>>>>> MatMumpsSetIcntl,
>>>>>> ?FP=7fff342cb920
>>>>>> MatMumpsSetIcntl,
>>>>>> ?FP=7fff342cba20
>>>>>> MatMumpsSetIcntl,
>>>>>> ?FP=7fff342cbb20
>>>>>> MatMumpsSetIcntl,
>>>>>> ?FP=7fff342cbc20
>>>>>> MatMumpsSetIcntl,
>>>>>> ?FP=7fff342cbd20
>>>>>> MatMumpsSetIcntl,
>>>>>> ?FP=7fff342cbe20
>>>>>> MatMumpsSetIcntl,
>>>>>> ?FP=7fff342cbf20
>>>>>> ... (Hundreds of MatMumpsSetIcntl) ...
>>>>>>
>>>>>> What can I do about that?
>>>>>>
>>>>>> Regards,
>>>>>> Alexander
>>>>
>>>>
>
-- next part --
A non-text attachment was scrubbed...
Name: ex57f.F
Type: application/octet-stream
Size: 13471 bytes
Desc: not available
URL:
<http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20111201/0dc52308/attachment.obj>
Thanks. If this is the case, will the option MAT_IGNORE_ZERO_ENTRIES be helpful?
Xiangdong
On Thu, Dec 1, 2011 at 8:00 PM, Barry Smith wrote:
>
> On Dec 1, 2011, at 6:44 PM, Xiangdong Liang wrote:
>
>> Hello everyone,
>>
>> I have a question about whether zeros will change the sparse pattern.
>>
The checks in the code are lines like
if (value == 0.0 && ignorezeroentries && (is == ADD_VALUES)) continue;
hence changing the source code to ignore small values is not difficult. But I
think this is a terribly cheesy hack and would hope people would not do this.
Barry
On De
On Dec 1, 2011, at 8:46 PM, Xiangdong Liang wrote:
> Thanks. If this is the case, will the option MAT_IGNORE_ZERO_ENTRIES be
> helpful?
What do you mean by "be helpful"? Sometimes you do want those "zero"
locations represented and sometimes you don't, so sometimes you would use the
option
Hello everyone,
I have a question about whether zeros will change the sparse pattern.
Suppose I am generating two sparse matrices A and B. A is generated
with exactly number of nonzeros, say 10 nnz per row. When I generate
matrix B, I specify 12 nonzeros, but two of them are given zero values
by M
On Dec 1, 2011, at 6:44 PM, Xiangdong Liang wrote:
> Hello everyone,
>
> I have a question about whether zeros will change the sparse pattern.
> Suppose I am generating two sparse matrices A and B. A is generated
> with exactly number of nonzeros, say 10 nnz per row. When I generate
> matrix B,
introduce an increase in the pattern.
>
>
>BTW: if you then load the matrix via a binary file to MATLAB it may be
> that MATLAB silently removes those locations so you don't see them.
> Barry
>
> > Thanks.
> >
> > Best,
> > Xiangdong
>
>
---
write
accessors so that I can do the following:
Vector a (PETSC_COMM_WORLD, PETSC_DECIDE,10); // make a petsc vector of global
length 10 using my own wrappers
for (int i=0; ihttp://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20111201/9db53cf2/attachment.htm>
On Dec 1, 2011, at 4:48 PM, Mohammad Mirzadeh wrote:
> Hi guys,
>
> I am having a conceptual problem trying to design my vector wrappers for my
> application. The problem is some of vector functionality in PETSc require
> extra function calls which makes it hard to design a good wrapper. One
re they begin their
experiments is infinitely more interesting than any results to which their
experiments lead.
-- Norbert Wiener
-- next part --
An HTML attachment was scrubbed...
URL:
<http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20111201/1110a18f/attachment.htm>
omehow nullify the pointer to prevent accidental use of pointer?
> Since all pointers all private members of my class I assume this is safe
> then unless these two functions are actually doing some extra work?
> >
> > 2) Is there any better way of writing these wrappers and get around the
> problem of having to call "supplemental" functions safely?
> >
> > Sorry if my lack of knowledge is causing all these confusion.
> >
> > Thanks,
> > Mohammad
> >
> >
>
>
-- next part --
An HTML attachment was scrubbed...
URL:
<http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20111201/349256ac/attachment.htm>
;> problem of having to call "supplemental" functions safely?
>>
>> Sorry if my lack of knowledge is causing all these confusion.
