| Time | Revision | Author | Filename |
| 2009/07/23 05:29:59 | 1.2 | insomnia52 | examples/charm++/Molecular3D/src_namd/jsplugin.c |
| got what hopefully is a fixed version of jsplugin.c | |||
| 2009/07/22 05:35:08 | 1.174 | gzheng | src/scripts/configure.in |
| 2009/07/22 05:35:07 | 1.140 | gzheng | src/scripts/configure |
| test if can build shared library when mpi library is linked | |||
| 2009/07/21 17:12:17 | 1.54 | gzheng | src/libs/ck-libs/ampi/ampif.C |
| added fortran interface for AMPI_Setmigratable | |||
| 2009/07/21 14:26:15 | 2.45 | gioachin | src/conv-ccs/conv-ccs.c |
| fixed long standing (9 years!) bug in CCS when running without charmrun (like in MPI build) | |||
| 2009/07/21 09:18:01 | 1.33 | gzheng | src/conv-core/cputopology.C |
| make sure -1 is returned when all topo function APIs are not supported | |||
| 2009/07/21 09:02:27 | 1.3 | gzheng | src/arch/mpi-linux-x86_64/cc-mpicxx.sh |
| fixed a typo in previous checkin | |||
| 2009/07/21 08:21:52 | 1.32 | gzheng | src/conv-core/cputopology.C |
| a bug in numUniqNodes() that returns wrong num of physical nodes when called multiple times | |||
| 2009/07/21 05:49:11 | 1.118 | gzheng | src/arch/mpi/machine.c |
| remove the hack for autobuild on vmi | |||
| 2009/07/20 11:42:02 | 1.494 | idooley2 | src/scripts/Makefile |
| 2009/07/20 11:42:02 | 1.94 | idooley2 | src/scripts/Make.depends |
| 2009/07/20 11:42:01 | 1.5 | idooley2 | src/ck-cp/pathHistory.h |
| 2009/07/20 11:42:01 | 1.1 | idooley2 | src/ck-cp/pathHistory.ci |
| 2009/07/20 11:42:01 | 1.4 | idooley2 | src/ck-cp/pathHistory.C |
| 2009/07/20 11:42:01 | 1.7 | idooley2 | src/ck-cp/controlPoints.h |
| 2009/07/20 11:42:01 | 1.6 | idooley2 | src/ck-cp/controlPoints.ci |
| 2009/07/20 11:42:01 | 1.22 | idooley2 | src/ck-cp/controlPoints.C |
| 2009/07/20 11:42:00 | 2.171 | idooley2 | src/ck-core/init.C |
| Moving the "critical path" history code into its own module which is included in the LIBCK_CORE library. By default the critical path functionality is all ifdef'ed out. | |||
| 2009/07/20 08:32:13 | 1.3 | gzheng | src/arch/mpi-sol-x86_64/conv-mach.sh |
| recognize icpc | |||
| 2009/07/20 08:31:32 | 1.10 | gzheng | src/arch/mpi-linux-x86_64/conv-mach.sh |
| similar fix for intel compiler | |||
| 2009/07/20 08:30:34 | 1.2 | gzheng | src/arch/mpi-linux-x86_64/cc-mpicxx.sh |
| recognize intel compiler in the mpicxx wrapper and add 64 bit compiler option to intel compiler. | |||
| 2009/07/18 17:40:38 | 1.16 | phil | src/conv-com/petable.C |
| Fix some precedence and cast issues flagged by warnings | |||
| 2009/07/18 15:28:53 | 1.67 | gioachin | src/conv-core/memory.c |
| Made memory module using GNU hooks to stop complaining about casting pointers of different types: added corrected signatures and wrapper functions. | |||
| 2009/07/18 14:53:58 | 2.4 | gioachin | src/conv-core/memory-gnu-internal.h |
| Avoid compiler complains by undefining __THROW before redefining it. | |||
| 2009/07/18 12:42:11 | 1.33 | idooley2 | tools/projections/projections/gui/Timeline/WindowControls.java |
| 2009/07/18 12:42:11 | 1.17 | idooley2 | tools/projections/projections/gui/Timeline/TimelineWindow.java |
| 2009/07/18 12:42:11 | 1.6 | idooley2 | tools/projections/projections/gui/Timeline/MessageStructures.java |
| 2009/07/18 12:42:11 | 1.21 | idooley2 | tools/projections/projections/gui/Timeline/MainPanel.java |
| 2009/07/18 12:42:11 | 1.41 | idooley2 | tools/projections/projections/gui/Timeline/EntryMethodObject.java |
| 2009/07/18 12:42:10 | 1.68 | idooley2 | tools/projections/projections/gui/Timeline/Data.java |
| Adding new options to disable loading of idle regions and of message sends. This could make viewing of large numbers of objects faster. | |||
| 2009/07/18 12:41:46 | 1.361 | idooley2 | tools/projections/bin/projections.jar |
| 2009/07/18 12:41:46 | 1.14 | idooley2 | tools/projections/bin/projections |
| Adding new options to disable loading of idle regions and of message sends. This could make viewing of large numbers of objects faster. | |||
| 2009/07/18 07:43:10 | 1.8 | phil | doc/charm++/checkpoint.tex |
| Correct typo in docs marco -> macro | |||
| 2009/07/16 19:38:37 | 1.24 | gioachin | src/util/ckhashtable.h |
| 2009/07/16 19:38:37 | 1.16 | gioachin | src/util/ckhashtable.C |
|
Added warning to remember to delete the iterator after finished using it.
Added C binding for deleting the iterator (function called CkHashtableDestroyIterator) Changed return type of CkHashtable::remove to int | |||
| 2009/07/16 07:16:29 | 1.48 | idooley2 | src/ck-com/ComlibManager.C |
| Disabling a periodic routine that was intended for debugging. | |||
| 2009/07/16 07:07:35 | 1.1 | bhatele | examples/charm++/jacobi1d/jacobi1d.ci |
| 2009/07/16 07:07:35 | 1.1 | bhatele | examples/charm++/jacobi1d/jacobi1d.C |
| 2009/07/16 07:07:35 | 1.1 | bhatele | examples/charm++/jacobi1d/Makefile |
| 1d decomposition of 2d jacobi | |||
| 2009/07/16 06:18:49 | 1.18 | idooley2 | tools/projections/projections/misc/LogEntryData.java |
| 2009/07/16 06:18:48 | 1.32 | idooley2 | tools/projections/projections/gui/Timeline/WindowControls.java |
| 2009/07/16 06:18:48 | 1.1 | idooley2 | tools/projections/projections/gui/Timeline/SolidColorJPanel.java |
| 2009/07/16 06:18:48 | 1.5 | idooley2 | tools/projections/projections/gui/Timeline/SaveImage.java |
| 2009/07/16 06:18:48 | 1.1 | idooley2 | tools/projections/projections/gui/Timeline/Render2by2PanelGrid.java |
| Added better screenshots to Timeline(all PE's timelines are rendered, not just the visible ones). | |||
| 2009/07/16 06:18:25 | 1.360 | idooley2 | tools/projections/bin/projections.jar |
| Added better screenshots to Timeline(all PE's timelines are rendered, not just the visible ones). | |||
| 2009/07/15 17:26:56 | 1.207 | gzheng | src/conv-core/threads.c |
| in stackcopy threads, fix a bug in retrieving the actual pointers on the stack. When the thread is currently running, it should use the current stack (stackbase), while when thread is sleeping, use the saved stack. | |||
| 2009/07/15 08:43:14 | 1.2 | idooley2 | tests/charm++/commtest/eachtomany/runtillcrash.pl |
| 2009/07/15 08:43:14 | 1.2 | idooley2 | tests/charm++/commtest/eachtomany/eachtomany.C |
| Enabling a fourth test within the program. | |||
| 2009/07/15 08:27:13 | 1.1 | idooley2 | tests/charm++/commtest/eachtomany/runtillcrash.pl |
| 2009/07/15 08:27:13 | 1.1 | idooley2 | tests/charm++/commtest/eachtomany/eachtomany.ci |
| 2009/07/15 08:27:13 | 1.1 | idooley2 | tests/charm++/commtest/eachtomany/eachtomany.C |
| 2009/07/15 08:27:13 | 1.1 | idooley2 | tests/charm++/commtest/eachtomany/Makefile |
| Adding a test program that uses the EachToManyMulticast strategy. | |||
| 2009/07/15 08:25:56 | 1.13 | idooley2 | src/conv-com/convcomlibstrategy.h |
| 2009/07/15 08:25:56 | 1.12 | idooley2 | src/conv-com/convcomlibstrategy.C |
| 2009/07/15 08:25:56 | 1.13 | idooley2 | src/conv-com/convcomlibmanager.h |
| 2009/07/15 08:25:56 | 1.27 | idooley2 | src/conv-com/convcomlibmanager.C |
| 2009/07/15 08:25:55 | 1.20 | idooley2 | src/ck-com/EachToManyMulticastStrategy.C |
| 2009/07/15 08:25:55 | 1.13 | idooley2 | src/ck-com/ComlibStrategy.h |
| 2009/07/15 08:25:55 | 1.20 | idooley2 | src/ck-com/ComlibStrategy.C |
| 2009/07/15 08:25:55 | 1.23 | idooley2 | src/ck-com/ComlibManager.h |
| 2009/07/15 08:25:55 | 1.14 | idooley2 | src/ck-com/ComlibManager.ci |
| 2009/07/15 08:25:55 | 1.47 | idooley2 | src/ck-com/ComlibManager.C |
| Fixing more bugs in the bracketed strategies including: 1) destination and source element lists are now handled correctly 2) error modes are handled and checked in a more correct manner in the protocol functions instead of beginIteration. | |||
| 2009/07/15 03:33:16 | 1.175 | gzheng | src/ck-ldb/CentralLB.C |
| compare avail_vector to NULL (xlc may not like it compares with 0) | |||
| 2009/07/14 17:38:41 | 1.67 | phil | src/ck-ldb/lbdb.h |
| Remove artifacts of a header being C-includable | |||
| 2009/07/13 17:13:27 | 2.85 | phil | src/ck-core/cklocation.h |
| 2009/07/13 17:13:26 | 2.128 | phil | src/ck-core/cklocation.C |
| 2009/07/13 17:13:26 | 1.141 | phil | src/ck-core/ckarray.h |
|
Core: Interface to ask Chare Array proxies how many elements are local
Location manager now has a simple means to count the number of records it has corresponding to locally-held indices. Array proxies expose this functionality through their interface. // How many elements of each associated array are local to this PE? // If this returns n, and there are k associated arrays, that // means k*n elements are living here unsigned int CkLocMgr::numLocalElements(); | |||
| 2009/07/13 16:55:57 | 2.84 | gzheng | src/ck-core/cklocation.h |
| another inconsistence of bool and CmiBool for isAsyncEvacuate() fixed | |||
| 2009/07/13 10:05:50 | 2.83 | gzheng | src/ck-core/cklocation.h |
| change asyncEvacuate from bool to CmiBool | |||
| 2009/07/13 10:03:47 | 1.56 | gzheng | src/scripts/conv-autoconfig.h.in |
| 2009/07/13 10:03:47 | 1.173 | gzheng | src/scripts/configure.in |
| 2009/07/13 10:03:47 | 1.139 | gzheng | src/scripts/configure |
|
more detailed info output into charmconfig.out.
a better test for compiling shared lib | |||
| 2009/07/13 09:52:04 | 1.6 | gzheng | src/libs/ck-libs/armci/Makefile |
| pass OPTS to charmc | |||
| 2009/07/13 05:14:22 | 1.325 | gzheng | src/libs/ck-libs/ampi/ampi.C |
|
fixed couple bug in MPI_Ssend.
1. prevent other mpi_irecv awakes the sleeping ampi thread before ssend finishes. 2. prevent a reference to an invalid msg pointer after generic() calling inorder(), which may free the message. | |||
| 2009/07/11 16:58:30 | 1.9 | phil | doc/charm++/nodegroups.tex |
| 2009/07/11 16:58:30 | 1.15 | phil | doc/charm++/groups.tex |
| Update manual to reflect that group classes should inherit from CBase_GroupName | |||
| 2009/07/11 13:08:11 | 1.14 | idooley2 | src/conv-com/routerstrategy.C |
| 2009/07/11 13:08:11 | 1.11 | idooley2 | src/conv-com/convcomlibstrategy.C |
| 2009/07/11 13:08:11 | 1.12 | idooley2 | src/conv-com/convcomlibmanager.h |
| 2009/07/11 13:08:10 | 1.7 | idooley2 | src/ck-com/EachToManyMulticastStrategy.h |
| 2009/07/11 13:08:10 | 1.19 | idooley2 | src/ck-com/EachToManyMulticastStrategy.C |
| 2009/07/11 13:08:10 | 1.12 | idooley2 | src/ck-com/ComlibStrategy.h |
| 2009/07/11 13:08:10 | 1.19 | idooley2 | src/ck-com/ComlibStrategy.C |
| 2009/07/11 13:08:10 | 1.13 | idooley2 | src/ck-com/ComlibManager.ci |
| 2009/07/11 13:08:09 | 1.46 | idooley2 | src/ck-com/ComlibManager.C |
| Fixing some comlib bugs that could possibly occur due to the way the src and dest elements were handled. | |||
| 2009/07/10 10:59:40 | 1.31 | bhatele | src/conv-core/cputopology.C |
| the bynodes array has to be populated always | |||
| 2009/07/10 10:52:48 | 1.30 | bhatele | src/conv-core/cputopology.C |
| a different way of getting cpu topology on Cray machines | |||
| 2009/07/10 08:49:53 | 1.10 | bhatele | src/util/CrayNid.c |
| cpu topology info should now be available always | |||
| 2009/07/10 07:50:48 | 1.17 | insomnia52 | examples/charm++/Molecular3D/src_charm/Patch.C |
| changed load balancing to start from step 0 instead of step 1 | |||
| 2009/07/10 07:21:43 | 2.77 | gzheng | src/ck-core/ckreduction.C |
| removed a debug print | |||
| 2009/07/10 07:00:08 | 1.55 | gzheng | src/scripts/conv-autoconfig.h.in |
| 2009/07/10 07:00:08 | 1.172 | gzheng | src/scripts/configure.in |
| 2009/07/10 07:00:08 | 1.138 | gzheng | src/scripts/configure |
| check if complier allows declaration of varsize array | |||
| 2009/07/10 06:54:20 | 2.48 | gzheng | src/ck-core/ckreduction.h |
| 2009/07/10 06:54:20 | 2.76 | gzheng | src/ck-core/ckreduction.C |
| 2009/07/10 06:54:20 | 2.26 | gzheng | src/ck-core/ckarrayreductionmgr.C |
| better tracing event deps in array/group and nodegroup reduction for bigsim | |||
| 2009/07/10 06:48:23 | 1.1 | idooley2 | tests/charm++/commtest/eachtomany-simple/eachtomany.ci |
| 2009/07/10 06:48:22 | 1.1 | idooley2 | tests/charm++/commtest/eachtomany-simple/eachtomany.C |
| 2009/07/10 06:48:22 | 1.1 | idooley2 | tests/charm++/commtest/eachtomany-simple/Makefile |
| Adding a very simple test of EachToManyMulticastStrategy | |||
| 2009/07/10 04:39:41 | 1.59 | gzheng | src/langs/bluegene/bigsim_logs.C |
| in timeline pup subrotuine, now creates all timelog at first during unpack, this is to make backward/forward dep works and avoid CkQ re-allocate buffer. | |||
| 2009/07/09 12:13:52 | 1.324 | emenese2 | src/libs/ck-libs/ampi/ampi.C |
| Fixed a bug with message logging protocol. Variable _currentObj wasn't updated before AMPI_Send. | |||
| 2009/07/09 10:24:23 | 1.2 | bhatele | src/arch/mpi-crayxt/conv-mach.h |
| topology collection is now turned off by default on Cray machines | |||
| 2009/07/09 10:23:24 | 1.3 | bhatele | src/util/XT4Torus.h |
| 2009/07/09 10:23:23 | 1.29 | bhatele | src/util/TopoManager.h |
| 2009/07/09 10:23:23 | 1.18 | bhatele | src/util/TopoManager.C |
| 2009/07/09 10:23:23 | 1.9 | bhatele | src/util/CrayNid.c |
|
topology code should now work on Kraken (XT5)
I am still going to turn it off by default | |||
| 2009/07/09 08:14:28 | 1.59 | phil | src/util/cklists.h |
| Add method CkVec::clear() to improve STL compatibility | |||
| 2009/07/09 07:24:05 | 1.29 | idooley2 | src/conv-core/cputopology.C |
| Switching from CkAssert to CmiAssert, and adding brackets around the for loop(there was a bug here caused by the last checkin). | |||
| 2009/07/09 06:22:56 | 1.28 | idooley2 | src/conv-core/cputopology.C |
| Adding an assertion to do bounds checking. This assertion fails on runs on kraken. | |||
| 2009/07/09 05:10:27 | 1.10 | insomnia52 | examples/charm++/Molecular3D/src_charm/mol3d.ci |
| 2009/07/09 05:10:27 | 1.10 | insomnia52 | examples/charm++/Molecular3D/src_charm/Patch.h |
| 2009/07/09 05:10:27 | 1.16 | insomnia52 | examples/charm++/Molecular3D/src_charm/Patch.C |
| 2009/07/09 05:10:27 | 1.3 | insomnia52 | examples/charm++/Molecular3D/src_charm/Main.h |
| 2009/07/09 05:10:27 | 1.13 | insomnia52 | examples/charm++/Molecular3D/src_charm/Main.C |
| 2009/07/09 05:10:27 | 1.10 | insomnia52 | examples/charm++/Molecular3D/src_charm/Compute.C |
| made load balancing and its instrumentation blocking, as well as a few other minor changes | |||
| 2009/07/08 11:22:19 | 1.29 | phil | src/libs/ck-libs/tcharm/tcharm_impl.h |
| 2009/07/08 11:22:19 | 1.17 | phil | src/libs/ck-libs/tcharm/tcharm.h |
|
TCharm/Emulator: Improve tracing to allow short-cutting library calls
TCharm provides a tracing facility to let developers see when library code is entered and exited. This commit extends that facility so that the BigSim emulator logs those entries and exits as events, and makes the exit event depend on the entrance. Specifically, this lets BigNetSim detect the call and return of AMPI functions, and enables it to model them differently than the sequence of messages they generate. | |||
| 2009/07/08 11:07:28 | 1.56 | phil | src/langs/bluegene/bigsim_logs.h |
| 2009/07/08 11:07:28 | 1.21 | phil | src/langs/bluegene/bigsim_api.C |
| Bigsim Emulator: Constify another char* to quiet warnings | |||
| 2009/07/08 10:08:48 | 1.7 | idooley2 | src/ck-com/OneTimeMulticastStrategy.h |
| 2009/07/08 10:08:48 | 1.11 | idooley2 | src/ck-com/OneTimeMulticastStrategy.C |
| 2009/07/08 10:08:48 | 1.18 | idooley2 | src/ck-com/ComlibStrategy.C |
| 2009/07/08 10:08:48 | 1.5 | idooley2 | src/ck-com/ComlibStats.h |
| 2009/07/08 10:08:48 | 1.16 | idooley2 | src/ck-com/ComlibSectionInfo.C |
| 2009/07/08 10:08:48 | 1.22 | idooley2 | src/ck-com/ComlibManager.h |
| Fixing projections message tracing for [nokeep] multicast messages. Projections message tracing for non [nokeep] messages is still broken. | |||
| 2009/07/08 09:09:37 | 1.52 | mokos | src/libs/ck-libs/pose/pose.h |
| 2009/07/08 09:09:37 | 1.42 | mokos | src/libs/ck-libs/pose/pose.C |
|
-Added _POSE_external_invoke (created by Eric) as a #define function to allow invocation of poser event entry methods by non-sim objects--only use this if absolutely necessary
-Added global _POSE_SEQUENTIAL for use by _POSE_external_invoke to specify whether or not sequential POSE is in use | |||
| 2009/07/08 06:44:57 | 2.85 | gzheng | src/ck-perf/trace.h |
| prompt the previous fix. Make sure no trace call is called if there is no tracing modules linked. | |||
| 2009/07/08 06:05:59 | 1.3 | gzheng | examples/charm++/X10/X10_lib.C |
| 2009/07/08 06:05:59 | 1.4 | gzheng | examples/charm++/X10/Makefile |
| minor changes in CkVec to fix the memory bug | |||
