GSoC

newer
GSoC 2014 - Full Text Search

older
[Move] BOOST_NO_RVALUE_REFERENCES...

Utsav Tiwary

1 Feb 2014 1 Feb '14

10:31 a.m.

Are there existing libraries for the implementation of finite difference method to solve boundary value problems ? I've been working on it lately using MPI.

Show replies by date

Larry Evans

1 Feb 1 Feb

1:18 p.m.

On 02/01/14 04:31, Utsav Tiwary wrote:

...

Are there existing libraries for the implementation of finite difference method to solve boundary value problems ? I've been working on it lately using MPI.

odeint was reviewed around Sept. 2012.

Utsav Tiwary

2:47 p.m.

Does it involve the use of parallel programming ? If not, is doing it a good idea ? On Sat, Feb 1, 2014 at 6:48 PM, Larry Evans wrote:

...

On 02/01/14 04:31, Utsav Tiwary wrote:

...
Are there existing libraries for the implementation of finite difference method to solve boundary value problems ? I've been working on it lately using MPI.

odeint was reviewed around Sept. 2012.

_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/ mailman/listinfo.cgi/boost

Larry Evans

3:46 p.m.

On 02/01/14 08:47, Utsav Tiwary wrote:

...

Does it involve the use of parallel programming ? If not, is doing it a good idea ?

I don't know. I'd suggest contacting the author or reading the docs ( I don't know where they are. You'll have to search the mailing list archives, or maybe wait till someone more knowledgeable replies :( )

...

On Sat, Feb 1, 2014 at 6:48 PM, Larry Evans wrote:

...
On 02/01/14 04:31, Utsav Tiwary wrote:

...
Are there existing libraries for the implementation of finite difference method to solve boundary value problems ? I've been working on it lately using MPI.

odeint was reviewed around Sept. 2012.

_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/ mailman/listinfo.cgi/boost

_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

Utsav Tiwary

5:29 p.m.

Hopefully someone will reply ... thanks :) On Sat, Feb 1, 2014 at 9:16 PM, Larry Evans wrote:

...

On 02/01/14 08:47, Utsav Tiwary wrote:

...
Does it involve the use of parallel programming ? If not, is doing it a good idea ?

I don't know. I'd suggest contacting the author or reading the docs ( I don't know where they are. You'll have to search the mailing list archives, or maybe wait till someone more knowledgeable replies :( )

...
On Sat, Feb 1, 2014 at 6:48 PM, Larry Evans wrote:

On 02/01/14 04:31, Utsav Tiwary wrote:

...
Are there existing libraries for the implementation of finite difference

...
method to solve boundary value problems ? I've been working on it lately using MPI.

odeint was reviewed around Sept. 2012.

_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/ mailman/listinfo.cgi/boost

_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/ mailman/listinfo.cgi/boost

_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/ mailman/listinfo.cgi/boost

Karsten Ahnert

7:19 p.m.

On 02/01/2014 03:47 PM, Utsav Tiwary wrote:

...

Does it involve the use of parallel programming ? If not, is doing it a good idea ?

There was a GSOC last year about parallelization of odeint with MPI and OpenMP [1]. I think it was successful, although the code has not yet been merged into the odeint trunk. Parallelization with GPUs already works with the help of Thrust, VexCL or ViennaCL. odeint is a library devoted to solve the initial value problem of ODEs. I think building shooting methods for boundary value problems on top of odeint would be a good idea. And we also would like to mentor a GSOC project about implicit methods in odeint. There are extremely important to solve PDEs. [1] https://github.com/BoostGSoC/odeint-v2

...

On Sat, Feb 1, 2014 at 6:48 PM, Larry Evans wrote:

...
On 02/01/14 04:31, Utsav Tiwary wrote:

...
Are there existing libraries for the implementation of finite difference method to solve boundary value problems ? I've been working on it lately using MPI.

odeint was reviewed around Sept. 2012.

_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/ mailman/listinfo.cgi/boost

_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

Larry Evans

8:31 p.m.

On 02/01/14 13:19, Karsten Ahnert wrote:

...

On 02/01/2014 03:47 PM, Utsav Tiwary wrote:

...
Does it involve the use of parallel programming ? If not, is doing it a good idea ?

There was a GSOC last year about parallelization of odeint with MPI and OpenMP [1]. I think it was successful, although the code has not yet been merged into the odeint trunk. Parallelization with GPUs already works with the help of Thrust, VexCL or ViennaCL.

odeint is a library devoted to solve the initial value problem of ODEs. I think building shooting methods for boundary value problems on top of odeint would be a good idea. And we also would like to mentor a GSOC project about implicit methods in odeint. There are extremely important to solve PDEs.

Back in June, 2011, petros posted a question to the user's list about multi-array and pdes: http://thread.gmane.org/gmane.comp.lib.boost.user/68017 Later, he posted a question about rotating the arrays axes: http://article.gmane.org/gmane.comp.lib.boost.user/68718/match=multi_array+r... to help solve a pde using ADI (Alternating Direction Implicit): http://article.gmane.org/gmane.comp.lib.boost.user/71096 In response, I prototyped a method using ADI and mentioned it here: http://article.gmane.org/gmane.comp.lib.boost.user/71047 HTH. -regards, Larry

Christopher Kormanyos

2 Feb 2 Feb

11 a.m.

...

And we also would like to mentor a GSOC project about implicit methods in odeint. There are extremely important to solve PDEs.

