Commits

commits@lists.einsteintoolkit.org

2 participants
6997 discussions

[svn:einsteintoolkit] NullDecomp/branches/tapir/ (Rev. 17)
by bela＠caltech.edu 02 Jan '14

02 Jan '14

User: szilagyi Date: 2014/01/02 09:45 PM Modified: /branches/tapir/ interface.ccl Log: make clear which variables do not need timelevels File Changes: Directory: /branches/tapir/ =========================== File [modified]: interface.ccl Delta lines: +2 -2 =================================================================== --- branches/tapir/interface.ccl 2013-12-28 04:57:10 UTC (rev 16) +++ branches/tapir/interface.ccl 2014-01-03 03:45:13 UTC (rev 17) @@ -6,12 +6,12 @@ private: -CCTK_REAL kern[2] TYPE=ARRAY TIMELEVELS=1 dim=2 size=\ +CCTK_REAL kern[2] TYPE=ARRAY dim=2 size=\ (NullGrid::N_ang_pts_inside_eq+2*(NullGrid::N_ang_ev_outside_eq+NullGrid::N_ang_stencil_size)),\ (NullGrid::N_ang_pts_inside_eq+2*(NullGrid::N_ang_ev_outside_eq+NullGrid::N_ang_stencil_size)) \ GHOSTSIZE=NullGrid::N_ang_ghost_pts,NullGrid::N_ang_ghost_pts "kernel of the integral" -CCTK_REAL area TYPE=ARRAY TIMELEVELS=1 dim=2 size=\ +CCTK_REAL area TYPE=ARRAY dim=2 size=\ (NullGrid::N_ang_pts_inside_eq+2*(NullGrid::N_ang_ev_outside_eq+NullGrid::N_ang_stencil_size)),\ (NullGrid::N_ang_pts_inside_eq+2*(NullGrid::N_ang_ev_outside_eq+NullGrid::N_ang_stencil_size)) \ GHOSTSIZE=NullGrid::N_ang_ghost_pts,NullGrid::N_ang_ghost_pts "integral algorithms internal variable"

1 0

[svn:einsteintoolkit] NullSHRExtract/branches/tapir/src/ (Rev. 42)
by bela＠caltech.edu 02 Jan '14

02 Jan '14

User: szilagyi Date: 2014/01/02 09:40 PM Modified: /branches/tapir/src/ NullSHRE_ExtractWT.F90 Log: remove unused symbol File Changes: Directory: /branches/tapir/src/ =============================== File [modified]: NullSHRE_ExtractWT.F90 Delta lines: +0 -2 =================================================================== --- branches/tapir/src/NullSHRE_ExtractWT.F90 2013-12-28 04:57:46 UTC (rev 41) +++ branches/tapir/src/NullSHRE_ExtractWT.F90 2014-01-03 03:40:06 UTC (rev 42) @@ -38,8 +38,6 @@ CCTK_INT, save :: reduce_handle character(len=500) :: message -#define dt cctk_delta_time - !--------------------------------------------------------------------------- ! world-tube, and analytic part of the coordinate transformation !---------------------------------------------------------------------------

1 0

[svn:einsteintoolkit] NullEvolve/branches/tapir/src/ (Rev. 19)
by bela＠caltech.edu 02 Jan '14

02 Jan '14

User: szilagyi Date: 2014/01/02 09:34 PM Modified: /branches/tapir/src/ make.code.defn Log: will no longer compile NullEvol_cfl_test.F90 and NullEvol_TimeStep.F90 The first of these is poorly documented and has use of magic numbers such as '76' in index positions. The second is setting cctk_delta_time and updating cctk_time from within fortran by simply setting values to these as if they were ordinary variables. File Changes: Directory: /branches/tapir/src/ =============================== File [modified]: make.code.defn Delta lines: +2 -2 =================================================================== --- branches/tapir/src/make.code.defn 2013-12-28 04:57:15 UTC (rev 18) +++ branches/tapir/src/make.code.defn 2014-01-03 03:34:36 UTC (rev 19) @@ -2,9 +2,9 @@ # $Header$ # Source files in this directory -SRCS = NullEvol_Bdry.F90 NullEvol_cfl_test.F90 NullEvol_Evol.F90 \ +SRCS = NullEvol_Bdry.F90 NullEvol_Evol.F90 \ NullEvol_hyper_beta.F90 NullEvol_hyper_u.F90 NullEvol_hyper_q.F90 NullEvol_hyper_w.F90 \ - NullEvol_InitialData.F90 NullEvol_TimeStep.F90 NullEvol_Mask.F90 NullEvol_DissipMask.F90 \ + NullEvol_InitialData.F90 NullEvol_Mask.F90 NullEvol_DissipMask.F90 \ NullEvol_Pulse.F90 NullEvol_Step.F90 NullEvol_sYlm.F90 NullEvol_InitialSlice.F90 \ NullEvol_Diag.F90 NullEvol_DiagMod.F90

1 0

[McLachlan]McLachlan annotated tag, ET_2013_11_v0, created. ET_2013_11_v0
by Carpet git repositories 27 Dec '13

