[petsc-dev] aij and baij data access

Jose David Bermeol jbermeol at purdue.edu
Tue Feb 4 10:01:54 CST 2014


Looking the code for MatGetRowIJ_SeqAIJ_Inode_Nonsymmetric, it is copying the data and I want to avoid that. So, is there other way to do this??

----- Original Message -----
From: "Matthew Knepley" <knepley at gmail.com>
To: "Jed Brown" <jed at jedbrown.org>
Cc: "Jose David Bermeol" <jbermeol at purdue.edu>, "petsc-dev" <petsc-dev at mcs.anl.gov>
Sent: Tuesday, February 4, 2014 9:21:34 AM
Subject: Re: [petsc-dev] aij and baij data access




On Tue, Feb 4, 2014 at 7:41 AM, Jed Brown < jed at jedbrown.org > wrote: 


Matthew Knepley < knepley at gmail.com > writes: 

> On Mon, Feb 3, 2014 at 10:22 AM, Jose David Bermeol < jbermeol at purdue.edu >wrote: 
> 
>> Hi, Is it possible to access to the raw data of aij matrices(row indices, 
>> column indices and data). I need to used for and optimize algorithm, this 
>> data will keep constant all the time. So is it cossible with the public 
>> interface you have or not?? 
>> 
> 
> This is a question for petsc-maint. 

I disagree; there is no reason to keep such questions private. 



Let me rephrase: This is a question for petsc-maint (my default suggestion so as not to force 
people into the public) or petsc-users, not petsc-dev which is for development discussion. 




> http://www.mcs.anl.gov/petsc/petsc-dev/docs/manualpages/Mat/MatGetColumnIJ.html 

You actually want MatGetRowIJ because MatGetColumnIJ is expensive with a 
row-based format like AIJ. 



Yes, bad click. 


Matt 


> http://www.mcs.anl.gov/petsc/petsc-dev/docs/manualpages/Mat/MatSeqAIJGetArray.html 




-- 
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 



More information about the petsc-dev mailing list