Cvodes 2.9.0 release

there is a new sundials 2.7.0 release which includes cvodes 2.9.0. The changelog is pasted below. I cannot see any real change except “minor corrections to the solver”.

I don’t think it should be a huge effort to upgrade, but then I am not sure what it buys us. Any thoughts?

Sebastian

Changelog:
Two additional nvector implementations were added { one for Hypre (parallel) vectors, and one for
PetSC vectors. These additions are accompanied by additions to various interface functions and to
user documentation.
Each nvector module now includes a function, N VGetVectorID, that returns the nvector
module name.
A bug was xed in the interpolation functions used in solving backward problems for adjoint
sensitivity analysis.
For each linear solver, the various solver performance counters are now initialized to 0 in both the
solver specication function and in solver linit function. This ensures that these solver counters are
initialized upon linear solver instantiation as well as at the beginning of the problem solution.
A memory leak was xed in the banded preconditioner interface. In addition, updates were done
to return integers from linear solver and preconditioner ‘free’ functions.
The Krylov linear solver Bi-CGstab was enhanced by removing a redundant dot product. Various
additions and corrections were made to the interfaces to the sparse solvers KLU and SuperLU MT,
including support for CSR format when using KLU.
In interpolation routines for backward problems, added logic to bypass sensitivity interpolation if
input sensitivity argument is NULL.
New examples were added for use of sparse direct solvers within sensitivity integrations and for
use of openMP.
Minor corrections and additions were made to the cvodes solver, to the examples, to installationrelated
les, and to the user documentation.

I’d prefer to keep up with the latest stable versions
of external packages. Isn’t this relevant:

A bug was xed in the interpolation functions used in solving backward problems for adjoint
sensitivity analysis.

as well as:

Minor corrections and additions were made to the cvodes solver

  • Bob

Hi!

Sure, we can upgrade.

The interpolation stuff does not affect us - this is for adjoint sensitivity analysis, we do forward sensitivity analysis from CVODES.

But yes, the minor correction and additions thing is something which does affect us.

Let me look into this then.

Sebastian