switched arch-cuda-double to use -arch=sm_13 instead of -arch=sm_20Hg-commit: 9a86c66b249d936d2787213e53659810731c1b26
adding double precision support to CUDA functionsHg-commit: 294f7a41af16bfec7a0e477ef9cae82674a2e141
fixed example ./configure options for iphonefixed mg bug when using galerkin option, extra destroy of matrixHg-commit: 3310b90f9082a619b25067b13e844bc2e71e0674
changed arch-cuda to use the -x flag for CUDA 3.1 Also added event logging for copying vectors up and down to GPU and began creating subclass for SeqCUDA to move code from Seq class.Hg-commit: c1a
changed arch-cuda to use the -x flag for CUDA 3.1 Also added event logging for copying vectors up and down to GPU and began creating subclass for SeqCUDA to move code from Seq class.Hg-commit: c1aa6f2e29f76ec64ce854bf916a7edd016eb261
show more ...
allow user to set global sizes in MATSHELL and then live with PETSc's choice of local sizesHg-commit: 6164fe92c28030e5bb9f281de75aabf4ac6c1f65
totally messed up cuda configure stuffHg-commit: 35a68003d3768b062f1b14ee6609aec9d10bd705
improved documentation for iPhone buildHg-commit: 7da5aa57562b2c07087e31bb5f62f39571c523c0
xcodebuilder.py now builds iphone petsc libraryadded config/examples for iphoneHg-commit: 8c20a6c119ea450fccae14cbb7683190dc6f4345
move untested config examples to 'old', and reorganize a bit.Hg-commit: f0a2b4b4ad1b1fb718aa480ab740c7aa68eb794b
aix, cray-xt5 configs.Hg-commit: 141b28ce78185162694fdffc0d736ac90ed4c74d
bgp updatesHg-commit: fa3b4527e996d8f5d74f4310cb813f445b6c83c1
valgrind with nemesis is hanging [unless latest valgrind-3.5 is used].Since mpich2+sock is valgrind clean on linux - revert back to it forthis nightly build.Hg-commit: 4fa8d10004f8e7a7fa924677a3b
valgrind with nemesis is hanging [unless latest valgrind-3.5 is used].Since mpich2+sock is valgrind clean on linux - revert back to it forthis nightly build.Hg-commit: 4fa8d10004f8e7a7fa924677a3b01536ab1540d3
Enable valgrind nightlybuild.buildtest script looks if PETSC_ARCH has the string 'valgrind' andautomatically enables mpiexec with valgrind.Hg-commit: ca5c833d0e17662641d3aa2e3a36ccb48bef4996
use nemesis for mpich2/valgrind testHg-commit: cbb178cf2c32a6a6639aaa56076fc8d5e9101a23
update mac buildsHg-commit: 9d477b777dc21a6743fd93df4b76f0196139701d
update osx 10.6 build with fortranHg-commit: 0e290ad12186dad06ae8cb1fa7fcb145312f9288
add osx 10.6 nightly buildHg-commit: a130e2b725c05f4d8ceea6de7d5ac224419ca7e0
fix config/examples/*.py with void-p and long-long changesHg-commit: 6fd5329625d0bf1a982a7882ed352e46501e447b
disable superlu from complex tests [as the tests break]Hg-commit: cf1af4303e0439f952343a77958c6741e1ea60b5
disable hypre/sundials builds on opensolaris - erratic build failuresHg-commit: 9a6f144923cbff60a92b76e8dfe1ccfbe6f45cb2
add opensolaris, freebsd to nightlybuildsHg-commit: 6dc5ec67929352243bc01d1372f0603b42fd17b3
removed --with-f90-interface from .pyhave-mpi-long-long replaced with known-mpi-long-longHg-commit: 092c346148105fb02be19b25c9b871a1db1561c9
changed --sizeof_XXXX and --bits_per_byte to the standard --known-sizeof-XXXX and --known-bits-per-byteHg-commit: 2fcd8591d5a21c123da85ced5a6eb1001cb0bd15
changed --with-memcmp-ok and --with-bad-memcmp to a single --known-memcmp-ok which is set by user or --with-batch to tell if ok 1 else 0changed --with-mpi-shared to --known-mpi-shared to indicate w
changed --with-memcmp-ok and --with-bad-memcmp to a single --known-memcmp-ok which is set by user or --with-batch to tell if ok 1 else 0changed --with-mpi-shared to --known-mpi-shared to indicate we know that it is shared or we know it is not sharedHg-commit: 87c224900faf57adb70bce0dcda3674560a49f51
fix c++ build with VC6/MPICH2Hg-commit: 0baa68b7d21fe5f9f53f940be5cb1d34629175f3
1...<<313233343536