KSPCGUseSingleReduction#
Merge the two inner products needed in CG into a single MPI_Allreduce() call.
Synopsis#
#include "petscksp.h"
PetscErrorCode KSPCGUseSingleReduction(KSP ksp, PetscBool flg)
Logically Collective on ksp
Input Parameters#
ksp - the iterative context
flg - turn on or off the single reduction
Options Database#
-ksp_cg_single_reduction
- Merge inner products into single MPI_Allreduce
The algorithm used in this case is described as Method 1 in Lapack Working Note 56, “Conjugate Gradient Algorithms with Reduced Synchronization Overhead Distributed Memory Multiprocessors”, by E. F. D’Azevedo, V. L. Eijkhout, and C. H. Romine, December 3, 1999. V. Eijkhout credits the algorithm initially to Chronopoulos and Gear.
It requires two extra work vectors than the conventional implementation in PETSc.
See also KSPPIPECG, KSPPIPECR, and KSPGROPPCG that use non-blocking reductions.
See Also#
Level#
intermediate
Location#
Implementations#
KSPCGUseSingleReduction_CG in src/ksp/ksp/impls/cg/cg.c
Index of all KSP routines
Table of Contents for all manual pages
Index of all manual pages