| Time | Revision | Author | Filename |
| 2008-05-12 00:37:24 | 1.86 | gzheng | src/ck-ldb/LBDatabase.h |
| 2008-05-12 00:37:24 | 1.7 | gzheng | src/ck-ldb/LBDatabase.ci |
| 2008-05-12 00:37:24 | 1.74 | gzheng | src/ck-ldb/LBDatabase.C |
| avail_vector is a per node array, which cause race condition in smp version. Use initnode call and lock to protect it. | |||
| 2008-05-11 18:41:05 | 1.1 | jetley | tests/charm++/loadbalancing/jacobi2D-periodic-lb/Makefile |
| 2008-05-11 18:41:05 | 1.1 | jetley | tests/charm++/loadbalancing/jacobi2D-periodic-lb/Jacobi2D.h |
| 2008-05-11 18:41:05 | 1.1 | jetley | tests/charm++/loadbalancing/jacobi2D-periodic-lb/Jacobi2D.ci |
| 2008-05-11 18:41:05 | 1.1 | jetley | tests/charm++/loadbalancing/jacobi2D-periodic-lb/Jacobi2D.C |
| anytime migration test | |||
| 2008-05-10 23:17:55 | 2.64 | gzheng | src/ck-core/init.h |
| added a function to clear pending messages queue (set to NULL) for GroupTable. | |||
| 2008-05-09 17:10:47 | 1.49 | wilmarth | src/libs/ck-libs/ParFUM/bulk_adapt_ops.C |
| 2008-05-09 17:10:47 | 1.3 | wilmarth | src/libs/ck-libs/ParFUM/bulk_adapt.h |
| 2008-05-09 17:10:47 | 1.4 | wilmarth | src/libs/ck-libs/ParFUM/bulk_adapt.C |
| Adding high-level refinement code. Currently exposes a bug in region locking. | |||
| 2008-05-09 13:15:36 | 2.7 | gzheng | src/conv-core/conv-config.h |
| automatically define macro CMK_64BIT when CMK_AMD64 is defined | |||
| 2008-05-09 13:13:02 | 1.3 | gzheng | src/arch/multicore-linux64/conv-mach.h |
| set optimized config | |||
| 2008-05-09 13:12:00 | 1.22 | gzheng | src/ck-core/ckmemcheckpoint.C |
| fixed compiler warnings for gcc 4.2.3 | |||
| 2008-05-09 13:09:00 | 2.339 | gzheng | src/arch/net/machine.c |
| fix the previous checkin: the locking was too finer grain in the case of broadcast. I suspect somewhere in megatest, we rely on the fact that the converse broadcast needs to be ordered. | |||
| 2008-05-09 12:02:49 | 1.312 | idooley2 | tools/projections/bin/projections.jar |
| 2008-05-09 12:02:49 | 1.1 | idooley2 | tools/projections/Makefile-old |
| 2008-05-09 12:02:48 | 1.92 | idooley2 | tools/projections/Makefile |
| Removed an extraneous print statement. Now Makefile just calls ant. | |||
| 2008-05-09 09:43:04 | 1.3 | gzheng | src/libs/ck-libs/ParFUM/bulk_adapt.C |
| added missing return statement. | |||
| 2008-05-09 02:31:25 | 2.158 | gzheng | src/ck-core/ck.C |
| ignore immeidate msg for QD purpose | |||
| 2008-05-08 16:02:53 | 1.5 | gzheng | src/arch/net-linux-amd64/conv-mach-smp.h |
|
set CMK_USE_TLS_THREAD to 1 by default now. This however is only activated when charm config check if __thread does compile.
Use PCQUEUE_LOCK by default. Leave the OS malloc default COMMENTED for now. Open the comment to use OS malloc as default. | |||
| 2008-05-08 15:55:49 | 2.338 | gzheng | src/arch/net/machine.c |
| make locking finer grain. | |||
| 2008-05-08 02:21:42 | 1.39 | gzheng | src/ck-ldb/BaseLB.h |
| fixed compiler warnings for gcc 4.2.3 | |||
| 2008-05-08 02:11:52 | 2.337 | gzheng | src/arch/net/machine.c |
| make sure only comm thread calls handleImmediateMsg in smp mode. | |||
| 2008-05-08 01:33:57 | 2.336 | gzheng | src/arch/net/machine.c |
| make sure only comm thread executes immediate message in smp mode. | |||
| 2008-05-08 01:32:57 | 2.25 | gzheng | src/ck-core/qd.C |
| 2008-05-08 01:32:56 | 2.157 | gzheng | src/ck-core/ck.C |
| fixed a subtle bug in QD in smp mode. now immediate messages are invisible to QD. | |||
| 2008-05-08 01:16:04 | 2.17 | gzheng | src/ck-core/cklocation.ci |
| 2008-05-08 01:16:04 | 1.28 | gzheng | src/ck-core/ckarray.ci |
| two init calls changed to initproc, because CpvInitialize needs to be called on every rank for SMP. | |||
| 2008-05-07 20:08:54 | 1.3 | gioachin | java/charm/debug/preference/Preference.java |
| - moved configuration and Executions to XML files | |||
| 2008-05-07 20:08:08 | 1.4 | gioachin | java/charm/debug/AllocationTreePanel.java |
| *** empty log message *** | |||
| 2008-05-07 20:07:44 | 1.2 | gioachin | java/charm/debug/GdbProcess.java |
| prevent crash if debugging output file cannot be written | |||
| 2008-05-07 19:59:32 | 1.4 | gioachin | java/charm/debug/preference/Execution.java |
| 2008-05-07 19:59:31 | 1.7 | gioachin | java/charm/debug/ParamsDialog.java |
| 2008-05-07 19:59:31 | 1.22 | gioachin | java/charm/debug/ParDebug.java |
|
- created attach mode to attach to running applications (if they have been started with server support)
- the port number of an Execution is not overwritten when a random one is chosen - created working directory parameter (option -dir) - fixed problem with -p parameter being ignored - moved configuration and Executions to XML files | |||
| 2008-05-07 19:57:24 | 1.1 | gioachin | java/charm/debug/preference/PyFilter.java |
| 2008-05-07 19:57:24 | 1.2 | gioachin | java/charm/debug/PythonDialog.java |
| more work done | |||
| 2008-05-07 19:46:27 | 1.2 | gioachin | java/charm/debug/inspect/TypedefType.java |
| 2008-05-07 19:46:27 | 1.3 | gioachin | java/charm/debug/inspect/GenericType.java |
| 2008-05-07 19:46:27 | 1.3 | gioachin | java/charm/debug/inspect/DataType.java |
| added a few utility functions | |||
| 2008-05-07 19:45:02 | 1.3 | gioachin | java/charm/util/ReflectiveXML.java |
| enhanced description | |||
| 2008-05-07 10:46:09 | 1.141 | gzheng | src/scripts/configure.in |
| 2008-05-07 10:46:08 | 1.108 | gzheng | src/scripts/configure |
| rename mpi.h under charm/include when testing native MPI. | |||
| 2008-05-07 02:19:13 | 1.8 | bohm | src/ck-com/RingMulticastStrategy.C |
| fix bad delete | |||
| 2008-05-06 23:03:12 | 2.312 | gzheng | src/conv-core/converse.h |
| macro for atomic increment and decrement. | |||
| 2008-05-06 15:29:03 | 1.21 | wilmarth | src/libs/ck-libs/ParFUM/parallel_part.C |
| 2008-05-06 15:29:03 | 1.21 | wilmarth | src/libs/ck-libs/ParFUM/bulk_adapt_ops.h |
| 2008-05-06 15:29:03 | 1.48 | wilmarth | src/libs/ck-libs/ParFUM/bulk_adapt_ops.C |
| 2008-05-06 15:29:03 | 1.2 | wilmarth | src/libs/ck-libs/ParFUM/bulk_adapt.h |
| 2008-05-06 15:29:02 | 1.2 | wilmarth | src/libs/ck-libs/ParFUM/bulk_adapt.C |
| 2008-05-06 15:29:02 | 1.31 | wilmarth | src/libs/ck-libs/ParFUM/Makefile |
|
Noises off in parallel_part. First bulk_adapt refinement algorithm (almost) ready. | |||
| 2008-05-06 15:26:01 | 1.20 | wilmarth | src/libs/ck-libs/multiphaseSharedArrays/msa-distArray.h |
| 2008-05-06 15:26:01 | 1.30 | wilmarth | src/libs/ck-libs/multiphaseSharedArrays/msa-DistPageMgr.h |
| Turning off noise. | |||
| 2008-05-06 14:57:46 | 1.110 | gzheng | src/arch/mpi/machine.c |
|
fixed a bug in multiqueue scheme, comm thread has its own queue as well for immediate message.
