[petsc-users] "Must select a target sorting criterion if using shift-and-invert"

[Jose E. Roman]

This was added in 3.8 to check the common case when people incorrectly sets shift-and-invert with EPS_SMALLEST_MAGNITUDE. To compute smallest eigenvalues with shift-and-invert the correct way is to set target=0 and which=EPS_TARGET_MAGNITUDE. See for instance 
http://slepc.upv.es/documentation/current/src/eps/examples/tutorials/ex13.c.html

Jose


> El 21 oct 2017, a las 1:51, Matthew Knepley <knepley at gmail.com> escribió:
> 
> On Fri, Oct 20, 2017 at 7:43 PM, Kong, Fande <fande.kong at inl.gov> wrote:
> Hi All,
> 
> I am trying to solve a generalized eigenvalue problem (using SLEPc) with "-eps_type krylovschur -st_type sinvert". I got an error message: "Must select a target sorting criterion if using shift-and-invert".
> 
> Not sure how to proceed.  I do not quite understand this sentence.
> 
> You need to know how to choose the shift. So for instance you want the smallest eigenvalues, or the closest to zero, etc.
> I don't know the options, but they are in the manual.
> 
>    Matt
>  
> 
> Fande,
> 
> 
> 
> -- 
> 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
> 
> https://www.cse.buffalo.edu/~knepley/