On Tue, Mar 18, 2008 at 5:00 PM, Matt Funk <[email protected]> wrote:

Hi,

i am trying out the MG preconditioner in PETSc and am encountering some issues. In fact when i use the MG preconditioner i get a segmentation violation. This does not happen with other preconditioners.

If you just do something like -pc_type mg, then you will definitely get an error. PC_MG is just the skeleton for a multigrid preconditioner. MG relies on knowledge of both the operator and domain, which is in general not available to PETSc. If you want something that just works, I suggest configuring with --download-ml or --download-hypre or --download-prometheus and running with the AMG provided in those packages. If you are using a DA, then you can use DMMG which provides the extra information. Matt

Running Petsc in debug mode with the -info option did not seem to produce any useful data (at least to me)

I thought i'd ask if someone encountered this before i use valgrind. I am using 2.3.3-p8. Or does anyone have a hint on where i could start looking? Again, what bothers me is that the other preconditioner do not throw any sort of error.

thanks mat

-- 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 Tue, Mar 25, 2008 at 6:15 PM, Matt Funk <[email protected]> wrote:

Sorry to bug the list with this.

i was just wondering how the possible use of PETSC in a commerical environment is/should be handled? I read the copyright notification on the website as well as the FAQs, but i didn't see much to that question.

If i am too blind to see it, maybe someone can point me to where i can read about it?

I think the license is clear. You can use it for whatever commercial purpose you want. Its already in a few commercial codes. Thanks, Matt

thanks mat -- 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 Tue, Mar 25, 2008 at 8:54 PM, Vijay S. Mahadevan <[email protected]> wrote:

Hi,

I am trying to use the SNESSetApplicationContext function to set an object as a user application context. This is being done after the SNESSolve was already called once on the SNES object, using a NULL application context.

For some reason, after the SNESSetApplicationContext call, the application context on SNES is not being set. I am wondering if maybe I need to set some flag to tell the SNES object to include this new context in further residual and Jacobian calls. I saw the code for the SNESSetApplicationContext function and it doesn't look like it but may be I am missing something ?

There is a misunderstanding. The application context is separate from both the Function context and the Jacobian context. If you want to change these, you must call SetFunction/Jacobian() again. Matt

Thanks for the help.

Vijay

No virus found in this outgoing message. Checked by AVG. Version: 7.5.519 / Virus Database: 269.22.0/1342 - Release Date: 3/25/2008 10:26 AM

-- 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

Matt, The first method from my previous email works fine. If I use SNES(Set/Get)ApplicationContext, I can store and retrieve user objects from the SNES object. And this is what I wanted. Thanks for the help and clarifying my mistake. Vijay

-----Original Message----- From: [email protected] [mailto:[email protected]] On Behalf Of Matthew Knepley Sent: Tuesday, March 25, 2008 9:16 PM To: [email protected] Subject: Re: SNESSetApplicationContext usage

On Tue, Mar 25, 2008 at 8:54 PM, Vijay S. Mahadevan <[email protected]> wrote:

...

Hi,

I am trying to use the SNESSetApplicationContext function to set an object as a user application context. This is being done after the SNESSolve was already called once on the SNES object, using a NULL application context.

For some reason, after the SNESSetApplicationContext call, the application context on SNES is not being set. I am wondering if maybe I need to set some flag to tell the SNES object to include this new context in further residual and Jacobian calls. I saw the code for the SNESSetApplicationContext function and it doesn't look like it but may be I am missing something ?

There is a misunderstanding. The application context is separate from both the Function context and the Jacobian context. If you want to change these, you must call SetFunction/Jacobian() again.

Matt

...

Thanks for the help.

Vijay

-- 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

No virus found in this outgoing message. Checked by AVG. Version: 7.5.519 / Virus Database: 269.22.0/1342 - Release Date: 3/25/2008 10:26 AM

Maybe i should clarify a little more. I guess what i am looking for is something that lets me split up the memory allocation within a given processor. The example you guys often show (for example on MatCreateMPIAIJ manual page) is where the local process owns 3 rows. But consider for example if the local process owns 1000 rows. Right now (as i can see it) i have two options. Do not allocate memory at all and use MatSetValues and allocate memory per row. This is very slow as you guys state on your manual pages. The other option is to allocate all 1000 rows at a time which means extra overhead. However, what i would like to be able to (as an example), make 2 calls for an allocation of 200 rows and a second call for an allocation of the remaining 800 rows. I do not know if this is possible as i don't know whether all the local stuff is stored in on continuous array? Anyway, i hope this clarifies my question a little more. thank mat On Wednesday 26 March 2008 14:31, Matt Funk wrote:

Hi,

in order to create a sparse MPIMatrix (with preallocated memory) it is necessary to give it the number of nonzero entries per row for the local and off-diagonal submatrix.

The way i do things right now is that i need to allocate the entire matrix. However, my domain is decomposed into boxes where one or more boxes reside on the current processor.