fixed a backward compatibility issue with MPI_Init_thread call. | |||
| 2008-05-06 14:18:53 | 1.7 | gzheng | src/arch/mpi-linux-amd64/conv-mach-smp.h |
| mistakenly turned memory fence on by default, use pcqueue lock now. | |||
| 2008-05-06 13:21:54 | 1.5 | idooley2 | src/arch/net-darwin-amd64/conv-mach.h |
| Removed duplicate line for IDLE_POLL | |||
| 2008-05-06 13:17:42 | 1.13 | idooley2 | src/arch/net/machine-pxshm.c |
| Now an error message only prints after 10 unsuccessful calls to shm_open, instead of 1. | |||
| 2008-05-06 12:10:52 | 1.40 | gzheng | src/scripts/conv-autoconfig.h.in |
| 2008-05-06 12:10:52 | 1.140 | gzheng | src/scripts/configure.in |
| 2008-05-06 12:10:52 | 1.107 | gzheng | src/scripts/configure |
| check if building charm on MPI, and check if MPI_Init_thread is supported. | |||
| 2008-05-06 10:32:04 | 2.105 | gzheng | src/ck-core/cklocation.C |
| oops, fixed the error in previous checkin. | |||
| 2008-05-06 02:14:41 | 2.104 | gzheng | src/ck-core/cklocation.C |
| revert the previous change temporarily so that user defined arraymap works again. | |||
| 2008-05-05 18:31:43 | 1.4 | idooley2 | src/arch/net-darwin-amd64/conv-mach.h |
| 2008-05-05 18:31:43 | 1.12 | idooley2 | src/arch/net/machine-pxshm.c |
| Adding support for OSSpinLock in the POSIX shared memory layer. Now there are three options for synchronization. | |||
| 2008-05-05 18:06:28 | 1.224 | gioachin | src/xlat-i/xi-symbol.C |
| changed the definition of setMemoryStatus. Now it supports recursive entry method calls (like inlined methods) | |||
| 2008-05-05 18:06:09 | 2.311 | gioachin | src/conv-core/converse.h |
| changed the definition of setMemoryStatus. Now it supports recursive entry method calls (like inlined methods) | |||
| 2008-05-05 18:04:43 | 2.16 | gioachin | src/ck-core/ckcallback.h |
|
fixed memory leak when using CkCallbackResumeThread.
Now the incoming message is not retrieved by the user. The only way to retrieve a message from a CkCallbackResumeThread is by passing in a pointer; this pointer will be pointing at the data upon return. When using directly a CkCallback of CkCallback::resumeThread type, thread_delay can still be used | |||
| 2008-05-05 18:01:31 | 1.14 | gioachin | src/conv-core/memory-charmdebug.c |
| fix bug in AllocationTree | |||
| 2008-05-05 17:55:12 | 1.30 | idooley2 | tools/projections/projections/gui/Timeline/EntryMethodObject.java |
| 2008-05-05 17:55:11 | 1.15 | idooley2 | tools/projections/projections/gui/NoiseMinerWindow.java |
| 2008-05-05 17:55:11 | 1.30 | idooley2 | tools/projections/projections/analysis/NoiseMiner.java |
| 2008-05-05 17:55:11 | 1.1 | idooley2 | tools/projections/lib/jfreechart-1.0.9.jar |
| 2008-05-05 17:55:11 | 1.1 | idooley2 | tools/projections/lib/jfreechart-1.0.9-experimental.jar |
| 2008-05-05 17:55:11 | 1.1 | idooley2 | tools/projections/lib/jfreechart-1.0.9-applet-demo.jar |
| 2008-05-05 17:55:11 | 1.1 | idooley2 | tools/projections/lib/jcommon-1.0.12.jar |
| 2008-05-05 17:55:10 | 1.1 | idooley2 | tools/projections/lib/itext-2.0.6.jar |
| 2008-05-05 17:55:00 | 1.311 | idooley2 | tools/projections/bin/projections.jar |
| 2008-05-05 17:55:00 | 1.8 | idooley2 | tools/projections/build.xml |
| 2008-05-05 17:55:00 | 1.4 | idooley2 | tools/projections/README |
| Adding some new charts to NoiseMiner | |||
| 2008-05-05 17:51:38 | 1.25 | idooley2 | tools/projections/projections/analysis/StsReader.java |
| Fixed a minor bug with entry methods not showing up in timeline | |||
| 2008-05-05 15:46:54 | 1.6 | gzheng | src/arch/mpi-linux-amd64/conv-mach-smp.h |
|
switch to gettimeofday instead of using CmiWallTimer() for MPI.
Using MPI call to CmiWallTimer may need lock (when mpich does not support multithreading), which is not sufficient. | |||
| 2008-05-05 15:44:11 | 1.3 | idooley2 | src/arch/net-darwin-amd64/conv-mach.h |
| Optimization for Mac | |||
| 2008-05-05 15:32:46 | 1.2 | idooley2 | src/arch/net-darwin-amd64/conv-mach.h |
| Adding better support for Mac | |||
| 2008-05-05 15:29:12 | 1.109 | gzheng | src/arch/mpi/machine.c |
|
define MULTI_SENDQUEUE, when used, use sendmsg queue for each worker thread.