27 Dec '13

This is an automated email from the git hooks/post-receive script. It was generated because a ref change was pushed to the repository containing the project "McLachlan". The annotated tag, ET_2013_11_v0 has been created at 018731f08ec5767d6ba47841744b49182211b3df (tag) tagging 18fe14998dac0391740f14a2bc5a8a2f014eb36b (commit) replaces ET_2013_05_v0 tagged by Frank Löffler on Fri Dec 27 22:59:34 2013 -0600 - Log ----------------------------------------------------------------- create tag ET_2013_11_v0 -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.12 (GNU/Linux) iQIcBAABCAAGBQJSvlq2AAoJEOkzpip+I59ksBQP/2aqGYKcgBeqSGw1/fSWJ2yV D/Zlcm3+LvvKfA0DqmApqwrxC+cUniwuTQxnSYMfECSJT3wjlaMhSCIIIDP3ndbF q1Qw+4OWEbyBqC1QeWhm19RaCfBSeaBakWg3ts3dc6GjrxO8v0eDt7q0YUiUtHFB n4ghP2UI3R95ZKdaF7Ds521nqp8O19n0xN9dTn2ZRzbhLwFV5/auWXx/5c6FmgW1 wf5sFnbKtBuISp6X4yKmvkd7mBFP/7mTLs1F228XIBV0cas3Db9DeSs7Tz0jMmNZ xgEpKQbYipq7mhxNJwtia99grpjkzrpBhMPQQPcEOa7twB3k6lqqgiDpmQZsS48E eRqdZRhfCE+iPGJf9f/EcfrS7gjBLe1uKGih9ZHluaKNo0i6hQduwIYf//j7Hz8O 5uYZmo07KwKhmtDeJj5mpwZ5muM3eBFlYQoJf2jS4IR9je8i4x2vnkB9RypRiDKb GFyt37yA05IZqLH+KMNzP7MCPkukqJhW4Fh9ae48O33CpqHzP1bBSMtEQFtGybeH 9pDdrvxRON/LxRf18Z4DmFX7g8eeUI9Mk26MQwHMPq8/QQLnDyD8WbNE26nRuK9K 4DyF2aomufltQNwRBYx+z95iSeLEzMAeqHLSfKu/XolNnVK7n0tAYCj7/l0anCO/ 0ffdTAZmn3m0vRz/AiHA =yxvq -----END PGP SIGNATURE----- Frank Löffler (1): par/CCZ4: remove parfiles because they don't work and never worked. Ian Hinder (13): Regenerate McLachlan (Kranc commit caf12ddfb11ccd9a7bf822cc8ec0d7691cded0aa) Regenerate ML_BSSN and ML_CCZ4 (Kranc commit 296e00b7388c67e7267116b58c3270c22dd132da) Regenerate code (Kranc commit 10daf4a83a783a812c7b5a3aba0df115f0eac98f) Regenerate code (Kranc commit 04da452636ed006901602e924900ea71378f55b5) Regenerate code (Kranc commit 9b459add6db98620df418ef12d61c932cec9bf05) Regenerate ML_ADM, ML_ADMConstraints and ML_ADMQuantities with fixed Jacobian code Regenerate code (Kranc commit 01478e8bf1bc509f67a3f8fa2c18021dc9b975f3) Regenerate code (Kranc commit 5238aa95745f46501d3f1ac657a02af0b0a44477) Regenerate code (Kranc commit 311d24d35eab704df7153748d0f4f205eb1c081b) Regenerate code (Kranc commit 0b09169bacd7995d28b56a574bd060ebb243ca6f) Regenerate code (Kranc commit da423f55aa544d5648c839195ca3407f0763cfa5) Regenerate code (Kranc commit 3237af2566109f099ee78a7effb511a65dc18ae0) Regenerate code (Kranc commit 20e7a8feb560e494efe6b00db07af8b350fc9783) ----------------------------------------------------------------------- hooks/post-receive -- McLachlan

1 0

[ianhinder/Kranc]
by knarrff 27 Dec '13

27 Dec '13

Branch: refs/tags/ET_2013_11_v0 Home: https://github.com/ianhinder/Kranc

1 0

[Carpet] annotated tag, ET_2013_11_v0, created. ET_2013_11_v0
by Carpet git repositories 27 Dec '13

