aboutsummaryrefslogtreecommitdiff
path: root/Carpet
Commit message (Collapse)AuthorAge
* Replace Cactus complex number type with C/C++ complex numbersErik Schnetter2013-03-08
| | | | | | | | | | | | | Map CCTK_COMPLEX to "double complex" in C, and "complex<double>" in C++. (It is already mapped to "double complex" in Fortran.) Update type definitions. Re-implement Cactus complex number math functions by calling the respective C functions. Update thorn that access real and imaginary parts of complex numbers to use standard-conforming methods instead.
* CarpetIOHDF5: remove index files along with checkpoint filesRoland Haas2013-02-28
|
* CarpetIOASCII: Use CCTK_BUILTIN_UNREACHABLEErik Schnetter2013-02-27
|
* Carpet: New aliased function GetRegriddingEpochsErik Schnetter2013-02-26
|
* CarpetLib: Set interpolate_from_buffer_zones to "no" by defaultErik Schnetter2013-02-18
|
* CarpetLib: Correct implementation of "don't interpolate from buffer zones"Erik Schnetter2013-02-18
|
* CarpetLib: Store level-wide set of buffer zonesErik Schnetter2013-02-18
|
* Revert "CarpetLib: Change default to interpolate_from_buffer_zones=no"Erik Schnetter2013-02-12
| | | | This reverts commit c290150324c2dd50e98cd6e35bbcd733e6f3b5a4.
* CarpetEvolutionMask: limit buffer_mask test to one processRoland Haas2013-02-11
|
* CarpetEvolutionMask: compute a mask indicating where buffers areRoland Haas2013-02-11
| | | | | | | this is s first implementatoion of this functionality. Eventually I want to extend it to label buffer points by the last MoL substep at which the grid point needs to have a RHS computed. Zero will mean "all the time" large numbers will mean "never".
* CarpetReduce: Improve code layoutErik Schnetter2013-02-09
|
* LoopControl: Correct parallelisation errors (mark variables volatile)Erik Schnetter2013-02-09
|
* LoopControl: Share cache lines for SMT threadsErik Schnetter2013-02-09
|
* LoopControl: Output statistics to a file instead of stdoutErik Schnetter2013-02-08
|
* LoopControl: Improve OpenMP parallelisationErik Schnetter2013-02-08
|
* LoopControl: Simplify STL data type usageErik Schnetter2013-02-08
| | | | Use string instead of const char*, use vector instead of list.
* LoopControl: Remove trivial vector lengths for j and k directionErik Schnetter2013-02-08
|
* CarpetLib: Do not memoize OpenMP thread idErik Schnetter2013-02-08
|
* Carpet: Beautify screen outputErik Schnetter2013-02-08
|
* CarpetLib: Change default to interpolate_from_buffer_zones=noErik Schnetter2013-02-08
|
* LoopControl: Remove restrict qualifiers from shared data structuresErik Schnetter2013-02-08
|
* LoopControl: Reduce number of arguments passed to selftest routineErik Schnetter2013-02-08
|
* Merge branch 'master' of carpetcode.org:carpetFrank Löffler2013-02-07
|\
| * CarpetLib: clear boundary slice offsets in dh::regrid()Roland Haas2013-01-29
| | | | | | | | | | this avoids assert() failures when regrid is called on a level that did not change it seems.
| * CarpetRegrid2: Correct another error in handling min_fractionErik Schnetter2013-01-25
| |
| * CarpetRegrid2: Output all verbose messages with veryverbose as wellErik Schnetter2013-01-25
| |
| * LoopControl: Avoid compiler warningsErik Schnetter2013-01-24
| |
| * LoopControl: Add debug assert statements when array padding is enabledErik Schnetter2013-01-24
| |
| * LoopControl: Correct Fortran OpenMP definitionsErik Schnetter2013-01-24
| | | | | | | | Also clean up code.
| * CarpetRegrid2: Correct min_fractionErik Schnetter2013-01-24
| | | | | | | | A condition was reversed, inverting the meaning of min_fraction, possibly leading to quite inefficient grid structures.
| * CarpetLib: Implement and correct head, tail, init, last in the vect classErik Schnetter2013-01-24
| |
* | CarpetRegrid2: Testsuite for trivial domains in z directionFrank Löffler2013-02-07
| |
* | CarpetRegrid2: Allow pseudo-2D domainsFrank Löffler2013-02-07
|/ | | | | | | Currently CarpetRegrid2 contains two asserts that fail for domains that are effectively 2D (but defines as 3D arrays, e.g. (nx,ny,1). This patch makes sure these domains don't trigger the assert by checking for ilower==iupper.
* CarpetLib, LoopControl: Correct required/optional capabilitiesErik Schnetter2013-01-22
|
* LoopControl: Rename macro arguments to help avoid name clashesErik Schnetter2013-01-21
|
* LoopControl: More self-checkingErik Schnetter2013-01-21
|
* LoopControl: Correct Fortran macros; provide Fortran wrappersErik Schnetter2013-01-21
|
* LoopControl: Update Fortran macros to follow cctk_Loop APIErik Schnetter2013-01-20
|
* LoopControl: Add assert statements to loop macros to catch indexing errorsErik Schnetter2013-01-20
|
* CarpetLib: Rename argument to vect::index from lsh to ashErik Schnetter2013-01-20
|
* Carpet: Use cctk_ash where necessary in poisoning and checksummingErik Schnetter2013-01-20
|
* Carpet: Return cctk_ash instead of cctk_lsh in CCTK_ArrayGroupSizeBErik Schnetter2013-01-20
|
* CarpetIOASCII: Update test results after domain decomposition changeErik Schnetter2013-01-20
|
* LoopControl: Correct indexing error with vectorisationErik Schnetter2013-01-20
|
* LoopControl: Store loop bounds in the provided variablesErik Schnetter2013-01-19
| | | | Store the loop bounds in the provided variables, not in lc_imin and lc_imax.
* Carpet: Correct load-balancing algorithm to really prefer splitting in the z ↵Erik Schnetter2013-01-19
| | | | direction
* CarpetIOBasic: Simplify code and avoid compiler warnings (again)Erik Schnetter2013-01-19
|
* CarpetLib: Disable DGFE operators if building without DGFE to avoid compiler ↵Erik Schnetter2013-01-19
| | | | warnings
* CarpetLib: Use C++ wrapper madd instead of mul and addErik Schnetter2013-01-18
|
* CarpetLib: Explicitly convert values to signedErik Schnetter2013-01-18
|