Use MPI_Init_thread instead of MPI_Init to retrieve the level the multi-threaded support by the MPI library. | |||
| 2008-05-05 11:08:06 | 1.2 | bohm | src/libs/ck-libs/pencilfft/pencilfft.h |
| 2008-05-05 11:08:06 | 1.2 | bohm | src/libs/ck-libs/pencilfft/Makefile |
| add include of sfftw.h | |||
| 2008-05-04 10:06:42 | 1.4 | gzheng | src/arch/net-linux-amd64/conv-mach-smp.h |
| added a placeholder for CMK_USE_TLS_THREAD | |||
| 2008-05-04 10:05:33 | 1.33 | gzheng | src/arch/util/machine-smp.c |
| move CmiState to __thread when using TLS. This change how worker/comm threads are initialized. | |||
| 2008-05-04 09:51:57 | 1.38 | gzheng | src/util/pup_util.C |
| 2008-05-04 09:51:57 | 1.54 | gzheng | src/util/pup.h |
| fixed gcc 4.2 compiler warnings. | |||
| 2008-05-04 09:50:53 | 1.6 | gzheng | src/xlat-i/sdag/CStateVar.h |
| 2008-05-04 09:50:53 | 1.91 | gzheng | src/xlat-i/xi-symbol.h |
| 2008-05-04 09:50:53 | 1.223 | gzheng | src/xlat-i/xi-symbol.C |
| 2008-05-04 09:50:53 | 1.15 | gzheng | src/xlat-i/xi-main.C |
| get rid of most compiler warnings for gcc 4.2 | |||
| 2008-05-03 23:12:18 | 2.308 | gzheng | src/conv-core/convcore.c |
| remove Sayantan's change in scheduling using macro definition of Cpv directly, which does not allow different implementation of Cpv. | |||
| 2008-05-02 23:38:17 | 2.310 | gzheng | src/conv-core/converse.h |
| a complete working version of CpvAccess using __thread. also support CpvAccessOther(). To use this Cpv, __thread must be detected at autoconf time, and user must specify CMK_USE_TLS_THREAD. | |||
| 2008-05-02 15:06:51 | 1.11 | schak | src/arch/net/machine-pxshm.c |
| Fixed a small bug in the fence based code | |||
| 2008-05-02 14:18:12 | 1.10 | schak | src/arch/net/machine-pxshm.c |
| Fixed the lock less fence based version | |||
| 2008-05-02 11:33:51 | 1.3 | gzheng | doc/projections/fig/timeline.eps |
| 2008-05-02 11:33:51 | 1.2 | gzheng | doc/projections/fig/apoa1_128_userEventsView.eps |
| 2008-05-02 11:33:50 | 1.2 | gzheng | doc/projections/fig/NoiseMiner2.eps |
| reduced file size | |||
| 2008-05-02 09:44:33 | 1.108 | gzheng | src/arch/mpi/machine.c |
| print MPI version number | |||
| 2008-05-01 22:11:38 | 1.9 | emenese2 | src/arch/net/machine-pxshm.c |
| Fence-based synchronization code added. | |||
| 2008-05-01 17:17:03 | 2.53 | gzheng | src/ck-core/ckcheckpoint.C |
| skip "sync" call, as it seems to hang cobalt.ncsa. | |||
| 2008-04-30 13:52:52 | 1.17 | idooley2 | tools/projections/projections/misc/MultiRunData.java |
| 2008-04-30 13:52:52 | 1.7 | idooley2 | tools/projections/projections/gui/Timeline/MessagePanel.java |
| 2008-04-30 13:52:52 | 1.5 | idooley2 | tools/projections/projections/gui/Timeline/MessageCanvas.java |
| 2008-04-30 13:52:52 | 1.29 | idooley2 | tools/projections/projections/gui/Timeline/EntryMethodObject.java |
| 2008-04-30 13:52:52 | 1.42 | idooley2 | tools/projections/projections/gui/Timeline/Data.java |
| 2008-04-30 13:52:52 | 1.2 | idooley2 | tools/projections/projections/gui/Timeline/ColorChooser.java |
| 2008-04-30 13:52:52 | 1.40 | idooley2 | tools/projections/projections/gui/graph/Graph.java |
| 2008-04-30 13:52:52 | 1.23 | idooley2 | tools/projections/projections/gui/TimeProfileWindow.java |
| 2008-04-30 13:52:52 | 1.26 | idooley2 | tools/projections/projections/gui/StlPanel.java |
| 2008-04-30 13:52:52 | 1.43 | idooley2 | tools/projections/projections/gui/ProfileWindow.java |
| 2008-04-30 13:52:51 | 1.10 | idooley2 | tools/projections/projections/gui/PieChartWindow.java |
| 2008-04-30 13:52:51 | 1.35 | idooley2 | tools/projections/projections/gui/HistogramWindow.java |
| 2008-04-30 13:52:51 | 1.14 | idooley2 | tools/projections/projections/gui/GraphData.java |
| 2008-04-30 13:52:51 | 1.9 | idooley2 | tools/projections/projections/gui/GraphAttributesWindow.java |
| 2008-04-30 13:52:51 | 1.38 | idooley2 | tools/projections/projections/gui/CommWindow.java |
| 2008-04-30 13:52:51 | 1.14 | idooley2 | tools/projections/projections/gui/CommTimeWindow.java |
| 2008-04-30 13:52:51 | 1.73 | idooley2 | tools/projections/projections/gui/Analysis.java |
| 2008-04-30 13:52:51 | 1.23 | idooley2 | tools/projections/projections/analysis/UsageCalc.java |
| 2008-04-30 13:52:51 | 1.24 | idooley2 | tools/projections/projections/analysis/StsReader.java |
| 2008-04-30 13:52:51 | 1.29 | idooley2 | tools/projections/projections/analysis/NoiseMiner.java |
| 2008-04-30 13:52:51 | 1.31 | idooley2 | tools/projections/projections/analysis/LogReader.java |
| 2008-04-30 13:52:51 | 1.61 | idooley2 | tools/projections/projections/analysis/LogLoader.java |
| 2008-04-30 13:52:51 | 1.6 | idooley2 | tools/projections/projections/analysis/CallTable.java |
| 2008-04-30 13:52:50 | 1.7 | idooley2 | tools/projections/projections/analysis/CallGraph.java |
| 2008-04-30 13:52:43 | 1.310 | idooley2 | tools/projections/bin/projections.jar |
| 2008-04-30 13:52:43 | 1.7 | idooley2 | tools/projections/build.xml |
| Modified many things to support non-contiguous Entry IDs. Now they can be any integers. They must be listed in the STS file, but they need not be numbered from 0 to n. Also I added a new image to the build.xml file. | |||
| 2008-04-30 10:24:07 | 2.17 | gioachin | src/ck-core/msgalloc.C |
| recording messages allocated with CkCopyMsg as messages in charmdebug | |||
| 2008-04-30 02:03:13 | 1.30 | gzheng | src/libs/ck-libs/ParFUM/Makefile |
| change it so that it does not make parfum libs every time typing make, this is needed for autobuild when libs have to be built on frontend node instead of on compute nodes. | |||
| 2008-04-30 00:13:26 | 1.14 | kpattab2 | examples/charm++/Molecular/main.C |
| 2008-04-30 00:13:26 | 1.2 | kpattab2 | examples/charm++/Molecular/common.h |
| 2008-04-30 00:13:26 | 1.2 | kpattab2 | examples/charm++/Molecular/cell.C |
| 2008-04-30 00:13:26 | 1.5 | kpattab2 | examples/charm++/Molecular/Makefile |
|
Changes to Physics part by Sanjay.
