1 /*$Id: fmg.c,v 1.26 2001/08/21 21:03:20 bsmith Exp $*/ 2 /* 3 Full multigrid using either additive or multiplicative V or W cycle 4 */ 5 #include "src/ksp/pc/impls/mg/mgimpl.h" 6 7 EXTERN int MGMCycle_Private(MG *,PetscTruth*); 8 9 /* 10 MGFCycle_Private - Given an MG structure created with MGCreate() runs 11 full multigrid. 12 13 Iput Parameters: 14 . mg - structure created with MGCreate(). 15 16 Note: This may not be what others call full multigrid. What we 17 do is restrict the rhs to all levels, then starting 18 on the coarsest level work our way up generating 19 initial guess for the next level. This provides an 20 improved preconditioner but not a great improvement. 21 */ 22 #undef __FUNCT__ 23 #define __FUNCT__ "MGFCycle_Private" 24 int MGFCycle_Private(MG *mg) 25 { 26 int i,l = mg[0]->levels,ierr; 27 PetscScalar zero = 0.0; 28 29 PetscFunctionBegin; 30 /* restrict the RHS through all levels to coarsest. */ 31 for (i=l-1; i>0; i--){ 32 ierr = MatRestrict(mg[i]->restrct,mg[i]->b,mg[i-1]->b);CHKERRQ(ierr); 33 } 34 35 /* work our way up through the levels */ 36 ierr = VecSet(&zero,mg[0]->x);CHKERRQ(ierr); 37 for (i=0; i<l-1; i++) { 38 ierr = MGMCycle_Private(&mg[i],PETSC_NULL);CHKERRQ(ierr); 39 ierr = MatInterpolate(mg[i+1]->interpolate,mg[i]->x,mg[i+1]->x);CHKERRQ(ierr); 40 } 41 ierr = MGMCycle_Private(&mg[l-1],PETSC_NULL);CHKERRQ(ierr); 42 PetscFunctionReturn(0); 43 } 44 45 /* 46 MGKCycle_Private - Given an MG structure created with MGCreate() runs 47 full Kascade MG solve. 48 49 Iput Parameters: 50 . mg - structure created with MGCreate(). 51 52 Note: This may not be what others call Kascadic MG. 53 */ 54 #undef __FUNCT__ 55 #define __FUNCT__ "MGKCycle_Private" 56 int MGKCycle_Private(MG *mg) 57 { 58 int i,l = mg[0]->levels,ierr; 59 PetscScalar zero = 0.0; 60 61 PetscFunctionBegin; 62 /* restrict the RHS through all levels to coarsest. */ 63 for (i=l-1; i>0; i--){ 64 ierr = MatRestrict(mg[i]->restrct,mg[i]->b,mg[i-1]->b);CHKERRQ(ierr); 65 } 66 67 /* work our way up through the levels */ 68 ierr = VecSet(&zero,mg[0]->x);CHKERRQ(ierr); 69 for (i=0; i<l-1; i++) { 70 if (mg[i]->eventsolve) {ierr = PetscLogEventBegin(mg[i]->eventsolve,0,0,0,0);CHKERRQ(ierr);} 71 ierr = KSPSetRhs(mg[i]->smoothd,mg[i]->b);CHKERRQ(ierr); 72 ierr = KSPSetSolution(mg[i]->smoothd,mg[i]->x);CHKERRQ(ierr); 73 ierr = KSPSolve(mg[i]->smoothd);CHKERRQ(ierr); 74 if (mg[i]->eventsolve) {ierr = PetscLogEventEnd(mg[i]->eventsolve,0,0,0,0);CHKERRQ(ierr);} 75 ierr = MatInterpolate(mg[i+1]->interpolate,mg[i]->x,mg[i+1]->x);CHKERRQ(ierr); 76 } 77 if (mg[l-1]->eventsolve) {ierr = PetscLogEventBegin(mg[l-1]->eventsolve,0,0,0,0);CHKERRQ(ierr);} 78 ierr = KSPSetRhs(mg[l-1]->smoothd,mg[l-1]->b);CHKERRQ(ierr); 79 ierr = KSPSetSolution(mg[l-1]->smoothd,mg[l-1]->x);CHKERRQ(ierr); 80 ierr = KSPSolve(mg[l-1]->smoothd);CHKERRQ(ierr); 81 if (mg[l-1]->eventsolve) {ierr = PetscLogEventEnd(mg[l-1]->eventsolve,0,0,0,0);CHKERRQ(ierr);} 82 83 PetscFunctionReturn(0); 84 } 85 86 87