| 2009/07/07 18:00:55 | 1.323 | gzheng | src/libs/ck-libs/ampi/ampi.C |
| set Ireq's src rank from actual received message. | |||
| 2009/07/07 17:22:44 | 1.35 | gzheng | src/libs/ck-libs/ampi/ampif.h |
| increase MPI_STATUS_SIZE because status structure has a pointer which may be 64 bit | |||
| 2009/07/07 11:56:15 | 2.82 | insomnia52 | src/ck-core/cklocation.h |
| 2009/07/07 11:56:15 | 2.127 | insomnia52 | src/ck-core/cklocation.C |
| Fixed problem with load balancer instrumentation in presence of local entry methods that call the same object recursively. | |||
| 2009/07/07 10:01:17 | 1.5 | wesolwsk | src/util/SSE-Float.h |
| 2009/07/07 10:01:17 | 1.6 | wesolwsk | src/util/SSE-Double.h |
| Removed the if statements from the constructors to improve performance | |||
| 2009/07/07 09:00:27 | 1.20 | gioachin | doc/projections/manual.tex |
| fix problem with compilation | |||
| 2009/07/07 07:02:05 | 1.31 | idooley2 | tools/projections/projections/gui/Timeline/UserEventObject.java |
| Fixing a bug which produces an extra newline character at the end of user event names. | |||
| 2009/07/07 07:01:40 | 1.359 | idooley2 | tools/projections/bin/projections.jar |
| Fixing a bug which produces an extra newline character at the end of user event names. | |||
| 2009/07/07 05:50:38 | 1.186 | gzheng | src/ck-core/ckarray.C |
| a minor change | |||
| 2009/07/07 05:29:33 | 1.15 | insomnia52 | examples/charm++/Molecular3D/src_charm/Patch.C |
| 2009/07/07 05:29:33 | 1.9 | insomnia52 | examples/charm++/Molecular3D/src_charm/Compute.C |
| minor changes | |||
| 2009/07/06 19:28:17 | 1.34 | gzheng | src/libs/ck-libs/ampi/ampif.h |
|
fixes to allow FLASH3 to compile with AMPI:
define two constants that were missing fixed MPI_PROC_NULL definition to be consistent with C header file. | |||
| 2009/07/06 06:45:44 | 1.156 | emenese2 | src/arch/net/charmrun/charmrun.c |
| Minor fix for message logging | |||
| 2009/07/06 06:44:31 | 1.6 | abecker | src/libs/ck-libs/ParFUM/partition.C |
| 2009/07/06 06:44:31 | 1.57 | abecker | src/libs/ck-libs/ParFUM/ParFUM_internals.h |
| 2009/07/06 06:44:31 | 1.19 | abecker | src/libs/ck-libs/ParFUM/ParFUM.C |
|
Process Metis face graph flag in FEM_Init
The command line argument specifying the use of face neighbor graphs instead of node neighbor graphs was processed at partition time. This can confuse programs when they're parsing arguments in init(). Instead, process our flag in FEM_Init so that the user code never sees it. | |||
| 2009/07/06 04:48:53 | 1.6 | insomnia52 | examples/charm++/Molecular3D/src_physics/nonbonded.h |
| 2009/07/06 04:48:53 | 1.14 | insomnia52 | examples/charm++/Molecular3D/src_charm/Patch.C |
| 2009/07/06 04:48:53 | 1.8 | insomnia52 | examples/charm++/Molecular3D/src_charm/Compute.C |
| made multicast manager reset after migration | |||
| 2009/07/03 16:33:36 | 2.126 | gzheng | src/ck-core/cklocation.C |
| double start/stop local method, which makes instrucmenaiton of local entry method totally wrong. | |||
| 2009/07/03 07:35:17 | 1.9 | insomnia52 | examples/charm++/Molecular3D/src_charm/mol3d.ci |
| 2009/07/03 07:35:16 | 1.9 | insomnia52 | examples/charm++/Molecular3D/src_charm/Patch.h |
| 2009/07/03 07:35:16 | 1.13 | insomnia52 | examples/charm++/Molecular3D/src_charm/Patch.C |
| 2009/07/03 07:35:16 | 1.12 | insomnia52 | examples/charm++/Molecular3D/src_charm/Main.C |
| 2009/07/03 07:35:16 | 1.4 | insomnia52 | examples/charm++/Molecular3D/src_charm/Compute.h |
| 2009/07/03 07:35:16 | 1.7 | insomnia52 | examples/charm++/Molecular3D/src_charm/Compute.C |
| 2009/07/03 07:35:16 | 1.7 | insomnia52 | examples/charm++/Molecular3D/compile/Makefile.config |
| 2009/07/03 07:35:16 | 1.4 | insomnia52 | examples/charm++/Molecular3D/compile/Makefile |
| 2009/07/03 07:35:16 | 1.3 | insomnia52 | examples/charm++/Molecular3D/README.tex |
| corrected some load balancing issues | |||
| 2009/07/02 11:54:13 | 1.10 | idooley2 | src/ck-com/OneTimeMulticastStrategy.C |
| Fixing bug in new multicast strategy. | |||
| 2009/07/02 11:08:25 | 1.6 | idooley2 | src/ck-com/OneTimeMulticastStrategy.h |
| 2009/07/02 11:08:25 | 1.9 | idooley2 | src/ck-com/OneTimeMulticastStrategy.C |
| 2009/07/02 11:08:25 | 1.12 | idooley2 | src/ck-com/ComlibManager.ci |
| Adding a new node aware multicast strategy that sends to PEs within a node along a ring. | |||
| 2009/07/02 09:45:16 | 1.11 | gzheng | src/arch/common/conv-mach-ifort.sh |
| use -fpic instead of -fPIC to be consistent with icc options | |||
| 2009/07/02 09:44:32 | 1.33 | gzheng | src/libs/ck-libs/ampi/ampif.h |
| declare MPI_Register as fortran subroutine instead of function to make it consistent. | |||
| 2009/07/02 08:18:16 | 1.8 | idooley2 | src/ck-com/OneTimeMulticastStrategy.C |
| Adding better debug printouts. | |||
| 2009/07/02 08:16:44 | 1.2 | idooley2 | tests/charm++/commtest/section-multicast-test-6/test.ci |
| 2009/07/02 08:16:44 | 1.4 | idooley2 | tests/charm++/commtest/section-multicast-test-6/test.C |
| 2009/07/02 08:16:44 | 1.3 | idooley2 | tests/charm++/commtest/section-multicast-test-6/Makefile |
| Using QD for startup to ensure that comlib has been initialized before we use it. | |||
| 2009/07/02 06:55:32 | 1.7 | idooley2 | src/ck-com/OneTimeMulticastStrategy.C |
| Minor changes to Node aware multicast strategy. | |||
| 2009/07/01 18:40:04 | 1.226 | gzheng | src/scripts/charmc |
| only skip libso when CMK_NO_BUILD_SHARED is true. | |||
| 2009/07/01 11:46:03 | 1.5 | idooley2 | src/ck-com/OneTimeMulticastStrategy.h |
| 2009/07/01 11:46:03 | 1.6 | idooley2 | src/ck-com/OneTimeMulticastStrategy.C |
| Fixing some bugs in the new node aware multicast strategy. | |||
| 2009/07/01 08:53:41 | 1.143 | gzheng | src/langs/bluegene/blue.C |
| reenforce the check to make sure number of target nodes are greater than no of physical nodes | |||
| 2009/07/01 06:49:42 | 1.4 | idooley2 | src/ck-com/OneTimeMulticastStrategy.h |
| 2009/07/01 06:49:42 | 1.5 | idooley2 | src/ck-com/OneTimeMulticastStrategy.C |
| 2009/07/01 06:49:41 | 1.11 | idooley2 | src/ck-com/ComlibManager.ci |
| Adding a new node aware one time multicast strategy. | |||
| 2009/07/01 06:45:35 | 1.27 | idooley2 | src/conv-core/cputopology.C |
| 2009/07/01 06:45:35 | 2.346 | idooley2 | src/conv-core/converse.h |
| Adding another cpu topology accessor method that returns the first PE for a specified node. | |||
| 2009/06/30 14:02:14 | 1.152 | phil | src/libs/ck-libs/ampi/ampiimpl.h |
|
AMPI: Quiet some warnings about silly return types
Returning 'const int' is the same as returning 'int', and ICC likes to tell us so. | |||
| 2009/06/30 10:00:09 | 1.322 | gzheng | src/libs/ck-libs/ampi/ampi.C |
| in checkpointing functions, call AMPIAPI after AMPI_Barrier(). which will also call AMPIAPI to swap global data and causes confusions. | |||
| 2009/06/30 09:58:04 | 1.28 | gzheng | src/ck-core/ckmemcheckpoint.C |
| minor change in debug prints | |||
| 2009/06/30 09:56:29 | 2.26 | gzheng | src/conv-core/global-elfgot.C |
|
in case of using isomalloc, global data segment is malloc'ed from isomalloc heap.
Call CmiIsomallocFree in destructor instead of free(). This seems to fix a crash in in-memory checkpoint | |||
| 2009/06/30 06:11:40 | 1.21 | phil | src/ck-com/ComlibManager.h |
| 2009/06/30 06:11:40 | 1.45 | phil | src/ck-com/ComlibManager.C |
| 2009/06/30 06:11:39 | 1.12 | phil | src/conv-com/convcomlibstrategy.h |
| 2009/06/30 06:11:39 | 1.10 | phil | src/conv-com/convcomlibstrategy.C |
|
Clean up some compiler warnings
Const-ify some more string literals, to quiet warnings. | |||
| 2009/06/29 12:39:54 | 1.74 | bhatele | src/libs/ck-libs/tcharm/tcharm.C |
| the warnings need to be printed from pe 0 only | |||
| 2009/06/29 12:39:12 | 2.132 | bhatele | src/ck-perf/trace-projections.C |
| the warning needs to printed from pe 0 only | |||
| 2009/06/29 09:53:01 | 1.11 | phil | src/conv-com/convcomlibmanager.h |
|
Fix comlib-related build failure under bgxlC
The grand comlib checkin exposed an ICE in the BlueGene version of xlC. I narrowed the cause in convcomlibstrategy.C to the line in Strategy::Strategy() that called insertStrategy() to register with the manager. This commit modifies the manager to 'friend' the entire Strategy class, rather than just the constructor. Checked compilation with GCC and xlC on BlueGene, and GCC and ICC on 64-bit lab machines. Things may yet break under the abominable Solaris cc. | |||
| 2009/06/26 11:28:51 | 1.18 | bhatele | src/langs/bluegene/bigsim_network.h |
| fix the BG/P latency formula | |||
| 2009/06/26 07:00:40 | 1.5 | gzheng | src/arch/common/conv-mach-gfortran.sh |
| silent errors when gfortran lib path can not be automatically located | |||
| 2009/06/25 12:01:40 | 1.67 | idooley2 | tools/projections/projections/gui/Timeline/Data.java |
| Fixing bug where bottom portion of timeline is obscured by the scrollbar. | |||
| 2009/06/25 12:01:17 | 1.358 | idooley2 | tools/projections/bin/projections.jar |
| Fixing bug where bottom portion of timeline is obscured by the scrollbar. | |||
| 2009/06/25 09:10:58 | 1.8 | insomnia52 | examples/charm++/Molecular3D/src_charm/Patch.h |
| 2009/06/25 09:10:58 | 1.12 | insomnia52 | examples/charm++/Molecular3D/src_charm/Patch.C |
| 2009/06/25 09:10:58 | 1.6 | insomnia52 | examples/charm++/Molecular3D/src_charm/Compute.C |
| added instrumentation control | |||
| 2009/06/25 09:00:00 | 1.11 | idooley2 | src/ck-com/ComlibStrategy.h |
| 2009/06/25 09:00:00 | 1.17 | idooley2 | src/ck-com/ComlibStrategy.C |
| Fixing return type incompatibilities. | |||
| 2009/06/25 08:50:55 | 1.16 | idooley2 | src/ck-com/ComlibStrategy.C |
| Fixing delivery bug in CharmStrategy::deliverToIndices. Adding some documentation notes. | |||
| 2009/06/25 07:55:48 | 1.4 | idooley2 | src/ck-com/OneTimeMulticastStrategy.C |
| Adding some user events to trace the multicasts. | |||
| 2009/06/25 07:09:18 | 1.5 | insomnia52 | examples/charm++/Molecular3D/src_physics/nonbonded.h |
| fixed memory leak | |||
| 2009/06/25 02:36:16 | 1.321 | gzheng | src/libs/ck-libs/ampi/ampi.C |
| fixed the previous change for non-bigsim | |||
| 2009/06/24 20:32:39 | 1.320 | gzheng | src/libs/ck-libs/ampi/ampi.C |
|
make sure recv event depends on the right "ampi::generic" event, and does not depend on directly on the event that wakes up the thread
CVS ---------------------------------------------------------------------- | |||
| 2009/06/24 12:25:22 | 2.84 | ramv | src/ck-perf/trace.h |
| Fixed a typo in prev commit introduced after testing :) | |||
| 2009/06/24 12:11:13 | 2.83 | ramv | src/ck-perf/trace.h |
| Exposed a per-entry-method memleak when testing with valgrind. Committing Chee Wai's fix that should restore projections' dont use, dont pay guarantee | |||
| 2009/06/24 11:34:04 | 1.4 | insomnia52 | examples/charm++/Molecular3D/src_physics/defs.h |
| 2009/06/24 11:34:04 | 1.8 | insomnia52 | examples/charm++/Molecular3D/src_charm/mol3d.ci |
| 2009/06/24 11:34:04 | 1.11 | insomnia52 | examples/charm++/Molecular3D/src_charm/Patch.C |
| 2009/06/24 11:34:04 | 1.11 | insomnia52 | examples/charm++/Molecular3D/src_charm/Main.C |
| changed load balancing intrevals | |||
| 2009/06/24 10:38:25 | 1.3 | idooley2 | src/ck-com/OneTimeMulticastStrategy.h |
| 2009/06/24 10:38:25 | 1.3 | idooley2 | src/ck-com/OneTimeMulticastStrategy.C |
| Simplify set of parameters to determineNextHopPEs(). Add more documentation. | |||
| 2009/06/24 09:55:46 | 1.2 | idooley2 | src/ck-com/OneTimeMulticastStrategy.h |
| 2009/06/24 09:55:46 | 1.2 | idooley2 | src/ck-com/OneTimeMulticastStrategy.C |
| 2009/06/24 09:55:46 | 1.10 | idooley2 | src/ck-com/ComlibManager.ci |
| Adding some new multicast strategies. | |||
| 2009/06/24 09:40:29 | 2.344 | idooley2 | src/conv-core/convcore.c |
| Fixing bug where messages are sent if list of negative length is specified. | |||
| 2009/06/24 06:57:16 | 1.73 | emenese2 | src/libs/ck-libs/tcharm/tcharm.C |
| If fault tolerance flag is set, the assignment of array elements is round robin. | |||
| 2009/06/23 18:04:45 | 2.131 | chaomei | src/ck-perf/trace-projections.C |
| 2009/06/23 18:04:45 | 2.83 | chaomei | src/ck-perf/trace-common.C |
| comment out some CkPrintfs which are believed to be used for debugging purpose. | |||
| 2009/06/23 11:20:16 | 1.7 | insomnia52 | examples/charm++/Molecular3D/src_charm/mol3d.ci |
| 2009/06/23 11:20:16 | 1.10 | insomnia52 | examples/charm++/Molecular3D/src_charm/Patch.C |
| 2009/06/23 11:20:16 | 1.10 | insomnia52 | examples/charm++/Molecular3D/src_charm/Main.C |
| 2away stuff | |||
| 2009/06/23 10:48:33 | 2.125 | emenese2 | src/ck-core/cklocation.C |
| Set to round robin the default assignment of array elements to processors when __FAULT__ flag is turned on. | |||
| 2009/06/23 10:47:31 | 1.174 | emenese2 | src/ck-ldb/CentralLB.C |
| Included an extra parameter in the pup function. | |||
| 2009/06/23 06:13:16 | 1.19 | idooley2 | doc/projections/manual.tex |
| Adding documentation for new user event types. | |||
| 2009/06/23 05:39:54 | 2.23 | idooley2 | src/conv-core/conv-trace.h |
| 2009/06/23 05:39:53 | 2.82 | idooley2 | src/ck-perf/trace.h |
| 2009/06/23 05:39:53 | 2.130 | idooley2 | src/ck-perf/trace-projections.C |
| 2009/06/23 05:39:53 | 2.76 | idooley2 | src/ck-perf/trace-projections.h |
| 2009/06/23 05:39:53 | 2.24 | idooley2 | src/ck-perf/trace-common.h |
| 2009/06/23 05:39:53 | 2.82 | idooley2 | src/ck-perf/trace-common.C |
| Adding a new type of user event tracing for projections. The events can be bracketed, contain a user supplied string, and are associated with a user event id (so they are colored like standard user events). | |||
| 2009/06/22 17:47:41 | 1.17 | idooley2 | tools/projections/projections/misc/LogEntryData.java |
| 2009/06/22 17:47:40 | 1.13 | idooley2 | tools/projections/projections/analysis/ProjDefs.java |
| 2009/06/22 17:47:40 | 1.66 | idooley2 | tools/projections/projections/analysis/LogLoader.java |
| 2009/06/22 17:47:40 | 1.23 | idooley2 | tools/projections/projections/analysis/LogEntry.java |
| 2009/06/22 17:47:40 | 1.33 | idooley2 | tools/projections/projections/analysis/GenericLogReader.java |
| 2009/06/22 17:47:40 | 1.30 | idooley2 | tools/projections/projections/gui/Timeline/UserEventObject.java |
| 2009/06/22 17:47:40 | 1.3 | idooley2 | tools/projections/projections/gui/Timeline/ChooseUserEventsWindow.java |
| Adding a new type of bracketed user event that contains a user supplied string. | |||
| 2009/06/22 17:47:15 | 1.357 | idooley2 | tools/projections/bin/projections.jar |
| Adding a new type of bracketed user event that contains a user supplied string. | |||
| 2009/06/22 09:50:50 | 1.8 | idooley2 | tools/projections/projections/streaming/MultiSeriesHandler.java |
| Adjusting window positions for streaming tool. | |||
| 2009/06/22 09:50:26 | 1.356 | idooley2 | tools/projections/bin/projections.jar |
| Adjusting window positions for streaming tool. | |||
| 2009/06/22 07:44:59 | 1.493 | idooley2 | src/scripts/Makefile |
| 2009/06/22 07:44:58 | 1.1 | idooley2 | src/ck-com/OneTimeMulticastStrategy.h |
| 2009/06/22 07:44:58 | 1.3 | idooley2 | src/ck-com/comlib.h |
| 2009/06/22 07:44:58 | 1.7 | idooley2 | src/ck-com/RingMulticastStrategy.h |
| 2009/06/22 07:44:58 | 1.3 | idooley2 | src/ck-com/RectMulticastStrategy.h |
| 2009/06/22 07:44:58 | 1.3 | idooley2 | src/ck-com/RectMulticastStrategy.C |
| 2009/06/22 07:44:58 | 1.93 | idooley2 | src/scripts/Make.depends |
| 2009/06/22 07:44:57 | 1.1 | idooley2 | src/ck-com/OneTimeMulticastStrategy.C |
| 2009/06/22 07:44:57 | 1.10 | idooley2 | src/ck-com/ComlibStrategy.h |
| 2009/06/22 07:44:57 | 1.15 | idooley2 | src/ck-com/ComlibStrategy.C |
| 2009/06/22 07:44:57 | 1.15 | idooley2 | src/ck-com/ComlibSectionInfo.C |
| 2009/06/22 07:44:57 | 1.9 | idooley2 | src/ck-com/ComlibManager.ci |
| 2009/06/22 07:44:57 | 1.2 | idooley2 | src/ck-com/MulticastStrategy.h |
| 2009/06/22 07:44:57 | 1.2 | idooley2 | src/ck-com/MulticastStrategy.C |
| 2009/06/22 07:44:57 | 1.6 | idooley2 | src/ck-com/MultiRingMulticast.h |
| 2009/06/22 07:44:57 | 1.10 | idooley2 | src/ck-com/DirectMulticastStrategy.h |
| Adding a new simple Comlib Multicast strategy that does not cache array section information on destination processors. | |||
| 2009/06/22 07:22:35 | 1.75 | gzheng | src/arch/net-aix-ppc/conv-mach.h |
| set to use netpoll | |||
| 2009/06/19 09:51:25 | 1.29 | gioachin | java/bin/charmdebug.jar |
| *** empty log message *** | |||
| 2009/06/19 09:42:31 | 1.37 | gioachin | java/charm/debug/ParDebug.java |
|
Fixed problems with working directory not being detected correctly.
