Lines Matching refs:in

5 Poisson Equation in PHASTA. If you want to start to modify or add
13 you continue. row-wise sparse storage is used everywhere in this
20 *** Files related to AMG in the ../common/ and ../incompressible/
39 **** in AMG directory ******
46 every function call (though it's still tedious in some places). 
49 dimentions. This is used, as shown in this file later, in submatrices
52   Then we have our multilevel AMG stored in amg_A_* arrays as sparse
54 are same as ilwork and nlwork, but in higher levels with smaller size.
63 CF-spliting is done in a seperate function ramg_CFsplit. Other than the
75 "coarse first, fine second". This order will be used in Gauss-Seidel
80 by group in parallel. For each node, there is a "group id" assigned to
84 of the neighbour rank. (see ramg_initbcflag in ramg_paratools.f). This
86 after each coarsening, the information is carried on to higher level in
89 every "if (afmap....)" in ramg_CFsplit you will see that.
109 Also in ramg_control, we do lu decomposition for the coarsest level, if
115 in ramg_window/ramg_winct), if it does not go down, we start restart
128 Pay attention to diagonal scaling in V-cycle, at the finest level, the
137 be done in extract and use the vector everytime.) 
149 another rule of thumb: use less "if" in a loop, that will increase
157 Must read the section of GGB in Chun Sun's thesis. ARPACK/PARPACK is
158 used through a user-specified AP-product. here in ggb_av.
171 1) You may want to do lu-decomposition first in the setup, then in GGB
172 apply, do a simple back-substitution. (like we did in coarsest solver
183 Note in the code we use "I" instead of "Q".
185 When you do sparse matrix-matrix product, don't attempt to do it in one
200 I'm writing things in ramg_cheby.f here too, coz they are all polynomial
203 set to 10 in input.config. You can change it to the ratio of coarse/fine
210 global_lhs is a function that reads in lhsP (in sparse format so there
220 representations (lives in 2 or more tasks), only one will be considered.
228 attention to scaling and corresponding function calls in ramg_tools.f
231 matrices. Everything is similar to plain CG, but smaller in size because
241 Okay that's it, again, be very carefull in this file. 
247 Only thing to note is in Gauss-Seidel, we follow red-black smoothing,