Working Demo! | |||
| 2008-04-29 22:47:45 | 1.9 | bohm | src/arch/bluegenep/machine.c |
| update CmiBarrier to match changes in converse.h | |||
| 2008-04-29 17:05:43 | 1.11 | cheelee | tools/projections/projections/gui/MainSummaryGraphPanel.java |
| 2008-04-29 17:05:35 | 1.309 | cheelee | tools/projections/bin/projections.jar |
| Prettying it up. | |||
| 2008-04-29 17:00:35 | 1.1 | cheelee | tools/projections/projections/images/noSummary.jpg |
| Ooops. | |||
| 2008-04-29 16:59:51 | 1.10 | cheelee | tools/projections/projections/gui/MainSummaryGraphPanel.java |
| 2008-04-29 16:59:50 | 1.3 | cheelee | tools/projections/projections/gui/BackGroundImagePanel.java |
| 2008-04-29 16:59:50 | 1.6 | cheelee | tools/projections/projections/analysis/ProjMain.java |
| 2008-04-29 16:59:43 | 1.308 | cheelee | tools/projections/bin/projections.jar |
| 2008-04-29 16:59:42 | 1.91 | cheelee | tools/projections/Makefile |
| Sorry, I just had to do it ... | |||
| 2008-04-29 15:29:20 | 1.1 | gzheng | src/arch/net-linux-ppc/cc-gcc64.sh |
| 2008-04-29 15:29:20 | 1.1 | gzheng | src/arch/net-linux-ppc/cc-gcc64.h |
| ported 64 bit version to new turing linux partition. | |||
| 2008-04-29 14:18:34 | 1.39 | gzheng | src/scripts/conv-autoconfig.h.in |
| 2008-04-29 14:18:34 | 1.139 | gzheng | src/scripts/configure.in |
| 2008-04-29 14:18:33 | 1.106 | gzheng | src/scripts/configure |
| added test for __thread | |||
| 2008-04-29 14:02:50 | 1.51 | gioachin | src/conv-core/memory.c |
| fixed problem when using CMK_OPTIMIZE | |||
| 2008-04-29 13:51:51 | 1.10 | gioachin | src/libs/ck-libs/pythonCCS/Makefile |
| fixed include problem | |||
| 2008-04-29 10:12:23 | 1.2 | gioachin | java/charm/util/ReflectiveXML.java |
| Added description | |||
| 2008-04-28 17:05:03 | 1.6 | bohm | src/conv-core/cmidirect.h |
| Add CmiDirect_readyMark and CmiDirect_readyPollq as split versions of CmiDirect_ready | |||
| 2008-04-28 17:04:03 | 2.50 | bohm | src/arch/net/machine-ibverbs.c |
|
add split version of CmiDirect_ready -> CmiDirect_readyMark and CmiDirect_readyPollQ
which mark the OOB byte pattern, and insert the handle into the polling queue respectively. This allows the marking to be done as early as is valid, while allowing us to shift the polling to as late as is sensible. It also erases a race condition wherein puts could occur before a late ready call was made and not be detected. Now the put can occur any time after Mark, and be detected by readyPollQ. | |||
| 2008-04-28 17:03:49 | 1.1 | jetley | examples/armci/putTest/put.c |
| 2008-04-28 17:03:49 | 1.1 | jetley | examples/armci/putTest/Makefile |
| put latency test | |||
| 2008-04-28 14:24:36 | 1.5 | idooley2 | tools/projections/projections/analysis/ProjMain.java |
| now a dialog box is displayed at startup if no command line arguments are given. It will search the current directory for any sts files, prompting the user to open one of the sts files it finds. It ignores an ".sum.sts" files it encounters. It will display a dialog box for each sts file found until the user clicks "yes" for one of them. | |||
| 2008-04-28 05:45:25 | 1.1 | gioachin | java/charm/util/ReflectiveXML.java |
|
utility to save the state of an object to XML format and restore it later.
Currently it supports generic classes, primitive types, and one-dimensional arrays. | |||
| 2008-04-28 05:44:00 | 1.3 | gioachin | java/build.xml |
|
adding charm.util to the charmdebug package
making charmdebug build with debugging symbols by default | |||
| 2008-04-25 15:05:04 | 2.49 | schak | src/arch/net/machine-ibverbs.c |
| check | |||
| 2008-04-25 14:35:15 | 1.46 | abecker | src/libs/ck-libs/Makefile |
| added ParFUM dependency on msa, added ParFUM-Tops to LIBS | |||
| 2008-04-24 21:07:54 | 1.4 | bhatele | examples/charm++/ckdirect/matmul3d/matmul3d.h |
| 2008-04-24 21:07:54 | 1.4 | bhatele | examples/charm++/ckdirect/matmul3d/matmul3d.ci |
| 2008-04-24 21:07:54 | 1.4 | bhatele | examples/charm++/ckdirect/matmul3d/matmul3d.C |
| *** empty log message *** | |||
| 2008-04-24 18:39:47 | 1.13 | gioachin | java/charm/debug/ServThread.java |
| 2008-04-24 18:39:47 | 1.21 | gioachin | java/charm/debug/ParDebug.java |
| fixed a bug in breakpoints. some code that updated the GUI was not running inside the dispatch thread creating problems. | |||
| 2008-04-24 17:17:08 | 1.50 | gioachin | src/conv-core/memory.c |
| fixed missing return statements | |||
| 2008-04-24 15:50:48 | 1.11 | idooley2 | tools/projections/projections/gui/Timeline/UserEventObject.java |
| 2008-04-24 15:50:48 | 1.72 | idooley2 | tools/projections/projections/gui/Analysis.java |
| 2008-04-24 15:50:48 | 1.23 | idooley2 | tools/projections/projections/analysis/StsReader.java |
| 2008-04-24 15:50:39 | 1.307 | idooley2 | tools/projections/bin/projections.jar |
| Fixed problem with user events that don't show up in the sts file. Now they will be displayed, but will be colored white. | |||
| 2008-04-24 13:37:19 | 1.12 | bhatele | examples/charm++/topology/matmul3d/matmul3d.C |
| 2008-04-24 13:37:19 | 1.2 | bhatele | examples/charm++/topology/matmul3d/Makefile |
| dgemm support | |||
| 2008-04-24 13:34:24 | 1.3 | bhatele | examples/charm++/ckdirect/matmul3d/matmul3d.h |
| 2008-04-24 13:34:24 | 1.3 | bhatele | examples/charm++/ckdirect/matmul3d/matmul3d.ci |
| 2008-04-24 13:34:24 | 1.3 | bhatele | examples/charm++/ckdirect/matmul3d/matmul3d.C |
| 2008-04-24 13:34:24 | 1.4 | bhatele | examples/charm++/ckdirect/matmul3d/Makefile |
| iterative version | |||
| 2008-04-24 13:28:50 | 1.2 | bhatele | examples/charm++/ckdirect/matmul3d/matmul3d.h |
| 2008-04-24 13:28:50 | 1.2 | bhatele | examples/charm++/ckdirect/matmul3d/matmul3d.ci |
| 2008-04-24 13:28:50 | 1.3 | bhatele | examples/charm++/ckdirect/matmul3d/Makefile |
| 2008-04-24 13:28:49 | 1.2 | bhatele | examples/charm++/ckdirect/matmul3d/matmul3d.C |
| *** empty log message *** | |||
| 2008-04-24 13:13:07 | 1.6 | bhatele | examples/charm++/topology/matmul3d/matmul3d.h |
| 2008-04-24 13:13:07 | 1.6 | bhatele | examples/charm++/topology/matmul3d/matmul3d.ci |
| 2008-04-24 13:13:07 | 1.11 | bhatele | examples/charm++/topology/matmul3d/matmul3d.C |
| synced with the CkDirect version | |||
| 2008-04-24 02:32:14 | 1.190 | gzheng | src/conv-core/threads.c |
| fixed the previous checkin for uJcontext | |||
| 2008-04-24 02:03:17 | 1.5 | kpattab2 | examples/charm++/Molecular/main.h |
| 2008-04-24 02:03:17 | 1.8 | kpattab2 | examples/charm++/Molecular/main.ci |
| 2008-04-24 02:03:17 | 1.13 | kpattab2 | examples/charm++/Molecular/main.C |
| 2008-04-24 02:03:17 | 1.1 | kpattab2 | examples/charm++/Molecular/common.h |
| 2008-04-24 02:03:17 | 1.1 | kpattab2 | examples/charm++/Molecular/cell.h |
| 2008-04-24 02:03:16 | 1.1 | kpattab2 | examples/charm++/Molecular/cell.ci |
| 2008-04-24 02:03:16 | 1.1 | kpattab2 | examples/charm++/Molecular/cell.C |
| 2008-04-24 02:03:16 | 1.4 | kpattab2 | examples/charm++/Molecular/Makefile |
|
Ready for demo - Final MD Example
Functionality split into many files | |||
| 2008-04-24 02:02:51 | 1.189 | gzheng | src/conv-core/threads.c |
|
an important fix for context version of thread implementation for amd64.