Fixed problem with memory leak search on client running on 32 bits. | |||
| 2009/06/19 09:41:35 | 1.10 | gioachin | java/charm/debug/inspect/InspectPanel.java |
| 2009/06/19 09:41:35 | 1.5 | gioachin | java/charm/debug/pdata/MsgTypeInfo.java |
| 2009/06/19 09:41:35 | 1.4 | gioachin | java/charm/debug/pdata/MsgPList.java |
| 2009/06/19 09:41:35 | 1.5 | gioachin | java/charm/debug/pdata/MsgInfo.java |
| Improved visualization of messages in debugger | |||
| 2009/06/19 09:40:58 | 1.9 | gioachin | java/charm/debug/CpdUtil.java |
| Do not close the debugger if an error is detected in the application, simply terminate the remote application | |||
| 2009/06/19 09:39:56 | 1.3 | gioachin | java/charm/debug/MemoryDialog.java |
| Added preferences for memory View | |||
| 2009/06/19 09:38:49 | 1.5 | gioachin | java/charm/debug/preference/Preference.java |
| Added preferences for memory View | |||
| 2009/06/19 09:38:33 | 1.6 | gioachin | java/charm/util/ReflectiveXML.java |
| Modified default object creation while reading from file: the fields are NOT reinitialized to zero after the default constructor has set them. | |||
| 2009/06/19 07:38:19 | 1.1 | gzheng | examples/bigsim/tools/changelog/changelog.C |
| 2009/06/19 07:38:19 | 1.1 | gzheng | examples/bigsim/tools/changelog/bgTrace3 |
| 2009/06/19 07:38:19 | 1.1 | gzheng | examples/bigsim/tools/changelog/bgTrace2 |
| 2009/06/19 07:38:18 | 1.1 | gzheng | examples/bigsim/tools/changelog/bgTrace1 |
| 2009/06/19 07:38:18 | 1.1 | gzheng | examples/bigsim/tools/changelog/bgTrace0 |
| 2009/06/19 07:38:18 | 1.1 | gzheng | examples/bigsim/tools/changelog/bgTrace |
| 2009/06/19 07:38:18 | 1.1 | gzheng | examples/bigsim/tools/changelog/Makefile |
| a new example of how to change bgtrace logs/ | |||
| 2009/06/19 04:08:48 | 1.3 | skumar2 | src/arch/mpi-bluegenep/conv-mach.sh |
| Reversing order of libraries to suit V1R4. | |||
| 2009/06/18 18:02:40 | 1.319 | gzheng | src/libs/ck-libs/ampi/ampi.C |
| mark MPI_Send into a separate event | |||
| 2009/06/18 17:57:05 | 2.34 | gzheng | src/ck-perf/trace-bluegene.h |
| 2009/06/18 17:57:05 | 2.36 | gzheng | src/ck-perf/trace-bluegene.C |
| added const to strings to silent compiler warnings. | |||
| 2009/06/18 17:55:11 | 1.55 | gzheng | src/langs/bluegene/bigsim_logs.h |
| minor change to ascii format of bgTrace | |||
| 2009/06/18 10:28:18 | 1.10 | idooley2 | src/ck-perf/trace-summaryBOC.h |
| 2009/06/18 10:28:18 | 1.62 | idooley2 | src/ck-perf/trace-summary.h |
| 2009/06/18 10:28:17 | 1.12 | idooley2 | src/ck-perf/trace-summary.ci |
| 2009/06/18 10:28:17 | 1.111 | idooley2 | src/ck-perf/trace-summary.C |
| Removing CCS tracing code from trace-summary. The working, good version is in trace-utilization. | |||
| 2009/06/18 10:13:39 | 1.142 | chaomei | src/langs/bluegene/blue.C |
| 2009/06/18 10:13:39 | 1.57 | chaomei | src/langs/bluegene/bigsim_timing.C |
| A small bug fix for the previous check in where the non-out-of-core emulation case is forgoten to be considered. | |||
| 2009/06/18 09:50:47 | 1.20 | chaomei | src/langs/bluegene/bigsim_proc.C |
| A small change for defining the out-of-core emulation macro | |||
| 2009/06/18 09:47:42 | 1.141 | chaomei | src/langs/bluegene/blue.C |
| 2009/06/18 09:47:42 | 1.56 | chaomei | src/langs/bluegene/bigsim_timing.C |
| Fixed the bug of generating projection logs when out-of-core emulation and isomalloc are used together. | |||
| 2009/06/18 06:46:28 | 1.72 | gzheng | src/libs/ck-libs/tcharm/tcharm.C |
| update due to change in TRACE_BG_AMPI_BREAK macro for bigsim | |||
| 2009/06/17 17:47:11 | 1.492 | abecker | src/scripts/Makefile |
| remove redundant dependCaseSensitive rule | |||
| 2009/06/17 17:26:50 | 2.33 | gzheng | src/ck-perf/trace-bluegene.h |
| break MPI_Waitall event so that it does not directly depend on the previous event (i.e. the thread wakeup event, which may be part of the Ireq->wait block). Waitall build up its own event dependencies based on recv'ed messages. | |||
| 2009/06/17 17:24:56 | 1.318 | gzheng | src/libs/ck-libs/ampi/ampi.C |
| better tracing for recv a message and the thread wakeup event. | |||
| 2009/06/17 12:36:38 | 2.32 | chaomei | src/ck-perf/trace-bluegene.h |
| Some macros' definition are missing when CMK_OPTIMIZED is turned on. | |||
| 2009/06/17 11:53:18 | 1.4 | idooley2 | src/ck-cp/arrayRedistributor.h |
| Fixing bug where stuff was executed after migrateMe was called. | |||
| 2009/06/17 06:29:20 | 1.5 | idooley2 | src/ck-cp/controlPoints.ci |
| 2009/06/17 06:29:20 | 1.3 | idooley2 | src/ck-cp/arrayRedistributor.h |
|
Fixing a race condition that occurs when redistributing the data. A new assertion was added to identify a further potential race
condition that is not yet handled correctly. | |||
| 2009/06/16 11:38:20 | 1.17 | bhatele | src/langs/bluegene/bigsim_network.h |
| 2009/06/16 11:38:20 | 1.22 | bhatele | src/langs/bluegene/bigsim_init.C |
| network dimensions will be now set both in the emulator and the simulator | |||
| 2009/06/16 09:10:19 | 1.21 | idooley2 | src/ck-cp/controlPoints.C |
| Switching default behaviors of control points. | |||
| 2009/06/16 06:59:42 | 1.5 | abecker | src/libs/ck-libs/ParFUM/partition.C |
|
change back to node neighbors by default
Now node neighbors are once again used as the default for ParFUM partitioning; this will always give correct results. Face neighbors can be selected using +Parfum_face_neighbor_graph, and will give better quality partitions with less memory and computation, but will fail for some classes of input meshes. | |||
| 2009/06/15 16:49:58 | 1.24 | abecker | src/arch/bluegenep/machine.c |
| 2009/06/15 16:49:58 | 1.22 | abecker | java/charm/debug/ServThread.java |
| reverted accidental cvs changes from incorrect synchronization | |||
| 2009/06/15 16:06:14 | 1.4 | abecker | src/libs/ck-libs/ParFUM/partition.C |
|
partition based on face neighbors rather than node
The graph we were passing to metis was constructed using the rule "any elements which share a node are neighbors", which produces an unnecessarily large and low-quality graph. The new default rule is "any elements which share a face are neighbors". This requires correctly oriented elements, and does not correctly detect the difference between quad meshes and tet meshes (we assume tet). To restore the old behavior, use the flag +Parfum_node_neighbor_graph. | |||
| 2009/06/15 16:05:36 | 1.32 | abecker | src/langs/charj/src/charj/translator/CharjEmitter.g |
| fixed output for 'boolean' and 'byte' | |||
| 2009/06/15 16:05:33 | 1.1 | abecker | src/langs/charj/src/charj/translator/TreeTraverser.java |
| 2009/06/15 16:05:33 | 1.21 | abecker | src/langs/charj/src/charj/translator/Translator.java |
| 2009/06/15 16:05:33 | 1.1 | abecker | src/langs/charj/src/charj/translator/ASTHTMLPrinter.java |
| added tree traversal framework and HTML AST output | |||
| 2009/06/15 16:05:24 | 1.22 | abecker | src/langs/charj/src/charj/translator/CharjSemantics.g |
| 2009/06/15 16:05:24 | 1.31 | abecker | src/langs/charj/src/charj/translator/CharjEmitter.g |
| 2009/06/15 16:05:24 | 1.4 | abecker | src/langs/charj/tests/unit/TODO |
| fixed codegen for dotted primary expressions | |||
| 2009/06/15 16:05:19 | 1.21 | abecker | src/langs/charj/src/charj/translator/CharjSemantics.g |
| 2009/06/15 16:05:19 | 1.30 | abecker | src/langs/charj/src/charj/translator/CharjEmitter.g |
| 2009/06/15 16:05:19 | 1.3 | abecker | src/langs/charj/tests/unit/TODO |
| 2009/06/15 16:05:19 | 1.1 | abecker | src/langs/charj/tests/unit/SimpleChareArray.cj |
| 2009/06/15 16:05:19 | 1.1 | abecker | src/langs/charj/tests/unit/SimpleChare.cj |
| 2009/06/15 16:05:18 | 1.26 | abecker | src/langs/charj/src/charj/translator/Charj.stg |
| 2009/06/15 16:05:18 | 1.27 | abecker | src/langs/charj/src/charj/translator/Charj.g |
| nicer syntax for declaring chare,group,etc subtypes | |||
| 2009/06/15 16:05:09 | 1.25 | abecker | src/langs/charj/src/charj/translator/Charj.stg |
| local var decls don't have colon-suffix modifiers | |||
| 2009/06/15 16:05:06 | 1.20 | abecker | src/langs/charj/src/charj/translator/CharjSemantics.g |
| 2009/06/15 16:05:06 | 1.29 | abecker | src/langs/charj/src/charj/translator/CharjEmitter.g |
| 2009/06/15 16:05:06 | 1.24 | abecker | src/langs/charj/src/charj/translator/Charj.stg |
| 2009/06/15 16:05:06 | 1.26 | abecker | src/langs/charj/src/charj/translator/Charj.g |
|
improved handling of declaration modifiers
Made lists of modifiers optional instead of mandatory but possibly empty, which simplifies code generation. Differentiated between declarations of class fields and local variables. Added default access specifier to fields that don't have one, to prevent them from picking up the access modifier of previously specified variables in the C++ code. | |||
| 2009/06/15 16:05:04 | 1.20 | abecker | src/langs/bluegene/bigsim_api.C |
| change back so the BgStartLogByName when creating a dummy event does not automatically assign objID. It is done selectively in trace-bluegene now | |||
| 2009/06/15 16:04:58 | 1.2 | abecker | src/langs/charj/tests/unit/Makefile |
| 2009/06/15 16:04:57 | 1.20 | abecker | src/langs/charj/src/charj/translator/Translator.java |
| 2009/06/15 16:04:57 | 1.10 | abecker | src/langs/charj/src/charj/Main.java |
| 2009/06/15 16:04:57 | 1.9 | abecker | src/langs/charj/bin/charjc |
| 2009/06/15 16:04:56 | 1.12 | abecker | src/langs/charj/Makefile |
| added --translate-only flag, to skip C++ compilation step | |||
| 2009/06/15 16:04:53 | 1.25 | abecker | src/langs/charj/src/charj/translator/Charj.g |
|
made embed block rule a fragment to prevent unwanted matches
delimiting blocks with single braces instead of double necessitates tweaks to the EMBED_BLOCK rule to prevent it from pre-empting brace matching in other production rules. | |||
| 2009/06/15 16:04:51 | 1.19 | abecker | src/langs/charj/src/charj/translator/CharjSemantics.g |
| 2009/06/15 16:04:51 | 1.28 | abecker | src/langs/charj/src/charj/translator/CharjEmitter.g |
| 2009/06/15 16:04:51 | 1.23 | abecker | src/langs/charj/src/charj/translator/Charj.stg |
| 2009/06/15 16:04:50 | 1.24 | abecker | src/langs/charj/src/charj/translator/Charj.g |
| fixed output templates to match grammar updates | |||
| 2009/06/15 16:04:49 | 1.18 | abecker | src/langs/charj/src/charj/translator/CharjSemantics.g |
| fixed tree rules to match update grammar structure and token defs | |||
| 2009/06/15 16:04:46 | 1.23 | abecker | src/langs/charj/src/charj/translator/Charj.g |
| removed new operator for anonymous classes | |||
| 2009/06/15 16:04:44 | 1.22 | abecker | src/langs/charj/src/charj/translator/Charj.g |
| fixed embed block and for loop update rule | |||
| 2009/06/15 16:04:42 | 1.21 | abecker | src/langs/charj/src/charj/translator/Charj.g |
| fixed enum production rules | |||
| 2009/06/15 16:04:40 | 1.17 | abecker | src/langs/charj/src/charj/translator/CharjSemantics.g |
| 2009/06/15 16:04:40 | 1.27 | abecker | src/langs/charj/src/charj/translator/CharjEmitter.g |
| 2009/06/15 16:04:39 | 1.20 | abecker | src/langs/charj/src/charj/translator/Charj.g |
|
greatly simplified the grammar rules
Eliminated a lot of redundant rules that are necessary for LL(1) parsing but aren't needed in antlr. This simplifies the grammar quite a bit and enhances readability. Also removed some java-isms that we won't be supporting (at least not yet), like reflection and wildcard bound types. | |||
| 2009/06/15 16:04:38 | 1.19 | abecker | src/langs/bluegene/bigsim_api.C |
| assign objID from parent log | |||
| 2009/06/15 16:04:34 | 1.2 | abecker | src/langs/charj/tests/unit/FieldModifiers.cj |
| added another test case for field modifiers | |||
| 2009/06/15 16:04:32 | 1.19 | abecker | src/langs/charj/src/charj/translator/Charj.g |
| simplified handling of generic type argument lists | |||
| 2009/06/15 16:04:30 | 1.26 | abecker | src/langs/charj/src/charj/translator/CharjEmitter.g |
| 2009/06/15 16:04:30 | 1.18 | abecker | src/langs/charj/src/charj/translator/Charj.g |
| Fixed bracket escaping in <expr> templates | |||
| 2009/06/15 16:04:24 | 1.8 | abecker | src/langs/charj/src/charj/translator/SymbolTable.java |
| reformatting | |||
| 2009/06/15 16:04:22 | 1.25 | abecker | src/langs/charj/src/charj/translator/CharjEmitter.g |
| 2009/06/15 16:04:22 | 1.22 | abecker | src/langs/charj/src/charj/translator/Charj.stg |
| added templated for array construction codegen | |||
| 2009/06/15 16:04:17 | 1.16 | abecker | src/langs/charj/src/charj/translator/CharjSemantics.g |
| 2009/06/15 16:04:17 | 1.24 | abecker | src/langs/charj/src/charj/translator/CharjEmitter.g |
| 2009/06/15 16:04:17 | 1.1 | abecker | src/langs/charj/tests/unit/FieldModifiers.cj |
| 2009/06/15 16:04:16 | 1.21 | abecker | src/langs/charj/src/charj/translator/Charj.stg |
| 2009/06/15 16:04:16 | 1.17 | abecker | src/langs/charj/src/charj/translator/Charj.g |
| split field modifiers into their own unit test | |||
| 2009/06/15 16:04:13 | 1.23 | abecker | src/langs/charj/src/charj/translator/CharjEmitter.g |
| 2009/06/15 16:04:13 | 1.20 | abecker | src/langs/charj/src/charj/translator/Charj.stg |
| 2009/06/15 16:04:13 | 1.2 | abecker | src/langs/charj/tests/unit/ForLoop.cj |
| added expression generation, fixed for loop init codegen | |||
| 2009/06/15 16:04:11 | 1.1 | abecker | src/langs/charj/tests/unit/Label.cj |
| 2009/06/15 16:04:10 | 1.22 | abecker | src/langs/charj/src/charj/translator/CharjEmitter.g |
| 2009/06/15 16:04:10 | 1.19 | abecker | src/langs/charj/src/charj/translator/Charj.stg |
| added labeled statements and fixed switch body code-gen | |||
| 2009/06/15 16:04:07 | 1.15 | abecker | src/langs/charj/src/charj/translator/CharjSemantics.g |
| fixed token name | |||
