Increased auxiliary term performance

[landinjm]

1. Auxiliary terms have dirichlet boundaries applied every increment regardless of input boundary conditions
2. NaN solutions are checked every increment

Fixing these to more closely resemble the process for explicit terms decreases computation time greatly. These changes have been tested on the Cahn-Hilliard app to produce identical results with an ~25% decrease in computation time.

[david-montiel-t]

Thank you!

On Mon, Aug 28, 2023 at 6:10 PM landinjm @.***> wrote:

@.**** commented on this pull request.

In src/matrixfree/solveIncrement.cc https://github.com/prisms-center/phaseField/pull/197#discussion_r1308008356 :

• if (currentIncrement%userInputs.skip_print_steps==0){
• if (!numbers::is_finite(solutionSet[fieldIndex]->l2_norm())){
• snprintf(buffer, sizeof(buffer), "ERROR: field '%s' solution is NAN. exiting.\n\n",
• fields[fieldIndex].name.c_str());
• pcout<<buffer;
• exit(-1);
• }

Done

— Reply to this email directly, view it on GitHub https://github.com/prisms-center/phaseField/pull/197#discussion_r1308008356, or unsubscribe https://github.com/notifications/unsubscribe-auth/AD3FA44D634LNBA352X56JTXXUJN7ANCNFSM6AAAAAA37XX23Q . You are receiving this because you commented.Message ID: @.***>