Hm, some update in the OrdinaryDiffEq.jl library broke this. So, in the test suite for the successful_step function the parallel ODE fails to convey that the steps take have failed. Perhaps this is related to the fact that we pass in a custom norm internalnorm = inorm when we create a parallel ODE system, here
|
pds = CoupledODEs(prob, ds.diffeq; internalnorm = inorm) |
but honestly I am not sure. What I am sure about is that the test fails:
|
pds = ParallelDynamicalSystem(deepcopy(ode), [u0,u0]) |
which is why in the current code this is commented out. We will need to test whether this is truly an OrdinaryDiffEq.jl problem by simply making a MWE with an ODEProblem with a parallel rule for two states and passing in a modified internalnorm as a keyword argument. If it is an OrdinaryDiffEq.jl problem, we open an issue there with our MWE. If not, then we have to fix it here.
Hm, some update in the OrdinaryDiffEq.jl library broke this. So, in the test suite for the
successful_stepfunction the parallel ODE fails to convey that the steps take have failed. Perhaps this is related to the fact that we pass in a custom norminternalnorm = inormwhen we create a parallel ODE system, hereDynamicalSystemsBase.jl/src/derived_systems/parallel_systems.jl
Line 63 in 89fcd8b
but honestly I am not sure. What I am sure about is that the test fails:
DynamicalSystemsBase.jl/test/successful_step.jl
Line 21 in 89fcd8b
which is why in the current code this is commented out. We will need to test whether this is truly an OrdinaryDiffEq.jl problem by simply making a MWE with an
ODEProblemwith a parallel rule for two states and passing in a modifiedinternalnormas a keyword argument. If it is an OrdinaryDiffEq.jl problem, we open an issue there with our MWE. If not, then we have to fix it here.