27 Dec '13

This is an automated email from the git hooks/post-receive script. It was generated because a ref change was pushed to the repository containing the project "Carpet — Adaptive Mesh Refinement for the Cactus Framework <http://carpetcode.org/>". The annotated tag, ET_2013_11_v0 has been created at 742f55da2047ca7cf03ffe91afdca2ed8d0c1e2e (tag) tagging 390fe95c7756d5102226ccb48eadbcee62e1a52d (commit) replaces carpet-6-init tagged by Frank Löffler on Fri Dec 27 22:58:48 2013 -0600 - Log ----------------------------------------------------------------- create tag ET_2013_11_v0 -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.12 (GNU/Linux) iQIcBAABCAAGBQJSvlqIAAoJEOkzpip+I59k4+YP/0MCgAnnnouuN5BTtdJNoJuM FNbwjaKH/9h0rifVBwjZDpyTjRtoeIfGQ6CZ8B6qXTwl4BmoTGxjwAyamPVQbmvH tPeDxj5xLhlwalq0Cdlh7lS2CEOd5oxM68r43caAI1mZ6V2NiV1UPKI9ipqGbhoq Q2JbGzFT8rVFowAMJJZlL2vqiNb5z1BVUAdecjFcEqRW36JtbgYoVuT+DgE9L+80 HM9+Dd+gCPXKDz0v1xkA4kKA3fKR8rmiFtLIA3ImdlPEBvNmTif0JlRptRi7yf3r ER7GRqLAy2xKgrKR+n/r/20j7IZpNpBdGeptz6/cJP7Hm8x9xl+GAxM/e4165acV N+VgwBio4c8R2g91OFq0GnP6hKXndCB3mo00NCJvZTQAZCRYbXQAzp7s37nKQCOG IirZ/KVnws0dBNGGuuXhHsDGru6e6kkvutXjrOwjhSAUmSIJqtzStBXOHiXw2Dym lS1EmDyLNncZE9WfFfvg8taBJfnkJo9xcJr62xzZou4N7jUYqlNnQSTD5qfZL0Hp yTZohf/SVax3lP6uktJNCO5/qbtrfE7Xq0A9ymbm1skCSKfDuRj0QuYOR6AmLlOT cinGJzbKpe/bwKG+1o6RpOLqA7E7lVOvbfumFO2BES9slGSjiwjqgEp0DxlutEni E4Zl0US/5xEj2khObx2q =EhEs -----END PGP SIGNATURE----- Anton Khirnov (2): CarpetIOHDF5: remove a superfluous H5Pclose call. CarpetIOHDF5: avoid variable shadowing. Christian Reisswig (3): Introduce parameter "additional_overlap_zones" to control the amount of additional (user-requested) overlap zones. CarpetInterp: New ENO2 interpolator. Merge branch 'master' of carpetcode.org:carpet David Radice (1): CarpetIOHDF5: Sliced HDF5 output index file not stored correctly Erik Schnetter (312): CarpetLib: Allow assigning a region to itself CarpetLib: Allow (again) merging regions that contain process trees Carpet: Disable storage for point_classification CarpetLib: Implement bboxset::contains(vect) CarpetLib: Store level_boxes in dh structure CarpetLib: Do not interpolate from buffer zones, if requested CarpetIOASCII: Pass vectors by reference CarpetIOASCII: White space change CarpetIOASCII: Implement using grid::coordinates to output coordinates Carpet: Remove outdated backtrace function declarations CarpetLib: Clean up backtrace generation Carpet: Correct errors with adaptive time stepping LoopControl: Change printstats default to "no" Remove Carpet's mechanisms to deal with how to call isnan CarpetIOF5: Correct calls to Carpet::Timer Carpet: Correct #ifdefs in CarpetSimpleMPIDatatypeLength Carpet: Requirements: Add aliased function to invalidate variables Carpet: Re-indent timer macros CarpetLib: Remove restrict qualifier from pointer cast Carpet: Correct errors in adaptive step size handling Carpet: Requirements: Warn about unnecessary syncs Carpet: Requirements: Add precondition to clause_t::active_on_timelevel CarpetRegrid2: Perform domain decomposition only when a refinement level changes CarpetRegrid2: Correct snapping to coarse grid for cell centering Carpet, CarpetRegrid2: When comparing refined regions treat non-existing levels as empty Carpet: Execute CPINITIAL only after initial data setup, not after recovering Carpet, CarpetLib: Move #define _GNU_SOURCE to correct file CarpetLib: Update cycle.h from upstream CarpetIOF5: Free HDF5 data when closing files Carpet: Add test cases checking cctk_time during initial data setup Carpet: Set NPROCS for kasner test cases Carpet: Correct comment CarpetRegrid2: Use "isfinite" instead of "not isnan" Carpet: Correct switch statement in CarpetMPIDatatype CarpetLib: Use correct MPI reduction operator for product reductions CarpetLib: Add self-checks to MPI reduction operators CarpetReduce: Avoid duplicate local variables CarpetReduce: Use self-defined MPI operators for complex numbers CarpetReduce: Do not access NULL pointer CarpetMask: Add missing OpenMP reduction clause CarpetLib: Define dh members tallying the size of refinement boundaries LoopControl: Add missing copyprivate CarpetIONirvana: Require MPI Allow padding in transport operators CarpetLib: Re-enable OpenMP collapse clauses CarpetLib: Set up all timers on all processes all the time CarpetLib: Make th::time_interpolation_during_regridding public CarpetIOF5: Various improvements Carpet: Simplify code Carpet: Turn time level / prolongation operator warning into an info message CarpetLib: Support padding in enovol prolongation operator Carpet: Re-apply init_fill_timelevels after postregridinitial CarpetLib: Use "process" instead of "processor" in comments and output CarpetLib: Add default case to switch statement CarpetLib: Correct process number in bbox output Carpet: Initialise TimerNode fields in correct order Carpet: Avoid unused and uninitialised variables Carpet: Check return value of dup() CarpetRegrid2: Correct output of refinement level during AMR CarpetRegrid2: More verbose info output during regridding CarpetLib: Check HRSCC_GLL_ELEMENT_HH instead of HRSCC_HH CarpetLib, CycleClock: Move cycle counting clock into its own thorn Carpet: Use dup2 instead of dup do control destination file descriptor CarpetLib: Ensure allocated memory is aligned for vectorisation CycleClock: Reduce wait time for determining cycle tick CycleClock: Determine elapsed time via integer arithmetic CycleClock: Optimise x86-64 PGI implementation CycleClock: Support Blue Gene/Q CycleClock: Abort compile if architecture is not supported CycleClock: Undo debugging change Carpet: Remove backup files CarpetReduce: Remove empty file Carpet: Initialise all timing variables CarpetReduce: Support array padding (cctk_ash) CarpetReduce: Clean up code defining reduction weight masks CycleClock: Handle Cray compiler CarpetLib: Rewrite array padding LoopControl: Rewrite LoopControl: Circumvent Cray compiler bug LoopControl: Run tests only with 2 processes CarpetLib: