[petsc-users] About GMRES Solver in PETSc

[Chih-Hao Chen]

Hello,


Thanks for the quick reply.

So if I specify matrix A and RHS vector b in parallel forms,

ie. having them distributed across several cores,

when I using the functions like MatMult, VecAYPX and etc.,

all the operations would be executed in parallel too?

Thanks very much.


Best,

Chih-Hao

________________________________
From: Hong <hzhang at mcs.anl.gov>
Sent: 18 January 2017 15:24:41
To: Chih-Hao Chen
Cc: petsc-users at mcs.anl.gov
Subject: Re: [petsc-users] About GMRES Solver in PETSc

Yes.
Hong

On Wed, Jan 18, 2017 at 2:22 PM, Chih-Hao Chen <chih-hao.chen2 at mail.mcgill.ca<mailto:chih-hao.chen2 at mail.mcgill.ca>> wrote:

Hello,


Sorry for this bother.

I"ve checked  GMRES solver in your libary supports parallel computing in your website:

http://www.mcs.anl.gov/petsc/documentation/linearsolvertable.html.<http://www.mcs.anl.gov/petsc/documentation/linearsolvertable.html>

Summary of Sparse Linear Solvers Available from PETSc<http://www.mcs.anl.gov/petsc/documentation/linearsolvertable.html>
www.mcs.anl.gov<http://www.mcs.anl.gov>
Summary of Sparse Linear Solvers Available from PETSc Requests and contributions welcome

As we all know the key algorithm inside GMRES solver is Gram-Schmidt process,

so I guess if users specify their matrix A and right-hand-side vector b in a linear equation (Ax=b),

the Gram-Schmidt process will be executed in parallel,

including matrix-vector multiplications and orthogonalization between the basis vectors.

Am I correct?

Thanks very much.


Best,

Chih-Hao


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20170118/f8d1c060/attachment.html>