>>
>> Thanks,
>> Mohammad
>>
>>
>>
>
>
> --
> 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 part --
An HTML attachment was scrubbed...
URL:
<http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20111201/34c4f649/attachment-0001.htm>
th is the following: I am trying to write
accessors so that I can do the following:
Vector a (PETSC_COMM_WORLD, PETSC_DECIDE,10); // make a petsc vector of
global length 10 using my own wrappers
for (int i=0; ihttp://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20111201/e91b2754/attachment.htm>
I have tried to use the following, this was before I had both you're
emails (Barry, Matthew) ;-)
1) First I set this up
PetscInt size = 7;
MatCreate(MPI_COMM_SELF,&LHS);
MatSetSizes(LHS,size,size,size,size);
MatSetType(LHS,MATSEQDENSE); //**NOTE THIS LINE
My approach will be a million times faster.
Barry
On Dec 1, 2011, at 8:50 AM, Matthew Knepley wrote:
> On Thu, Dec 1, 2011 at 8:44 AM, Parker, Andrew (UK Filton) baesystems.com> wrote:
> I have tried to use the following, this was before I had both you're
> emails (Barry, Matthew) ;-)
>
>>>
>> Yes. The same matrices yields a speedup of 4-6x when solving the system
>> on the GPU.
>>
>
> Please confirm that the correct routine by running wth -info and sending
> the output.
>
> Please send the output of -log_summary so I can confirm the results.
>
> You can run KSP ex4 and reproduce my results where I see a 5.5x speedup on
> the GTX285
>
> I am not sure what to look for in those outputs. I have uploaded the
output of running my assembly program with -info and -log_summary, and the
output of running ex4 with -log_summary. See
http://folk.uio.no/fredva/assembly_info.txt
http://folk.uio.no/fredva/assembly_log_summary.txt
http://folk.uio.no/fredva/ex4_log_summary.txt
Trying this on a different machine now, I actually see some speedup. 3rd
order Poisson on 300x300 assembles in 0.211 sec on GPU and 0.4232 sec on
CPU. For 1st order and 1000x1000 mesh, I go from 0.31 sec to 0.205 sec.
I have tried to increase the mesh size to see if the speedup increases, but
I hit the bad_alloc error pretty quick.
For a problem of that size, should I expect even more speedup? Please let
me know if you need any more output from test runs on my machine.
--
Fredrik
-- next part --
An HTML attachment was scrubbed...
URL:
<http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20111201/460366cb/attachment.htm>
y of Chicago
5735 S. Ellis Avenue
Chicago, IL 60637
---
Phone: 630-252-1229
Fax: 630-252-5986
-- next part --
An HTML attachment was scrubbed...
URL:
<http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20111201/a1803c73/attachment.htm>
-252-1229
Fax: 630-252-5986
-- next part --
An HTML attachment was scrubbed...
URL:
<http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20111201/0caf9b48/attachment.htm>
-- next part --
An HTML attachment was scrubbed...
URL:
<http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20111201/5506b483/attachment.htm>
Hong,
This should be enough to reproduce the problem:
call KSPCreate(PETSC_COMM_WORLD,ksp,ierr);CHKERRQ(ierr)
call
KSPSetOperators(ksp,A,A,DIFFERENT_NONZERO_PATTERN,ierr);CHKERRQ(ierr)
call KSPSetType(ksp,KSPPREONLY,ierr);CHKERRQ(ierr)
call KSPGetPC(ksp,pc,ierr);CHKERRQ(ierr)
call
****
> >
> >
> >
> >
> > --
> > 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
>
>
>
--
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 part --
An HTML attachment was scrubbed...
URL:
<http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20111201/4275a819/attachment.htm>
On Dec 1, 2011, at 7:05 AM, Matthew Knepley wrote:
> You can get this effect using PCBJACOBI, setting the number of local blocks,
> and using -sub_pc_type lu -sub_ksp_type preonly.
Do not do this. It will give bad performance since it explicitly pulls out
each block with MatGetSubMatrices()
which their
experiments lead.
-- Norbert Wiener
-- next part --
An HTML attachment was scrubbed...
URL:
<http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20111201/01fa51b5/attachment.htm>
dispatching to the CUDA version. The regular
>>>> version just calls MatSetValues() in a loop. Are you
>>>> using a SEQAIJCUSP matrix?
>>>>
>>> Yes. The same matrices yields a speedup of 4-6x when solving the system
>>> on the GPU.
>&g
24 matches
Mail list logo