Remove unused variables CarpetIOBasic: Simplify code and avoid compiler warnings CarpetIOHDF5: Remove unused variables, initialise other variable to avoid compiler warnings CarpetIOHDF5: Make variable unsigned to avoid compiler warning CarpetLib: Explicitly convert values to signed CarpetLib: Use C++ wrapper madd instead of mul and add CarpetLib: Disable DGFE operators if building without DGFE to avoid compiler warnings CarpetIOBasic: Simplify code and avoid compiler warnings (again) Carpet: Correct load-balancing algorithm to really prefer splitting in the z direction LoopControl: Store loop bounds in the provided variables LoopControl: Correct indexing error with vectorisation CarpetIOASCII: Update test results after domain decomposition change Carpet: Return cctk_ash instead of cctk_lsh in CCTK_ArrayGroupSizeB Carpet: Use cctk_ash where necessary in poisoning and checksumming CarpetLib: Rename argument to vect::index from lsh to ash LoopControl: Add assert statements to loop macros to catch indexing errors LoopControl: Update Fortran macros to follow cctk_Loop API TestLoopControl: New thorn TestLoopControl: Use module loopcontrol LoopControl: Correct Fortran macros; provide Fortran wrappers LoopControl: More self-checking TestLoopControl: Correct errors in testing ghost zones LoopControl: Rename macro arguments to help avoid name clashes CarpetLib, LoopControl: Correct required/optional capabilities CarpetLib: Implement and correct head, tail, init, last in the vect class CarpetRegrid2: Correct min_fraction LoopControl: Correct Fortran OpenMP definitions LoopControl: Add debug assert statements when array padding is enabled LoopControl: Avoid compiler warnings CarpetRegrid2: Output all verbose messages with veryverbose as well CarpetRegrid2: Correct another error in handling min_fraction LoopControl: Reduce number of arguments passed to selftest routine LoopControl: Remove restrict qualifiers from shared data structures CarpetLib: Change default to interpolate_from_buffer_zones=no Carpet: Beautify screen output CarpetLib: Do not memoize OpenMP thread id LoopControl: Remove trivial vector lengths for j and k direction LoopControl: Simplify STL data type usage LoopControl: Improve OpenMP parallelisation CarpetProlongateTest: Correct default parameter values for interpolator test LoopControl: Output statistics to a file instead of stdout LoopControl: Share cache lines for SMT threads LoopControl: Correct parallelisation errors (mark variables volatile) CarpetReduce: Improve code layout Revert "CarpetLib: Change default to interpolate_from_buffer_zones=no" CarpetLib: Store level-wide set of buffer zones CarpetLib: Correct implementation of "don't interpolate from buffer zones" CarpetLib: Set interpolate_from_buffer_zones to "no" by default Carpet: New aliased function GetRegriddingEpochs CarpetIOASCII: Use CCTK_BUILTIN_UNREACHABLE CarpetIntegrateTest: Use ReflectionSymmetry for bitant mode Replace Cactus complex number type with C/C++ complex numbers CarpetProlongateTest: Increase test case tolerances for higher order prolongation CarpetIntegrateTest: Modernize test parameter file Carpet: Add CactusTimer functions to get all clock values of a timer Carpet: Output timer tree at shutdown CycleClock: Add parameter to disable registering a Cactus clock LoopControl: Output statistics into per-process files, not into a single file Carpet: Output all clocks in the timer tree Carpet: Use C++ style cast to avoid compiler warning Carpet: Correct output of load balancing statistics CarpetReduce: Don't require a valid cctkGH unless really needed CarpetReduce: New parameter min_max_time_interpolation LoopControl: Simplify #if sequence LoopControl: Improve OpenMP parallelisation of get_num_fine_threads CarpetIO*: Output complex numbers as pair of real numbers Requirements: New thorn Replace abs() by fabs() Requirements: Use vector instead of set to store ignored variables Requirements: White space cleanup Carpet: Call Requirements before, not after the respective actions Requirements: Beautify code Requirements: Remove dependency on defs.hh Requirements: Remove duplication from info messages Requirements: Add todo-note to invalidate non-written regions in partial projections Requirements: Break lines to make code more readable Requirements: Minor code re-shuffling and cleanup Carpet, Requirements: Pass timelevel_offset Requirements: Some source code cleanup Requirements: Add Frank's html generator script Requirements: Rename parameter print_changes to output_changes Requirements: Simplify logic to detect changes in gridpoint_t Requirements: Warn about Requirements.hh being publicly visible Requirements: Re-order if statement (no change in meaning) Requirements: Modify iteration arguments in Requirements.hh Carpet: Do not restrict variables marked as "don't restrict" Requirements: Remove some duplicate code Requirements: Correct handling of timelevels in BeforeRoutine and AfterRoutine Carpet: Don't call Requirements when only asking about allocated storage Requirements: Don't use identifiers starting with an underscore Requirements: Unify order of #include statements Requirements: Change location output format Requirements: Output iteration number in warnings and error messages Requirements: Change location output order; output map before timelevel Requirements: Use location_t to output warnings and errors CarpetLib: Remove unused/unfinished bboxtree class Carpet: Exit restriction early if no groups should be restricted Carpet: Small performance improvement in SyncGroupsByDir Carpet: Take tapering into account during postrestrict Requirements: Whitespace change CarpetLib: Vectors is currently required, not optional CarpetReduce: Offset reduction weights when periodic boundaries are used CarpetLib: Correct alignment mechanism when allocating memory CarpetIOF5: Correct calculating the I/O processor numbers CarpetIOF5: Do not output irrelevant "Writing simulation metadata" messages CarpetInterp2: Convert