Commit Graph

10042 Commits

Author SHA1 Message Date
Ulrike Yang
dcac0cf07b removed hard-coded print statement 2017-09-29 13:06:28 -07:00
Ulrike Yang
d890cf8593 updated changes 2017-09-29 11:07:25 -07:00
Rob Falgout
c8af600c15 Updating version number for 2.12.1 2017-09-29 08:57:18 -07:00
Rob Falgout
fdd5309e14 Updating CHANGELOG for 2.12.1 release 2017-09-29 08:51:38 -07:00
liruipengblue@gmail.com
a979213e9e fixed RNG seed == 0 issue and changed the corresponding regression test .saved file 2017-08-21 12:55:49 -07:00
liruipengblue@gmail.com
21d33f737e fixed the problem in ij.c when being used with ``-vardifconv -print'' 2017-08-21 11:23:35 -07:00
liruipengblue@gmail.com
308fd1b655 add the new file to CMakelist
int -> HYPRE_Int
2017-08-18 16:29:22 -07:00
liruipengblue@gmail.com
1da38c94eb new vardifconv with the discretization and examples in Ruge-Stuben's paper 2017-08-18 13:41:05 -07:00
Ulrike M. Yang
df03349fb1 fixed compile bug 2017-08-11 13:30:55 -07:00
liruipengblue@gmail.com
9047f85bf6 restore the commented lines with valgrind 2017-07-28 17:18:30 -07:00
liruipengblue@gmail.com
a59513bf34 1. change the default conv-diff in ij to what it used to be, add cmd-line options
2. fixed issues in machine-tux.sh ("./")
3. make AUTOTEST happy again
2017-07-28 17:05:25 -07:00
liruipengblue@gmail.com
e997fee01c Merge branch 'master' of ssh://mystash.llnl.gov:7999/hypre/hypre 2017-07-28 00:21:06 -07:00
liruipengblue@gmail.com
6e7cfd60fc added more options (upwind,centered,forward,backward) for FD discretization of Conv-Diff problem 2017-07-28 00:17:30 -07:00
Rob Falgout
6cf6cde605 Merge branch 'master' of ssh://mystash.llnl.gov:7999/hypre/hypre 2017-07-05 12:20:37 -07:00
Rob Falgout
4155a54145 Updated machine name list to include rztopaz and syrah 2017-07-05 12:19:20 -07:00
Rob Falgout
9e076b7853 Removing rzzeus and rzmerl and adding rztopaz and syrah 2017-07-05 11:30:51 -07:00
Ulrike Yang
d4fab7c148 fixed a memory leak 2017-06-30 14:03:29 -07:00
Ulrike Yang
5f53bdc738 Changed relaxtype 7 (weighted Jacobi using matvec) to be usable for GPUs 2017-06-22 10:49:27 -07:00
Ulrike Yang
2bfb229528 fixed issue 1471 2017-06-22 10:46:48 -07:00
Falgout, Rob
e65ffe10ba Merge pull request #9 in HYPRE/hypre from pull-8 to master
Merging Joe Koning's pull request into master

* commit '68bbece9348dd3001474fe8fbd65ba40b6f0084c':
  Put the FEI HAVE_MLI and HAVE_SUPERLU defs in configure
  Configure improvement for FEI when MLI and Superlu are not required.
2017-06-20 14:35:23 -07:00
Rob Falgout
68bbece934 Put the FEI HAVE_MLI and HAVE_SUPERLU defs in configure 2017-06-16 17:21:29 -07:00
Falgout, Rob
f972cf18ca Merge pull request #8 in HYPRE/hypre from ~KONING1/hypre:master to pull-8
* commit 'a6b70edcac69b3e094ebd58b5921546cce560d39':
  Configure improvement for FEI when MLI and Superlu are not required.
2017-06-16 15:58:56 -07:00
Rob Falgout
d44409fc87 Moved autotest scripting to another repository and made a few other changes 2017-06-16 13:48:44 -07:00
Joseph M. Koning
a6b70edcac Configure improvement for FEI when MLI and Superlu are not required. 2017-06-16 12:50:40 -07:00
Ulrike Yang
9c86d5099b Merge branch 'master' of ssh://mystash.llnl.gov:7999/hypre/hypre 2017-05-24 16:21:36 -07:00
Ulrike Yang
50be6ed76a changed CG to GMRES for nonsymmetric problem 2017-05-24 16:20:20 -07:00
Rob Falgout
4b43f8451e Added './' in front of executables in two example runtest tests 2017-05-24 13:18:48 -07:00
Rob Falgout
13527db5ce Trivial change to clean up check-double autotest script test 2017-05-19 10:08:43 -07:00
Falgout, Rob
bbacaa7af9 Merge pull request #6 in HYPRE/hypre from single-dev to master
Adding support for single and longdouble precision

