From schnetter at cct.lsu.edu Sun Jan 10 13:29:44 2010 From: schnetter at cct.lsu.edu (Erik Schnetter) Date: Sun, 10 Jan 2010 13:29:44 -0600 Subject: [Patches] Run MoL_PostStep after recovery Message-ID: <446AA32C-695C-4CA2-875E-DAC2DF5B3073@cct.lsu.edu> I suggest to run MoL_PostStep after recovery. This will ensure that all dependent variables are calculated, and boundary conditions are applied correctly. (Dependent variables are e.g. the ADMBase variables which may not be checkpointed.) MoL_PostStep is already run after initial data have been set up, after restriction, and after regridding. -erik -- Erik Schnetter http://www.cct.lsu.edu/~eschnett/ -------------- next part -------------- A non-text attachment was scrubbed... Name: mol.diff Type: application/octet-stream Size: 1092 bytes Desc: not available Url : http://www.cactuscode.org/pipermail/patches/attachments/20100110/ab075b57/attachment.obj From schnetter at cct.lsu.edu Mon Jan 11 14:02:51 2010 From: schnetter at cct.lsu.edu (Erik Schnetter) Date: Mon, 11 Jan 2010 14:02:51 -0600 Subject: [Patches] CactusUtils/TerminationTrigger Message-ID: <833AACF1-1C1C-4C33-AC02-B23F96EBD95C@cct.lsu.edu> I suggest to add a new thorn to CactusUtils: TerminationTrigger. This thorn watches the elapsed walltime. If only n minutes are left before the some limit set by the user, it triggers termination of the simulation. Termination is also triggered if a special file with a special content exists. This thorn was written by Christian Ott based on an original idea by Steve White. -erik -- Erik Schnetter http://www.cct.lsu.edu/~eschnett/ -------------- next part -------------- A non-text attachment was scrubbed... Name: TerminationTrigger.tar.gz Type: application/x-gzip Size: 9342 bytes Desc: not available Url : http://www.cactuscode.org/pipermail/patches/attachments/20100111/8251afb5/attachment-0001.gz From schnetter at cct.lsu.edu Mon Jan 18 14:18:37 2010 From: schnetter at cct.lsu.edu (Erik Schnetter) Date: Mon, 18 Jan 2010 14:18:37 -0600 Subject: [Patches] TimerReport: Output more than one clock Message-ID: <2B18B574-49B6-4A69-AF51-36834D662089@cct.lsu.edu> Currently, thorn CactusUtils/TimerReport outputs one clock only for each timer. This clock can be selected with the parameter all_timers_clock. This makes the output much more readable, since there may be many clocks. The enclosed patch adds a new array parameter all_timers_clocks[] that allows multiple clocks to be specified. They are output into a different file each, with the file name containing the clock name. This affects the output_all_timers, output_all_timers_together, and output_all_timers_readable functionality. Implementation-wise, there is now a for loop surrounding the output code that traverses all clocks. This makes is possible to output e.g. the wall time as well as several PAPI clocks. -erik -- Erik Schnetter http://www.cct.lsu.edu/~eschnett/ -------------- next part -------------- A non-text attachment was scrubbed... Name: timerreport.diff Type: application/octet-stream Size: 25353 bytes Desc: not available Url : http://www.cactuscode.org/pipermail/patches/attachments/20100118/2263a0b4/attachment.obj From knarf at cct.lsu.edu Tue Jan 26 16:31:09 2010 From: knarf at cct.lsu.edu (Frank Loeffler) Date: Tue, 26 Jan 2010 16:31:09 -0600 Subject: [Patches] CactusEinstein: Excision in HydroBase Message-ID: <20100126223109.GP18541@numrel07.cct.lsu.edu> Hi, the attched patch adds excision definitions for HydroBase: - New parameter "hydro_excision" to turn the storage for the excision mask on/off (default off) - Excision mask (being INTEGER for the moment, might change to BYTE later) - Initialization of the mask to 'normal' - Headerfile with definitions of the msak values: HYDROEXCISION_NORMAL / HYDROEXCISION_EXCISED - schedule group definition in POSTSTEP to schedle thorns setting the mask before that and thorns using the mask after it Frank -------------- next part -------------- Index: interface.ccl =================================================================== RCS file: /cactusdevcvs/CactusEinstein/HydroBase/interface.ccl,v retrieving revision 1.11 diff -u -r1.11 interface.ccl --- interface.ccl 8 Dec 2009 19:46:35 -0000 1.11 +++ interface.ccl 26 Jan 2010 21:07:15 -0000 @@ -4,6 +4,10 @@ inherits: InitBase +USES INCLUDE: HydroBase.h + +INCLUDE HEADER: HydroBase.h in HydroBase.h + public: # These variables correspond to the Valencia formulation @@ -34,3 +38,6 @@ CCTK_REAL Y_e type = GF Timelevels = 3 tags='ProlongationParameter="HydroBase::prolongation_type" tensortypealias="Scalar" interpolator="matter"' "electron fraction" CCTK_REAL Bvec[3] type = GF Timelevels = 3 tags='ProlongationParameter="HydroBase::prolongation_type" tensortypealias="U" interpolator="matter"' "Magnetic field components B^i" + +# This may become CCTK_BYTE when Carpet enables it by default +CCTK_INT hydro_excision_mask type = GF Timelevels = 1 tags='Prolongation="None" checkpoint="no"' "Mask for hydro excision" Index: param.ccl =================================================================== RCS file: /cactusdevcvs/CactusEinstein/HydroBase/param.ccl,v retrieving revision 1.7 diff -u -r1.7 param.ccl --- param.ccl 7 Dec 2009 17:12:39 -0000 1.7 +++ param.ccl 26 Jan 2010 21:07:15 -0000 @@ -53,3 +53,8 @@ { "none" :: "Evolution for Bvec is disabled" } "none" + +BOOLEAN hydro_excision "Turn on of off (default) storage for hydro excision" +{ +} "no" + Index: schedule.ccl =================================================================== RCS file: /cactusdevcvs/CactusEinstein/HydroBase/schedule.ccl,v retrieving revision 1.8 diff -u -r1.8 schedule.ccl --- schedule.ccl 7 Dec 2009 17:12:39 -0000 1.8 +++ schedule.ccl 26 Jan 2010 21:07:15 -0000 @@ -46,6 +46,11 @@ } } +if (hydro_excision) +{ + STORAGE: hydro_excision_mask +} + schedule group HydroBase_Initial at CCTK_INITIAL \ after (ADMBase_InitialData ADMBase_InitialGauge \ IOUtil_RecoverIDFromDatafiles) \ @@ -137,3 +142,16 @@ LANG: C } "Set magnetic field to 0" } + +if (hydro_excision) +{ + schedule GROUP HydroBase_ExcisionHasBeenSet at CCTK_PostStep + { + } "Group to schedule thorns changing the mask before and thorns using the mask after" + + schedule HydroBase_InitExcisionMask in HydroBase_Initial + { + LANG: C + } "Initialize hydro excision mask to 'no excision everywhere'" +} + Index: src/HydroBase.h =================================================================== RCS file: src/HydroBase.h diff -N src/HydroBase.h --- /dev/null 1 Jan 1970 00:00:00 -0000 +++ src/HydroBase.h 26 Jan 2010 21:07:15 -0000 @@ -0,0 +1,2 @@ +#define HYDROEXCISION_NORMAL 0 +#define HYDROEXCISION_EXCISED 1 Index: src/Initialisation.c =================================================================== RCS file: /cactusdevcvs/CactusEinstein/HydroBase/src/Initialisation.c,v retrieving revision 1.3 diff -u -r1.3 Initialisation.c --- src/Initialisation.c 7 Dec 2009 17:12:39 -0000 1.3 +++ src/Initialisation.c 26 Jan 2010 21:07:15 -0000 @@ -1,7 +1,7 @@ #include #include #include - +#include "HydroBase.h" void HydroBase_Zero (CCTK_ARGUMENTS) @@ -166,3 +166,16 @@ "Unsupported parameter value for InitBase::initial_data_setup_method"); } } + +void HydroBase_InitExcisionMask (CCTK_ARGUMENTS) +{ + DECLARE_CCTK_ARGUMENTS; + DECLARE_CCTK_PARAMETERS; + + int const np = cctk_lsh[0] * cctk_lsh[1] * cctk_lsh[2]; + + #pragma omp parallel for + for (int i=0; i References: <20091027200122.GE22099@numrel07.cct.lsu.edu> Message-ID: This patch makes Cactus add run-time linker directory paths for every linker directory path that it uses to build the executable. This means that, at run time, the same shared library should be used that was also found when linking the executable. It introduces as new configuration variable RUNDIR_PREFIX which defaults to "-Wl,-rpath,", and which can be changed e.g. to "-R" if appropriate. -erik -- Erik Schnetter http://www.cct.lsu.edu/~eschnett/ -------------- next part -------------- A non-text attachment was scrubbed... Name: libdir.diff Type: application/octet-stream Size: 1746 bytes Desc: not available Url : http://www.cactuscode.org/pipermail/patches/attachments/20100131/dc3bef91/attachment.obj