* Is the idea ripe enough for formulation? * Is there a committed mentor? Might the potential mentor formally add the project to the list here: https://svn.boost.org/trac/boost/wiki/SoC2014 Or is the potential project already there, and I'm too dumb to recognize it? Cheers, Chris. On Saturday, February 1, 2014 9:31 PM, Larry Evans wrote: On 02/01/14 13:19, Karsten Ahnert wrote:

...

On 02/01/2014 03:47 PM, Utsav Tiwary wrote:

...
Does it involve the use of parallel programming ? If not, is doing it a good idea ?

There was a GSOC last year about parallelization of odeint with MPI and OpenMP [1]. I think it was successful, although the code has not yet been merged into the odeint trunk. Parallelization with GPUs already works with the help of Thrust, VexCL or ViennaCL.

odeint is a library devoted to solve the initial value problem of ODEs. I think building shooting methods for boundary value problems on top of odeint would be a good idea. And we also would like to mentor a GSOC project about implicit methods in odeint. There are extremely important to solve PDEs.

Utsav Tiwary

12:59 p.m.

No one has yet agreed to mentor the project idea. The idea does seem ripe though. On Sun, Feb 2, 2014 at 4:30 PM, Christopher Kormanyos wrote:

...

...
And we also would like to mentor a GSOC project about implicit methods in odeint. There are extremely important to solve PDEs.

* Is the idea ripe enough for formulation? * Is there a committed mentor?

Might the potential mentor formally add the project to the list here:

https://svn.boost.org/trac/boost/wiki/SoC2014

Or is the potential project already there, and I'm too dumb to recognize it?

Cheers, Chris.

On Saturday, February 1, 2014 9:31 PM, Larry Evans < cppljevans@suddenlink.net> wrote:

On 02/01/14 13:19, Karsten Ahnert wrote:

...
On 02/01/2014 03:47 PM, Utsav Tiwary wrote:

...
Does it involve the use of parallel programming ? If not, is doing it a good idea ?

There was a GSOC last year about parallelization of odeint with MPI and OpenMP [1]. I think it was successful, although the code has not yet been merged into the odeint trunk. Parallelization with GPUs already works with the help of Thrust, VexCL or ViennaCL.

odeint is a library devoted to solve the initial value problem of ODEs. I think building shooting methods for boundary value problems on top of odeint would be a good idea. And we also would like to mentor a GSOC project about implicit methods in odeint. There are extremely important to solve PDEs.

Back in June, 2011, petros posted a question to the user's list about multi-array and pdes:

http://thread.gmane.org/gmane.comp.lib.boost.user/68017

Later, he posted a question about rotating the arrays axes:

http://article.gmane.org/gmane.comp.lib.boost.user/68718/match=multi_array+r...

to help solve a pde using ADI (Alternating Direction Implicit):

http://article.gmane.org/gmane.comp.lib.boost.user/71096

In response, I prototyped a method using ADI and mentioned it here:

http://article.gmane.org/gmane.comp.lib.boost.user/71047

HTH.

-regards,

Larry

_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

Karsten Ahnert

2:50 p.m.

On 02/02/2014 12:00 PM, Christopher Kormanyos wrote:

...

...
And we also would like to mentor a GSOC project about implicit methods in odeint. There are extremely important to solve PDEs.

* Is the idea ripe enough for formulation?

Yes, we had several GSOC project proposals about implicit ODE solvers last year but none of them had been accepted. Is this a problem for applications this year?

...

* Is there a committed mentor?

Might the potential mentor formally add the project to the list here:

https://svn.boost.org/trac/boost/wiki/SoC2014

I would volunteer to mentor this project. Of course, I will add the description to the ideas page.

...

On Saturday, February 1, 2014 9:31 PM, Larry Evans wrote:

On 02/01/14 13:19, Karsten Ahnert wrote:

...
On 02/01/2014 03:47 PM, Utsav Tiwary wrote:

...
Does it involve the use of parallel programming ? If not, is doing it a good idea ?

There was a GSOC last year about parallelization of odeint with MPI and OpenMP [1]. I think it was successful, although the code has not yet been merged into the odeint trunk. Parallelization with GPUs already works with the help of Thrust, VexCL or ViennaCL.

odeint is a library devoted to solve the initial value problem of ODEs. I think building shooting methods for boundary value problems on top of odeint would be a good idea. And we also would like to mentor a GSOC project about implicit methods in odeint. There are extremely important to solve PDEs.

Back in June, 2011, petros posted a question to the user's list about multi-array and pdes:

http://thread.gmane.org/gmane.comp.lib.boost.user/68017

Later, he posted a question about rotating the arrays axes:

http://article.gmane.org/gmane.comp.lib.boost.user/68718/match=multi_array+r...

to help solve a pde using ADI (Alternating Direction Implicit):

http://article.gmane.org/gmane.comp.lib.boost.user/71096

In response, I prototyped a method using ADI and mentioned it here:

http://article.gmane.org/gmane.comp.lib.boost.user/71047

HTH.

-regards,

Larry

_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

Utsav Tiwary

3:08 p.m.

Thankyou for showing interest in the idea Mr. Ahnert. I would like to know what pre-requisite knowledge do you expect any candidate to have prior to starting to work on this project ? On Sun, Feb 2, 2014 at 8:20 PM, Karsten Ahnert < karsten.ahnert@googlemail.com> wrote:

...

On 02/02/2014 12:00 PM, Christopher Kormanyos wrote:

...
...
And we also would like to mentor a GSOC project about implicit methods in odeint. There are extremely important to solve PDEs.

* Is the idea ripe enough for formulation?

Yes, we had several GSOC project proposals about implicit ODE solvers last year but none of them had been accepted. Is this a problem for applications this year?

...
* Is there a committed mentor?