makecontext call only takes INTEGER arguments, instead we pass pointers. had to fix this to allow makecontext to work with memory os. | |||
| 2008-04-23 22:10:19 | 1.451 | gioachin | src/scripts/Makefile |
| missed a space | |||
| 2008-04-23 22:07:45 | 1.4 | kpattab2 | examples/charm++/Molecular/main.h |
| 2008-04-23 22:07:44 | 1.12 | kpattab2 | examples/charm++/Molecular/main.C |
| MD Examples - Particles assigned initial velocity | |||
| 2008-04-23 22:06:28 | 1.1 | idooley2 | src/arch/net-darwin-amd64/conv-mach.sh |
| 2008-04-23 22:06:28 | 1.1 | idooley2 | src/arch/net-darwin-amd64/conv-mach.h |
| 2008-04-23 22:06:28 | 1.1 | idooley2 | src/arch/net-darwin-amd64/conv-mach-smp.sh |
| 2008-04-23 22:06:28 | 1.1 | idooley2 | src/arch/net-darwin-amd64/conv-mach-smp.h |
| 2008-04-23 22:06:28 | 1.1 | idooley2 | src/arch/net-darwin-amd64/conv-mach-pxshm.sh |
| 2008-04-23 22:06:26 | 1.1 | idooley2 | src/arch/net-darwin-amd64/conv-mach-pxshm.h |
| Adding a 64-bit version for Intel Macs. | |||
| 2008-04-23 21:20:07 | 1.222 | gioachin | src/xlat-i/xi-symbol.C |
| 2008-04-23 21:20:07 | 1.1 | gioachin | src/util/crc32.h |
|
Adding routines to compute crc checksums on memory regions.
Fixed bug in charmxi for debugging puppers, when pupping parameter marshalled arrays. Adding compile time option to charmxi generated files to compute and check CRC fields for all memory allocation. Works only in conjuction with "-memory charmdebug". Modified definition of "user allocated memory". Before it was all the memory allocated inside an entry method, now it only refers to user-defined entry methods. | |||
| 2008-04-23 21:15:37 | 1.49 | gioachin | src/conv-core/memory.c |
| 2008-04-23 21:15:37 | 2.3 | gioachin | src/conv-core/memory-verbose.c |
| 2008-04-23 21:15:37 | 2.16 | gioachin | src/conv-core/memory-paranoid.c |
| 2008-04-23 21:15:37 | 2.1 | gioachin | src/conv-core/memory-os-wrapper.C |
| 2008-04-23 21:15:37 | 2.3 | gioachin | src/conv-core/memory-leak.c |
| 2008-04-23 21:15:37 | 2.11 | gioachin | src/conv-core/memory-isomalloc.c |
| 2008-04-23 21:15:37 | 1.200 | gioachin | src/scripts/charmc |
| 2008-04-23 21:15:37 | 1.450 | gioachin | src/scripts/Makefile |
|
Wrapping memory-os with the other memory libraries available: paranoid, charmdebug, verbose, leak, isomalloc.
The new libraries are called libmemory-os-<wrapper_library> An extra file and library, libmemory-os-wrapper, is compiled to provide runtime support to this feature. charmc automatically includes this library if the user requests a library wrapping the OS malloc (like "-memory os-leak" Charm build process does not depend on these libraries to successfully build. | |||
| 2008-04-23 21:10:50 | 1.13 | gioachin | src/conv-core/memory-charmdebug.c |
|
adding support for CRC checked memory.
added a few pointer casts. fixed problem when collecting stack traces, it crashed if there was not a single starting function for all stacks. this is not true when user-level threads are present. | |||
| 2008-04-23 21:08:22 | 2.43 | gioachin | src/arch/net/machine-eth.c |
| 2008-04-23 21:08:21 | 2.307 | gioachin | src/conv-core/convcore.c |
| modified definition of "user allocated memory". Before it referred to the memory allocate while delivering a message. Now it refers to memory allocated while delivering messages to user-defined entry methods. | |||
| 2008-04-23 15:56:59 | 1.3 | kpattab2 | examples/charm++/Molecular/main.h |
| 2008-04-23 15:56:59 | 1.7 | kpattab2 | examples/charm++/Molecular/main.ci |
| 2008-04-23 15:56:59 | 1.11 | kpattab2 | examples/charm++/Molecular/main.C |
| 2008-04-23 15:56:59 | 1.3 | kpattab2 | examples/charm++/Molecular/Makefile |
| Molecular Dynamics Example - With LiveViz | |||
| 2008-04-23 13:15:17 | 1.55 | gzheng | src/arch/uth/machine.c |
| a dummy implementation of CmiBarrier() | |||
| 2008-04-23 10:47:29 | 1.41 | idooley2 | tools/projections/projections/gui/Timeline/Data.java |
| 2008-04-23 10:47:29 | 1.60 | idooley2 | tools/projections/projections/analysis/LogLoader.java |
| 2008-04-23 10:47:21 | 1.306 | idooley2 | tools/projections/bin/projections.jar |
| Fix for a race condition in parallel file reading for timeline. | |||
| 2008-04-23 03:51:52 | 1.6 | kpattab2 | examples/charm++/Molecular/main.ci |
| 2008-04-23 03:51:52 | 1.10 | kpattab2 | examples/charm++/Molecular/main.C |
| Multi-Step working - NAN needs to be fixed | |||
| 2008-04-22 23:33:14 | 2.37 | gzheng | src/conv-core/isomalloc.c |
|
on some machines, isomalloc memory regions on different nodes
can be different. use +isomalloc_sync to calculate the intersect of all memory regions on all nodes. This simple-minded implementation is to use file (on NFS) to calculate the intersect across all processors. | |||
| 2008-04-22 23:27:55 | 2.309 | gzheng | src/conv-core/converse.h |
| CmiBarrier() now returns a status to tell if it is implemnented. | |||
| 2008-04-22 23:26:57 | 2.42 | gzheng | src/arch/net/machine-eth.c |
|
finally a udp version of CmiBarrier() implemented.
However, this only works when +netpoll, this is because otherwise an interrupt-based implementation is tricky. | |||
| 2008-04-22 23:25:53 | 2.65 | gzheng | src/arch/net/machine-gm.c |
| 2008-04-22 23:25:53 | 2.70 | gzheng | src/arch/net/machine-dgram.c |
| CmiBarrier() now returns a status to tell if it is implemnented. | |||
| 2008-04-22 23:24:31 | 1.38 | gzheng | src/arch/vmi/machine.h |
| 2008-04-22 23:24:30 | 1.54 | gzheng | src/arch/vmi/machine.c |
| CmiBarrier() now returns a status to tell if it is implemnented. | |||
| 2008-04-22 23:24:12 | 1.7 | gzheng | src/arch/template/machine.c |
| CmiBarrier() now returns a status to tell if it is implemnented. | |||
| 2008-04-22 23:23:34 | 1.107 | gzheng | src/arch/mpi/machine.c |
| CmiBarrier() now returns a status to tell if it is implemnented. | |||
| 2008-04-22 23:22:45 | 1.4 | gzheng | src/arch/bluegenel/machine.c |
| 2008-04-22 23:22:45 | 1.27 | gzheng | src/arch/bluegenel/bglmachine.C |
| CmiBarrier() now returns a status. | |||
| 2008-04-22 20:32:12 | 1.9 | kpattab2 | examples/charm++/Molecular/main.C |
| MD Version - Fixed a Bug. | |||
| 2008-04-22 20:03:09 | 1.8 | kpattab2 | examples/charm++/Molecular/main.C |
|
Atom Updates - Working.