some loop bounds to int CarpetIOASCII: Indentation change CarpetLib: Add some comments explaining implementation CarpetLib: Add constructors to create a vector from a low-dimensional vector and a scalar CarpetLib: Rename vect::zip to vect::vzip CarpetLib: Change meaning of vect::count to count true (non-zero) elements CarpetLib: Only check cache line size if it is used CarpetLib: More verbose error output if sanity check fails CarpetLib: Don't explicitly #include cctk_Config.h CarpetLib: Don't define static_assert CarpetLib: Correct isnorma() for complex numbers CarpetLib: Add routines for integer division that always round down CarpetLib: Disable unused refluxing face descriptions CarpetLib: Add missing #include files to cacheinfo.cc CarpetLib: Correct some type declarations for 4d grid structures CarpetLib: Remove now outdated definitions of idiv and imod CarpetLib: Provide idiv and imod implementations for vect<CCTK_REAL,...> CarpetIO*: New parameter CarpetIO*::out_group_separator CarpetLib: Provide rint, round, trunc, and lrint for vectors CarpetLib: Correct vect::count CarpetLib: Use C++ standard methods to compare containers CarpetLib: Use imod instead of % operator in bbox class CarpetLib: Remove superfluous check that all indices are non-negative CarpetLib: New function bbox::is_poison to check for poison CarpetLib: Replace more uses of operator% by calls to imod CarpetLib: New function bbox::anti_contracted_for CarpetLib: New class bboxset2 LoopControl: Implement automatic performance improvements TestBBoxSet2: New thorn Requirements: Re-order #include statements Requirements: Enforce 80 character line length limit Requirements: Check return value of CCTK_TraverseString CarpetIOF5: Implement lower-dimensional output (i.e. 1d lines, 2d slices) Carpet: Use fabs() instead of abs() Carpet: Provide global instead of process-local timer output Carpet: Disable automatic synchronisation after restriction Carpet: Remove unused, broken code Carpet: Output list of compute nodes only in verbose mode CarpetLib: Disable bboxset2 by default Requirements: Do not mark current timelevel as invalid after time level cycling CarpetLib: Provide manual look-alikes for certain C++11 features CarpetLib: Instantiate some templates that are necessary on some systems CarpetLib: Don't use run-together right angle brackets in templates CarpetLib: Whitespace change CarpetLib: Re-implement part of bboxset2 for pre-C++11 compilers CarpetLib: Add optional dependency on Boost CarpetLib: Update GetCacheInfo1 API LoopControl: Update GetCacheInfo1 API CarpetIOF5: Use CCTK_BUILTIN_UNREACHABLE instead of __builtin_unreachable CarpetLib: Use CCTK_BUILTIN_UNREACHABLE instead of __builtin_unreachable CarpetReduce: Use CCTK_ERROR CarpetReduce: Add test case for reduction weights on periodic domains LoopControl: Do not output statistics by default LoopControl: Turn hard-coded parameters into Cactus parameters LoopControl: Use random() instead of rand() LoopControl: Sort statistics output LoopControl: Describe unoptimized time with respect to actual time, not vice versa LoopControl: Improve error output handling LoopControl: Correct logic that chooses tile sizes LoopControl: Correct example parameter file descriptions LoopControl: Add example parameter file for dynamic optimizations CarpetLib: New parameter use_loopcontrol_in_operators CarpetLib: Use OpenMP collapse clause in Fortran transport operators LoopControl: Declare some variables as potentially unused CarpetLib: Calculate stepped buffer zones only in debug mode CarpetLib: Do not use LoopControl for restriction operators when parallelizing via gdata.cc LoopControl: Use C++ constructure to initialize lc_descr_t LoopControl: Whitespace change LoopControl: Add OpenMP barrier after loop ends LoopControl: Correct check whether vectorization is enabled Carpet: Support group tag "map0group" Carpet: Correct "map-0-group" implementation Carpet: Do not use dynamic_cast CarpetLib: Store registered gh, dh, th, gf, data etc. via sets, not via lists CarpetLib: Add comment in gdata class CarpetLib: Remove some outdated, commented-out code Carpet: Add aliased function Driver_TotalMemoryUsed CarpetLib: Use posix_memalign to allocate aligned memory CycleClock: Automatically start clock when it is created CarpetRegrid2: Always set do_recompose when adaptive_refiment is true Carpet: Store amount of allocated memory in grid scalars Carpet: Turn timer calls into no-ops when timer hierarchy does not exist Carpet: Release all memory when shutting down CarpetLib: Free ggf data structures in ggf, not in gf<T> CarpetLib: Ensure that gh/dh/th destructors are called in the right order CarpetLib: Ensure oldstorage is empty in ggf destructor Carpet: Remove aliased function Driver_TotalMemoryUsed LoopControl: Correct indentation in statistics screen output LoopControl: Output LoopControl's memory usage CarpetLib: Do not reallocate communication buffers; instead, keep them around Timers: Move all timer-related code into a new thorn Timers Timers: Add parameters to disable all timers CarpetLib: Some harmless code cleanup CarpetLib: Small code simplification CarpetLib: Use CCTK_Error LoopControl: Make explore_eagerly_before_iteration and settle_after_iteration work LoopControl: Turn fine thread communicators into global (static) variables LoopControl: Introduce memory pool for lc_params and lc_setup LoopControl: Optimize strlen LoopControl: Output final statistics also if periodic statistics were requested Carpet: Output timers before they are destructed at Shutdown CarpetLib: Correct type error for nbytes in memory allocation CarpetLib: Use malloc (again) instead of posix_memalign LoopControl: Avoid non-standard C++ feature LoopControl: Update Fortran macros CarpetLib: Optimize memory allocation when message_count_multiplier>1 CarpetLib: Correct memory allocation error in communication buffers LoopControl: Initialize num_smt_threads explicitly during loop startup LoopControl: Explicitly check for