* commit '6a39fbf9c64d7f466eb874ae30502fca25a15991':
  Fixed utilities header files so 'headers' script works correctly
  Added single/longdouble to cmake and added regression tests to tux autotest
  Modified saved files for TEST_single and TEST_longdouble
  modified to use hypre's random.c routines instead of C rand().
  Fixed a memory leak and 'check-double' regression test issue
  Added new test folders and tests for single and longdouble cases.
  Minor modifications dlartg.c to avoid integer overflow issues when longdouble is enabled
  Added new constants to compute machine tolerances/thresholds based on precision
  Modified to change constant dt_inf to be consistent with single precision
  Added float.h include to HYPRE_utilities.h
  Updated to define floating point constants to be consistent for single precision arithmetic representation. Also did some minor changes of hypre_MPI_DOUBLE to hypre_MPI_REAL in several places for Euclid, pilut, and Parasails.
  Fixed additional bugs with --enable-single and --enable-longdouble
  Fixed -tol option for struct driver
  This compiles, but generates the wrong results for HYPRE_SINGLE
  Preliminary implementation of single precision capability
2017-05-18 11:11:09 -07:00
Rob Falgout
6a39fbf9c6 Merge branch 'master' into single-dev 2017-05-18 08:31:49 -07:00
Jacob Schroder
4d6bd49acc Changing "0.0" to "0.0d0" in ex12f.f, because Fortran doesn't
always pass 0.0 correctly to C.  That is, it can be passed as a single
or double, and this caused a regression test failure for me.  Specifying
it explicitly as a double fixes the issue.  ex5f.f already does this.
2017-05-17 08:49:48 -07:00
Ulrike Yang
5362a60e7b fixed type 2017-05-17 08:38:58 -07:00
Ulrike Yang
b1a427b610 added new functions to Fortran interface 2017-05-17 08:36:38 -07:00
Ulrike Yang
d69f4f83d9 changed MAlloc to CAlloc to avoid valgrind warnings 2017-05-16 16:21:41 -07:00
Ulrike Yang
9d6afbf211 Merge branch 'master' of ssh://mystash.llnl.gov:7999/hypre/hypre 2017-05-16 14:25:09 -07:00
Rob Falgout
08d6833cd2 Fixed autotest scripts to work when '.' is not in the user's path 2017-05-16 11:26:35 -07:00
Ulrike Yang
fef660be27 fixed a memory leak and changed/added a regression test 2017-05-16 09:57:04 -07:00
Rob Falgout
f4274a1486 Fixed utilities header files so 'headers' script works correctly 2017-05-16 08:16:37 -07:00
Rob Falgout
e758460272 Added single/longdouble to cmake and added regression tests to tux autotest 2017-05-15 17:37:19 -07:00
Rob Falgout
b2e2eedc4d Merge branch 'master' into single-dev 2017-05-15 16:26:16 -07:00
Rob Falgout
2b8442ba6d Fixing the cmake compile 2017-05-15 10:05:53 -07:00
Ulrike Yang
1ec8646871 added regression tests for new polynomial smoother settings 2017-05-12 10:05:26 -07:00
Ulrike Yang
3ded8759cb added regression tests for new additive options 2017-05-12 08:58:28 -07:00
Ulrike Yang
871729d68c added new Cheby routines 2017-05-11 15:38:54 -07:00
Ulrike Yang
4966c9b096 Merge branch 'master' of ssh://mystash.llnl.gov:7999/hypre/hypre 2017-05-11 14:41:25 -07:00
Ulrike Yang
e31bd6074e Moved various operations for polynomial smoothers into setup phase to improve performance and added new parameters to allow use of different polynomials 2017-05-11 14:39:22 -07:00
Daniel Osei-Kuffuor
d1472da148 Modified saved files for TEST_single and TEST_longdouble 2017-05-09 17:15:03 -07:00
Daniel Osei-Kuffuor
da65135af6 Merge branch 'master' into single-dev 2017-05-09 12:13:49 -07:00
Lu WANG
a34db78e8c clean the macro hypre_MatrixIndexMove to fix the mistake 2017-05-03 16:48:37 -07:00
Lu WANG
51f9646c6c change the macro MatrixIndexMove without GPU to avoid the warning on mac 2017-05-03 10:24:48 -07:00