Minor change the chebyshev filtering for state parallelization.
Description
Slight change the Chebyshev filtering for state parallelization. The solver was using the first eigenvalue of the local states (batch) for evaluating polynomial order criteria in optimal_polynomial_degree
. This creates a dependence on the state distribution, as batch size varies w.r.t. GPU vs CPU. To avoid this, the lowest eigenvalue of the system is included in the evaluation of the criteria
News snippet
Minor change for the Chebyshev filtering with state parallelization.
Checklist
-
I have checked that my code follows the Octopus coding standards -
I have added tests for all the new features added in this request.
Edited by Alex Buccheri