Might the potential mentor formally add the project to the list here:

https://svn.boost.org/trac/boost/wiki/SoC2014

I would volunteer to mentor this project. Of course, I will add the description to the ideas page.

...
On Saturday, February 1, 2014 9:31 PM, Larry Evans < cppljevans@suddenlink.net> wrote:

On 02/01/14 13:19, Karsten Ahnert wrote:

...
On 02/01/2014 03:47 PM, Utsav Tiwary wrote:

...
Does it involve the use of parallel programming ? If not, is doing it a good idea ?

There was a GSOC last year about parallelization of odeint with MPI and OpenMP [1]. I think it was successful, although the code has not yet been merged into the odeint trunk. Parallelization with GPUs already works with the help of Thrust, VexCL or ViennaCL.

odeint is a library devoted to solve the initial value problem of ODEs. I think building shooting methods for boundary value problems on top of odeint would be a good idea. And we also would like to mentor a GSOC project about implicit methods in odeint. There are extremely important to solve PDEs.

Back in June, 2011, petros posted a question to the user's list about multi-array and pdes:

http://thread.gmane.org/gmane.comp.lib.boost.user/68017

Later, he posted a question about rotating the arrays axes:

http://article.gmane.org/gmane.comp.lib.boost.user/68718/match=multi_array+r...

...
to help solve a pde using ADI (Alternating Direction Implicit):

http://article.gmane.org/gmane.comp.lib.boost.user/71096

In response, I prototyped a method using ADI and mentioned it here:

http://article.gmane.org/gmane.comp.lib.boost.user/71047

HTH.

-regards,

Larry

_______________________________________________ Unsubscribe & other changes:

http://lists.boost.org/mailman/listinfo.cgi/boost

...
_______________________________________________ Unsubscribe & other changes:

http://lists.boost.org/mailman/listinfo.cgi/boost

...
_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

Karsten Ahnert

7:54 p.m.

On 02/02/2014 04:08 PM, Utsav Tiwary wrote:

...

Thankyou for showing interest in the idea Mr. Ahnert. I would like to know what pre-requisite knowledge do you expect any candidate to have prior to starting to work on this project ?

I have added a description of the implicit routines project idea and basic requirements for possible candidates at https://svn.boost.org/trac/boost/wiki/SoC2014 It would also be good to have a basic idea how odeint works internally. If you are asking about the boundary value problem I need to know more about the idea and the methods to see wether I could mentor such a project. So a working description would be very useful.

...

On Sun, Feb 2, 2014 at 8:20 PM, Karsten Ahnert < karsten.ahnert@googlemail.com> wrote:

...
On 02/02/2014 12:00 PM, Christopher Kormanyos wrote:

...
...
And we also would like to mentor a GSOC project about implicit methods in odeint. There are extremely important to solve PDEs.

* Is the idea ripe enough for formulation?

Yes, we had several GSOC project proposals about implicit ODE solvers last year but none of them had been accepted. Is this a problem for applications this year?

...
* Is there a committed mentor?

Might the potential mentor formally add the project to the list here:

https://svn.boost.org/trac/boost/wiki/SoC2014

I would volunteer to mentor this project. Of course, I will add the description to the ideas page.

...
On Saturday, February 1, 2014 9:31 PM, Larry Evans < cppljevans@suddenlink.net> wrote:

On 02/01/14 13:19, Karsten Ahnert wrote:

...
On 02/01/2014 03:47 PM, Utsav Tiwary wrote:

...
Does it involve the use of parallel programming ? If not, is doing it a good idea ?

There was a GSOC last year about parallelization of odeint with MPI and OpenMP [1]. I think it was successful, although the code has not yet been merged into the odeint trunk. Parallelization with GPUs already works with the help of Thrust, VexCL or ViennaCL.

odeint is a library devoted to solve the initial value problem of ODEs. I think building shooting methods for boundary value problems on top of odeint would be a good idea. And we also would like to mentor a GSOC project about implicit methods in odeint. There are extremely important to solve PDEs.

Back in June, 2011, petros posted a question to the user's list about multi-array and pdes:

http://thread.gmane.org/gmane.comp.lib.boost.user/68017

Later, he posted a question about rotating the arrays axes:

http://article.gmane.org/gmane.comp.lib.boost.user/68718/match=multi_array+r...

...
to help solve a pde using ADI (Alternating Direction Implicit):

http://article.gmane.org/gmane.comp.lib.boost.user/71096

In response, I prototyped a method using ADI and mentioned it here:

http://article.gmane.org/gmane.comp.lib.boost.user/71047

HTH.

-regards,

Larry

_______________________________________________ Unsubscribe & other changes:

http://lists.boost.org/mailman/listinfo.cgi/boost

...
_______________________________________________ Unsubscribe & other changes:

http://lists.boost.org/mailman/listinfo.cgi/boost

...
_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

Utsav Tiwary

8:01 p.m.

Currently I am working on a semester long project to learn various numerical methods like Gauss Elimination, Gauss Jacobi, Gauss Seidel methods etc. to learn how to do the computations using MPI(Message Passing Interface). Later I will be learning how to solve boundary value problem using finite difference method using the same tool - MPI. The idea is to develop a library which could make use of parallel programming to achieve the objective of solving boundary value problems. I would love to hear suggestions regarding the same. On Mon, Feb 3, 2014 at 1:24 AM, Karsten Ahnert < karsten.ahnert@googlemail.com> wrote:

...

On 02/02/2014 04:08 PM, Utsav Tiwary wrote:

...
Thankyou for showing interest in the idea Mr. Ahnert. I would like to know what pre-requisite knowledge do you expect any candidate to have prior to starting to work on this project ?