Please check. | |||
| 2008-04-22 16:38:56 | 1.27 | idooley2 | smart-build.pl |
| Removed extraneous print statment. | |||
| 2008-04-22 16:26:28 | 1.26 | idooley2 | smart-build.pl |
| Combined the options for multicore/smp/pxshm into a single question. | |||
| 2008-04-22 16:01:27 | 1.1 | skumar2 | src/arch/bluegenep/charmrun |
| *** empty log message *** | |||
| 2008-04-22 15:51:44 | 1.9 | idooley2 | src/libs/ck-libs/pythonCCS/Makefile |
| Fixed a slight problem where some shells don't interpret a \n correctly. | |||
| 2008-04-22 15:39:19 | 1.1 | idooley2 | src/arch/net-darwin-x86/conv-mach-pxshm.sh |
| 2008-04-22 15:39:19 | 1.1 | idooley2 | src/arch/net-darwin-x86/conv-mach-pxshm.h |
| Added the posix shared memory support for darwin x86 | |||
| 2008-04-22 14:27:05 | 1.29 | abecker | src/libs/ck-libs/multiphaseSharedArrays/msa-DistPageMgr.h |
| 2008-04-22 14:27:05 | 1.11 | abecker | src/libs/ck-libs/multiphaseSharedArrays/msa-DistPageMgr.ci |
|
modified enrollment to determine number of enrolled pes, with the goal of fixing
split communicator case in parfum | |||
| 2008-04-22 11:58:34 | 1.2 | emenese2 | examples/charm++/Molecular/main.h |
| *** empty log message *** | |||
| 2008-04-22 11:58:09 | 1.7 | emenese2 | examples/charm++/Molecular/main.C |
| *** empty log message *** | |||
| 2008-04-22 03:14:07 | 1.5 | kpattab2 | examples/charm++/Molecular/main.ci |
| 2008-04-22 03:14:07 | 1.6 | kpattab2 | examples/charm++/Molecular/main.C |
| Small changes in interface - Updated by PK | |||
| 2008-04-21 17:59:18 | 1.8 | schak | src/arch/net/machine-pxshm.c |
| Addded correctness check at exit time while unliking shared memory objects | |||
| 2008-04-21 16:59:22 | 1.5 | emenese2 | examples/charm++/Molecular/main.C |
| Several features regarding Particle were included. | |||
| 2008-04-21 16:58:40 | 1.4 | emenese2 | examples/charm++/Molecular/main.ci |
| Support for Particles class added. | |||
| 2008-04-21 16:57:20 | 1.1 | emenese2 | examples/charm++/Molecular/main.h |
| *** empty log message *** | |||
| 2008-04-21 15:58:46 | 2.335 | schak | src/arch/net/machine.c |
| Commented out debug code checked in by mistake | |||
| 2008-04-21 13:59:22 | 2.103 | gzheng | src/ck-core/cklocation.C |
| when ckNew a multiple dimensional array with known sizes, the old implementation still uses hash table to distribute array elements which does not make sense. Changed it to round-robin fashion according to the serialization of array indices. | |||
| 2008-04-21 02:51:02 | 1.3 | kpattab2 | examples/charm++/Molecular/main.ci |
| 2008-04-21 02:51:02 | 1.4 | kpattab2 | examples/charm++/Molecular/main.C |
| 2008-04-21 02:51:02 | 1.2 | kpattab2 | examples/charm++/Molecular/Makefile |
| Final parallel structure - Ready to take physical forces! | |||
| 2008-04-21 01:17:27 | 1.3 | kpattab2 | examples/charm++/Molecular/main.C |
| Parallel structure - Working with multiple steps. | |||
| 2008-04-20 23:58:06 | 1.2 | kpattab2 | examples/charm++/Molecular/main.ci |
| 2008-04-20 23:58:06 | 1.2 | kpattab2 | examples/charm++/Molecular/main.C |
| Updating the latest main.C and main.ci files - One Step computation working | |||
| 2008-04-20 20:13:11 | 1.1 | kpattab2 | examples/charm++/Molecular/main.ci |
| 2008-04-20 20:13:11 | 1.1 | kpattab2 | examples/charm++/Molecular/main.C |
| 2008-04-20 20:13:11 | 1.1 | kpattab2 | examples/charm++/Molecular/Makefile |
| Simple charm example illustrating molecular dynamics | |||
| 2008-04-18 17:35:21 | 1.134 | schak | src/arch/net/charmrun/charmrun.c |
|
Inline function used global static variable.
xlc complained. So function is no longer inline | |||
| 2008-04-18 14:33:07 | 2.41 | gzheng | src/arch/net/machine-eth.c |
| beautify a little bit about the indent. | |||
| 2008-04-18 14:10:46 | 2.306 | gzheng | src/conv-core/convcore.c |
| a minor change for multicore vesion. | |||
| 2008-04-18 14:09:30 | 2.334 | gzheng | src/arch/net/machine.c |
| in multicore versio, worker thread can't go to sleep when idle because there is no comm thread. | |||
| 2008-04-18 14:07:27 | 1.49 | abecker | src/libs/ck-libs/ParFUM/adapt_adj.C |
| fixed errant delete | |||
| 2008-04-18 13:35:14 | 1.47 | wilmarth | src/libs/ck-libs/ParFUM/bulk_adapt_ops.C |
|
3D edge bisection seems to be working in parallel now, at least for
multiple VPs. Still need to test the strange trouble cases of crashes when using -memory paranoid and/or optimization. | |||
| 2008-04-18 00:51:03 | 1.17 | schak | src/libs/ck-libs/ParFUM/ParFUM_SA.C |
| Forgot to add sharedidxl on remote side of a lock request | |||
| 2008-04-17 18:25:05 | 1.14 | abecker | src/libs/ck-libs/ParFUM-Tops/ParFUM_TOPS.cc |
| insert cohesive now does node/element renumbering and sets new entities valid | |||
| 2008-04-17 18:24:01 | 1.48 | abecker | src/libs/ck-libs/ParFUM/ParFUM_internals.h |
| added parameter defaults for set_(in)valid | |||
| 2008-04-17 15:32:04 | 1.20 | gzheng | tools/netfem_client_linux/src/tkApp.cpp |
|
two changes:
1. fixed "Previous" and "Next" buttons position so that it does not change when timestep number changes. 2. when netfem conencts to the server too early before the server even has data initialized, it will wait until it gets meaningful data. | |||
| 2008-04-17 15:29:54 | 1.11 | gzheng | src/libs/ck-libs/netfem/netfem.C |
| fixed a bug when a request is received before netfem has any data to serve. | |||
| 2008-04-16 22:40:07 | 1.8 | idooley2 | tools/netfem_client_linux/src/Makefile |
| Now it builds against ubuntu vtk and tk packages. | |||
| 2008-04-16 22:04:39 | 1.7 | idooley2 | tools/netfem_client_linux/src/Makefile |
| Added an include for the tcl/tk headers | |||
| 2008-04-16 17:49:48 | 1.20 | wilmarth | src/libs/ck-libs/ParFUM/bulk_adapt_ops.h |
| 2008-04-16 17:49:47 | 1.46 | wilmarth | src/libs/ck-libs/ParFUM/bulk_adapt_ops.C |
| 2008-04-16 17:49:47 | 1.34 | wilmarth | src/libs/ck-libs/ParFUM/adapt_adj.h |
| 2008-04-16 17:49:47 | 1.47 | wilmarth | src/libs/ck-libs/ParFUM/ParFUM_internals.h |
| 2008-04-16 17:49:47 | 1.14 | wilmarth | src/libs/ck-libs/ParFUM/ParFUM_SA.h |
| 2008-04-16 17:49:47 | 1.10 | wilmarth | src/libs/ck-libs/ParFUM/ParFUM_SA.ci |
| 2008-04-16 17:49:47 | 1.16 | wilmarth | src/libs/ck-libs/ParFUM/ParFUM_SA.C |
|
Made it possible to check for insertion of a node by looking up in IDXL.