consistent loop arguments from the caller LoopControl: Add debug output to help track down assertion failure LoopControl: Initialize fine thread communicators ahead of time LoopControl: Disable debug checks CarpetInterp2: Explicitly convert values to CCTK_REAL CarpetRegrid2: Correct wording of error message Carpet: Introduce parameter max_time_levels Carpet: Modernize error reporting Carpet: Comment out unused routine Carpet: Initialise number of timelevels at the right time CarpetLib: Add sanity check for number of time levels Carpet: Make dependency on CycleClock optional CycleClock: Do not abort when no good clock is available CarpetLib: Define elapsed() when CycleClock provides no timer CarpetInterp2: Make code compile when CARPETINTERP2_CHECK is not set CarpetInter2: Avoid const warnings CarpetLib: Remove commented-out code CarpetLib: Remove unnecessary parentheses LoopControl: Avoid nans in statistics output Frank Löffler (16): CarpetIOHDF5: cast string constant to char array to get rid of compiler warning CarpetIOHDF5: comment unused variable (commented because code using it is also commented CarpetRegrid2: Allow pseudo-2D domains CarpetRegrid2: Testsuite for trivial domains in z direction Merge branch 'master' of carpetcode.org:carpet output more verbose info (variable index and nr of timelevels make gridpoint_t a class make bool members of gridpoint_t class private, provide accessors and use them Merge branch 'master' of carpetcode.org:carpet Let set_ accessors return whether something was changed Make set_* functions be void again, add location_t struct (unused so far) use parameter to enable output if state of variable changes Requirements: pass iteration down and print in debug output; also print when something changes (readable string) swap upper/lower case for location output Requirements: fix a couple of bugs in the debug output CycleClock: define CycleClock_Setup() without HAVE_TICK_COUNTER Ian Hinder (21): Support more than one hierarchy of timers Add a timer tree for measuring time spent in each Carpet mode Requirements: Add comments Requirements: Move clause class into separate files Requirements: Move clauses class and some utility functions to separate files Requirements: Move all_clauses class to a separate file Requirements: Move location class to a separate file Revert part of "Requirements: Beautify code" Requirements: Move gridpoint class to a separate file Requirements: Move all_state class to separate files Requirements: Move flag variables into gridpoint class Requirements: Pass all_clauses as a reference argument rather than using a global variable Requirements: Remove unneeded #includes Rename new max_time_levels parameter as max_timelevels to be consistent with other uses of "timelevels" in parameter names in Carpet CarpetInterp2: Add timers for fasterp interpolation CarpetInterp2: Add parameter "interp_barrier" to add a barrier before interpolating CarpetInterp2: Add a barrier after interpolating Timers: Output all clocks in XML timer tree files CarpetInterp2: Time only interpolation using Interpolate timer, and enable more timers for tracing Carpet: Add a barrier between prolongation and sync controlled by sync_barriers parameter Carpet: Remove accidentally-committed extra arguments to Timer constructor Peter Diener (3): Fix in order to compile with IBM compilers. Merge branch 'master' of carpetgit@carpetcode.org:carpet Reduce time spent in cycling timelevels if there is only one timelevel. Roland Haas (42): Carpet: fix mg level count typo during intialization CarpetIOASCII: do not output coordinate column labels for grid arrays in CarpetIOHDF5: output index files for sliced output if requested CarpetIOHDF5: close all HDF5 objects when output file is closed CarpetIOHDF5: fix check for missing datasets when called in FileReader mode CarpetLib: add volume averaged cc eno operator CarpetLib: support padding in non-OpenMP transport operator CarpetLib: clear boundary slice offsets in dh::regrid() CarpetProlongateTest: restrict tests to 2 processes, reduce regions, CarpetEvolutionMask: compute a mask indicating where buffers are CarpetEvolutionMask: limit buffer_mask test to one process CarpetIOHDF5: remove index files along with checkpoint files Requirements: add missing include file LoopControl: disable smt by default CarpetProlongateTest: add tests for higher order restriction CarpetIOHDF5: delay browsing of input files CarpetIOHDF5: support IO->alias option in reader CarpetIOHDF5: use simple_extents rather than storage size on disk CarpetIOHDF5: fix reading of new-style grid structures CarpetIOHDF5: remove two compiler warnings CarpetIOHDF5: check whether map exists before accessing it CarpetProlongateTest: update test data CarpetIOHDF5: correct output method name in INFO message Carpet: correct wording of error message about number of time levels CarpetLib: add some code for electric fence like functionality Carpet: use electric fence provided by CarpetLib CarpetEFenceTest: test electric fence in CarpetLib Carpet: add new group tag no_split_directions Carpet: fix no_split_directions implementation CarpetSlab: pass size of allocated chunk to memory object CarpetTest: make compile again CarpetTest: recreate test data in new CarpetIOASCII format CarpetTest: add no_split_in_directions tag Carpet: remove duplicate myproc variable in SetupGH CarpetLib: make electric_fence parameter restricted Carpet: skip CheckFence routine if electric_fence is off CarpetLib: use map instead of set for dh::gfs LoopControl: provide omp_get_max_threads fallback CarpetLib: code layout changes when using map for allgfs Revert "CarpetLib: make electric_fence parameter restricted" CarpetLib: introduce accessor in gdata to query electric_fence Carpet: use gdata::fence_is_energized to query electric_fence ----------------------------------------------------------------------- hooks/post-receive -- Carpet — Adaptive Mesh Refinement for the Cactus Framework <http://carpetcode.org/>