| 2009/06/15 16:04:03 | 1.14 | abecker | src/langs/charj/src/charj/translator/CharjSemantics.g |
| 2009/06/15 16:04:03 | 1.21 | abecker | src/langs/charj/src/charj/translator/CharjEmitter.g |
| 2009/06/15 16:04:03 | 1.18 | abecker | src/langs/charj/src/charj/translator/Charj.stg |
| 2009/06/15 16:04:03 | 1.16 | abecker | src/langs/charj/src/charj/translator/Charj.g |
| overhaul of statement codegen | |||
| 2009/06/15 16:03:59 | 1.20 | abecker | src/langs/charj/src/charj/translator/CharjEmitter.g |
| 2009/06/15 16:03:59 | 1.17 | abecker | src/langs/charj/src/charj/translator/Charj.stg |
| added lots of undelegated templates | |||
| 2009/06/15 16:03:44 | 1.34 | abecker | src/libs/ck-libs/ParFUM-Tops/ParFUM_TOPS.cc |
| added success assertions | |||
| 2009/06/15 16:03:05 | 1.21 | abecker | java/charm/debug/ServThread.java |
| reading stderr as well as stdout | |||
| 2009/06/15 16:02:52 | 1.23 | abecker | src/arch/bluegenep/machine.c |
| this count should be declared independently | |||
| 2009/06/15 16:02:48 | 1.33 | abecker | src/libs/ck-libs/ParFUM-Tops/ParFUM_TOPS.cc |
| added cudaFree for two arrays that weren't freed before | |||
| 2009/06/15 16:00:13 | 1.317 | abecker | src/libs/ck-libs/ampi/ampi.C |
| 2009/06/15 16:00:13 | 1.113 | abecker | src/libs/ck-libs/ampi/ampi.h |
| Fixed size=0 bug; added Test_cancelled and Rbsend | |||
| 2009/06/15 15:59:52 | 1.1 | abecker | src/langs/charj/tests/fib/Fib.h |
| 2009/06/15 15:59:51 | 1.11 | abecker | src/langs/charj/Makefile |
| restrucured tests, tweaked fib | |||
| 2009/06/15 15:59:30 | 1.2 | abecker | src/langs/charj/src/charj/translator/InterfaceSymbol.java |
| formatting | |||
| 2009/06/15 15:59:21 | 1.7 | abecker | src/langs/charj/src/charj/translator/SymbolTable.java |
| modified type lookup in symbol table | |||
| 2009/06/15 15:59:18 | 1.6 | abecker | src/langs/charj/src/charj/translator/SymbolTable.java |
| 2009/06/15 15:59:18 | 1.22 | abecker | src/xlat-i/xi-main.C |
| 2009/06/15 15:59:17 | 1.6 | abecker | src/langs/charj/src/charj/translator/PackageScope.java |
| 2009/06/15 15:59:17 | 1.5 | abecker | src/langs/charj/src/charj/translator/ClassSymbol.java |
| Changed package lookup mechanism to be more class oriented | |||
| 2009/06/15 15:59:08 | 1.15 | abecker | src/langs/charj/src/charj/translator/Charj.g |
| minor cleanup | |||
| 2009/06/15 15:59:06 | 1.14 | abecker | src/langs/charj/src/charj/translator/Charj.g |
| fixed accidental paste | |||
| 2009/06/15 15:59:03 | 1.13 | abecker | src/langs/charj/src/charj/translator/CharjSemantics.g |
| 2009/06/15 15:59:03 | 1.19 | abecker | src/langs/charj/src/charj/translator/CharjEmitter.g |
| 2009/06/15 15:59:03 | 1.16 | abecker | src/langs/charj/src/charj/translator/Charj.stg |
| 2009/06/15 15:59:03 | 1.13 | abecker | src/langs/charj/src/charj/translator/Charj.g |
| got rid of error message cruft, improved importing | |||
| 2009/06/15 15:58:57 | 1.15 | abecker | src/langs/charj/src/charj/translator/Charj.stg |
| started introducing namespaces in generated code | |||
| 2009/06/15 15:58:21 | 1.491 | abecker | src/scripts/Makefile |
| Added a new target "dependsCaseSensitive" which will work on Macs and possibly other systems that have case insensitive filesystems. | |||
| 2009/06/15 12:08:13 | 1.140 | bhatele | src/langs/bluegene/blue.C |
| 2009/06/15 12:08:13 | 1.16 | bhatele | src/langs/bluegene/bigsim_network.h |
| 2009/06/15 12:08:13 | 1.21 | bhatele | src/langs/bluegene/bigsim_init.C |
|
intraNodeLat and intraNodeBW moved to BigNetSim
and BlueGeneL changed to BlueGeneP | |||
| 2009/06/15 07:29:52 | 1.4 | insomnia52 | examples/charm++/Molecular3D/src_physics/nonbonded.h |
| 2009/06/15 07:29:52 | 1.3 | insomnia52 | examples/charm++/Molecular3D/src_physics/defs.h |
| 2009/06/15 07:29:52 | 1.6 | insomnia52 | examples/charm++/Molecular3D/src_charm/mol3d.ci |
| 2009/06/15 07:29:52 | 1.7 | insomnia52 | examples/charm++/Molecular3D/src_charm/Patch.h |
| 2009/06/15 07:29:52 | 1.9 | insomnia52 | examples/charm++/Molecular3D/src_charm/Patch.C |
| 2009/06/15 07:29:52 | 1.9 | insomnia52 | examples/charm++/Molecular3D/src_charm/Main.C |
| 2009/06/15 07:29:52 | 1.5 | insomnia52 | examples/charm++/Molecular3D/src_charm/Compute.C |
| added 2-away X | |||
| 2009/06/15 07:20:51 | 1.151 | gzheng | src/libs/ck-libs/ampi/ampiimpl.h |
| added an "log" field into ampiMsg for bigsim only | |||
| 2009/06/15 06:49:35 | 1.316 | gzheng | src/libs/ck-libs/ampi/ampi.C |
| make sure the bigsim event dependencies in MPI_Waitall are generated correctly in all cases ( msg received before or after mpi-Irecv posted) | |||
| 2009/06/15 06:47:32 | 2.31 | gzheng | src/ck-perf/trace-bluegene.h |
| reduce projections call interferes log generation timing. | |||
| 2009/06/15 06:46:20 | 2.35 | gzheng | src/ck-perf/trace-bluegene.C |
| prevent some of the log generation cost being count as user entry time | |||
| 2009/06/12 12:44:59 | 1.185 | emenese2 | src/ck-core/ckarray.C |
| Deleted a few CkPrintfs. | |||
| 2009/06/12 09:33:09 | 1.173 | emenese2 | src/ck-ldb/CentralLB.C |
| Commented out some CkPrintf instructions. | |||
| 2009/06/12 09:25:38 | 2.18 | gzheng | src/ck-core/ckcallback.h |
| avoid a "compiler bug" with optimization that does not like copy from CkArrayIndex to CkArrayIndexMax. Cast CkArrayIndex to ArrayIndexMax to workaround. | |||
| 2009/06/12 09:04:43 | 1.172 | emenese2 | src/ck-ldb/CentralLB.C |
| Fixed a bug in the message logging protocol. | |||
| 2009/06/12 09:04:03 | 1.184 | emenese2 | src/ck-core/ckarray.C |
| Fixed a bug in the message logging protocol. | |||
| 2009/06/11 08:36:58 | 1.13 | skumar2 | src/util/BGPTorus.h |
| Enable rank2network. | |||
| 2009/06/11 08:34:09 | 1.22 | skumar2 | src/arch/bluegenep/machine.c |
| Enabling V1R4 driver which will be released soon. | |||
| 2009/06/10 19:23:13 | 1.139 | gzheng | src/langs/bluegene/blue.C |
| call PAPI_library_init so that PAPI_query_event behaves correctly. | |||
| 2009/06/10 15:27:00 | 1.18 | gzheng | src/scripts/conv-config.sh |
| void double "64" in QT when CMK_CC64 is set. | |||
| 2009/06/10 11:06:03 | 1.138 | gzheng | src/langs/bluegene/blue.C |
| print total number of FP instructions, L1 cache misses on each emulating processors | |||
| 2009/06/10 11:05:00 | 2.343 | chaomei | src/conv-core/convcore.c |
|
cmiArgDebugFlag may not be initialized when CmiPrintf is called, so a SEGFAULT is caused.
This problem is observed in MPI-SMP layer when running an app with projections and specifying the value of "+traceroot". (the value of traceroot is printed out through CmiPrintf) | |||
| 2009/06/10 10:26:20 | 1.3 | gzheng | src/arch/common/conv-mach-papi.sh |
| added -Wl,-rpath to papi libdir for loading the papi.so | |||
| 2009/06/10 09:58:02 | 1.4 | gzheng | src/arch/net-linux-x86_64/conv-mach.sh |
| set CMK_CC64 to true to note this is a 64 bit compiler | |||
| 2009/06/10 09:49:40 | 1.2 | gzheng | src/arch/common/conv-mach-papi.sh |
| guess default directory | |||
| 2009/06/09 12:04:00 | 1.2 | idooley2 | src/ck-cp/arrayRedistributor.h |
| 2009/06/09 12:04:00 | 1.2 | idooley2 | src/ck-cp/arrayRedistributor.C |
| Adding some checks on the bounds of values passed into some functions. | |||
| 2009/06/09 11:00:18 | 1.40 | idooley2 | tools/projections/projections/gui/Timeline/EntryMethodObject.java |
| 2009/06/09 11:00:18 | 1.7 | idooley2 | tools/projections/projections/gui/Timeline/MainPanelForeground.java |
| 2009/06/09 11:00:17 | 1.66 | idooley2 | tools/projections/projections/gui/Timeline/Data.java |
| Fixing bug with drawing of message send lines. If an entry method had duration of 0us, then it would draw incorrectly. | |||
| 2009/06/09 10:59:56 | 1.355 | idooley2 | tools/projections/bin/projections.jar |
| Fixing bug with drawing of message send lines. If an entry method had duration of 0us, then it would draw incorrectly. | |||
| 2009/06/09 10:58:51 | 1.29 | idooley2 | tools/projections/projections/gui/Timeline/UserEventObject.java |
| 2009/06/09 10:58:51 | 1.83 | idooley2 | tools/projections/projections/gui/Analysis.java |
| Better handling of null user event names. | |||
| 2009/06/09 10:57:55 | 1.16 | idooley2 | tools/projections/projections/misc/LogEntryData.java |
| Adding more information to the log file viewer. | |||
| 2009/06/09 09:15:56 | 1.28 | gzheng | src/libs/ck-libs/multicast/ckmulticast.h |
| 2009/06/09 09:15:56 | 1.17 | gzheng | src/libs/ck-libs/multicast/ckmulticast.ci |
| added migratable attrib to ckmulticast group. This makes checkpointing possible.However, multicast group can not be directly pup'ed due to cookie stuff. After restart, the group will be empty, and the multicast tree needs to be rebuilt. | |||
| 2009/06/08 11:43:13 | 1.2 | insomnia52 | examples/charm++/Molecular3D/README.tex |
| updated readme file | |||
| 2009/06/08 11:41:03 | 1.4 | insomnia52 | examples/charm++/Molecular3D/src_namd/copy_namd.sh |
| 2009/06/08 11:41:03 | 1.1 | insomnia52 | examples/charm++/Molecular3D/README.tex |
| administrational changes | |||
| 2009/06/08 11:38:52 | 1.1 | insomnia52 | examples/charm++/Molecular3D/src_namd/jsplugin.c |
| this namd file had a bug involved with endinannes so we fixed it and attach it to the other files | |||
| 2009/06/08 07:23:50 | 1.3 | insomnia52 | examples/charm++/Molecular3D/src_physics/nonbonded.h |
| 2009/06/08 07:23:49 | 1.5 | insomnia52 | examples/charm++/Molecular3D/src_charm/mol3d.ci |
| 2009/06/08 07:23:49 | 1.6 | insomnia52 | examples/charm++/Molecular3D/src_charm/Patch.h |
| 2009/06/08 07:23:49 | 1.8 | insomnia52 | examples/charm++/Molecular3D/src_charm/Patch.C |
| 2009/06/08 07:23:49 | 1.8 | insomnia52 | examples/charm++/Molecular3D/src_charm/Main.C |
| 2009/06/08 07:23:49 | 1.3 | insomnia52 | examples/charm++/Molecular3D/src_charm/Compute.h |
| 2009/06/08 07:23:49 | 1.4 | insomnia52 | examples/charm++/Molecular3D/src_charm/Compute.C |
| added ckmulticast reductions and also input parameter for multicast tree branch factor | |||
| 2009/06/08 07:18:31 | 1.354 | idooley2 | tools/projections/bin/projections.jar |
| The log file viewer now interprets user events correctly. | |||
| 2009/06/07 16:16:00 | 1.15 | gzheng | tests/charm++/Makefile |
| disable commtest temporarily | |||
| 2009/06/07 02:44:39 | 2.30 | gzheng | src/ck-perf/trace-bluegene.h |
| fixed a bug in adding backward deps in AMPI_BREAK | |||
| 2009/06/06 18:26:42 | 1.12 | gzheng | examples/bigsim/tools/loadlog/loadlog.C |
| sanity check for the case when both incoming message and backward dep exist | |||
| 2009/06/06 18:25:57 | 1.315 | gzheng | src/libs/ck-libs/ampi/ampi.C |
| now make sure it won't heppen when an event has both incoming mesg and backward depes. This confuses bigNetsim a lot. | |||
| 2009/06/05 10:01:00 | 1.1 | idooley2 | examples/charm++/gaussSeidel3D/gaussSeidel3d.ci |
| 2009/06/05 10:01:00 | 1.1 | idooley2 | examples/charm++/gaussSeidel3D/gaussSeidel3d.C |
| 2009/06/05 10:01:00 | 1.1 | idooley2 | examples/charm++/gaussSeidel3D/Makefile |
| Adding a new Gauss Seidel example program. | |||
| 2009/06/05 09:29:43 | 1.18 | gzheng | src/langs/bluegene/bigsim_api.C |
| 2009/06/05 09:29:42 | 1.54 | gzheng | src/langs/bluegene/bigsim_logs.h |
| a minor change to BgWriteTraceSummary() so that one can specify a file name other than "bgTrace" | |||
| 2009/06/05 05:53:35 | 1.3 | gzheng | tests/charm++/commtest/section-multicast-test-6/test.C |
| random => rand(), as random() does not exist on windows | |||
| 2009/06/05 05:46:07 | 1.3 | gzheng | src/ck-cp/pathHistory.C |
| fixed a compiler error for uth-linux | |||
| 2009/06/05 02:37:43 | 1.11 | gzheng | src/conv-com/convcomlibstrategy.h |
| 2009/06/05 02:37:43 | 1.26 | gzheng | src/conv-com/convcomlibmanager.C |
| 2009/06/05 02:37:43 | 1.2 | gzheng | src/conv-com/StreamingStrategy.h |
| 2009/06/05 02:37:43 | 1.2 | gzheng | src/conv-com/StreamingStrategy.C |
| change Ckpv to Cpv in conv-com | |||
| 2009/06/04 17:30:13 | 1.314 | gzheng | src/libs/ck-libs/ampi/ampi.C |
| fixed a bug in MPI_Ssend when sending ack back by examing the message, the message could have been free'ed at that point. | |||
| 2009/06/04 16:58:32 | 2.59 | gzheng | src/ck-core/envelope.h |
| initialize envelop's ref counter to 0. | |||
| 2009/06/04 16:57:46 | 1.150 | gzheng | src/libs/ck-libs/ampi/ampiimpl.h |
| 2009/06/04 16:57:46 | 1.313 | gzheng | src/libs/ck-libs/ampi/ampi.C |
| 2009/06/04 16:57:46 | 1.112 | gzheng | src/libs/ck-libs/ampi/ampi.h |
| a real implementation of MPI_Ssend and MPI_Ssend_init, which requires receiver to send ack to sender | |||
| 2009/06/04 14:17:05 | 1.149 | chaomei | src/libs/ck-libs/ampi/ampiimpl.h |
| Bug fix for bigsim out-of-core emulation when generating forward/backward dependencies in trace logs. In particular, the "event" field in AmpiRequest and ATAReq needs to be pupped. | |||
| 2009/06/04 12:33:09 | 1.4 | idooley2 | src/ck-cp/pathHistory.h |
| Fixing minor problem with recent checkin. | |||
| 2009/06/04 12:30:08 | 1.3 | idooley2 | src/ck-cp/pathHistory.h |
| 2009/06/04 12:30:08 | 1.2 | idooley2 | src/ck-cp/pathHistory.C |
| Adding no-op versions of the critical path methods. | |||
| 2009/06/04 12:01:48 | 1.24 | bhatele | examples/charm++/topology/jacobi3d/jacobi3d.C |
| bug fix | |||
| 2009/06/04 11:50:17 | 1.5 | bhatele | examples/bigsim/sdag/jacobi3d/jacobi3d.C |
| minor bug fix by Isaac | |||
| 2009/06/04 08:20:06 | 1.44 | bhatele | src/ck-com/ComlibManager.C |
| commented out a printf | |||
| 2009/06/04 06:37:06 | 1.3 | bhatele | examples/bigsim/sdag/jacobi3d/jacobi3d.ci |
| 2009/06/04 06:37:06 | 1.4 | bhatele | examples/bigsim/sdag/jacobi3d/jacobi3d.C |
| use a feature of SDAG to keep track of messages across iterations | |||
| 2009/06/04 06:07:48 | 1.17 | idooley2 | tests/charm++/commtest/Makefile |
