Lines Matching refs:a

4 This document is a README file for coding parallel AMG for Pressure
7 document as a help.
26 do setup everytime, ramg_control handles that. ) 2) Build a loop for
43   We have to start from this file. This is a module file containing
44 global control variables, matrices, and vectors. To have a global module
63 CF-spliting is done in a seperate function ramg_CFsplit. Other than the
66 the highest value out from the array. So heap sort is a natural choice. 
68 After CF-split, we do a communication: from ramg_update_cfmap to
80 by group in parallel. For each node, there is a "group id" assigned to
82 processor. For a boundary node, if it is a "master", the rank of it is
83 the neighbour rank; if it is a "slave", the rank of it is the negative
92 Interpolation operator is a sparse matrix, but it's hard to build it at
96 attention to transpose algorithm, it's a bit complicated but as usual,
111 only need to do a back substitution, which saves time greatly,
112 especially when you have a bit larger coarest level.
125 possibilities (mode) to do a stand-alone Preconditioned CG by us, which
135 to directly give a scaled matrix. Another place is the allocation of
144 will dramatically decrease efficiency. Actually, a rule of thumb: break
148 into one loop, the math will change to a more complicated one. Oh,
149 another rule of thumb: use less "if" in a loop, that will increase
152 Now to parallel. lhsGP* is a sparse-communicated duplicate of lhsP
158 used through a user-specified AP-product. here in ggb_av.
160 generate_gmap creates a mapping and reverse mapping (gmap,grevmap) for a
165 In ggb_setup, there is a Allreduce with MPI_MAX, this is to search for a
172 apply, do a simple back-substitution. (like we did in coarsest solver
210 global_lhs is a function that reads in lhsP (in sparse format so there
219 considering triple boundaries, if a boundary node has 3 or more
222 which won't be a big issue for the overall performance. It is a direct
245 It's a wrapper for G-S, Jacobi, Cheby, and MLS. Nothing fancy here.
251 while it should be the same and symmetric). Let's suggest a smoother to
256 nothing fancy here. but there are  a lot of output tools you may want to