static char help[] = "Demonstrates adjoint sensitivity analysis for Reaction-Diffusion Equations.\n"; /* See ex5.c for details on the equation. This code demonestrates the TSAdjoint interface to a system of time-dependent partial differential equations. It computes the sensitivity of a component in the final solution, which locates in the center of the 2D domain, w.r.t. the initial conditions. The user does not need to provide any additional functions. The required functions in the original simulation are reused in the adjoint run. Runtime options: -forwardonly - run the forward simulation without adjoint -implicitform - provide IFunction and IJacobian to TS, if not set, RHSFunction and RHSJacobian will be used -aijpc - set the matrix used to compute the preconditioner to be aij (the Jacobian matrix can be of a different type such as ELL) */ #include "reaction_diffusion.h" #include #include PetscErrorCode InitialConditions(DM, Vec); PetscErrorCode InitializeLambda(DM da, Vec lambda, PetscReal x, PetscReal y) { PetscInt i, j, Mx, My, xs, ys, xm, ym; Field **l; PetscFunctionBegin; PetscCall(DMDAGetInfo(da, PETSC_IGNORE, &Mx, &My, PETSC_IGNORE, PETSC_IGNORE, PETSC_IGNORE, PETSC_IGNORE, PETSC_IGNORE, PETSC_IGNORE, PETSC_IGNORE, PETSC_IGNORE, PETSC_IGNORE, PETSC_IGNORE)); /* locate the global i index for x and j index for y */ i = (PetscInt)(x * (Mx - 1)); j = (PetscInt)(y * (My - 1)); PetscCall(DMDAGetCorners(da, &xs, &ys, NULL, &xm, &ym, NULL)); if (xs <= i && i < xs + xm && ys <= j && j < ys + ym) { /* the i,j vertex is on this process */ PetscCall(DMDAVecGetArray(da, lambda, &l)); l[j][i].u = 1.0; l[j][i].v = 1.0; PetscCall(DMDAVecRestoreArray(da, lambda, &l)); } PetscFunctionReturn(PETSC_SUCCESS); } int main(int argc, char **argv) { TS ts; /* ODE integrator */ Vec x; /* solution */ DM da; AppCtx appctx; Vec lambda[1]; PetscBool forwardonly = PETSC_FALSE, implicitform = PETSC_TRUE; PetscFunctionBeginUser; PetscCall(PetscInitialize(&argc, &argv, NULL, help)); PetscCall(PetscOptionsGetBool(NULL, NULL, "-forwardonly", &forwardonly, NULL)); PetscCall(PetscOptionsGetBool(NULL, NULL, "-implicitform", &implicitform, NULL)); appctx.aijpc = PETSC_FALSE; PetscCall(PetscOptionsGetBool(NULL, NULL, "-aijpc", &appctx.aijpc, NULL)); appctx.D1 = 8.0e-5; appctx.D2 = 4.0e-5; appctx.gamma = .024; appctx.kappa = .06; /* - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Create distributed array (DMDA) to manage parallel grid and vectors - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - */ PetscCall(DMDACreate2d(PETSC_COMM_WORLD, DM_BOUNDARY_PERIODIC, DM_BOUNDARY_PERIODIC, DMDA_STENCIL_STAR, 64, 64, PETSC_DECIDE, PETSC_DECIDE, 2, 1, NULL, NULL, &da)); PetscCall(DMSetFromOptions(da)); PetscCall(DMSetUp(da)); PetscCall(DMDASetFieldName(da, 0, "u")); PetscCall(DMDASetFieldName(da, 1, "v")); /* - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Extract global vectors from DMDA; then duplicate for remaining vectors that are the same types - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - */ PetscCall(DMCreateGlobalVector(da, &x)); /* - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Create timestepping solver context - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - */ PetscCall(TSCreate(PETSC_COMM_WORLD, &ts)); PetscCall(TSSetDM(ts, da)); PetscCall(TSSetProblemType(ts, TS_NONLINEAR)); PetscCall(TSSetEquationType(ts, TS_EQ_ODE_EXPLICIT)); /* less Jacobian evaluations when adjoint BEuler is used, otherwise no effect */ if (!implicitform) { PetscCall(TSSetType(ts, TSRK)); PetscCall(TSSetRHSFunction(ts, NULL, RHSFunction, &appctx)); PetscCall(TSSetRHSJacobian(ts, NULL, NULL, RHSJacobian, &appctx)); } else { PetscCall(TSSetType(ts, TSCN)); PetscCall(TSSetIFunction(ts, NULL, IFunction, &appctx)); if (appctx.aijpc) { Mat A, B; PetscCall(DMSetMatType(da, MATSELL)); PetscCall(DMCreateMatrix(da, &A)); PetscCall(MatConvert(A, MATAIJ, MAT_INITIAL_MATRIX, &B)); /* FIXME do we need to change viewer to display matrix in natural ordering as DMCreateMatrix_DA does? */ PetscCall(TSSetIJacobian(ts, A, B, IJacobian, &appctx)); PetscCall(MatDestroy(&A)); PetscCall(MatDestroy(&B)); } else { PetscCall(TSSetIJacobian(ts, NULL, NULL, IJacobian, &appctx)); } } /* - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Set initial conditions - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - */ PetscCall(InitialConditions(da, x)); PetscCall(TSSetSolution(ts, x)); /* Have the TS save its trajectory so that TSAdjointSolve() may be used */ if (!forwardonly) PetscCall(TSSetSaveTrajectory(ts)); /* - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Set solver options - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - */ PetscCall(TSSetMaxTime(ts, 200.0)); PetscCall(TSSetTimeStep(ts, 0.5)); PetscCall(TSSetExactFinalTime(ts, TS_EXACTFINALTIME_MATCHSTEP)); PetscCall(TSSetFromOptions(ts)); /* - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Solve ODE system - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - */ PetscCall(TSSolve(ts, x)); if (!forwardonly) { /* - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Start the Adjoint model - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - */ PetscCall(VecDuplicate(x, &lambda[0])); /* Reset initial conditions for the adjoint integration */ PetscCall(InitializeLambda(da, lambda[0], 0.5, 0.5)); PetscCall(TSSetCostGradients(ts, 1, lambda, NULL)); PetscCall(TSAdjointSolve(ts)); PetscCall(VecDestroy(&lambda[0])); } /* - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Free work space. All PETSc objects should be destroyed when they are no longer needed. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - */ PetscCall(VecDestroy(&x)); PetscCall(TSDestroy(&ts)); PetscCall(DMDestroy(&da)); PetscCall(PetscFinalize()); return 0; } /* ------------------------------------------------------------------- */ PetscErrorCode InitialConditions(DM da, Vec U) { PetscInt i, j, xs, ys, xm, ym, Mx, My; Field **u; PetscReal hx, hy, x, y; PetscFunctionBegin; PetscCall(DMDAGetInfo(da, PETSC_IGNORE, &Mx, &My, PETSC_IGNORE, PETSC_IGNORE, PETSC_IGNORE, PETSC_IGNORE, PETSC_IGNORE, PETSC_IGNORE, PETSC_IGNORE, PETSC_IGNORE, PETSC_IGNORE, PETSC_IGNORE)); hx = 2.5 / (PetscReal)Mx; hy = 2.5 / (PetscReal)My; /* Get pointers to vector data */ PetscCall(DMDAVecGetArray(da, U, &u)); /* Get local grid boundaries */ PetscCall(DMDAGetCorners(da, &xs, &ys, NULL, &xm, &ym, NULL)); /* Compute function over the locally owned part of the grid */ for (j = ys; j < ys + ym; j++) { y = j * hy; for (i = xs; i < xs + xm; i++) { x = i * hx; if (PetscApproximateGTE(x, 1.0) && PetscApproximateLTE(x, 1.5) && PetscApproximateGTE(y, 1.0) && PetscApproximateLTE(y, 1.5)) u[j][i].v = PetscPowReal(PetscSinReal(4.0 * PETSC_PI * x), 2.0) * PetscPowReal(PetscSinReal(4.0 * PETSC_PI * y), 2.0) / 4.0; else u[j][i].v = 0.0; u[j][i].u = 1.0 - 2.0 * u[j][i].v; } } /* Restore vectors */ PetscCall(DMDAVecRestoreArray(da, U, &u)); PetscFunctionReturn(PETSC_SUCCESS); } /*TEST build: depends: reaction_diffusion.c requires: !complex !single test: args: -ts_max_steps 10 -ts_monitor -ts_adjoint_monitor -da_grid_x 20 -da_grid_y 