I have added a description of the implicit routines project idea and basic requirements for possible candidates at

https://svn.boost.org/trac/boost/wiki/SoC2014

It would also be good to have a basic idea how odeint works internally.

If you are asking about the boundary value problem I need to know more about the idea and the methods to see wether I could mentor such a project. So a working description would be very useful.

...
On Sun, Feb 2, 2014 at 8:20 PM, Karsten Ahnert < karsten.ahnert@googlemail.com> wrote:

...
On 02/02/2014 12:00 PM, Christopher Kormanyos wrote:

...
...
And we also would like to mentor a GSOC project about implicit methods in odeint. There are extremely important to solve PDEs.

* Is the idea ripe enough for formulation?

Yes, we had several GSOC project proposals about implicit ODE solvers last year but none of them had been accepted. Is this a problem for applications this year?

...
* Is there a committed mentor?

Might the potential mentor formally add the project to the list here:

https://svn.boost.org/trac/boost/wiki/SoC2014

I would volunteer to mentor this project. Of course, I will add the description to the ideas page.

...
On Saturday, February 1, 2014 9:31 PM, Larry Evans < cppljevans@suddenlink.net> wrote:

On 02/01/14 13:19, Karsten Ahnert wrote:

...
On 02/01/2014 03:47 PM, Utsav Tiwary wrote:

...
Does it involve the use of parallel programming ? If not, is doing

it a

...
...
...
...
...
good idea ?

There was a GSOC last year about parallelization of odeint with MPI and OpenMP [1]. I think it was successful, although the code has not yet been merged into the odeint trunk. Parallelization with GPUs already works with the help of Thrust, VexCL or ViennaCL.

odeint is a library devoted to solve the initial value problem of ODEs. I think building shooting methods for boundary value problems on top of odeint would be a good idea. And we also would like to mentor a GSOC project about implicit methods in odeint. There are extremely important to solve PDEs.

Back in June, 2011, petros posted a question to the user's list about multi-array and pdes:

http://thread.gmane.org/gmane.comp.lib.boost.user/68017

Later, he posted a question about rotating the arrays axes:

http://article.gmane.org/gmane.comp.lib.boost.user/68718/match=multi_array+r...

...
...
to help solve a pde using ADI (Alternating Direction Implicit):

http://article.gmane.org/gmane.comp.lib.boost.user/71096

In response, I prototyped a method using ADI and mentioned it here:

http://article.gmane.org/gmane.comp.lib.boost.user/71047

HTH.

-regards,

Larry

_______________________________________________ Unsubscribe & other changes:

http://lists.boost.org/mailman/listinfo.cgi/boost

...
_______________________________________________ Unsubscribe & other changes:

http://lists.boost.org/mailman/listinfo.cgi/boost

...
_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

Karsten Ahnert

8:07 p.m.

On 02/02/2014 09:01 PM, Utsav Tiwary wrote:

...

Currently I am working on a semester long project to learn various numerical methods like Gauss Elimination, Gauss Jacobi, Gauss Seidel methods etc. to learn how to do the computations using MPI(Message Passing Interface). Later I will be learning how to solve boundary value problem using finite difference method using the same tool - MPI. The idea is to develop a library which could make use of parallel programming to achieve the objective of solving boundary value problems. I would love to hear suggestions regarding the same.

What exactly do you mean by boundary value problems? And how are finite difference are related to this problem?

...

On Mon, Feb 3, 2014 at 1:24 AM, Karsten Ahnert < karsten.ahnert@googlemail.com> wrote:

...
On 02/02/2014 04:08 PM, Utsav Tiwary wrote:

...
Thankyou for showing interest in the idea Mr. Ahnert. I would like to know what pre-requisite knowledge do you expect any candidate to have prior to starting to work on this project ?

I have added a description of the implicit routines project idea and basic requirements for possible candidates at

https://svn.boost.org/trac/boost/wiki/SoC2014

It would also be good to have a basic idea how odeint works internally.

If you are asking about the boundary value problem I need to know more about the idea and the methods to see wether I could mentor such a project. So a working description would be very useful.

...
On Sun, Feb 2, 2014 at 8:20 PM, Karsten Ahnert < karsten.ahnert@googlemail.com> wrote:

...
On 02/02/2014 12:00 PM, Christopher Kormanyos wrote:

...
...
And we also would like to mentor a GSOC project about implicit methods in odeint. There are extremely important to solve PDEs.

* Is the idea ripe enough for formulation?

Yes, we had several GSOC project proposals about implicit ODE solvers last year but none of them had been accepted. Is this a problem for applications this year?

...
* Is there a committed mentor?

Might the potential mentor formally add the project to the list here:

https://svn.boost.org/trac/boost/wiki/SoC2014

I would volunteer to mentor this project. Of course, I will add the description to the ideas page.

...
On Saturday, February 1, 2014 9:31 PM, Larry Evans < cppljevans@suddenlink.net> wrote:

On 02/01/14 13:19, Karsten Ahnert wrote:

...
On 02/01/2014 03:47 PM, Utsav Tiwary wrote: > Does it involve the use of parallel programming ? If not, is doing

it a

...
...
...
...
> good idea ?

There was a GSOC last year about parallelization of odeint with MPI and OpenMP [1]. I think it was successful, although the code has not yet been merged into the odeint trunk. Parallelization with GPUs already works with the help of Thrust, VexCL or ViennaCL.

odeint is a library devoted to solve the initial value problem of ODEs. I think building shooting methods for boundary value problems on top of odeint would be a good idea. And we also would like to mentor a GSOC project about implicit methods in odeint. There are extremely important to solve PDEs.