This solves the multiple insertion problem. Also fixed an array index bug. Currently a failure to lock properly is the bug, and there are printfs checked in with this version so Sayantan can see what is happening. | |||
| 2008-04-16 03:42:53 | 1.2 | gzheng | src/arch/multicore-cygwin/conv-mach.h |
| fixed flags so that multicore-cygwin works. | |||
| 2008-04-16 03:42:16 | 2.333 | gzheng | src/arch/net/machine.c |
| print status when +commthread is used. | |||
| 2008-04-16 03:38:53 | 2.332 | gzheng | src/arch/net/machine.c |
| move _Cmi_noprocforcommthread to convcore | |||
| 2008-04-16 03:38:02 | 2.305 | gzheng | src/conv-core/convcore.c |
| fixed a compialtion error in previous checkin | |||
| 2008-04-16 02:58:43 | 1.32 | gzheng | src/arch/util/machine-smp.c |
| fix thread head count when having comm thread in multicore version for thread creation and barrier operations. | |||
| 2008-04-16 02:57:00 | 2.331 | gzheng | src/arch/net/machine.c |
|
a new runtime flag +commthread for multicore version to fire a comm thread.
When comm thread is fired, set _Cmi_noprocforcommthread to 1 to put it into sleep mode. | |||
| 2008-04-16 02:53:09 | 2.304 | gzheng | src/conv-core/convcore.c |
| remove redundent initialization of global var_Cmi_noprocforcommthread, so that it allows multicore version to set a different value. | |||
| 2008-04-16 01:07:36 | 1.29 | abecker | src/libs/ck-libs/ParFUM/Makefile |
| added phony targets to fix headers hack and make sure libraries are always rebuilt when necessary | |||
| 2008-04-15 20:13:08 | 1.6 | gioachin | examples/charm++/jacobi2d-iter/jacobi2d.C |
|
added flexibility to modify number of iterations (third input parameter)
printing the time taken by each iteration | |||
| 2008-04-15 16:34:07 | 1.15 | wilmarth | src/libs/ck-libs/ParFUM/import.h |
| 2008-04-15 16:34:07 | 1.25 | wilmarth | src/libs/ck-libs/ParFUM/import.C |
| 2008-04-15 16:34:07 | 1.3 | wilmarth | src/libs/ck-libs/ParFUM/global_numbering.C |
| Made this code take a communicator in as parameter to be more generic. | |||
| 2008-04-15 13:24:41 | 1.19 | wilmarth | src/libs/ck-libs/ParFUM/bulk_adapt_ops.h |
| 2008-04-15 13:24:41 | 1.45 | wilmarth | src/libs/ck-libs/ParFUM/bulk_adapt_ops.C |
| 2008-04-15 13:24:40 | 1.33 | wilmarth | src/libs/ck-libs/ParFUM/adapt_adj.h |
| 2008-04-15 13:24:40 | 1.48 | wilmarth | src/libs/ck-libs/ParFUM/adapt_adj.C |
| 2008-04-15 13:24:40 | 1.13 | wilmarth | src/libs/ck-libs/ParFUM/ParFUM_SA.h |
| 2008-04-15 13:24:40 | 1.9 | wilmarth | src/libs/ck-libs/ParFUM/ParFUM_SA.ci |
| 2008-04-15 13:24:40 | 1.15 | wilmarth | src/libs/ck-libs/ParFUM/ParFUM_SA.C |
| 2008-04-15 13:24:40 | 1.28 | wilmarth | src/libs/ck-libs/ParFUM/Makefile |
| *** empty log message *** | |||
| 2008-04-14 22:58:47 | 1.7 | schak | src/arch/net-linux-amd64/conv-mach-ibverbs.h |
| Fast start macro | |||
| 2008-04-14 22:58:23 | 1.133 | schak | src/arch/net/charmrun/charmrun.c |
| 2008-04-14 22:58:22 | 2.330 | schak | src/arch/net/machine.c |
| 2008-04-14 22:58:22 | 2.48 | schak | src/arch/net/machine-ibverbs.c |
| Fast start for net ibverbs | |||
| 2008-04-14 21:29:40 | 1.6 | abecker | src/libs/ck-libs/ParFUM-Tops/ParFUM_TOPS_Types.h |
| 2008-04-14 21:29:40 | 1.3 | abecker | src/libs/ck-libs/ParFUM-Tops/ParFUM_TOPS_Iterators.cc |
| 2008-04-14 21:29:40 | 1.29 | abecker | src/libs/ck-libs/ParFUM-Tops/ParFUM_TOPS.h |
| 2008-04-14 21:29:40 | 1.13 | abecker | src/libs/ck-libs/ParFUM-Tops/ParFUM_TOPS.cc |
| *** empty log message *** | |||
| 2008-04-13 01:41:46 | 1.18 | schak | src/libs/ck-libs/pythonCCS/PythonCCS.C |
| Server creation timings | |||
| 2008-04-12 23:04:02 | 1.13 | abecker | src/libs/ck-libs/ParFUM/mesh_adjacency.C |
| removed gratuitous printfs | |||
| 2008-04-12 22:43:39 | 1.5 | abecker | src/libs/ck-libs/ParFUM-Tops/ParFUM_TOPS_Types.h |
| 2008-04-12 22:43:39 | 1.2 | abecker | src/libs/ck-libs/ParFUM-Tops/ParFUM_TOPS_Iterators.cc |
| 2008-04-12 22:43:39 | 1.28 | abecker | src/libs/ck-libs/ParFUM-Tops/ParFUM_TOPS.h |
| 2008-04-12 22:43:39 | 1.12 | abecker | src/libs/ck-libs/ParFUM-Tops/ParFUM_TOPS.cc |
| 2008-04-12 22:43:39 | 1.14 | abecker | src/libs/ck-libs/ParFUM-Tops/Makefile |
|
merged in new facet/cohesive support, with associated update to ptops-explicit
app | |||
| 2008-04-12 12:30:35 | 1.7 | emenese2 | src/arch/net/machine-pxshm.c |
| Flag for stats printing turned off. | |||
| 2008-04-11 18:35:56 | 1.47 | abecker | src/libs/ck-libs/ParFUM/adapt_adj.C |
| fixed edge adjacency bug | |||
| 2008-04-11 14:33:25 | 1.6 | emenese2 | src/arch/net/machine-pxshm.c |
|
Simple bug fixed.