1 0

[barrywardell/EinsteinExact]
by knarrff 27 Dec '13

27 Dec '13

Branch: refs/tags/ET_2013_11_v0 Home: https://github.com/barrywardell/EinsteinExact

1 0

numrel/LSUThorns/PeriodicCarpet SVN commit 21: /tags/ET_2013_11_v0/
by knarf＠cct.lsu.edu 27 Dec '13

27 Dec '13

User: knarf Date: 2013/12/27 10:51 PM Added: /tags/ET_2013_11_v0/ Log: ET_2013_11_v0 release tag

1 0

[svn:einsteintoolkit] incoming/PITTNullCode/SphericalHarmonicReconGen/src/ (Rev. 106)
by rhaas＠tapir.caltech.edu 27 Dec '13

27 Dec '13

User: rhaas Date: 2013/12/27 04:16 PM Modified: /PITTNullCode/SphericalHarmonicReconGen/src/ startup.cc Log: accept path parameter with or without trailing "/" File Changes: Directory: /PITTNullCode/SphericalHarmonicReconGen/src/ ======================================================= File [modified]: startup.cc Delta lines: +14 -8 =================================================================== --- PITTNullCode/SphericalHarmonicReconGen/src/startup.cc 2013-12-27 22:13:16 UTC (rev 105) +++ PITTNullCode/SphericalHarmonicReconGen/src/startup.cc 2013-12-27 22:16:05 UTC (rev 106) @@ -66,8 +66,7 @@ // open database, i.e. scan input file containing the harmonic coefficients CCTK_INFO("Opening and scanning file containing harmonic coefficients for the Cauchy variables."); - string pathname = path; - CCTK_INFO(string(string("Trying to open ASCII files in ")+pathname).c_str()); + CCTK_VInfo(CCTK_THORNSTRING,"Trying to open ASCII files in \"%s\".", path); // get all different databases-names vector<string> fnames; @@ -122,17 +121,24 @@ } if (found < 0) { - CCTK_INFO(string(string("Parsing ")+pathname+fnames[i]).c_str()); + string fullname = path; + if (!fullname.empty() && fullname[fullname.size()-1] != '/') { + // an empty string means that the filenames contain full paths + fullname.append("/"); + } + fullname.append(fnames[i]); + + CCTK_VInfo(CCTK_THORNSTRING, "Parsing %s", fullname.c_str()); if (CCTK_EQUALS(format, "ASCII")) - SHR::db[i] = new SHR::SPH_db_ASCII(pathname+fnames[i], verbose, cached_timesteps+1, column_time, column_iteration, column_radius); + SHR::db[i] = new SHR::SPH_db_ASCII(fullname, verbose, cached_timesteps+1, column_time, column_iteration, column_radius); if (CCTK_EQUALS(format, "DAT")) - SHR::db[i] = new SHR::SPH_db_DAT(pathname+fnames[i], verbose, cached_timesteps+1, lmaxInFile, column_time, column_iteration, column_radius, column_lmax, column_n_variables, column_data, false); + SHR::db[i] = new SHR::SPH_db_DAT(fullname, verbose, cached_timesteps+1, lmaxInFile, column_time, column_iteration, column_radius, column_lmax, column_n_variables, column_data, false); if (CCTK_EQUALS(format, "DAT-v2")) - SHR::db[i] = new SHR::SPH_db_DAT(pathname+fnames[i], verbose, cached_timesteps+1, lmaxInFile, column_time, column_iteration, column_radius, column_lmax, column_n_variables, column_data, true); + SHR::db[i] = new SHR::SPH_db_DAT(fullname, verbose, cached_timesteps+1, lmaxInFile, column_time, column_iteration, column_radius, column_lmax, column_n_variables, column_data, true); if (CCTK_EQUALS(format, "SpEC-H5")) - SHR::db[i] = new SHR::SPH_db_SpEC_H5(pathname+fnames[i], verbose, cached_timesteps+1); + SHR::db[i] = new SHR::SPH_db_SpEC_H5(fullname, verbose, cached_timesteps+1); if (CCTK_EQUALS(format, "SpEC-H5-v2")) - SHR::db[i] = new SHR::SPH_db_SpEC_H5_v2(pathname+fnames[i], verbose, cached_timesteps+1); + SHR::db[i] = new SHR::SPH_db_SpEC_H5_v2(fullname, verbose, cached_timesteps+1); // check if timestep-size found in file matches characteristic timestep! if (fabs(SHR::db[i]->delta_t()-CCTK_DELTA_TIME) > 1e-8)

