[petsc-dev] Deflated Krylov solvers for PETSc

[Jie Chen]

I second this. How many subdomains for a 1 billion unknown problem?

Jie


----- Original Message -----
From: "Jed Brown" <jedbrown at mcs.anl.gov>
To: "For users of the development version of PETSc" <petsc-dev at mcs.anl.gov>
Sent: Friday, March 1, 2013 11:43:35 PM
Subject: Re: [petsc-dev] Deflated Krylov solvers for PETSc


An alternative description for the method in the paper is "two-level unsmoothed aggregation applied to the ASM-preconditioned operator". 


When there are many processes/subdomains, the dimension of the coarse space can be huge. 


Extreme eigenvectors would be more useful, but you would still need a very large number to scale to many subdomains.