[petsc-users] SNES for 1D problem?

[Barry Smith]

> On Mar 8, 2017, at 4:24 PM, Adrian Croucher <a.croucher at auckland.ac.nz> wrote:
> 
> hi
> 
> If part of my code needs a one-dimensional nonlinear solver (i.e. to solve for a single scalar variable), which for some problems will be called many times, does it still make sense to use SNES for that? or are there overheads from the vector machinery that make it less efficient for that case than rolling your own simple solver?

   Should be ok if you are using the same SNES object. If you create the SNES object fresh for each one; no reason you would do this; that might be noticeable.

   Barry

> 
> - Adrian
> 
> -- 
> Dr Adrian Croucher
> Senior Research Fellow
> Department of Engineering Science
> University of Auckland, New Zealand
> email: a.croucher at auckland.ac.nz
> tel: +64 (0)9 923 4611
>