| Disabling each-to-many test until it is fixed. | |||
| 2009/06/04 05:43:14 | 1.2 | gzheng | tests/charm++/commtest/eachtomany-test-2/Makefile |
| pass TESTOPTS | |||
| 2009/06/04 05:42:43 | 1.2 | gzheng | tests/charm++/commtest/section-multicast-test-6/Makefile |
| pass TESTOPTS | |||
| 2009/06/04 05:41:35 | 1.2 | gzheng | tests/charm++/commtest/section-multicast-test/Makefile |
| pass TESTOPTS | |||
| 2009/06/04 05:32:39 | 1.14 | gzheng | tests/charm++/Makefile |
| skip commtest for bigsim | |||
| 2009/06/04 03:41:24 | 1.92 | gzheng | src/scripts/Make.depends |
| updated | |||
| 2009/06/03 19:18:52 | 1.1 | gzheng | src/libs/ck-libs/ampi/ampif90 |
| 2009/06/03 19:18:52 | 1.1 | gzheng | src/libs/ck-libs/ampi/ampif77 |
| 2009/06/03 19:18:52 | 1.1 | gzheng | src/libs/ck-libs/ampi/ampiCC |
| 2009/06/03 19:18:52 | 1.41 | gzheng | src/libs/ck-libs/ampi/Makefile |
| renamed original mpi scripts | |||
| 2009/06/03 14:00:16 | 1.117 | gzheng | src/arch/mpi/machine.c |
| progress engine for SMP was empty, change it to call communicationServer when in comm thread | |||
| 2009/06/03 13:07:45 | 1.3 | bhatele | examples/bigsim/sdag/jacobi3d/jacobi3d.C |
| 2009/06/03 13:07:45 | 1.3 | bhatele | examples/bigsim/sdag/jacobi3d/Makefile |
| optimizations | |||
| 2009/06/03 11:41:25 | 1.43 | idooley2 | src/ck-com/ComlibManager.C |
| Removing a periodic call used for debugging. | |||
| 2009/06/03 11:30:50 | 1.2 | bhatele | examples/bigsim/sdag/jacobi3d/jacobi3d.ci |
| 2009/06/03 11:30:50 | 1.2 | bhatele | examples/bigsim/sdag/jacobi3d/jacobi3d.C |
| 2009/06/03 11:30:50 | 1.2 | bhatele | examples/bigsim/sdag/jacobi3d/Makefile |
| working example now | |||
| 2009/06/03 11:22:31 | 1.16 | idooley2 | tests/charm++/commtest/Makefile |
| Adding each-to-many commlib test. | |||
| 2009/06/03 11:06:24 | 1.10 | idooley2 | src/conv-com/convcomlibmanager.h |
| 2009/06/03 11:06:24 | 1.25 | idooley2 | src/conv-com/convcomlibmanager.C |
| 2009/06/03 11:06:23 | 1.20 | idooley2 | src/ck-com/ComlibManager.h |
| 2009/06/03 11:06:23 | 1.8 | idooley2 | src/ck-com/ComlibManager.ci |
| 2009/06/03 11:06:23 | 1.42 | idooley2 | src/ck-com/ComlibManager.C |
| 2009/06/03 11:06:22 | 1.91 | idooley2 | src/scripts/Make.depends |
|
Moving some comlib startup code to an initproc routine. This code needs to execute before the user program's mainchare executes, so it cannot be
performed inside the comlib mainchare. | |||
| 2009/06/03 09:40:40 | 1.13 | idooley2 | tests/charm++/Makefile |
|
Reenabling commtest now that there are three test programs that use the new comlib. These test programs have successfully run on two
test platforms. | |||
| 2009/06/03 09:37:48 | 1.2 | idooley2 | tests/charm++/commtest/eachtomany-test-2/eachtomany.C |
| Removing some of the print statements from the test. | |||
| 2009/06/03 09:31:57 | 1.2 | idooley2 | tests/charm++/commtest/section-multicast-test-6/test.C |
| Removing print statements and reducing number of iterations in test | |||
| 2009/06/03 08:16:52 | 1.18 | idooley2 | doc/projections/manual.tex |
| Updating manual to note a newly discovered problem. | |||
| 2009/06/03 07:44:51 | 1.11 | idooley2 | tools/projections/build.xml |
| Forcing source and target java versions to be 1.5. | |||
| 2009/06/03 07:00:33 | 1.353 | idooley2 | tools/projections/bin/projections.jar |
| Recompiled the jar file, incase it isn't up to date right now. | |||
| 2009/06/03 06:08:40 | 1.40 | gzheng | src/libs/ck-libs/ampi/Makefile |
| removed mpicc since it is same as mpiCC. Makefile will create ampicc from mpiCC . | |||
| 2009/06/03 05:57:14 | 1.5 | gzheng | src/libs/ck-libs/parmetis/Makefile.in |
| mpicc to ampicc | |||
| 2009/06/03 05:51:37 | 1.36 | gzheng | src/libs/ck-libs/ParFUM/Makefile |
| mpiCC to ampiCC | |||
| 2009/06/03 05:50:24 | 1.52 | gzheng | src/libs/ck-libs/fem/Makefile |
| mpiCC to ampiCC | |||
| 2009/06/03 05:48:48 | 1.14 | gzheng | src/libs/ck-libs/datatransfer/Makefile |
| updated Makefile | |||
| 2009/06/03 05:33:23 | 1.39 | gzheng | src/libs/ck-libs/ampi/Makefile |
| do not provide mpicc named scripts | |||
| 2009/06/03 05:15:17 | 1.9 | gzheng | src/arch/mpi-linux-x86_64/conv-mach.sh |
| give an error and abort when ampiCC is called instead of the native mpiCC | |||
| 2009/06/02 18:48:49 | 1.38 | gzheng | src/libs/ck-libs/ampi/Makefile |
|
copy mpicc, etc to ampicc for alternatives.
having same name as native mpicc creates problems when building charm on mpi | |||
| 2009/06/02 15:56:38 | 1.1 | idooley2 | tests/charm++/commtest/eachtomany-test-2/runTests.pl |
| 2009/06/02 15:56:38 | 1.1 | idooley2 | tests/charm++/commtest/eachtomany-test-2/eachtomany.ci |
| 2009/06/02 15:56:38 | 1.1 | idooley2 | tests/charm++/commtest/eachtomany-test-2/eachtomany.C |
| 2009/06/02 15:56:38 | 1.1 | idooley2 | tests/charm++/commtest/eachtomany-test-2/eachtomany-Group.ci |
| 2009/06/02 15:56:38 | 1.1 | idooley2 | tests/charm++/commtest/eachtomany-test-2/eachtomany-Group.C |
| 2009/06/02 15:56:37 | 1.1 | idooley2 | tests/charm++/commtest/eachtomany-test-2/Makefile |
| 2009/06/02 15:56:37 | 1.15 | idooley2 | tests/charm++/commtest/Makefile |
| Adding a new each-to-many comlib test program. This test program is not yet built or tested by "make test" | |||
| 2009/06/02 15:38:44 | 1.1 | idooley2 | tests/charm++/commtest/section-multicast-test/test.ci |
| 2009/06/02 15:38:44 | 1.1 | idooley2 | tests/charm++/commtest/section-multicast-test/test.C |
| 2009/06/02 15:38:44 | 1.1 | idooley2 | tests/charm++/commtest/section-multicast-test/runTests.pl |
| 2009/06/02 15:38:44 | 1.1 | idooley2 | tests/charm++/commtest/section-multicast-test/Makefile |
| 2009/06/02 15:38:44 | 1.1 | idooley2 | tests/charm++/commtest/section-multicast-test-6/test.ci |
| 2009/06/02 15:38:44 | 1.1 | idooley2 | tests/charm++/commtest/section-multicast-test-6/test.C |
| 2009/06/02 15:38:44 | 1.1 | idooley2 | tests/charm++/commtest/section-multicast-test-6/Makefile |
| 2009/06/02 15:38:44 | 1.1 | idooley2 | tests/charm++/commtest/section-multicast-test-6/runTests.pl |
| 2009/06/02 15:38:43 | 1.14 | idooley2 | tests/charm++/commtest/Makefile |
| Adding two new comlib tests of section multicasts. These tests should work with the new version of comlib. | |||
| 2009/06/02 11:20:40 | 1.312 | idooley2 | src/libs/ck-libs/ampi/ampi.C |
| Disabling commlib entirely for AMPI. In the future it will be reenabled once it has been fully tested. | |||
| 2009/06/02 11:08:42 | 1.13 | idooley2 | tests/charm++/commtest/Makefile |
| Disabling compilation and running of these tests. The tests as they are updated will be reenabled. | |||
| 2009/06/02 10:52:13 | 1.13 | idooley2 | tests/charm++/commtest/comlib/Makefile |
| Disabling old comlib tests. New tests will be added and enabled in the future. | |||
| 2009/06/02 09:53:55 | 1.26 | gzheng | src/conv-core/cputopology.C |
| a new way of calculating number of uniq nodes without the strong assumption of charm assigning node seq no. This should make it work with topomanager | |||
| 2009/06/02 09:22:15 | 1.2 | gzheng | src/arch/mpi-sol-x86_64/conv-mach.sh |
| removed -lmpich | |||
| 2009/06/02 09:19:53 | 1.9 | gzheng | src/arch/mpi-linux-ia64/conv-mach.sh |
| removed -lmpich | |||
| 2009/06/02 08:59:27 | 2.342 | phil | src/conv-core/convcore.c |
| Clean up some commenting errors that caused autobuild failures | |||
| 2009/06/02 07:53:05 | 1.12 | gzheng | src/arch/mpi-linux/conv-mach.sh |
| remove -lmpich since mpiCC is used for linking | |||
| 2009/06/02 05:59:56 | 1.25 | gzheng | src/conv-core/cputopology.C |
|
fixed a bug in cray, bluegene specific code
change comm thread to call network progress engine. | |||
| 2009/06/02 05:58:56 | 1.30 | gzheng | src/conv-core/cpuaffinity.c |
| use a lock to protect a node level counter that comm thread uses to exit from its waiting loop | |||
| 2009/06/02 02:24:02 | 1.120 | gzheng | src/libs/ck-libs/fem/fem.C |
| use lock to protect PUPable which is smp safe. | |||
| 2009/06/02 02:23:27 | 1.42 | gzheng | src/util/pup_util.C |
| revert the previous change since it crashes marshall test in megatest | |||
| 2009/06/01 17:35:08 | 1.41 | gzheng | src/util/pup_util.C |
| 2009/06/01 17:35:07 | 2.341 | gzheng | src/conv-core/convcore.c |
| revert the previous change. Basically pup_util.C can be compiled outside of charm as standalone utility, so it can not be smp-aware. The previous bug is only partially resolved. Only machines that support __thread is free of this pup/smp bug | |||
| 2009/06/01 10:44:52 | 1.1 | bhatele | examples/bigsim/sdag/jacobi3d/jacobi3d.ci |
| 2009/06/01 10:44:52 | 1.1 | bhatele | examples/bigsim/sdag/jacobi3d/jacobi3d.C |
| 2009/06/01 10:44:52 | 1.1 | bhatele | examples/bigsim/sdag/jacobi3d/Makefile |
| jacobi3d in sdag | |||
| 2009/06/01 07:19:07 | 1.40 | gzheng | src/util/pup_util.C |
| 2009/06/01 07:19:06 | 2.340 | gzheng | src/conv-core/convcore.c |
| a better fix I think. Create a pup global init function which is called from ConverseCommonInit(). Inside which, a cpvinitilize function is called on the PUP_reg table variable. | |||
| 2009/06/01 07:08:17 | 1.39 | gzheng | src/util/pup_util.C |
|
PUP_regEntry is not SMP safe!
A simple fix to use TLS to make PUP_regEntry private to each processor works. However, since there is pup is written generally not allowing Cpv varaibles, I have not completely fix this problem. | |||
| 2009/06/01 06:03:10 | 1.29 | idooley2 | src/libs/ck-libs/ParFUM/import.C |
| 2009/06/01 06:03:10 | 1.148 | idooley2 | src/libs/ck-libs/ampi/ampiimpl.h |
| 2009/06/01 06:03:09 | 2.339 | idooley2 | src/conv-core/convcore.c |
| Commenting out #warning preprocessor statements. Windows compiler doesn't support these. | |||
| 2009/06/01 05:10:38 | 1.24 | idooley2 | src/conv-com/convcomlibmanager.C |
| Adding missing return statements. | |||
| 2009/05/31 08:34:31 | 1.41 | idooley2 | src/ck-com/ComlibManager.C |
| An attempt to fix a problem that occurs when compiling net-sol and other versions. For some reason the syntax was not parsing. | |||
| 2009/05/31 07:05:39 | 1.14 | idooley2 | src/ck-com/ComlibSectionInfo.C |
| Removed #warning preprocessor command that is not liked by windows compiler. | |||
| 2009/05/30 08:04:29 | 2.58 | idooley2 | src/ck-core/envelope.h |
| 2009/05/30 08:04:29 | 2.164 | idooley2 | src/ck-core/charm++.h |
| Adding include for envelope.h into charm++.h, and moving one inline function back into charm++ after the include. | |||
| 2009/05/29 08:00:30 | 2.57 | idooley2 | src/ck-core/envelope.h |
| 2009/05/29 08:00:29 | 2.97 | idooley2 | src/ck-core/charm.h |
| Moving an enumeration of Charm++ message types into charm.h from envelope.h. This way the use of these types in generated .def.h files will not cause a compiler error. | |||
| 2009/05/29 07:59:09 | 1.15 | idooley2 | src/xlat-i/sdag/CSdagConstruct.C |
| 2009/05/29 07:59:09 | 1.10 | idooley2 | src/xlat-i/sdag/CParsedFile.h |
| Critical path generation for sdag is now optional. It would fail previously if charm++ was not built with critical path generation enabled. | |||
| 2009/05/29 06:23:34 | 1.4 | idooley2 | tests/util/check.C |
| Including envelope.h | |||
| 2009/05/29 06:18:44 | 1.58 | gzheng | src/util/pup.h |
| some changes so that pup.h can be used in standalone mode outside of charm. | |||
| 2009/05/28 19:22:38 | 1.9 | skumar2 | src/conv-com/convcomlibmanager.h |
| Fixing ICE from the XL compiler. | |||
| 2009/05/28 18:47:41 | 1.21 | skumar2 | src/arch/bluegenep/machine.c |
| Adding a flag to throttle broadcast. | |||
| 2009/05/28 13:06:37 | 1.4 | kunzman | examples/charm++/cell/md/selfCompute.ci |
| 2009/05/28 13:06:37 | 1.4 | kunzman | examples/charm++/cell/md/patch.h |
| 2009/05/28 13:06:37 | 1.4 | kunzman | examples/charm++/cell/md/patch.ci |
| 2009/05/28 13:06:37 | 1.7 | kunzman | examples/charm++/cell/md/patch.C |
| 2009/05/28 13:06:37 | 1.5 | kunzman | examples/charm++/cell/md/pairCompute.ci |
| 2009/05/28 13:06:37 | 1.5 | kunzman | examples/charm++/cell/md/md_config.h |
| 2009/05/28 13:06:37 | 1.8 | kunzman | examples/charm++/cell/md/main.h |
| 2009/05/28 13:06:37 | 1.5 | kunzman | examples/charm++/cell/md/main.C |
| Updated code to refect changes to simd.h. In particular, the removal of the vector size from the names of vector types and operations. | |||
| 2009/05/28 13:04:55 | 1.15 | kunzman | src/util/simd.h |
| Changed the type and operation names to not include the vector size. This is being done for portability between platforms with vectors of different sizes. | |||
| 2009/05/28 11:06:32 | 1.490 | idooley2 | src/scripts/Makefile |
| 2009/05/28 11:06:31 | 1.90 | idooley2 | src/scripts/Make.depends |
| Missed a dependency in the makefile. | |||
| 2009/05/28 10:59:46 | 1.1 | idooley2 | src/ck-perf/trace-utilization.h |
| 2009/05/28 10:59:46 | 1.1 | idooley2 | src/ck-perf/trace-utilization.ci |
| 2009/05/28 10:59:46 | 1.1 | idooley2 | src/ck-perf/trace-utilization.C |
| 2009/05/28 10:59:45 | 1.225 | idooley2 | src/scripts/charmc |
| 2009/05/28 10:59:45 | 1.489 | idooley2 | src/scripts/Makefile |
| Adding a new trace module that gathers sum-detail style utilization profiles and sends them over CCS to a visualization client. | |||
| 2009/05/28 10:45:05 | 1.1 | idooley2 | src/ck-core/envelope-path.h |
| Forgot to check this file in earlier. | |||
| 2009/05/28 10:28:12 | 1.7 | idooley2 | tools/projections/projections/streaming/StreamingTool.java |
| 2009/05/28 10:28:12 | 1.6 | idooley2 | tools/projections/projections/streaming/StartupDialogBox.java |
| 2009/05/28 10:28:11 | 1.7 | idooley2 | tools/projections/projections/streaming/MultiSeriesHandler.java |
| Detailed view for streaming tool. | |||
| 2009/05/28 10:27:50 | 1.352 | idooley2 | tools/projections/bin/projections.jar |
| Detailed view for streaming tool. | |||