20 output_file: output/ex5adj_1.out test: suffix: 2 nsize: 2 args: -ts_max_steps 10 -ts_time_step 10 -ts_monitor -ts_adjoint_monitor -ksp_monitor_short -da_grid_x 20 -da_grid_y 20 -ts_trajectory_dirname Test-dir -ts_trajectory_file_template test-%06D.cp test: suffix: 3 nsize: 2 args: -ts_max_steps 10 -ts_time_step 10 -ts_adjoint_monitor_draw_sensi output_file: output/empty.out test: suffix: 4 nsize: 2 args: -ts_max_steps 10 -ts_time_step 10 -ts_monitor -ts_adjoint_monitor -ksp_monitor_short -da_grid_x 20 -da_grid_y 20 -snes_fd_color output_file: output/ex5adj_2.out test: suffix: 5 nsize: 2 args: -ts_max_steps 10 -implicitform 0 -ts_type rk -ts_rk_type 4 -ts_monitor -ts_adjoint_monitor -da_grid_x 20 -da_grid_y 20 -snes_fd_color output_file: output/ex5adj_1.out test: suffix: knl args: -ts_max_steps 10 -ts_monitor -ts_adjoint_monitor -ts_trajectory_type memory -ts_trajectory_solution_only 0 -malloc_hbw -ts_trajectory_use_dram 1 output_file: output/empty.out requires: knl test: suffix: sell nsize: 4 args: -forwardonly -ts_max_steps 10 -ts_monitor -snes_monitor_short -dm_mat_type sell -pc_type none output_file: output/ex5adj_sell_1.out test: suffix: aijsell nsize: 4 args: -forwardonly -ts_max_steps 10 -ts_monitor -snes_monitor_short -dm_mat_type aijsell -pc_type none output_file: output/ex5adj_sell_1.out test: suffix: sell2 nsize: 4 args: -forwardonly -ts_max_steps 10 -ts_monitor -snes_monitor_short -dm_mat_type sell -pc_type mg -pc_mg_levels 2 -mg_coarse_pc_type sor output_file: output/ex5adj_sell_2.out test: suffix: aijsell2 nsize: 4 args: -forwardonly -ts_max_steps 10 -ts_monitor -snes_monitor_short -dm_mat_type aijsell -pc_type mg -pc_mg_levels 2 -mg_coarse_pc_type sor output_file: output/ex5adj_sell_2.out test: suffix: sell3 nsize: 4 args: -forwardonly -ts_max_steps 10 -ts_monitor -snes_monitor_short -dm_mat_type sell -pc_type mg -pc_mg_levels 2 -mg_coarse_pc_type bjacobi -mg_levels_pc_type bjacobi output_file: output/ex5adj_sell_3.out test: suffix: sell4 nsize: 4 args: -forwardonly -implicitform -ts_max_steps 10 -ts_monitor -snes_monitor_short -dm_mat_type sell -pc_type mg -pc_mg_levels 2 -mg_coarse_pc_type bjacobi -mg_levels_pc_type bjacobi output_file: output/ex5adj_sell_4.out test: suffix: sell5 nsize: 4 args: -forwardonly -ts_max_steps 10 -ts_monitor -snes_monitor_short -dm_mat_type sell -aijpc output_file: output/ex5adj_sell_5.out test: suffix: aijsell5 nsize: 4 args: -forwardonly -ts_max_steps 10 -ts_monitor -snes_monitor_short -dm_mat_type aijsell output_file: output/ex5adj_sell_5.out test: suffix: sell6 args: -ts_max_steps 10 -ts_monitor -ts_adjoint_monitor -ts_trajectory_type memory -ts_trajectory_solution_only 0 -dm_mat_type sell -pc_type jacobi output_file: output/ex5adj_sell_6.out test: suffix: sell7 args: -ts_max_steps 10 -ts_monitor -ts_adjoint_monitor -ts_trajectory_type memory -ts_trajectory_solution_only 0 -dm_mat_type sellcuda -dm_vec_type cuda -pc_type jacobi output_file: output/ex5adj_sell_6.out requires: cuda test: suffix: gamg1 args: -pc_type gamg -pc_gamg_esteig_ksp_type gmres -pc_gamg_esteig_ksp_max_it 10 -ksp_monitor_short -ts_max_steps 2 -ts_monitor -ts_adjoint_monitor -ts_trajectory_type memory -ksp_rtol 1e-2 -pc_gamg_use_sa_esteig 0 output_file: output/ex5adj_gamg.out test: suffix: gamg2 args: -pc_type gamg -pc_gamg_esteig_ksp_type gmres -pc_gamg_esteig_ksp_max_it 10 -ksp_monitor_short -ts_max_steps 2 -ts_monitor -ts_adjoint_monitor -ts_trajectory_type memory -ksp_use_explicittranspose -ksp_rtol 1e-2 -pc_gamg_use_sa_esteig 0 output_file: output/ex5adj_gamg.out TEST*/