A pipelined conjugate gradient method from Bill Gropp There method has two reductions, one of which is overlapped with the matrix-vector product and one of which is
overlapped with the preconditioner.
See also KSPPIPECG, which has only a single reduction that overlaps both the matrix-vector product and the preconditioner.
MPI configuration may be necessary for reductions to make asynchronous progress, which is important for performance of pipelined methods.
See the FAQ on the PETSc website for details.
Pieter Ghysels, Universiteit Antwerpen, Intel Exascience lab Flanders
KSPCreate(), KSPSetType(), KSPPIPECG, KSPPIPECR, KSPPGMRES, KSPCG, KSPCGUseSingleReduction()
Index of all KSP routines
Table of Contents for all manual pages
Index of all manual pages