So, i was wondering if it is possible to allocate the matrix memory per box. Right now what i do is this: 1) iterate over the whole domain (every single box) and extract the info i need from each box for memory allocation. 2) call MatCreateMPIAIJ to create the matrix/allocate memory. 3) iterate over the whole domain again to get the info to insert the values into the matrix.

Is it possible to instead do something like: 1) get info for values and memory allocation needed per box 2) allocate memory for x number of rows in global matrix and insert values for them (via MatSetValues) 3) start at 1) till every box in the domain is covered.

The problem in 2) is that i can insert the values, but there is no memory preallocated for it, so it would be slow. If i could allocate it, i think it would solve my problem.

I hope i am making sense. By the way, i cannot use the DA construct because i have overlapping values.

thanks mat

Well, it is of course possible that i am looking at this wrong. as to your question what it would save: - case 1) : allocating all local rows on the local process at one time: in order to determine the information for the nonzero entries for the local on and off diagonal matrices i have to iterate over my entire domain and do logic for each node to determine whether my surrounding nodes (for the Laplacian) are on the local processor or not and/or whether they are even in the domain. Because i have to do the same thing again once i get to the point where i insert values into the matrix this is really redundant, but necessary to determine the space to allocate. i would like to eliminate this. -case 2) : do not preallocate anything very slow, and not really an option. so, what it would gain me would be the best compromise between memory allocation performance and logic overhead. I.e. the logic while iterating through the domain layout it executed only once instead of twice. I hope it makes sense... but i guess it is not possible anyway. too bad ... thanks though mat On Wednesday 26 March 2008 17:07, Matthew Knepley wrote:

This is not possible with the current implementation. I am not sure why you would want to do it. It seems to me to save nothing. Do you have some sort of performance problem?

Matt

On Wed, Mar 26, 2008 at 5:01 PM, Matt Funk <[email protected]> wrote:

...

Maybe i should clarify a little more. I guess what i am looking for is something that lets me split up the memory allocation within a given processor.

The example you guys often show (for example on MatCreateMPIAIJ manual page) is where the local process owns 3 rows.

But consider for example if the local process owns 1000 rows. Right now (as i can see it) i have two options. Do not allocate memory at all and use MatSetValues and allocate memory per row. This is very slow as you guys state on your manual pages. The other option is to allocate all 1000 rows at a time which means extra overhead.

However, what i would like to be able to (as an example), make 2 calls for an allocation of 200 rows and a second call for an allocation of the remaining 800 rows.

I do not know if this is possible as i don't know whether all the local stuff is stored in on continuous array?

Anyway, i hope this clarifies my question a little more.

thank mat

On Wednesday 26 March 2008 14:31, Matt Funk wrote:

...

Hi,

in order to create a sparse MPIMatrix (with preallocated memory) it is necessary to give it the number of nonzero entries per row for the local and off-diagonal submatrix.

The way i do things right now is that i need to allocate the entire matrix. However, my domain is decomposed into boxes where one or more boxes reside on the current processor.

So, i was wondering if it is possible to allocate the matrix memory per box. Right now what i do is this: 1) iterate over the whole domain (every single box) and extract the info i need from each box for memory allocation. 2) call MatCreateMPIAIJ to create the matrix/allocate memory. 3) iterate over the whole domain again to get the info to insert the values into the matrix.

Is it possible to instead do something like: 1) get info for values and memory allocation needed per box 2) allocate memory for x number of rows in global matrix and insert values for them (via MatSetValues) 3) start at 1) till every box in the domain is covered.

The problem in 2) is that i can insert the values, but there is no memory preallocated for it, so it would be slow. If i could allocate it, i think it would solve my problem.

I hope i am making sense. By the way, i cannot use the DA construct because i have overlapping values.

thanks mat

Matt, I do not see any specific dis-advantage in your looping over the boxes and determining the preallocation and then looping again to set the values? Surely there is no performance issue? Having support for this in PETSc would be a very dramatic change, perhaps if we understand better the problem perhaps we can come up with an alternative solution to your problem. Barry On Mar 26, 2008, at 3:31 PM, Matt Funk wrote:

Hi,

in order to create a sparse MPIMatrix (with preallocated memory) it is necessary to give it the number of nonzero entries per row for the local and off-diagonal submatrix.

The way i do things right now is that i need to allocate the entire matrix. However, my domain is decomposed into boxes where one or more boxes reside on the current processor.

So, i was wondering if it is possible to allocate the matrix memory per box. Right now what i do is this: 1) iterate over the whole domain (every single box) and extract the info i need from each box for memory allocation. 2) call MatCreateMPIAIJ to create the matrix/allocate memory. 3) iterate over the whole domain again to get the info to insert the values into the matrix.

Is it possible to instead do something like: 1) get info for values and memory allocation needed per box 2) allocate memory for x number of rows in global matrix and insert values for them (via MatSetValues) 3) start at 1) till every box in the domain is covered.

The problem in 2) is that i can insert the values, but there is no memory preallocated for it, so it would be slow. If i could allocate it, i think it would solve my problem.

I hope i am making sense. By the way, i cannot use the DA construct because i have overlapping values.

thanks mat