Back in June, 2011, petros posted a question to the user's list about multi-array and pdes:

http://thread.gmane.org/gmane.comp.lib.boost.user/68017

Later, he posted a question about rotating the arrays axes:

http://article.gmane.org/gmane.comp.lib.boost.user/68718/match=multi_array+r...

...
...
to help solve a pde using ADI (Alternating Direction Implicit):

http://article.gmane.org/gmane.comp.lib.boost.user/71096

In response, I prototyped a method using ADI and mentioned it here:

http://article.gmane.org/gmane.comp.lib.boost.user/71047

HTH.

-regards,

Larry

_______________________________________________ Unsubscribe & other changes:

http://lists.boost.org/mailman/listinfo.cgi/boost

...
_______________________________________________ Unsubscribe & other changes:

http://lists.boost.org/mailman/listinfo.cgi/boost

...
_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

Utsav Tiwary

8:13 p.m.

Wikipedia explains it concisely and clearly http://en.wikipedia.org/wiki/Boundary_value_problem Also, to get a brief idea about Finite difference Method - http://www.stanford.edu/~fringer/teaching/numerical_methods_02/handouts/lect... . This is a good link to start with. Pages 7 -10 explain the method. On Mon, Feb 3, 2014 at 1:37 AM, Karsten Ahnert < karsten.ahnert@googlemail.com> wrote:

...

On 02/02/2014 09:01 PM, Utsav Tiwary wrote:

...
Currently I am working on a semester long project to learn various numerical methods like Gauss Elimination, Gauss Jacobi, Gauss Seidel methods etc. to learn how to do the computations using MPI(Message Passing Interface). Later I will be learning how to solve boundary value problem using finite difference method using the same tool - MPI. The idea is to develop a library which could make use of parallel programming to achieve the objective of solving boundary value problems. I would love to hear suggestions regarding the same.

What exactly do you mean by boundary value problems? And how are finite difference are related to this problem?

...
On Mon, Feb 3, 2014 at 1:24 AM, Karsten Ahnert < karsten.ahnert@googlemail.com> wrote:

...
On 02/02/2014 04:08 PM, Utsav Tiwary wrote:

...
Thankyou for showing interest in the idea Mr. Ahnert. I would like to know what pre-requisite knowledge do you expect any candidate to have prior

to

...
...
...
starting to work on this project ?

I have added a description of the implicit routines project idea and basic requirements for possible candidates at

https://svn.boost.org/trac/boost/wiki/SoC2014

It would also be good to have a basic idea how odeint works internally.

If you are asking about the boundary value problem I need to know more about the idea and the methods to see wether I could mentor such a project. So a working description would be very useful.

...
On Sun, Feb 2, 2014 at 8:20 PM, Karsten Ahnert < karsten.ahnert@googlemail.com> wrote:

...
On 02/02/2014 12:00 PM, Christopher Kormanyos wrote:

...
> And we also would like to mentor a GSOC project about implicit > methods in odeint. There are extremely important to solve PDEs.

* Is the idea ripe enough for formulation?

Yes, we had several GSOC project proposals about implicit ODE solvers last year but none of them had been accepted. Is this a problem for applications this year?

...
* Is there a committed mentor?

Might the potential mentor formally add the project to the list here:

https://svn.boost.org/trac/boost/wiki/SoC2014

I would volunteer to mentor this project. Of course, I will add the description to the ideas page.

...
On Saturday, February 1, 2014 9:31 PM, Larry Evans < cppljevans@suddenlink.net> wrote:

On 02/01/14 13:19, Karsten Ahnert wrote: > On 02/01/2014 03:47 PM, Utsav Tiwary wrote: >> Does it involve the use of parallel programming ? If not, is doing

it a

...
...
...
>> good idea ? > > There was a GSOC last year about parallelization of odeint with MPI and > OpenMP [1]. I think it was successful, although the code has not yet > been merged into the odeint trunk. Parallelization with GPUs already > works with the help of Thrust, VexCL or ViennaCL. > > odeint is a library devoted to solve the initial value problem of ODEs. > I think building shooting methods for boundary value problems on top of > odeint would be a good idea. And we also would like to mentor a GSOC > project about implicit methods in odeint. There are extremely important > to solve PDEs. > Back in June, 2011, petros posted a question to the user's list about multi-array and pdes:

http://thread.gmane.org/gmane.comp.lib.boost.user/68017

Later, he posted a question about rotating the arrays axes:

http://article.gmane.org/gmane.comp.lib.boost.user/68718/match=multi_array+r...

...
...
...
...
to help solve a pde using ADI (Alternating Direction Implicit):

http://article.gmane.org/gmane.comp.lib.boost.user/71096

In response, I prototyped a method using ADI and mentioned it here:

http://article.gmane.org/gmane.comp.lib.boost.user/71047

HTH.

-regards,

Larry

_______________________________________________ Unsubscribe & other changes:

http://lists.boost.org/mailman/listinfo.cgi/boost

...
_______________________________________________ Unsubscribe & other changes:

http://lists.boost.org/mailman/listinfo.cgi/boost

...
_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

Christopher Kormanyos

3:15 p.m.

...

Yes, we had several GSOC project proposals about implicit ODE solvers last year but none of them had been accepted. Is this a problem for applications this year?

Last year was my first year participating as mentor, so I am no real expert. Nonetheless, we had 10 projects, but we received only 7 funded slots. This year we have (together with your project) so far 3 projects, and a lot of good ideas. So it looks like we may have fewer projects. I do not know how this will influence our application or the success thereof.