| 2009/05/28 10:15:30 | 1.8 | idooley2 | src/xlatcpm/conv-cpm.c |
| 2009/05/28 10:15:29 | 1.224 | idooley2 | src/scripts/charmc |
| 2009/05/28 10:15:29 | 1.488 | idooley2 | src/scripts/Makefile |
| 2009/05/28 10:15:28 | 1.16 | idooley2 | src/libs/ck-libs/tcharm/tcharm.h |
| 2009/05/28 10:15:28 | 1.71 | idooley2 | src/libs/ck-libs/tcharm/tcharm.C |
| 2009/05/28 10:15:28 | 1.89 | idooley2 | src/scripts/Make.depends |
| 2009/05/28 10:15:27 | 1.10 | idooley2 | src/libs/ck-libs/datatransfer/paralleltransfer.C |
| 2009/05/28 10:15:27 | 1.51 | idooley2 | src/libs/ck-libs/pose/pose.h |
| 2009/05/28 10:15:27 | 1.41 | idooley2 | src/libs/ck-libs/pose/pose.C |
| 2009/05/28 10:15:27 | 1.48 | idooley2 | src/libs/ck-libs/pose/sim.C |
| 2009/05/28 10:15:27 | 1.94 | idooley2 | src/libs/ck-libs/pose/gvt.C |
| 2009/05/28 10:15:26 | 1.3 | idooley2 | src/libs/ck-libs/datatransfer/parallelsurfacetransfer.C |
| 2009/05/28 10:15:26 | 1.7 | idooley2 | src/libs/ck-libs/cache/CkCache.h |
| 2009/05/28 10:15:25 | 1.147 | idooley2 | src/libs/ck-libs/ampi/ampiimpl.h |
| 2009/05/28 10:15:25 | 1.311 | idooley2 | src/libs/ck-libs/ampi/ampi.C |
| 2009/05/28 10:15:24 | 1.66 | idooley2 | src/conv-core/memory.c |
| 2009/05/28 10:15:24 | 2.345 | idooley2 | src/conv-core/converse.h |
| 2009/05/28 10:15:24 | 2.338 | idooley2 | src/conv-core/convcore.c |
| 2009/05/28 10:15:23 | 1.13 | idooley2 | src/conv-com/routerstrategy.C |
| 2009/05/28 10:15:23 | 1.15 | idooley2 | src/conv-com/petable.C |
| 2009/05/28 10:15:23 | 1.4 | idooley2 | src/conv-com/treerouter.h |
| 2009/05/28 10:15:23 | 1.3 | idooley2 | src/conv-com/treerouter.C |
| 2009/05/28 10:15:23 | 1.5 | idooley2 | src/conv-com/routerstrategy.h |
| 2009/05/28 10:15:23 | 1.5 | idooley2 | src/conv-com/router.h |
| 2009/05/28 10:15:23 | 1.4 | idooley2 | src/conv-com/prefixrouter.h |
| 2009/05/28 10:15:23 | 1.2 | idooley2 | src/conv-com/prefixrouter.C |
| 2009/05/28 10:15:23 | 1.2 | idooley2 | src/conv-com/pipelinestrategy.h |
| 2009/05/28 10:15:23 | 1.2 | idooley2 | src/conv-com/pipelinestrategy.C |
| 2009/05/28 10:15:23 | 1.4 | idooley2 | src/conv-com/pipebroadcastconverse.h |
| 2009/05/28 10:15:23 | 1.16 | idooley2 | src/conv-com/pipebroadcastconverse.C |
| 2009/05/28 10:15:23 | 1.8 | idooley2 | src/conv-com/petable.h |
| 2009/05/28 10:15:23 | 1.1 | idooley2 | src/conv-com/nocomlib.c |
| 2009/05/28 10:15:23 | 1.2 | idooley2 | src/conv-com/hypercubetopology.h |
| 2009/05/28 10:15:22 | 1.10 | idooley2 | src/conv-com/convcomlibstrategy.h |
| 2009/05/28 10:15:22 | 1.9 | idooley2 | src/conv-com/convcomlibstrategy.C |
| 2009/05/28 10:15:22 | 1.8 | idooley2 | src/conv-com/convcomlibmanager.h |
| 2009/05/28 10:15:22 | 1.23 | idooley2 | src/conv-com/convcomlibmanager.C |
| 2009/05/28 10:15:22 | 1.2 | idooley2 | src/conv-com/hypercubetopology.C |
| 2009/05/28 10:15:22 | 1.1 | idooley2 | src/conv-com/hypercuberouter.h |
| 2009/05/28 10:15:22 | 1.1 | idooley2 | src/conv-com/hypercuberouter.C |
| 2009/05/28 10:15:22 | 1.5 | idooley2 | src/conv-com/gridrouter.h |
| 2009/05/28 10:15:22 | 1.8 | idooley2 | src/conv-com/gridrouter.C |
| 2009/05/28 10:15:22 | 1.3 | idooley2 | src/conv-com/graphrouter.h |
| 2009/05/28 10:15:22 | 1.5 | idooley2 | src/conv-com/graphrouter.C |
| 2009/05/28 10:15:22 | 1.2 | idooley2 | src/conv-com/dummyconversestrategy.h |
| 2009/05/28 10:15:22 | 1.2 | idooley2 | src/conv-com/dummyconversestrategy.C |
| 2009/05/28 10:15:22 | 1.2 | idooley2 | src/conv-com/convcomlib.h |
| 2009/05/28 10:15:22 | 1.1 | idooley2 | src/conv-com/StreamingStrategy.h |
| 2009/05/28 10:15:22 | 1.1 | idooley2 | src/conv-com/StreamingStrategy.C |
| 2009/05/28 10:15:21 | 1.1 | idooley2 | src/conv-com/MeshStreamingStrategy.h |
| 2009/05/28 10:15:21 | 1.1 | idooley2 | src/conv-com/MeshStreamingStrategy.C |
| 2009/05/28 10:15:21 | 1.4 | idooley2 | src/conv-com/3dgridrouter.h |
| 2009/05/28 10:15:21 | 1.7 | idooley2 | src/conv-com/3dgridrouter.C |
| 2009/05/28 10:15:21 | 2.34 | idooley2 | src/ck-perf/trace-bluegene.C |
| 2009/05/28 10:15:20 | 1.9 | idooley2 | src/ck-ldb/TopoLB.h |
| 2009/05/28 10:15:20 | 1.8 | idooley2 | src/ck-ldb/TopoCentLB.h |
| 2009/05/28 10:15:20 | 1.20 | idooley2 | src/ck-cp/controlPoints.C |
| 2009/05/28 10:15:19 | 2.81 | idooley2 | src/ck-core/cklocation.h |
| 2009/05/28 10:15:19 | 2.124 | idooley2 | src/ck-core/cklocation.C |
| 2009/05/28 10:15:19 | 1.140 | idooley2 | src/ck-core/ckarray.h |
| 2009/05/28 10:15:19 | 2.56 | idooley2 | src/ck-core/envelope.h |
| 2009/05/28 10:15:19 | 2.15 | idooley2 | src/ck-core/cksection.h |
| 2009/05/28 10:15:19 | 2.13 | idooley2 | src/ck-core/ckcallback.C |
| 2009/05/28 10:15:19 | 1.33 | idooley2 | src/ck-core/ckarray.ci |
| 2009/05/28 10:15:18 | 1.183 | idooley2 | src/ck-core/ckarray.C |
| 2009/05/28 10:15:18 | 2.163 | idooley2 | src/ck-core/charm++.h |
| 2009/05/28 10:15:17 | 1.2 | idooley2 | src/ck-com/comlib.h |
| 2009/05/28 10:15:17 | 1.6 | idooley2 | src/ck-com/RingMulticastStrategy.h |
| 2009/05/28 10:15:17 | 1.9 | idooley2 | src/ck-com/RingMulticastStrategy.C |
| 2009/05/28 10:15:17 | 1.2 | idooley2 | src/ck-com/RectMulticastStrategy.h |
| 2009/05/28 10:15:17 | 1.2 | idooley2 | src/ck-com/RectMulticastStrategy.C |
| 2009/05/28 10:15:17 | 1.6 | idooley2 | src/ck-com/PrioStreaming.h |
| 2009/05/28 10:15:17 | 1.4 | idooley2 | src/ck-com/PrioStreaming.C |
| 2009/05/28 10:15:17 | 1.8 | idooley2 | src/ck-com/PipeBroadcastStrategy.h |
| 2009/05/28 10:15:17 | 1.11 | idooley2 | src/ck-com/PipeBroadcastStrategy.C |
| 2009/05/28 10:15:17 | 1.2 | idooley2 | src/ck-com/NodeMulticast.h |
| 2009/05/28 10:15:17 | 1.2 | idooley2 | src/ck-com/NodeMulticast.C |
| 2009/05/28 10:15:17 | 1.1 | idooley2 | src/ck-com/MulticastStrategy.h |
| 2009/05/28 10:15:16 | 1.6 | idooley2 | src/ck-com/EachToManyMulticastStrategy.h |
| 2009/05/28 10:15:16 | 1.18 | idooley2 | src/ck-com/EachToManyMulticastStrategy.C |
| 2009/05/28 10:15:16 | 1.9 | idooley2 | src/ck-com/ComlibStrategy.h |
| 2009/05/28 10:15:16 | 1.14 | idooley2 | src/ck-com/ComlibStrategy.C |
| 2009/05/28 10:15:16 | 1.4 | idooley2 | src/ck-com/ComlibStats.h |
| 2009/05/28 10:15:16 | 1.1 | idooley2 | src/ck-com/MulticastStrategy.C |
| 2009/05/28 10:15:16 | 1.5 | idooley2 | src/ck-com/MultiRingMulticast.h |
| 2009/05/28 10:15:16 | 1.8 | idooley2 | src/ck-com/MultiRingMulticast.C |
| 2009/05/28 10:15:16 | 1.10 | idooley2 | src/ck-com/MsgPacker.h |
| 2009/05/28 10:15:16 | 1.7 | idooley2 | src/ck-com/MsgPacker.C |
| 2009/05/28 10:15:16 | 1.2 | idooley2 | src/ck-com/MPIStrategy.h |
| 2009/05/28 10:15:16 | 1.3 | idooley2 | src/ck-com/MPIStrategy.C |
| 2009/05/28 10:15:16 | 1.2 | idooley2 | src/ck-com/KDirectMulticastStrategy.h |
| 2009/05/28 10:15:16 | 1.2 | idooley2 | src/ck-com/KDirectMulticastStrategy.C |
| 2009/05/28 10:15:16 | 1.2 | idooley2 | src/ck-com/DummyStrategy.h |
| 2009/05/28 10:15:16 | 1.3 | idooley2 | src/ck-com/DummyStrategy.C |
| 2009/05/28 10:15:16 | 1.9 | idooley2 | src/ck-com/DirectMulticastStrategy.h |
| 2009/05/28 10:15:16 | 1.15 | idooley2 | src/ck-com/DirectMulticastStrategy.C |
| 2009/05/28 10:15:16 | 1.2 | idooley2 | src/ck-com/ComlibStats.C |
| 2009/05/28 10:15:16 | 1.9 | idooley2 | src/ck-com/ComlibSectionInfo.h |
| 2009/05/28 10:15:15 | 1.13 | idooley2 | src/ck-com/ComlibSectionInfo.C |
| 2009/05/28 10:15:15 | 1.19 | idooley2 | src/ck-com/ComlibManager.h |
| 2009/05/28 10:15:15 | 1.7 | idooley2 | src/ck-com/ComlibManager.ci |
| 2009/05/28 10:15:15 | 1.40 | idooley2 | src/ck-com/ComlibManager.C |
| 2009/05/28 10:15:15 | 1.1 | idooley2 | src/ck-com/ComlibModuleInterface.C |
| 2009/05/28 10:15:15 | 1.4 | idooley2 | src/ck-com/ComlibLearner.h |
| 2009/05/28 10:15:15 | 1.4 | idooley2 | src/ck-com/ComlibArrayListener.h |
| 2009/05/28 10:15:15 | 1.3 | idooley2 | src/ck-com/ComlibArrayListener.C |
| 2009/05/28 10:15:15 | 1.3 | idooley2 | src/ck-com/BroadcastStrategy.h |
| 2009/05/28 10:15:15 | 1.6 | idooley2 | src/ck-com/BroadcastStrategy.C |
| 2009/05/28 10:15:15 | 1.7 | idooley2 | src/ck-com/AAPLearner.h |
| 2009/05/28 10:15:15 | 1.8 | idooley2 | src/ck-com/AAPLearner.C |
| 2009/05/28 10:15:15 | 1.3 | idooley2 | src/ck-com/AAMLearner.h |
| 2009/05/28 10:15:15 | 1.6 | idooley2 | src/ck-com/AAMLearner.C |
| The new version of comlib! This version passed "make test" in charm/tests on order.cs.uiuc.edu. | |||
| 2009/05/28 09:41:31 | 1.155 | idooley2 | src/arch/net/charmrun/charmrun.c |
| Adding a command line option for specifying additional commands that get passed into gdb when running with ++debug. | |||
| 2009/05/28 09:12:41 | 1.14 | idooley2 | src/xlat-i/sdag/CSdagConstruct.C |
| 2009/05/28 09:12:41 | 1.9 | idooley2 | src/xlat-i/sdag/CParsedFile.h |
| 2009/05/28 09:12:40 | 1.22 | idooley2 | src/xlat-i/xi-util.h |
| 2009/05/28 09:12:40 | 1.9 | idooley2 | src/xlat-i/xi-util.C |
| 2009/05/28 09:12:40 | 1.257 | idooley2 | src/xlat-i/xi-symbol.C |
| 2009/05/28 09:12:39 | 1.223 | idooley2 | src/scripts/charmc |
| 2009/05/28 09:12:39 | 1.487 | idooley2 | src/scripts/Makefile |
| 2009/05/28 09:12:38 | 2.129 | idooley2 | src/ck-perf/trace-projections.C |
| 2009/05/28 09:12:38 | 2.75 | idooley2 | src/ck-perf/trace-projections.h |
| 2009/05/28 09:12:38 | 1.88 | idooley2 | src/scripts/Make.depends |
| 2009/05/28 09:12:37 | 2.19 | idooley2 | src/ck-core/msgalloc.C |
| 2009/05/28 09:12:37 | 1.2 | idooley2 | src/ck-cp/pathHistory.h |
| 2009/05/28 09:12:37 | 1.6 | idooley2 | src/ck-cp/controlPoints.h |
| 2009/05/28 09:12:37 | 1.19 | idooley2 | src/ck-cp/controlPoints.C |
| 2009/05/28 09:12:36 | 2.75 | idooley2 | src/ck-core/ckreduction.C |
| 2009/05/28 09:12:36 | 1.182 | idooley2 | src/ck-core/ckarray.C |
| 2009/05/28 09:12:36 | 2.55 | idooley2 | src/ck-core/envelope.h |
| 2009/05/28 09:12:36 | 2.173 | idooley2 | src/ck-core/ck.C |
| Implementing a two-phase critical path detection scheme. All the critical path detection code is disabled by a preprocessor macro by default. The new scheme is supported in SDAG and Charisma now. | |||
| 2009/05/28 08:19:34 | 1.1 | idooley2 | src/ck-cp/pathHistory.h |
| 2009/05/28 08:19:34 | 1.1 | idooley2 | src/ck-cp/pathHistory.C |
| 2009/05/28 08:19:34 | 1.5 | idooley2 | src/ck-cp/controlPoints.h |
| 2009/05/28 08:19:34 | 1.4 | idooley2 | src/ck-cp/controlPoints.ci |
| 2009/05/28 08:19:34 | 1.18 | idooley2 | src/ck-cp/controlPoints.C |
| 2009/05/28 08:19:34 | 1.1 | idooley2 | src/ck-cp/arrayRedistributor.h |
| 2009/05/28 08:19:34 | 1.1 | idooley2 | src/ck-cp/arrayRedistributor.C |
| Refactoring control point and critical path detection code. | |||
| 2009/05/28 07:09:29 | 1.14 | kunzman | src/util/simd.h |
| Updated some of the binary operations to correct errors. | |||
| 2009/05/27 11:17:54 | 1.13 | kunzman | src/util/simd.h |
| Fixed some bugs with the AltiVec version of the code. | |||
| 2009/05/27 10:08:50 | 1.12 | kunzman | src/util/simd.h |
| Fixed not operations. | |||
| 2009/05/27 09:44:57 | 2.81 | gzheng | src/ck-perf/trace-common.C |
| revert Phil's previous check in since it crashes on solaris and windows. | |||
| 2009/05/27 06:43:29 | 1.7 | insomnia52 | examples/charm++/Molecular3D/src_charm/Patch.C |
| 2009/05/27 06:43:28 | 1.6 | insomnia52 | examples/charm++/Molecular3D/compile/Makefile.config |
| made use of CkMulticast properly and made it update the section after every load balance event | |||
| 2009/05/26 12:29:03 | 1.3 | insomnia52 | examples/charm++/Molecular3D/testInsert/Makefile |
| 2009/05/26 12:29:03 | 1.4 | insomnia52 | examples/charm++/Molecular3D/src_charm/mol3d.ci |
| 2009/05/26 12:29:03 | 1.5 | insomnia52 | examples/charm++/Molecular3D/src_charm/Patch.h |
| 2009/05/26 12:29:03 | 1.6 | insomnia52 | examples/charm++/Molecular3D/src_charm/Patch.C |
| 2009/05/26 12:29:03 | 1.7 | insomnia52 | examples/charm++/Molecular3D/src_charm/Main.C |
| 2009/05/26 12:29:03 | 1.5 | insomnia52 | examples/charm++/Molecular3D/compile/Makefile.config |
| hopefully fixed bug and returned to normal insert patch | |||
| 2009/05/26 12:05:27 | 1.2 | insomnia52 | examples/charm++/Molecular3D/testInsert/testInsert.C |
| 2009/05/26 12:05:26 | 1.2 | insomnia52 | examples/charm++/Molecular3D/testInsert/Makefile |
| added doneinserting | |||
| 2009/05/26 11:53:31 | 1.1 | insomnia52 | examples/charm++/Molecular3D/testInsert/testInsert.h |
| 2009/05/26 11:53:31 | 1.1 | insomnia52 | examples/charm++/Molecular3D/testInsert/testInsert.ci |
| 2009/05/26 11:53:31 | 1.1 | insomnia52 | examples/charm++/Molecular3D/testInsert/testInsert.C |
| 2009/05/26 11:53:31 | 1.1 | insomnia52 | examples/charm++/Molecular3D/testInsert/Makefile |
| 2009/05/26 11:53:31 | 1.6 | insomnia52 | examples/charm++/Molecular3D/src_charm/Main.C |
| 2009/05/26 11:53:31 | 1.4 | insomnia52 | examples/charm++/Molecular3D/compile/Makefile.config |
| 2009/05/26 11:53:31 | 1.3 | insomnia52 | examples/charm++/Molecular3D/compile/Makefile |
| added test | |||
| 2009/05/26 10:37:10 | 1.2 | abecker | src/langs/charj/tests/unit/TODO |
| updated TODO | |||
| 2009/05/26 10:37:08 | 1.2 | abecker | src/langs/charj/tests/unit/ReservedWords.cj |
| 2009/05/26 10:37:07 | 1.2 | abecker | src/langs/charj/tests/unit/Conditional.cj |
| added some more test cases | |||
| 2009/05/26 10:37:05 | 1.9 | abecker | src/langs/charj/src/charj/Main.java |
| 2009/05/26 10:37:02 | 1.8 | abecker | src/langs/charj/bin/charjc |
| fixed arg name in main, simplified option handling | |||
| 2009/05/26 10:36:59 | 1.1 | abecker | src/langs/charj/tests/unit/ReservedWords.cj |
| 2009/05/26 10:36:59 | 1.2 | abecker | src/langs/charj/tests/unit/Empty.cj |
| tweaked template test, added test for use of reserved words | |||
| 2009/05/26 10:36:53 | 1.2 | abecker | src/langs/charj/.gitignore |
| improved vcs ignoring | |||
| 2009/05/26 10:36:51 | 1.12 | abecker | src/langs/charj/src/charj/translator/CharjSemantics.g |
