On Tue, Jan 19, 2021 at 10:16 AM Roland Richter <[email protected]> wrote:
Hei,
during recent experiments with the adaptive stepper in odeint I noticed that the step size used for each step can vary if I am using more than one thread at the same time (i.e. thread 1 can have a different step size than thread 2 when doing the same step). This is not suitable for my project, after the function calculating the right hand side relies on having the same step size for each thread.
Before starting deeper experiments with the adaptive steppers in PETSc, I therefore wanted to ask if those steppers synchronize the step size of the sub-steps between the different mpi processes at each step, or if not?
Yes, the timestep is the same across the communicator that you create the TS with. Thanks, Matt
Thanks!
Regards,
Roland
-- What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead. -- Norbert Wiener https://www.cse.buffalo.edu/~knepley/ <http://www.cse.buffalo.edu/~knepley/>