[petsc-dev] Line search

[Barry Smith]

> On Feb 15, 2017, at 2:09 PM, Mark Adams <mfadams at lbl.gov> wrote:
> 
> 
> > Its not a coincidence that it is equal to the FunctionEval time. That is the main cost.
> 
>    Mark,
> 
>     When Newton is in the quadratic basin then it will do one Function evaluation and one Jacobian evaluation per Newton step. The one function evaluation takes place inside the line search code (and the result is used also at the beginning of the next iteration). If Newton is not in the quadratic basin it may do several function evaluations inside the line search (to find a suitable decrease in the function norm); again basically all the work in the line search is the function evaluations, there is nothing else computationally expensive in a line search (maybe a matrix-vector product).
> 
 
  I understand, better to ask.

> 
> Barry, I am trying to understand what is going on in the code. First, someone is writing a paper with me and says we use 'line search', but they do not (they specify 'basic'), and I wanted to get it straight. Second, another project has expensive function evaluation, same cost as a Jacobian, it is an unusual model. I know I am not doing extra function evaluations (I would notice that), but I just want to make sure I understand exactly what the code is doing and that I have my nomenclature straight (I have never thought much about line search).
> 
> Thanks,
> 
>