Update to SUNDIALS 5.8.0 #1836

dweindl · 2022-07-22T12:01:54Z

See https://github.com/LLNL/sundials/releases/tag/v5.8.0 for changes.

codecov · 2022-07-22T12:26:29Z

Codecov Report

Merging #1836 (4146e11) into develop (e7b71d4) will increase coverage by 0.04%.
The diff coverage is 100.00%.

@@             Coverage Diff             @@
##           develop    #1836      +/-   ##
===========================================
+ Coverage    76.05%   76.09%   +0.04%     
===========================================
  Files           74       74              
  Lines        12681    12682       +1     
===========================================
+ Hits          9644     9651       +7     
+ Misses        3037     3031       -6

Flag	Coverage Δ
cpp	`72.73% <100.00%> (+0.07%)`	⬆️
petab	`59.44% <ø> (ø)`
python	`68.43% <ø> (ø)`

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
src/solver_cvodes.cpp	`70.23% <100.00%> (+0.21%)`	⬆️
src/sundials_matrix_wrapper.cpp	`80.16% <0.00%> (-0.41%)`	⬇️
src/model.cpp	`76.95% <0.00%> (-0.07%)`	⬇️
src/exception.cpp	`81.08% <0.00%> (+5.40%)`	⬆️
src/spline.cpp	`72.04% <0.00%> (+6.45%)`	⬆️

... that show the actual difference

dweindl · 2022-07-22T18:22:39Z

FAILED test_compare_conservation_laws_sbml.py::test_adjoint_pre_and_post_equilibration
FAILED test_preequilibration.py::test_equilibration_methods_with_adjoints

~~Looks like this breaks state reinitialization. I don't see any related change in SUNDIALS :/~~

Just had to update NlsRhsFun when changing cv_f.

FFroehlich

👍

FFroehlich · 2022-07-25T07:58:28Z

src/solver_cvodes.cpp

@@ -118,6 +118,8 @@ void CVodeSolver::initSteadystate(const realtype /*t0*/, const AmiVector &/*x0*/
       as re-calling init would unset solver settings. */
    auto cv_mem = static_cast<CVodeMem>(solver_memory_.get());
    cv_mem->cv_f = fxBdot_ss;
+


maybe add a comment that both functions must be set for simulation to work properly and that just setting cv_mem->cv_f will result in the previous function being used as alternate Nls function?