| 2009/05/26 10:36:51 | 1.12 | abecker | src/langs/charj/src/charj/translator/Charj.g |
| fixed up some very long lines | |||
| 2009/05/26 10:36:48 | 1.1 | abecker | src/langs/charj/tests/unit/TODO |
| 2009/05/26 10:36:47 | 1.1 | abecker | src/langs/charj/tests/unit/FuncArgs.cj |
| new tests, todo list | |||
| 2009/05/26 10:36:42 | 1.1 | abecker | src/langs/charj/tests/unit/Makefile |
| simple makefile for unit tests | |||
| 2009/05/26 10:36:39 | 1.1 | abecker | src/langs/charj/.gitignore |
| added .gitignore | |||
| 2009/05/26 10:36:36 | 1.1 | abecker | src/langs/charj/tests/unit/WhileLoop.cj |
| 2009/05/26 10:36:36 | 1.1 | abecker | src/langs/charj/tests/unit/Switch.cj |
| 2009/05/26 10:36:36 | 1.1 | abecker | src/langs/charj/tests/unit/SimpleClass.cj |
| 2009/05/26 10:36:36 | 1.1 | abecker | src/langs/charj/tests/unit/PrimitiveTypes.cj |
| 2009/05/26 10:36:36 | 1.1 | abecker | src/langs/charj/tests/unit/ForLoop.cj |
| 2009/05/26 10:36:36 | 1.1 | abecker | src/langs/charj/tests/unit/Empty.cj |
| 2009/05/26 10:36:36 | 1.1 | abecker | src/langs/charj/tests/unit/Declarations.cj |
| 2009/05/26 10:36:36 | 1.1 | abecker | src/langs/charj/tests/unit/Conditional.cj |
| 2009/05/26 10:36:35 | 1.1 | abecker | src/langs/charj/tests/unit/Assert.cj |
| added unit test directory | |||
| 2009/05/26 09:59:21 | 1.3 | insomnia52 | examples/charm++/Molecular3D/src_charm/mol3d.ci |
| 2009/05/26 09:59:21 | 1.4 | insomnia52 | examples/charm++/Molecular3D/src_charm/Patch.h |
| 2009/05/26 09:59:21 | 1.5 | insomnia52 | examples/charm++/Molecular3D/src_charm/Patch.C |
| 2009/05/26 09:59:21 | 1.5 | insomnia52 | examples/charm++/Molecular3D/src_charm/Main.C |
| 2009/05/26 09:59:21 | 1.3 | insomnia52 | examples/charm++/Molecular3D/compile/Makefile.config |
| modified patch placement scheme for testing | |||
| 2009/05/26 06:50:18 | 2.363 | gzheng | src/arch/net/machine.c |
| allow net version to work with ccs on PE0 | |||
| 2009/05/26 06:46:31 | 1.9 | gzheng | src/arch/multicore-linux64/conv-mach.h |
| set NODE_0_IS_CONVHOST to allow ccs working through PE0 | |||
| 2009/05/26 06:27:25 | 1.17 | gzheng | src/langs/bluegene/bigsim_api.C |
| change back so the BgStartLogByName when creating a dummy event does not automatically assign objID. It is done selectively in trace-bluegene now | |||
| 2009/05/26 06:03:28 | 1.13 | gzheng | src/xlat-i/sdag/CSdagConstruct.C |
| mark creating a sdag event as splitting an existing event | |||
| 2009/05/26 06:02:38 | 2.29 | gzheng | src/ck-perf/trace-bluegene.h |
| 2009/05/26 06:02:38 | 2.33 | gzheng | src/ck-perf/trace-bluegene.C |
| added a parameter in creating a dummy event, which indicate if the new event is just a split of an existing event. | |||
| 2009/05/25 08:51:09 | 1.17 | phil | src/ck-ldb/LBSimulation.C |
| 2009/05/25 08:51:09 | 1.88 | phil | src/ck-ldb/CentralLB.h |
| 2009/05/25 08:51:08 | 2.74 | phil | src/ck-core/ckreduction.C |
| 2009/05/25 08:51:08 | 2.123 | phil | src/ck-core/cklocation.C |
| Quiet a bunch more warnings | |||
| 2009/05/25 08:40:35 | 1.4 | insomnia52 | examples/charm++/Molecular3D/src_charm/Patch.C |
| 2009/05/25 08:40:35 | 1.4 | insomnia52 | examples/charm++/Molecular3D/src_charm/Main.C |
| fixed patch creation to not use a ridiculously large and unnecessary amount of memory | |||
| 2009/05/25 07:19:49 | 1.23 | jetley | examples/charm++/barnes-charm/barnes.h |
| 2009/05/25 07:19:49 | 1.20 | jetley | examples/charm++/barnes-charm/TreePiece.cpp |
| 2009/05/25 07:19:49 | 1.22 | jetley | examples/charm++/barnes-charm/ParticleChunk.cpp |
| 2009/05/25 07:19:49 | 1.26 | jetley | examples/charm++/barnes-charm/Main.cpp |
| removed comments | |||
| 2009/05/25 06:53:52 | 1.3 | insomnia52 | examples/charm++/Molecular3D/src_namd/copy_namd.sh |
| 2009/05/25 06:53:52 | 1.1 | insomnia52 | examples/charm++/Molecular3D/src_namd/Parameters.C |
| added this namd2 file in so it would not have to by modified every time | |||
| 2009/05/25 06:51:29 | 2.172 | phil | src/ck-core/ck.C |
|
Remove unused variables found by -Wall
They were even declared 'register', despite the non-criticality of this code for performance. | |||
| 2009/05/25 06:41:59 | 2.80 | phil | src/ck-perf/trace-common.C |
| Correct format string issues identified by -Wall | |||
| 2009/05/23 15:26:33 | 1.8 | jetley | examples/charm++/barnes-charm/defs.h |
| 2009/05/23 15:26:33 | 1.3 | jetley | examples/charm++/barnes-charm/code_io.C |
| 2009/05/23 15:26:33 | 1.22 | jetley | examples/charm++/barnes-charm/barnes.h |
| 2009/05/23 15:26:33 | 1.20 | jetley | examples/charm++/barnes-charm/barnes.ci |
| 2009/05/23 15:26:33 | 1.21 | jetley | examples/charm++/barnes-charm/ParticleChunk.cpp |
| 2009/05/23 15:26:33 | 1.16 | jetley | examples/charm++/barnes-charm/Makefile |
| 2009/05/23 15:26:33 | 1.25 | jetley | examples/charm++/barnes-charm/Main.cpp |
| log | |||
| 2009/05/22 13:44:42 | 1.4 | phil | src/libs/ck-libs/ampi/ampiProjections.h |
| Fix variable names in header to match source, calls | |||
| 2009/05/22 11:30:33 | 1.3 | insomnia52 | examples/charm++/Molecular3D/src_charm/Patch.h |
| 2009/05/22 11:30:33 | 1.3 | insomnia52 | examples/charm++/Molecular3D/src_charm/Patch.C |
| 2009/05/22 11:30:33 | 1.3 | insomnia52 | examples/charm++/Molecular3D/src_charm/Compute.C |
| simple load balancer handlers | |||
| 2009/05/21 17:37:59 | 1.154 | gzheng | src/arch/net/charmrun/charmrun.c |
| a new command line option "++no-xrdb" that tells charmrun not to find xrdb. This is to make ++debug works on turing compute nodes. | |||
| 2009/05/21 10:37:05 | 1.16 | gzheng | src/langs/bluegene/bigsim_api.C |
| assign objID from parent log | |||
| 2009/05/21 10:36:08 | 2.28 | gzheng | src/ck-perf/trace-bluegene.h |
| 2009/05/21 10:36:08 | 2.32 | gzheng | src/ck-perf/trace-bluegene.C |
| changes to propagate objID among SDAG depedence events. | |||
| 2009/05/21 10:31:27 | 1.5 | kunzman | examples/charm++/cell/jacobi/main.h |
| 2009/05/21 10:31:27 | 1.9 | kunzman | examples/charm++/cell/jacobi/main.C |
| 2009/05/21 10:31:27 | 1.8 | kunzman | examples/charm++/cell/jacobi/jacobi.C |
| 2009/05/21 10:31:25 | 1.7 | kunzman | examples/charm++/cell/jacobi/jacobi.h |
| Updated malloc calls to use Charm++ aligned malloc instead of malloc_aligned. | |||
| 2009/05/21 03:49:59 | 1.53 | gzheng | src/langs/bluegene/bigsim_logs.h |
| make "group" in BgMsgEntry public again | |||
| 2009/05/20 15:32:31 | 1.9 | idooley2 | src/ck-perf/trace-summaryBOC.h |
| 2009/05/20 15:32:31 | 1.110 | idooley2 | src/ck-perf/trace-summary.C |
| Making sum detail tracing only get intialized once. | |||
| 2009/05/20 10:58:21 | 1.52 | gzheng | src/langs/bluegene/bigsim_logs.h |
| 2009/05/20 10:58:21 | 1.58 | gzheng | src/langs/bluegene/bigsim_logs.C |
| 2009/05/20 10:58:21 | 1.15 | gzheng | src/langs/bluegene/bigsim_api.C |
| get rid of compiler warnings for compiling SDAG code for bigsim. | |||
| 2009/05/20 07:52:29 | 1.28 | bhatele | src/util/TopoManager.h |
| sync of charm and standalone versions | |||
| 2009/05/19 18:33:33 | 1.6 | idooley2 | tools/projections/projections/streaming/StreamingTool.java |
| 2009/05/19 18:33:33 | 1.5 | idooley2 | tools/projections/projections/streaming/StartupDialogBox.java |
| 2009/05/19 18:33:33 | 1.3 | idooley2 | tools/projections/projections/streaming/SingleSeriesHandler.java |
| 2009/05/19 18:33:33 | 1.6 | idooley2 | tools/projections/projections/streaming/MultiSeriesHandler.java |
| 2009/05/19 18:33:33 | 1.4 | idooley2 | tools/projections/projections/streaming/ByteParser.java |
| Added replay feature to sum detail CCS streaming tool. All incoming CCS messages can be dumped into a file, and a simple driver can replay them if needed. | |||
| 2009/05/19 17:33:08 | 1.5 | idooley2 | tools/projections/projections/streaming/StreamingTool.java |
| 2009/05/19 17:33:08 | 1.4 | idooley2 | tools/projections/projections/streaming/StartupDialogBox.java |
| 2009/05/19 17:33:08 | 1.5 | idooley2 | tools/projections/projections/streaming/MultiSeriesHandler.java |
| 2009/05/19 17:33:08 | 1.3 | idooley2 | tools/projections/projections/streaming/ByteParser.java |
| Changes to sum detail CCS streaming client. | |||
| 2009/05/19 17:21:47 | 1.351 | idooley2 | tools/projections/bin/projections.jar |
| More changes to CCS sum detail streaming tool. | |||
| 2009/05/19 09:44:08 | 1.1 | bhatele | src/arch/mpi-linux-mips64/conv-mach.sh |
| 2009/05/19 09:44:08 | 1.1 | bhatele | src/arch/mpi-linux-mips64/conv-mach.h |
| 2009/05/19 09:44:08 | 1.1 | bhatele | src/arch/mpi-linux-mips64/conv-mach-smp.sh |
| 2009/05/19 09:44:08 | 1.1 | bhatele | src/arch/mpi-linux-mips64/conv-mach-smp.h |
| new Charm++ port for SiCortex and other mips64 machines | |||
| 2009/05/19 09:35:19 | 1.20 | jetley | examples/charm++/barnes-charm/ParticleChunk.cpp |
| 2009/05/19 09:35:19 | 1.24 | jetley | examples/charm++/barnes-charm/Main.cpp |
| minor changes | |||
| 2009/05/18 12:36:46 | 1.4 | idooley2 | tools/projections/projections/streaming/StreamingTool.java |
| 2009/05/18 12:36:46 | 1.4 | idooley2 | tools/projections/projections/streaming/MultiSeriesHandler.java |
| 2009/05/18 12:36:45 | 1.2 | idooley2 | tools/projections/projections/streaming/ByteParser.java |
| Fixing bugs, adding features to the sum detail CCS streaming tool. | |||
| 2009/05/18 12:36:24 | 1.350 | idooley2 | tools/projections/bin/projections.jar |
| Fixing bugs, adding features to the sum detail CCS streaming tool. | |||
| 2009/05/18 12:32:21 | 1.109 | idooley2 | src/ck-perf/trace-summary.C |
| Fixing bugs in the Sum Detail CCS streaming. | |||
| 2009/05/18 07:41:02 | 1.15 | jetley | examples/charm++/barnes-charm/Makefile |
| proj | |||
| 2009/05/18 07:39:50 | 1.14 | jetley | examples/charm++/barnes-charm/Makefile |
| modified makefile so that CHARM_DIR can be changed | |||
| 2009/05/18 07:30:27 | 1.7 | jetley | examples/charm++/barnes-charm/defs.h |
| 2009/05/18 07:30:27 | 1.21 | jetley | examples/charm++/barnes-charm/barnes.h |
| 2009/05/18 07:30:27 | 1.19 | jetley | examples/charm++/barnes-charm/barnes.ci |
| 2009/05/18 07:30:27 | 1.19 | jetley | examples/charm++/barnes-charm/TreePiece.cpp |
| 2009/05/18 07:30:27 | 1.19 | jetley | examples/charm++/barnes-charm/ParticleChunk.cpp |
| 2009/05/18 07:30:27 | 1.13 | jetley | examples/charm++/barnes-charm/Makefile |
| 2009/05/18 07:30:27 | 1.23 | jetley | examples/charm++/barnes-charm/Main.cpp |
| added code for load balancing. | |||
| 2009/05/18 07:05:31 | 1.5 | gzheng | src/arch/vmi-linux-x86_64/conv-mach.sh |
| updated vmi path | |||
| 2009/05/15 12:41:46 | 1.11 | bhatele | examples/bigsim/charm/jacobi2d/jacobi2d.C |
| *** empty log message *** | |||
| 2009/05/15 08:35:29 | 1.108 | idooley2 | src/ck-perf/trace-summary.C |
| Removing bzero call which broke autobuild. Bzero is not portable. | |||
| 2009/05/15 07:36:12 | 1.3 | idooley2 | tools/projections/projections/streaming/StartupDialogBox.java |
| 2009/05/15 07:36:12 | 1.2 | idooley2 | tools/projections/projections/streaming/SingleSeriesHandler.java |
| 2009/05/15 07:36:12 | 1.3 | idooley2 | tools/projections/projections/streaming/MultiSeriesHandler.java |
| 2009/05/15 07:36:12 | 1.1 | idooley2 | tools/projections/projections/streaming/ByteParser.java |
| This version of the streaming tool better handles the more compressed types that can be provided from trace-summary.C. | |||
| 2009/05/15 07:32:37 | 1.107 | idooley2 | src/ck-perf/trace-summary.C |
| Now the summary detail gathering mechanism supports multiple types for the fields (to adjust level compression). | |||
| 2009/05/14 18:41:13 | 1.10 | bhatele | examples/bigsim/charm/jacobi2d/jacobi2d.C |
| *** empty log message *** | |||
| 2009/05/14 14:55:00 | 1.55 | bohm | src/libs/ck-libs/pose/etrans.pl |
|
Add DEBUG_POSE_INVOKE as a compile time ifdef which will output verbose
parameter information for each POSE invoke. Useful for figuring out when your invoke results in an mismatch between method message and object. | |||
| 2009/05/14 14:02:41 | 1.9 | bhatele | examples/bigsim/charm/jacobi2d/jacobi2d.C |
| bug fixes | |||
| 2009/05/14 10:45:38 | 1.61 | idooley2 | src/ck-perf/trace-summary.h |
| 2009/05/14 10:45:38 | 1.11 | idooley2 | src/ck-perf/trace-summary.ci |
| 2009/05/14 10:45:38 | 1.106 | idooley2 | src/ck-perf/trace-summary.C |
| Improving the CCS sum detail streaming reporting tool. | |||
| 2009/05/13 19:45:31 | 1.3 | bhatele | examples/bigsim/charm/jacobi2d/jacobi2d.ci |
| 2009/05/13 19:45:31 | 1.8 | bhatele | examples/bigsim/charm/jacobi2d/jacobi2d.C |
| implemented local barriers to get around bigsim problem | |||
| 2009/05/13 16:51:54 | 1.7 | bhatele | examples/bigsim/charm/jacobi2d/jacobi2d.C |
| proper BgPrintfs | |||
| 2009/05/13 11:07:07 | 1.24 | bhatele | src/conv-core/cputopology.C |
| this wouldn't compile with gnu compilers on XT4 | |||
| 2009/05/13 11:05:59 | 1.4 | bhatele | src/arch/bluegenep/cc-xlc.sh |
| this was needed for AMPI to compile correctly by using Charm's mpi.h | |||
| 2009/05/12 08:58:04 | 1.142 | gzheng | build |
| hide cuda as build option | |||
| 2009/05/12 06:38:17 | 2.337 | gioachin | src/conv-core/convcore.c |
| wrong comment style | |||
| 2009/05/11 18:15:46 | 2.80 | gioachin | src/ck-core/cklocation.h |
| 2009/05/11 18:15:46 | 2.122 | gioachin | src/ck-core/cklocation.C |
| 2009/05/11 18:15:46 | 1.139 | gioachin | src/ck-core/ckarray.h |
| 2009/05/11 18:15:46 | 1.181 | gioachin | src/ck-core/ckarray.C |
| 2009/05/11 18:15:46 | 2.162 | gioachin | src/ck-core/charm++.h |
| 2009/05/11 18:15:45 | 2.12 | gioachin | src/ck-core/charisma.h |
|
Split nInts of CkArrayIndex into two shorts: the first one still named nInts, the second one containing the dimensionality of the index. Valid only for 1-6D array indices.