...

I would volunteer to mentor this project. Of course, I will add the description to the ideas page.

Thanks Karsten. OK, wake up from winter hibernation, Boost! :-) Let's get a few more good projects with commited mentors up on our list! https://svn.boost.org/trac/boost/wiki/SoC2014 Cheers, Chris. On Sunday, February 2, 2014 3:50 PM, Karsten Ahnert wrote: On 02/02/2014 12:00 PM, Christopher Kormanyos wrote:

...

...
And we also would like to mentor a GSOC project about implicit methods in odeint. There are extremely important to solve PDEs.

* Is the idea ripe enough for formulation?

Yes, we had several GSOC project proposals about implicit ODE solvers last year but none of them had been accepted. Is this a problem for applications this year?

...

* Is there a committed mentor?

Might the potential mentor formally add the project to the list here:

https://svn.boost.org/trac/boost/wiki/SoC2014

I would volunteer to mentor this project. Of course, I will add the description to the ideas page.

...

On Saturday, February 1, 2014 9:31 PM, Larry Evans wrote: On 02/01/14 13:19, Karsten Ahnert wrote:

...
On 02/01/2014 03:47 PM, Utsav Tiwary wrote:

...
Does it involve the use of parallel programming ? If not, is doing it a good idea ?

There was a GSOC last year about parallelization of odeint with MPI and OpenMP [1]. I think it was successful, although the code has not yet been merged into the odeint trunk. Parallelization with GPUs already works with the help of Thrust, VexCL or ViennaCL.

odeint is a library devoted to solve the initial value problem of ODEs. I think building shooting methods for boundary value problems on top of odeint would be a good idea. And we also would like to mentor a GSOC project about implicit methods in odeint. There are extremely important to solve PDEs.

Back in June, 2011, petros posted a question to the user's list about multi-array and pdes:

http://thread.gmane.org/gmane.comp.lib.boost.user/68017

Later, he posted a question about rotating the arrays axes:

http://article.gmane.org/gmane.comp.lib.boost.user/68718/match=multi_array+r...

to help solve a pde using ADI (Alternating Direction Implicit):

http://article.gmane.org/gmane.comp.lib.boost.user/71096

In response, I prototyped a method using ADI and mentioned it here:

http://article.gmane.org/gmane.comp.lib.boost.user/71047

HTH.

-regards,

Larry

_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

...

_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

Boris Schäling

3 Feb 3 Feb

7:48 a.m.

On Sun, 02 Feb 2014 16:15:38 +0100, Christopher Kormanyos wrote:

...

[...]OK, wake up from winter hibernation, Boost! :-)

Let's get a few more good projects with commited mentors up on our list!

https://svn.boost.org/trac/boost/wiki/SoC2014

Thanks for pushing this a bit, Chris! The application period for GSoC 2014 starts in a few hours. If we (Boost) want to participate in this year's program again it's time to add all project ideas to https://svn.boost.org/trac/boost/wiki/SoC2014. Feel free to add ideas even if you know that you won't be able to work on them yourself. Don't think only about libraries but maybe unit tests or any utilities which could help Boost manage source code better after the switch from Subversion to Git. And if anyone is interested in becoming this year's GSoC admin - let me know! Boris

...

[...]

shubhv＠iitk.ac.in

1:42 p.m.

New subject: Regarding GSoC

Hey, I am a sophomore student of Computer Science Department, IIT Kanpur, India. I have been going through the boost libraries for quite some time now, and also have read through the proposed GSoC projects in 2014. I am especially interested in the Boost.Math Generalized Hyper geometric Functions. Though I have gone through the links provided under that project and am trying to become versed with Boost's Math library, I am finding it difficult to find a starting point to contributing. I would be really glad if someone could help me in the initial phase by directing me to the right starting point and guide me through. Thanks Shubhangee Verma

Christopher Kormanyos

8:27 p.m.

New subject: Regarding GSoC

Well, one potential first step would be to see if you can start working with the detailed sample code provided in the link on the GSoC projects page. It's a lot of heavy code, almost 1,000 lines, but it's clean and well commented. Try to compile the code. We just fixed a minor bug with GCC 4.8.1 today, so we should be good to go on that sample file for, say GCC 4.6-ish through GCC 4.8.2, and VC10 through VC12. I only used a few C++11 features such as lambdas. See if you can pull the file and get it to run on a compiler of your choice. Best regards, Chris. On Monday, February 3, 2014 2:42 PM, "shubhv@iitk.ac.in" wrote: Hey, I am a sophomore student of Computer Science Department, IIT Kanpur, India. I have been going through the boost libraries for quite some time now, and also have read through the proposed GSoC projects in 2014. I am especially interested in the Boost.Math Generalized Hyper geometric Functions. Though I have gone through the links provided under that project and am trying to become versed with Boost's Math library, I am finding it difficult to find a starting point to contributing. I would be really glad if someone could help me in the initial phase by directing me to the right starting point and guide me through. Thanks Shubhangee Verma _______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

David kiarie

4 Feb 4 Feb

9:49 a.m.

New subject: Regarding GSoC

I am also a student and I think what can help me most is an idea of what is needed next.What libraries need to developed next. Yes now we have the Boost.Math Generalized Hyper geometric Functions I feel that I really have nothing objective when I just compile the code and cross my fingers for errors to occur *I know they won't*. I bet you can advice me on how am I supposed to look into this,what do you as the author of the library thinks need more effort? Well that would be really helpfull On Mon, Feb 3, 2014 at 11:27 PM, Christopher Kormanyos wrote:

...