1 0

[svn:einsteintoolkit] incoming/PITTNullCode/SphericalHarmonicReconGen/src/ (Rev. 105)
by rhaas＠tapir.caltech.edu 27 Dec '13

27 Dec '13

User: rhaas Date: 2013/12/27 04:13 PM Modified: /PITTNullCode/SphericalHarmonicReconGen/src/ sph_database.cc Log: add error checking around HDF5 calls in particular abort when HDF5 file cannot be opened File Changes: Directory: /PITTNullCode/SphericalHarmonicReconGen/src/ ======================================================= File [modified]: sph_database.cc Delta lines: +69 -37 =================================================================== --- PITTNullCode/SphericalHarmonicReconGen/src/sph_database.cc 2013-12-27 22:09:03 UTC (rev 104) +++ PITTNullCode/SphericalHarmonicReconGen/src/sph_database.cc 2013-12-27 22:13:16 UTC (rev 105) @@ -1,11 +1,26 @@ #include "sph_database.hh" +#include <cassert> namespace SHR { using namespace std; +// check return code of HDF5 call and print a warning in case of an error +#define HDF5_ERROR(fn_call) HDF5_CALL(__LINE__, __FILE__, CCTK_THORNSTRING, \ + fn_call, #fn_call) +static int HDF5_CALL(const int line, const char* file, + const char* thornstring, int error_code, + const char* fn_call) +{ + if (error_code < 0) + { + CCTK_VError (line, file, thornstring, + "HDF5 call '%s' returned error code %d", + fn_call, error_code); + } + return error_code; +} - SPH_db_ASCII::SPH_db_ASCII(const string& fname_, const bool verbose_, const int cached_timesteps, @@ -899,6 +914,10 @@ { // open file file = H5Fopen(_fname.c_str(), H5F_ACC_RDONLY, H5P_DEFAULT); + if (file < 0) { + CCTK_VError(__LINE__, __FILE__, CCTK_THORNSTRING, + "Could not open file \"%s\".", _fname.c_str()); + } scan(); } @@ -907,7 +926,7 @@ SPH_db_SpEC_H5::~SPH_db_SpEC_H5() { // close file - H5Fclose(file); + HDF5_ERROR(H5Fclose(file)); } @@ -977,10 +996,12 @@ CCTK_WARN(0, "Requested timestep not in worldtube data file. Stopping!"); } - hid_t dataset = H5Dopen(file, (varname_table[varno][0]+"/"+steps[timestep].name+"/"+varname_table[varno][1]).c_str());//, H5P_DEFAULT); + hid_t dataset = HDF5_ERROR(H5Dopen(file, (varname_table[varno][0]+"/"+ + steps[timestep].name+"/"+ + varname_table[varno][1]).c_str())); vector<double> buffer((_lmax+1)*(_lmax+1)*2); - H5Dread(dataset, H5T_NATIVE_DOUBLE, H5S_ALL, H5S_ALL, H5P_DEFAULT, &buffer.front()); - H5Dclose(dataset); + HDF5_ERROR(H5Dread(dataset, H5T_NATIVE_DOUBLE, H5S_ALL, H5S_ALL, H5P_DEFAULT, &buffer.front())); + HDF5_ERROR(H5Dclose(dataset)); int c = 0; for (int l=0; l <= _lmax; ++l) { @@ -1005,11 +1026,11 @@ if (gname.find(step_basename) != string::npos) { // get time attribute double time; - hid_t group = H5Gopen(loc_id, name); - hid_t attribute = H5Aopen(group, time_attrib_name.c_str(), H5P_DEFAULT); - H5Aread(attribute, H5T_NATIVE_DOUBLE, &time); - H5Aclose(attribute); - H5Gclose(group); + hid_t group = HDF5_ERROR(H5Gopen(loc_id, name)); + hid_t attribute = HDF5_ERROR(H5Aopen(group, time_attrib_name.c_str(), H5P_DEFAULT)); + HDF5_ERROR(H5Aread(attribute, H5T_NATIVE_DOUBLE, &time)); + HDF5_ERROR(H5Aclose(attribute)); + HDF5_ERROR(H5Gclose(group)); SpEC_H5->steps.push_back(timestep_t(name, time)); } @@ -1033,7 +1054,9 @@ // pick one variable and browse through entire file // to find out how many (and what) cycles and times there are // (assuming that all variables are present at the same timesteps) - H5Literate_by_name (file, varname_table[0][0].c_str(), H5_INDEX_NAME, H5_ITER_NATIVE, NULL, H5iter, this, H5P_DEFAULT); + HDF5_ERROR(H5Literate_by_name (file, varname_table[0][0].c_str(), + H5_INDEX_NAME, H5_ITER_NATIVE, NULL, + H5iter, this, H5P_DEFAULT)); // sort 'steps' vector //sort(steps.begin(), steps.end()); @@ -1046,13 +1069,17 @@ } // get lmax from first variable and timestep - hid_t dataset = H5Dopen(file, (varname_table[0][0]+"/"+steps[0].name+"/"+varname_table[0][1]).c_str()); //, H5P_DEFAULT); + hid_t dataset = HDF5_ERROR(H5Dopen(file, (varname_table[0][0]+"/"+ + steps[0].name+"/"+ + varname_table[0][1]).c_str())); //cout << (varname_table[0][0]+"/"+steps[0].name+"/"+varname_table[0][1]) << endl; - hid_t dataspace = H5Dget_space(dataset); - hsize_t size = H5Sget_simple_extent_npoints(dataspace); - H5Sclose(dataspace); - H5Dclose(dataset); + hid_t dataspace = HDF5_ERROR(H5Dget_space(dataset)); + hsize_t size = HDF5_ERROR(H5Sget_simple_extent_npoints(dataspace)); + HDF5_ERROR(H5Sclose(dataspace)); + HDF5_ERROR(H5Dclose(dataset)); + assert (size % 2 == 0); + _lmax = sqrt(size/2)-1; _n_variables = 30; _n_spheres = 1; @@ -1110,6 +1137,10 @@ { // open file file = H5Fopen(_fname.c_str(), H5F_ACC_RDONLY, H5P_DEFAULT); + if (file < 0) { + CCTK_VError(__LINE__, __FILE__, CCTK_THORNSTRING, + "Could not open file \"%s\".", _fname.c_str()); + } scan(); } @@ -1118,7 +1149,7 @@ SPH_db_SpEC_H5_v2::~SPH_db_SpEC_H5_v2() { // close file - H5Fclose(file); + HDF5_ERROR(H5Fclose(file)); } @@ -1216,27 +1247,28 @@ for (int v=0; v < _n_variables; ++v) { - hid_t dataset = H5Dopen(file, varname_table[v].c_str()); + hid_t dataset = HDF5_ERROR(H5Dopen(file, varname_table[v].c_str())); const int nmodes = (_lmax+1)*(_lmax+1)*2; vector<double> buffer(nmodes*nsteps); hsize_t dimsm[1] = { buffer.size() }; - hid_t memspace = H5Screate_simple (1, dimsm, NULL); + hid_t memspace = HDF5_ERROR(H5Screate_simple (1, dimsm, NULL)); - hid_t dataspace = H5Dget_space(dataset); + hid_t dataspace = HDF5_ERROR(H5Dget_space(dataset)); hsize_t dims[2]; - H5Sget_simple_extent_dims (dataspace, dims, NULL); + HDF5_ERROR(H5Sget_simple_extent_dims (dataspace, dims, NULL)); hsize_t offset[2] = { timestep, 1 }; hsize_t count[2] = { nsteps, nmodes }; - H5Sselect_hyperslab (dataspace, H5S_SELECT_SET, offset, NULL, - count, NULL); + HDF5_ERROR(H5Sselect_hyperslab (dataspace, H5S_SELECT_SET, offset, NULL, + count, NULL)); - H5Dread(dataset, H5T_NATIVE_DOUBLE, memspace, dataspace, H5P_DEFAULT, &buffer.front()); + HDF5_ERROR(H5Dread(dataset, H5T_NATIVE_DOUBLE, memspace, dataspace, H5P_DEFAULT, + &buffer.front())); - H5Sclose(memspace); - H5Sclose(dataspace); - H5Dclose(dataset); + HDF5_ERROR(H5Sclose(memspace)); + HDF5_ERROR(H5Sclose(dataspace)); + HDF5_ERROR(H5Dclose(dataset)); for (int t=0; t < nsteps; ++t) { @@ -1279,11 +1311,11 @@ */ // get lmax from first variable and timestep - hid_t dataset = H5Dopen(file, "gxx.dat"); //, H5P_DEFAULT); + hid_t dataset = HDF5_ERROR(H5Dopen(file, "gxx.dat")); //cout << (varname_table[0][0]+"/"+steps[0].name+"/"+varname_table[0][1]) << endl; - hid_t dataspace = H5Dget_space(dataset); + hid_t dataspace = HDF5_ERROR(H5Dget_space(dataset)); hsize_t dims[2]; - H5Sget_simple_extent_dims (dataspace, dims, NULL); + HDF5_ERROR(H5Sget_simple_extent_dims (dataspace, dims, NULL)); // first dim is number of timesteps steps.resize(dims[0]); @@ -1293,20 +1325,20 @@ // select hyperslab in dataset hsize_t offset[2] = { 0, 0 }; hsize_t count[2] = { dims[0], 1 }; - H5Sselect_hyperslab (dataspace, H5S_SELECT_SET, offset, NULL, - count, NULL); + HDF5_ERROR(H5Sselect_hyperslab (dataspace, H5S_SELECT_SET, offset, NULL, + count, NULL)); // define dataspace of memory hsize_t dimsm[1] = { steps.size() }; hid_t memspace = H5Screate_simple (1, dimsm, NULL); // read "times" - H5Dread (dataset, H5T_NATIVE_DOUBLE, memspace, dataspace, - H5P_DEFAULT, &steps.front()); + HDF5_ERROR(H5Dread (dataset, H5T_NATIVE_DOUBLE, memspace, dataspace, + H5P_DEFAULT, &steps.front())); - H5Sclose(memspace); - H5Sclose(dataspace); - H5Dclose(dataset); + HDF5_ERROR(H5Sclose(memspace)); + HDF5_ERROR(H5Sclose(dataspace)); + HDF5_ERROR(H5Dclose(dataset)); _n_variables = 30; _n_spheres = 1;

1 0

Jump to page:

2026

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

Commits