treadway
ae68dc30e2
Using global variable ${AR} for ar command, AIX needed extra flags
...
to support 64-bit object files (AR="ar -X64")
2002-10-24 20:54:38 +00:00
painter
487294a329
change name of krylov library
2002-06-20 20:05:27 +00:00
falgout
b63d3b4951
Added prototype for BuildStencilMatrix.
2002-05-03 23:00:10 +00:00
falgout
ecb0295918
Added a fairly general stencil-based approach for building test matrices.
2002-04-04 19:08:35 +00:00
ulrikey
72f7edc43d
removed transpose.c, since it is now contained in csr_matop.c in the seq_mv
...
directory, and changed the calling sequence for hypre_CSRMatrixTranspose.c
in rap.
2001-08-28 16:25:17 +00:00
treadway
d43b4afa55
removed MAKEFLAGS references
2001-08-27 21:24:02 +00:00
treadway
cf57f32309
modified to except pmake option
2001-08-23 19:35:10 +00:00
painter
159aa18884
add argument to hypre_PCGFunctions
2001-08-23 18:28:44 +00:00
panayot
fc111309b6
another fix related to the case of zero num_faces;
2001-08-10 15:46:02 +00:00
panayot
c52fb73d09
Fixed minor memory leak (when num_faces = 0);
2001-08-10 15:25:17 +00:00
ulrikey
68ebc11e68
removed redundant definitions
2001-08-02 14:00:42 +00:00
ulrikey
fadaa57e7a
added the possibility to truncate A and P.
...
P can now be truncated by specifying a truncation factor through -Ptr <xxx>
and/or specifying the max. no. of elements per row desired through
-Pmx <nnn>
A can be truncated by specifying a truncation factor through -Atr <xxx>
Truncation for A is done by determining the max diagonal element , and
eliminating every off-diagonal ellement smaller than the max diag.elmt *trunc
factor and adding this element to the diagonal
Truncation for P is done by eliminating every element per row smaller than
the abs. max elmt of this row * trunc-factor, and renormalizing the row
after truncation (or determining the max. k elements per row , eliminating
the others and renormalizing)
2001-07-31 20:44:21 +00:00
ulrikey
ad49af592b
added symm. Gauss-Seidel (-rlx 2 in command line)
2001-07-31 20:36:26 +00:00
panayot
1f6828b7f8
An older version has been replaced with most recent "bug--free" one.
2001-07-18 20:39:13 +00:00
falgout
9a12c0b9f6
Not supposed to have Makefiles checked into repository.
2001-06-22 01:35:32 +00:00
ulrikey
0cde99f60c
new ij interface
2001-06-21 22:46:00 +00:00
treadway
64272662da
added LD_FLAGS
2001-06-18 17:45:11 +00:00
panayot
179a3a99e1
minimal change to call matrix_inverse insted of mat_inv.
2001-05-30 01:44:52 +00:00
panayot
3e7713ac57
pseudo inverse of a symmetric positive semi--definite matrix;
...
uses eigenvector decomposition of symmetric matrix;
2001-05-30 01:43:46 +00:00
panayot
deec1cf94c
V--cycle with GS smoothing; stationary solver;
2001-05-30 01:42:14 +00:00
panayot
d75c794dea
V--cycle with GS smoothing routine;
2001-05-30 01:41:16 +00:00
panayot
32feabf64a
sym_GS_pcg routine.
2001-05-30 01:40:37 +00:00
panayot
32aaec2601
setup routine for spectral AMGe; Colorado version.
2001-05-30 01:38:12 +00:00
panayot
7fa2d3b5a0
interpolation routine for spectral amge; Colorado version.
2001-05-30 01:37:13 +00:00
panayot
86c8617d81
Makefile for spectral amge; Colorado version.
2001-05-30 01:35:39 +00:00
panayot
357b8273f3
Driver for testing spectral amge method; Colorado version.
2001-05-30 01:34:37 +00:00
falgout
0fd11dbbc0
Moved these data files out of the repository and into a common directory.
2001-03-09 18:01:40 +00:00
panayot
e9b11ee06c
fix a bug in a temporary array i_dof_index;
2001-02-16 00:30:57 +00:00
panayot
473a5d80ae
builds interpolation for spectral agglomerate AMGe;
2001-01-30 17:11:03 +00:00
panayot
e9f7ff4705
multiplicative Schwarz smoother; minimal (element based) overlap.
2001-01-30 17:09:34 +00:00
panayot
b88c248b06
computes schur complements needed in the spectral AMGe interpolation;
2001-01-30 17:08:08 +00:00
panayot
c631a5207e
changed name: matinv to mat_inv;
2001-01-30 17:05:40 +00:00
panayot
ce2f8e2809
changed name: matinv to mat_inv
2001-01-30 17:04:43 +00:00
chtong
5ae8d92402
Add a third problem for amge.
2001-01-23 19:40:52 +00:00
chtong
c4208ab6b0
add a second test problem for amge.
2001-01-23 19:40:06 +00:00
chtong
7d807cd7d7
Add a test problem for amge.
2001-01-23 19:39:19 +00:00
ulrikey
2fe5d9666c
fixed a bug that caused the code to hang for small thresholds
2000-12-28 18:46:46 +00:00
panayot
1ccef22d01
some prints removed.
2000-12-15 18:08:55 +00:00
panayot
1b586d8370
small test example
2000-12-15 18:06:13 +00:00
panayot
02e8868454
large test example
2000-12-15 18:05:12 +00:00
panayot
bf642ff2c1
Large test example
2000-12-15 18:04:47 +00:00
panayot
b5a198a474
Large test example;
2000-12-15 18:01:20 +00:00
panayot
ecdc0e7f39
Added one more file: compute_dof_function.c
2000-12-15 17:48:32 +00:00
falgout
d9015e48a9
_matrix_vector=>_mv translations
2000-12-14 18:19:22 +00:00
falgout
3ac2d0d8d7
_linear_solvers=>_ls translations
2000-12-14 18:17:26 +00:00
falgout
020612ce4e
_linear_solvers=>_ls and _matrix_vector=>_mv translations
2000-12-14 18:16:54 +00:00
painter
4a0ef6c539
driver calls generic Krylov functions directly. The corresponding
...
functions in HYPRE_csr{gmres,pcg} do nothing but call the generic ones;
all but Create, Destroy, and diagonal scaling are now obsolete.
2000-12-14 16:35:23 +00:00
treadway
9c6831b48f
corrected include path
2000-12-12 21:41:41 +00:00
painter
c6aa0df18b
restore previous *.h files and definitions of matrix and vector types
2000-12-06 21:39:33 +00:00
treadway
75cfb0de3c
fixed syntax error in clean and veryclean, which showed up on osf make
2000-11-16 19:28:18 +00:00