Well, one potential first step would be to see if you can start working with the detailed sample code provided in the link on the GSoC projects page.

It's a lot of heavy code, almost 1,000 lines, but it's clean and well commented.

Try to compile the code.

We just fixed a minor bug with GCC 4.8.1 today, so we should be good to go on that sample file for, say GCC 4.6-ish through GCC 4.8.2, and VC10 through VC12.

I only used a few C++11 features such as lambdas.

See if you can pull the file and get it to run on a compiler of your choice.

Best regards, Chris.

On Monday, February 3, 2014 2:42 PM, "shubhv@iitk.ac.in" < shubhv@iitk.ac.in> wrote:

Hey, I am a sophomore student of Computer Science Department, IIT Kanpur, India. I have been going through the boost libraries for quite some time now, and also have read through the proposed GSoC projects in 2014. I am especially interested in the Boost.Math Generalized Hyper geometric Functions. Though I have gone through the links provided under that project and am trying to become versed with Boost's Math library, I am finding it difficult to find a starting point to contributing.

I would be really glad if someone could help me in the initial phase by directing me to the right starting point and guide me through.

Thanks

Shubhangee Verma

_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

Utsav Tiwary

3 Feb 3 Feb

3:32 p.m.

Last year the topic of the project mentored by Boost was - "*Boost uBlas offers a wide range of algorithms which can be applied on vectors or matrices. These calculations are mostly done on very substantial sized objects, which significantly increases the computation time. However, the time performance of a program is one of the most important criteria nowadays. The aim of our project is to enable the parallel execution thanks to the MPI technology*." I think the idea we are discussing in this thread is a good extension to the application of parallel computing and High Performance Computing framework to mathematical methods of solving boundary value problems. On Mon, Feb 3, 2014 at 1:18 PM, Boris Schäling wrote:

...

On Sun, 02 Feb 2014 16:15:38 +0100, Christopher Kormanyos < e_float@yahoo.com> wrote:

[...]OK, wake up from winter hibernation, Boost! :-)

...
Let's get a few more good projects with commited mentors up on our list!

https://svn.boost.org/trac/boost/wiki/SoC2014

Thanks for pushing this a bit, Chris! The application period for GSoC 2014 starts in a few hours. If we (Boost) want to participate in this year's program again it's time to add all project ideas to < https://svn.boost.org/trac/boost/wiki/SoC2014>. Feel free to add ideas even if you know that you won't be able to work on them yourself. Don't think only about libraries but maybe unit tests or any utilities which could help Boost manage source code better after the switch from Subversion to Git.

And if anyone is interested in becoming this year's GSoC admin - let me know!

Boris

[...]

...
_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/ mailman/listinfo.cgi/boost

Niall Douglas

6:37 p.m.

On 3 Feb 2014 at 8:48, Boris Schäling wrote:

...

And if anyone is interested in becoming this year's GSoC admin - let me know!

Are you saying you are resigning Boris? If so, has Andrew decided on if he is interested in retaking the role? Niall -- Currently unemployed and looking for work in Ireland. Work Portfolio: http://careers.stackoverflow.com/nialldouglas/

Boris Schäling

6:52 p.m.

On Mon, 03 Feb 2014 19:37:57 +0100, Niall Douglas wrote:

...

On 3 Feb 2014 at 8:48, Boris Schäling wrote:

...
And if anyone is interested in becoming this year's GSoC admin - let me know!

Are you saying you are resigning Boris? If so, has Andrew decided on

Hehe, not resigning really. :) But if anyone wants to take over that would be perfectly fine with me. I can offer to help out as a deputy admin if desired. Boris

...

[...]

Niall Douglas

7:31 p.m.

On 3 Feb 2014 at 19:52, Boris Schäling wrote:

...

...
Are you saying you are resigning Boris? If so, has Andrew decided on

Hehe, not resigning really. :) But if anyone wants to take over that would be perfectly fine with me. I can offer to help out as a deputy admin if desired.

Did you not see my earlier offer to you by private email that if and only if you don't want it, and if and only if Andrew doesn't want it, I would be willing. After all, unemployment has advantages! Niall -- Currently unemployed and looking for work in Ireland. Work Portfolio: http://careers.stackoverflow.com/nialldouglas/

Utsav Tiwary

8 Feb 8 Feb

2:06 p.m.

So , is someone willing to mentor this idea ? .. and any suggestions for changes/improvements that can be done to this project idea? On Tue, Feb 4, 2014 at 1:01 AM, Niall Douglas wrote:

...

On 3 Feb 2014 at 19:52, Boris Schäling wrote:

...
...
Are you saying you are resigning Boris? If so, has Andrew decided on

Hehe, not resigning really. :) But if anyone wants to take over that would be perfectly fine with me. I can offer to help out as a deputy admin if desired.

Did you not see my earlier offer to you by private email that if and only if you don't want it, and if and only if Andrew doesn't want it, I would be willing. After all, unemployment has advantages!

Niall

-- Currently unemployed and looking for work in Ireland. Work Portfolio: http://careers.stackoverflow.com/nialldouglas/

_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

Karsten Ahnert

2:32 p.m.

On 02/08/2014 03:06 PM, Utsav Tiwary wrote:

...

So , is someone willing to mentor this idea ? .. and any suggestions for changes/improvements that can be done to this project idea?

In principle I would be interested in mentoring such kind of project. But I need to know more about the idea and if it could be integrated into odeint. If I understand the difference method correctly there is no need for any ode solver. That is a difference compared to shooting methods which are build on top of a solver. I think it would be very good to build something general for boundary value problems, and which integrates with existing libraries and concepts.