Deleted some duplicated code (idx2str) | |||
| 2009/05/11 18:00:12 | 1.24 | gioachin | examples/ampi/Cjacobi3D/Makefile |
| fixed problem with parallel make | |||
| 2009/05/11 17:11:54 | 1.105 | idooley2 | src/ck-perf/trace-summary.C |
| Switching reduction of sum detail data to floating point utilization values and integer sizes/EPs. A single char isn't big enough for these reductions. | |||
| 2009/05/11 13:40:58 | 2.12 | gioachin | src/conv-ccs/ccs-server.c |
| Fixed define for VMI layers | |||
| 2009/05/11 10:01:32 | 1.43 | gzheng | src/arch/mpi/charmrun |
| find aprun first, using mpirun does not work on kraken | |||
| 2009/05/11 09:11:43 | 1.2 | idooley2 | tools/projections/projections/streaming/StartupDialogBox.java |
| 2009/05/11 09:11:43 | 1.2 | idooley2 | tools/projections/projections/streaming/MultiSeriesHandler.java |
| 2009/05/11 09:11:42 | 1.28 | idooley2 | tools/projections/projections/gui/Timeline/UserEventObject.java |
| Adding support for CCS streaming sum detail information. | |||
| 2009/05/11 09:11:21 | 1.349 | idooley2 | tools/projections/bin/projections.jar |
| Adding support for CCS streaming sum detail information. | |||
| 2009/05/11 09:10:23 | 1.8 | idooley2 | src/ck-perf/trace-summaryBOC.h |
| 2009/05/11 09:10:23 | 1.60 | idooley2 | src/ck-perf/trace-summary.h |
| 2009/05/11 09:10:23 | 1.10 | idooley2 | src/ck-perf/trace-summary.ci |
| 2009/05/11 09:10:23 | 1.104 | idooley2 | src/ck-perf/trace-summary.C |
| New functionality in CCS streaming of sum detail data for projections. The reduction combiner needs to be fixed, and then the sum detail information will be correct. | |||
| 2009/05/10 13:50:01 | 1.7 | idooley2 | src/ck-perf/trace-summaryBOC.h |
| 2009/05/10 13:50:01 | 1.59 | idooley2 | src/ck-perf/trace-summary.h |
| 2009/05/10 13:50:01 | 1.9 | idooley2 | src/ck-perf/trace-summary.ci |
| 2009/05/10 13:50:01 | 1.103 | idooley2 | src/ck-perf/trace-summary.C |
| Adding CCS handlers that export trace information to projections as the program runs. Currently simple summary utilization over time is exported, and the beginnings of some handlers that export sum-detail per EP utilization. | |||
| 2009/05/10 13:43:39 | 1.3 | idooley2 | tools/projections/projections/streaming/StreamingTool.java |
| 2009/05/10 13:43:39 | 1.1 | idooley2 | tools/projections/projections/streaming/StartupDialogBox.java |
| 2009/05/10 13:43:38 | 1.2 | idooley2 | tools/projections/projections/ccs/CcsThread.java |
| 2009/05/10 13:43:38 | 1.1 | idooley2 | tools/projections/projections/streaming/SingleSeriesHandler.java |
| 2009/05/10 13:43:38 | 1.1 | idooley2 | tools/projections/projections/streaming/MultiSeriesHandler.java |
| 2009/05/10 13:43:38 | 1.32 | idooley2 | tools/projections/projections/gui/MainMenuManager.java |
| Adding some CCS live streaming data views for projections. These gather data from teh tracemode summary module in charm++. | |||
| 2009/05/10 13:43:15 | 1.348 | idooley2 | tools/projections/bin/projections.jar |
| Adding some CCS live streaming data views for projections. These gather data from teh tracemode summary module in charm++. | |||
| 2009/05/08 20:50:35 | 1.14 | gzheng | src/langs/bluegene/bigsim_api.C |
| added a comment about seqno in BgWriteTimelines | |||
| 2009/05/08 20:44:23 | 1.51 | gzheng | src/langs/bluegene/bigsim_logs.h |
| 2009/05/08 20:44:23 | 1.13 | gzheng | src/langs/bluegene/bigsim_api.C |
| 2009/05/08 20:44:23 | 1.17 | gzheng | src/langs/bluegene/bigsim_read.C |
| minor changes in API writing bgtrace summary file, and fixed couple memory leaks | |||
| 2009/05/08 17:51:09 | 1.153 | gioachin | src/arch/net/charmrun/charmrun.c |
| 2009/05/08 17:51:09 | 2.11 | gioachin | src/conv-ccs/ccs-server.c |
| bug fix for net layer | |||
| 2009/05/08 13:23:44 | 1.2 | idooley2 | tools/projections/projections/streaming/StreamingTool.java |
| Tool now retrieves data. | |||
| 2009/05/08 13:04:49 | 1.8 | chaomei | tests/charm++/kNeighbor/kNeighbor.C |
| Added an implementation for all cores only having inter-node communication (#msgs is roughly O(p^2)) | |||
| 2009/05/08 11:47:12 | 2.44 | gioachin | src/conv-ccs/conv-ccs.c |
| 2009/05/08 11:47:12 | 2.10 | gioachin | src/conv-ccs/ccs-server.c |
| 2009/05/08 11:47:11 | 1.152 | gioachin | src/arch/net/charmrun/charmrun.c |
| 2009/05/08 11:47:11 | 2.336 | gioachin | src/conv-core/convcore.c |
| implemented forwarding of output streams for all machine layers that do not reimplement CMK_PRINTF (includes MPI). Given this, moved some common code from charmrun.c to ccs-server.c | |||
| 2009/05/08 11:44:48 | 2.161 | gioachin | src/ck-core/charm++.h |
| fixed bug on CkConditional using erroneous destructor | |||
| 2009/05/08 11:38:26 | 1.256 | gioachin | src/xlat-i/xi-symbol.C |
| fixed bug on CkConditional using erroneous destructor | |||
| 2009/05/07 13:57:59 | 1.6 | jetley | examples/charm++/barnes-charm/defs.h |
| 2009/05/07 13:57:59 | 1.2 | jetley | examples/charm++/barnes-charm/code_io.C |
| 2009/05/07 13:57:59 | 1.20 | jetley | examples/charm++/barnes-charm/barnes.h |
| 2009/05/07 13:57:59 | 1.18 | jetley | examples/charm++/barnes-charm/barnes.ci |
| 2009/05/07 13:57:59 | 1.18 | jetley | examples/charm++/barnes-charm/TreePiece.cpp |
| 2009/05/07 13:57:59 | 1.18 | jetley | examples/charm++/barnes-charm/ParticleChunk.cpp |
| 2009/05/07 13:57:58 | 1.12 | jetley | examples/charm++/barnes-charm/Makefile |
| 2009/05/07 13:57:58 | 1.22 | jetley | examples/charm++/barnes-charm/Main.cpp |
|
code now relies on QD to start tree building. it also allocates cells and leaves
from contiguous blocks within treepieces, instead of calling new each time makecell/leaf is invoked. bugs still remain. look up TMS logs. | |||
| 2009/05/07 10:29:55 | 1.5 | gzheng | src/arch/common/cc-xlc.sh |
| make sure xlc generate 32 bit code by using -q32 | |||
| 2009/05/06 20:56:39 | 1.19 | jetley | examples/charm++/barnes-charm/barnes.h |
| 2009/05/06 20:56:39 | 1.17 | jetley | examples/charm++/barnes-charm/TreePiece.cpp |
| 2009/05/06 20:56:39 | 1.17 | jetley | examples/charm++/barnes-charm/ParticleChunk.cpp |
|
modifications.
Code mostly revised. * Have to insert QD code and callback to doBuildTree(). * Also, pendingChildren isn't correctly set. Currently, it is incremented in sendParticlesToChildren(), which can be called multiple times for a single child (first time when created, and next time when it is sent a message). * Otherwise fine: if node is too fat and don't have parent, buffer messages till haveParent is set from recvRootFromParent. then, because wantToSplit, myRoot is created, all self and buffered particles are processed and sent to children. if node is too fat and have parent, we directly do the latter steps. | |||
| 2009/05/06 17:18:56 | 1.18 | jetley | examples/charm++/barnes-charm/barnes.h |
| 2009/05/06 17:18:56 | 1.17 | jetley | examples/charm++/barnes-charm/barnes.ci |
| 2009/05/06 17:18:56 | 1.16 | jetley | examples/charm++/barnes-charm/TreePiece.cpp |
| 2009/05/06 17:18:56 | 1.16 | jetley | examples/charm++/barnes-charm/ParticleChunk.cpp |
| 2009/05/06 17:18:55 | 1.11 | jetley | examples/charm++/barnes-charm/Makefile |
| 2009/05/06 17:18:55 | 1.21 | jetley | examples/charm++/barnes-charm/Main.cpp |
|
code to continue simulation for multiple iterations. treepieces are reused
and must be given roots by their parents. treepieces can only split their roots when they have received their parents' roots. the splitting and treebuilding must be delayed until they do. | |||
| 2009/05/06 10:48:17 | 1.6 | bhatele | examples/bigsim/charm/jacobi2d/jacobi2d.C |
| bgprintf added | |||
| 2009/05/06 06:11:48 | 1.7 | gioachin | src/libs/ck-libs/pythonCCS/charmdebug-python.C |
| 2009/05/06 06:11:48 | 1.20 | gioachin | src/libs/ck-libs/pythonCCS/PythonCCS.h |
| 2009/05/06 06:11:48 | 1.25 | gioachin | src/libs/ck-libs/pythonCCS/PythonCCS.C |
| Moved freezing condition for CharmDebug into the charmdebug specific CpdPython group | |||
| 2009/05/05 14:26:58 | 1.2 | abecker | src/libs/ck-libs/ParFUM-Iterators/ParFUM_Iterators_CUDA.h |
| renamed cuda iterators | |||
| 2009/05/05 13:24:21 | 1.42 | bhatele | src/langs/bluegene/blue_impl.h |
|
changed the bigsim mapping of processors to coordinates from ZYX to XYZ
this now matches BigNetSim's mapping | |||
| 2009/05/05 12:40:57 | 1.7 | chaomei | tests/charm++/kNeighbor/kNeighbor.C |
| Added node cyclic mapping for elements to test how well comm thread behaves when inter-node comm happens more frequently. | |||
| 2009/05/05 12:24:37 | 1.1 | abecker | src/libs/ck-libs/ParFUM-Iterators/libmoduleParFUM_Iterators.dep |
| added module dep file | |||
| 2009/05/05 12:20:46 | 1.1 | abecker | src/libs/ck-libs/ParFUM-Iterators/ParFUM_Iterators_Types.h |
| 2009/05/05 12:20:46 | 1.1 | abecker | src/libs/ck-libs/ParFUM-Iterators/ParFUM_Iterators_CUDA.h |
| 2009/05/05 12:20:46 | 1.1 | abecker | src/libs/ck-libs/ParFUM-Iterators/ParFUM_Iterators_CUDA.cu |
| 2009/05/05 12:20:46 | 1.1 | abecker | src/libs/ck-libs/ParFUM-Iterators/ParFUM_Iterators_CPU.cc |
| 2009/05/05 12:20:46 | 1.1 | abecker | src/libs/ck-libs/ParFUM-Iterators/ParFUM_Iterators.h |
| 2009/05/05 12:20:46 | 1.1 | abecker | src/libs/ck-libs/ParFUM-Iterators/ParFUM_Iterators.ci |
| 2009/05/05 12:20:46 | 1.1 | abecker | src/libs/ck-libs/ParFUM-Iterators/ParFUM_Iterators.cc |
| 2009/05/05 12:20:46 | 1.1 | abecker | src/libs/ck-libs/ParFUM-Iterators/ParFUM-Iterators.doxy |
| 2009/05/05 12:20:45 | 1.1 | abecker | src/libs/ck-libs/ParFUM-Iterators/Makefile |
| Added ParFUM iterators layer | |||
| 2009/05/05 09:07:02 | 1.21 | abecker | src/libs/ck-libs/ParFUM-Tops/Makefile |
| default optimization level changed | |||
| 2009/05/05 09:06:34 | 1.32 | abecker | src/libs/ck-libs/ParFUM-Tops/ParFUM_TOPS.cc |
| added success assertions | |||
| 2009/05/05 08:53:44 | 1.20 | idooley2 | src/libs/ck-libs/ParFUM-Tops/Makefile |
| Minor changes to makefile. | |||
| 2009/05/05 07:45:41 | 1.23 | bhatele | src/conv-core/cputopology.C |
| compilation error fixed | |||
| 2009/05/04 21:27:16 | 2.73 | gzheng | src/ck-core/ckreduction.C |
| 2009/05/04 21:27:16 | 2.160 | gzheng | src/ck-core/charm++.h |
| when reduction client is CkReductionClientBundle, pup storedCallback is not enough, because the c function param point to CkReductionClientBundle, which needs to be separately pup'ed. | |||
| 2009/05/04 21:10:07 | 1.171 | gzheng | src/ck-ldb/CentralLB.C |
| fixed a typo | |||
| 2009/05/04 21:03:50 | 1.151 | gzheng | src/arch/net/charmrun/charmrun.c |
| disable scalable-start's fork setting when restart a crashed node. | |||
| 2009/05/04 20:56:41 | 1.170 | gzheng | src/ck-ldb/CentralLB.C |
| fixed USE_REDUCTION macro redefinition when using FT | |||
| 2009/05/04 20:17:55 | 1.3 | bhatele | examples/bigsim/charm/jacobi2d/Makefile |
| *** empty log message *** | |||
| 2009/05/04 16:47:32 | 1.22 | bhatele | src/conv-core/cputopology.C |
|
the unscalable way of collecting topology will not be used on BG and XT machines
anymore. I also changed nodenum to nodeIDs | |||
| 2009/05/02 03:18:55 | 1.12 | gzheng | src/langs/bluegene/bigsim_api.C |
| 2009/05/02 03:18:55 | 1.55 | gzheng | src/langs/bluegene/bigsim_timing.C |
| moved new function BgSplitEntry out of bigsim API lib because otherwise it can not compile a standalone bigsim program due to calls to timers | |||
| 2009/05/01 20:03:34 | 1.180 | gzheng | src/ck-core/ckarray.C |
| a new way to split broadcast, instead of a chain, all broadcast delivery has a backward dep to the head event | |||
| 2009/05/01 20:02:34 | 1.50 | gzheng | src/langs/bluegene/bigsim_logs.h |
| 2009/05/01 20:02:34 | 1.57 | gzheng | src/langs/bluegene/bigsim_logs.C |
| 2009/05/01 20:02:34 | 1.11 | gzheng | src/langs/bluegene/bigsim_api.C |
| 2009/05/01 20:02:34 | 1.58 | gzheng | src/langs/bluegene/blue.h |
| a new split log bigsim API calls that takes an array of bglogs for backward deps | |||
| 2009/05/01 14:18:32 | 1.1 | jetley | examples/charm++/barnes-charm/util.C |
| 2009/05/01 14:18:32 | 1.1 | jetley | examples/charm++/barnes-charm/code_io.C |
| 2009/05/01 14:18:32 | 1.17 | jetley | examples/charm++/barnes-charm/barnes.h |
| 2009/05/01 14:18:32 | 1.15 | jetley | examples/charm++/barnes-charm/TreePiece.cpp |
| 2009/05/01 14:18:32 | 1.10 | jetley | examples/charm++/barnes-charm/Makefile |
| 2009/05/01 14:18:31 | 1.20 | jetley | examples/charm++/barnes-charm/Main.cpp |
|
made changes to child treepiece spawn procedure. Only pieces that are actually needed
are ever created. | |||
| 2009/04/30 20:10:59 | 1.1 | chaomei | src/arch/multicore-linux32/conv-mach.sh |
| 2009/04/30 20:10:59 | 1.1 | chaomei | src/arch/multicore-linux32/conv-mach.h |
| Initial check-ins for multicore-linux32 | |||
| 2009/04/30 14:09:31 | 1.16 | jetley | examples/charm++/barnes-charm/barnes.h |
| 2009/04/30 14:09:31 | 1.16 | jetley | examples/charm++/barnes-charm/barnes.ci |
| 2009/04/30 14:09:31 | 1.14 | jetley | examples/charm++/barnes-charm/TreePiece.cpp |
| 2009/04/30 14:09:31 | 1.15 | jetley | examples/charm++/barnes-charm/ParticleChunk.cpp |
| 2009/04/30 14:09:31 | 1.9 | jetley | examples/charm++/barnes-charm/Makefile |
| 2009/04/30 14:09:31 | 1.19 | jetley | examples/charm++/barnes-charm/Main.cpp |
|
All phases work for single iteration. Remove inserted treepieces and free
allocated memory in ParticleChunk::cleanup, etc. | |||
| 2009/04/30 13:10:06 | 1.3 | insomnia52 | examples/charm++/Molecular3D/src_charm/Main.C |
| fixed a bug, moved delete call down in Main function | |||
| 2009/04/30 13:09:22 | 1.2 | insomnia52 | examples/charm++/Molecular3D/compile/Makefile |
| added a few small things to the Makefile | |||
| 2009/04/30 10:35:39 | 1.6 | gzheng | src/QuickThreads/configure |
| print teh current configuration target | |||
| 2009/04/30 10:34:52 | 1.486 | gzheng | src/scripts/Makefile |
| when compiling under QT, sue $(MAKE) instead calling make directly | |||
| 2009/04/30 09:43:26 | 1.1 | gzheng | src/arch/net-aix-ppc/conv-mach-pxshm.sh |
| 2009/04/30 09:43:26 | 1.1 | gzheng | src/arch/net-aix-ppc/conv-mach-pxshm.h |
|
pxshm on powerpc. memory fence version of pxshm does not run, so use locking.
not thoroughly tested | |||
| 2009/04/30 07:03:37 | 1.1 | jetley | examples/charm++/barnes-charm/grav.C |
| added new gravity force calc routines | |||
| 2009/04/29 20:24:36 | 1.1 | gzheng | src/arch/multicore-aix-ppc/cc-xlc64.sh |
| 2009/04/29 20:24:35 | 1.1 | gzheng | src/arch/multicore-aix-ppc/cc-xlc64.h |
| config files for xlc64 | |||
| 2009/04/29 20:23:54 | 1.150 | gzheng | src/arch/net/charmrun/charmrun.c |
| mode changed to 0664 | |||
| 2009/04/29 20:21:17 | 1.149 | gzheng | src/arch/net/charmrun/charmrun.c |
| open with create needs third argument (mode), this is caught by intel v11 compiler | |||
| 2009/04/29 13:27:12 | 1.15 | jetley | examples/charm++/barnes-charm/barnes.h |
| 2009/04/29 13:27:12 | 1.15 | jetley | examples/charm++/barnes-charm/barnes.ci |
| 2009/04/29 13:27:12 | 1.13 | jetley | examples/charm++/barnes-charm/TreePiece.cpp |
| 2009/04/29 13:27:12 | 1.14 | jetley | examples/charm++/barnes-charm/ParticleChunk.cpp |
| 2009/04/29 13:27:12 | 1.8 | jetley | examples/charm++/barnes-charm/Makefile |
| 2009/04/29 13:27:12 | 1.18 | jetley | examples/charm++/barnes-charm/Main.cpp |
|
forces working. advancement of particles working. iteration continuation still
under construction. | |||
| 2009/04/29 07:44:41 | 1.21 | gzheng | src/conv-core/cputopology.C |
| new "+skip_cpu_topology" runtime option to turn off the process of getting cpu topology, which does not scale on very large machines. | |||
| 2009/04/29 05:45:48 | 1.23 | emenese2 | src/ck-ldb/HybridBaseLB.C |
| Debugging code removed. | |||