Niall Douglas

4:35 p.m.

On 8 Feb 2014 at 19:36, Utsav Tiwary wrote:

...

So , is someone willing to mentor this idea ? .. and any suggestions for changes/improvements that can be done to this project idea?

I think Utsav our problem is that we don't know: (a) what exactly are you proposing. (b) how you plan to solve it. (c) whether you have the technical nous to implement your solution. A big advantage of using one of the mentor supplied GSoC work items from https://svn.boost.org/trac/boost/wiki/SoC2014 is that the mentor knows the problem, knows the solution and knows what is needed from a student. Do remember that we auto-reject half of all GSoC student candidates because they simply don't have the skills at writing basic coherent English, let alone writing code. Last year we saw around 70 students apply, but I think we only seriously considered about 20 such was how awful their project proposals were. Utsav: here is what I suggest: look at the project proposals on https://svn.boost.org/trac/boost/wiki/SoC2014 and write up what you want to do in the same format and post it here. We'll be far better able to advise you and figure out a suitable mentor if your ideas are in that format. And thanks for showing the interest Utsav, we really like students who come early. Niall -- Currently unemployed and looking for work in Ireland. Work Portfolio: http://careers.stackoverflow.com/nialldouglas/

Utsav Tiwary

5:25 p.m.

Thank you so much Mr. Niall. Your answer clarified a lot of things. I have ideas and I shall try to pen them down clearly. Also, I do not deny that I am not the most consummate programmer you would find, but just the fact that I have a keen interest in programming and the FAQ that said that the applicants need not be the most skilled ones, but the more passionate ones have motivated me to apply for GSoC 2014. Thank you for appreciating my effort. Also, could you please specify the deadline you expect of me to put the project idea in the proper format so that I try to get it done as early as possible ? On Sat, Feb 8, 2014 at 10:05 PM, Niall Douglas wrote:

...

On 8 Feb 2014 at 19:36, Utsav Tiwary wrote:

...
So , is someone willing to mentor this idea ? .. and any suggestions for changes/improvements that can be done to this project idea?

I think Utsav our problem is that we don't know:

(a) what exactly are you proposing.

(b) how you plan to solve it.

(c) whether you have the technical nous to implement your solution.

A big advantage of using one of the mentor supplied GSoC work items from https://svn.boost.org/trac/boost/wiki/SoC2014 is that the mentor knows the problem, knows the solution and knows what is needed from a student. Do remember that we auto-reject half of all GSoC student candidates because they simply don't have the skills at writing basic coherent English, let alone writing code. Last year we saw around 70 students apply, but I think we only seriously considered about 20 such was how awful their project proposals were.

Utsav: here is what I suggest: look at the project proposals on https://svn.boost.org/trac/boost/wiki/SoC2014 and write up what you want to do in the same format and post it here. We'll be far better able to advise you and figure out a suitable mentor if your ideas are in that format.

And thanks for showing the interest Utsav, we really like students who come early.

Niall

-- Currently unemployed and looking for work in Ireland. Work Portfolio: http://careers.stackoverflow.com/nialldouglas/

_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

Niall Douglas

9 Feb 9 Feb

1:31 p.m.

On 8 Feb 2014 at 22:55, Utsav Tiwary wrote:

...

Thank you so much Mr. Niall. Your answer clarified a lot of things. I have ideas and I shall try to pen them down clearly. Also, I do not deny that I am not the most consummate programmer you would find, but just the fact that I have a keen interest in programming and the FAQ that said that the applicants need not be the most skilled ones, but the more passionate ones have motivated me to apply for GSoC 2014.

That is definitely the case here at Boost: we would prefer enthusiasm and hard workers to absolute skill, but you must understand that there is a minimum coding skill level. For example, if a student has never used std::vector<T> before, they probably don't have the minimum coding skill in C++ level we would need.

...

Thank you for appreciating my effort. Also, could you please specify the deadline you expect of me to put the project idea in the proper format so that I try to get it done as early as possible ?

I know you will need a lot of time to look over existing Boost libraries and to formalise the mathematics you are proposing. Usually most features added to Boost are already implemented in some other C or C++ or Python library out there, so a list of those and how they solved your problem (or if they didn't solve your problem) proves you did your prior art research correctly. Boost takes much more seriously those who can prove they have researched the problem before coming here with proposed solutions. Send whatever you write to this list by the 16th, so one week from now. Don't worry if you find writing down the work item hard, just send us whatever you write. Good luck! Niall --- Boost C++ Libraries Google Summer of Code 2014 admin https://svn.boost.org/trac/boost/wiki/SoC2014

3880

Age (days ago)

3888

Last active (days ago)

List overview

Download

28 comments

8 participants

participants (8)

Boris Schäling
Christopher Kormanyos
David kiarie
Karsten Ahnert
Larry Evans
Niall Douglas
shubhv＠iitk.ac.in
Utsav Tiwary

GSoC

Utsav Tiwary

Larry Evans

Utsav Tiwary

Larry Evans

Utsav Tiwary

Karsten Ahnert

Larry Evans

Christopher Kormanyos

Utsav Tiwary

Karsten Ahnert

Utsav Tiwary

Karsten Ahnert

Utsav Tiwary

Karsten Ahnert

Utsav Tiwary

Christopher Kormanyos

Boris Schäling

shubhv＠iitk.ac.in

Christopher Kormanyos

David kiarie

Utsav Tiwary

Niall Douglas

Boris Schäling

Niall Douglas

Utsav Tiwary

Karsten Ahnert

Niall Douglas

Utsav Tiwary

Niall Douglas

tags

participants (8)