Instruction order changed, before acquiring the lock, number of entries in the receive buffer is checked. | |||
| 2008-04-11 12:18:55 | 1.7 | gzheng | src/arch/net-linux-amd64/conv-mach.h |
| 2008-04-11 12:18:55 | 1.6 | gzheng | src/arch/net-linux-amd64/conv-mach-ibverbs.h |
| move the flag of dusabiling sync to ibverbs. | |||
| 2008-04-11 01:03:13 | 1.6 | schak | src/arch/net-linux-amd64/conv-mach.h |
| The sync command does not gel well with infiniband | |||
| 2008-04-11 00:54:11 | 2.47 | schak | src/arch/net/machine-ibverbs.c |
| Infiniband addition for multisend | |||
| 2008-04-11 00:53:46 | 2.308 | schak | src/conv-core/converse.h |
| 2008-04-11 00:53:45 | 2.303 | schak | src/conv-core/convcore.c |
| Fix mulitsend to work with Infiniband | |||
| 2008-04-11 00:49:32 | 1.33 | idooley2 | src/libs/ck-libs/ParFUM/mesh_modify.C |
| 2008-04-11 00:49:32 | 1.12 | idooley2 | src/libs/ck-libs/ParFUM/mesh_adjacency.C |
| 2008-04-11 00:49:32 | 1.46 | idooley2 | src/libs/ck-libs/ParFUM/ParFUM_internals.h |
| 2008-04-11 00:49:32 | 1.12 | idooley2 | src/libs/ck-libs/ParFUM/ParFUM.C |
| 2008-04-11 00:49:32 | 1.27 | idooley2 | src/libs/ck-libs/ParFUM/Makefile |
| Modifications to support multiple element types for new TOPS layer. | |||
| 2008-04-10 16:02:05 | 1.46 | abecker | src/libs/ck-libs/ParFUM/adapt_adj.C |
| removed duplicate adds of neighbor adjacencies | |||
| 2008-04-10 12:25:26 | 1.18 | wilmarth | src/libs/ck-libs/ParFUM/bulk_adapt_ops.h |
| 2008-04-10 12:25:26 | 1.44 | wilmarth | src/libs/ck-libs/ParFUM/bulk_adapt_ops.C |
| 2008-04-10 12:25:26 | 1.32 | wilmarth | src/libs/ck-libs/ParFUM/adapt_adj.h |
| 2008-04-10 12:25:26 | 1.45 | wilmarth | src/libs/ck-libs/ParFUM/adapt_adj.C |
|
I made all the adapt_adj functions accept an FEM_Mesh* meshPtr. For
some reason, the functions that I passed int meshID to had problems with TCharm. | |||
| 2008-04-09 21:58:12 | 1.5 | bhatele | examples/charm++/topology/jacobi3d/jacobi3d.ci |
| 2008-04-09 21:58:12 | 1.22 | bhatele | examples/charm++/topology/jacobi3d/jacobi3d.C |
| fixed | |||
| 2008-04-09 15:19:59 | 1.7 | schak | tests/charm++/pingpong/pingpong.ci |
| 2008-04-09 15:19:59 | 1.12 | schak | tests/charm++/pingpong/pingpong.C |
| Fixed the node group ckdirect version for the 1 processor case | |||
| 2008-04-09 15:19:07 | 2.46 | schak | src/arch/net/machine-ibverbs.c |
| Some more debug prints | |||
| 2008-04-09 15:18:05 | 1.4 | bhatele | examples/charm++/topology/jacobi3d/jacobi3d.ci |
| 2008-04-09 15:18:05 | 1.21 | bhatele | examples/charm++/topology/jacobi3d/jacobi3d.C |
| *** empty log message *** | |||
| 2008-04-09 02:47:23 | 1.20 | bhatele | examples/charm++/topology/jacobi3d/jacobi3d.C |
| chnaged 3d temp array to 1d | |||
| 2008-04-09 02:29:38 | 1.19 | bhatele | examples/charm++/topology/jacobi3d/jacobi3d.C |
| *** empty log message *** | |||
| 2008-04-08 17:16:25 | 1.16 | gioachin | src/libs/ck-libs/pythonCCS/PythonCCS.h |
| added constructors for CkCallbackPython | |||
| 2008-04-08 16:44:34 | 1.17 | wilmarth | src/libs/ck-libs/ParFUM/bulk_adapt_ops.h |
| 2008-04-08 16:44:34 | 1.43 | wilmarth | src/libs/ck-libs/ParFUM/bulk_adapt_ops.C |
| *** empty log message *** | |||
| 2008-04-08 14:40:22 | 1.5 | schak | src/conv-core/cmidirect.h |
| Recver key can be more than 32 bytes on some machines | |||
| 2008-04-08 14:36:07 | 2.45 | schak | src/arch/net/machine-ibverbs.c |
| Some more error reporting and detection | |||
| 2008-04-08 14:27:17 | 1.16 | wilmarth | src/libs/ck-libs/ParFUM/bulk_adapt_ops.h |
| 2008-04-08 14:27:17 | 1.42 | wilmarth | src/libs/ck-libs/ParFUM/bulk_adapt_ops.C |
| 2008-04-08 14:27:17 | 1.12 | wilmarth | src/libs/ck-libs/ParFUM/ParFUM_SA.h |
| 2008-04-08 14:27:17 | 1.8 | wilmarth | src/libs/ck-libs/ParFUM/ParFUM_SA.ci |
| 2008-04-08 14:27:17 | 1.14 | wilmarth | src/libs/ck-libs/ParFUM/ParFUM_SA.C |
|
Cleaned up some old functions to pass idxl values of shared nodes
instead of coords. | |||
| 2008-04-08 14:25:53 | 1.44 | abecker | src/libs/ck-libs/ParFUM/adapt_adj.C |
| fixed bunches of memory leaks | |||
| 2008-04-08 13:22:22 | 1.15 | wilmarth | src/libs/ck-libs/ParFUM/bulk_adapt_ops.h |
| 2008-04-08 13:22:22 | 1.41 | wilmarth | src/libs/ck-libs/ParFUM/bulk_adapt_ops.C |
| Fixed some memory leaks. | |||
| 2008-04-07 16:28:21 | 2.302 | gzheng | src/conv-core/convcore.c |
| fixed for the QD bug in bigsim | |||
| 2008-04-07 16:26:45 | 2.150 | gzheng | src/ck-core/init.C |
| a subtle bug for bigsim with QD: a QD may be invoked by a user from another PE before it is get initialized. This was because right now QD is a hack for bigsim that they are still in converse level. | |||
| 2008-04-07 02:02:26 | 1.11 | bohm | tests/charm++/pingpong/pingpong.C |
| make portable to CkDirect implementations other than ibverbs | |||
| 2008-04-07 01:39:27 | 1.4 | bohm | src/conv-core/cmidirect.h |
|
ibverbs layer is cranky about field order in the userhandle.
this fixes bug | |||
| 2008-04-04 18:27:26 | 1.43 | abecker | src/libs/ck-libs/ParFUM/adapt_adj.C |
| removed inlining from lookup functions because it breaks with -O3 | |||
| 2008-04-04 18:09:15 | 1.40 | wilmarth | src/libs/ck-libs/ParFUM/bulk_adapt_ops.C |
| 2008-04-04 18:09:15 | 1.42 | wilmarth | src/libs/ck-libs/ParFUM/adapt_adj.C |
| 2008-04-04 18:09:14 | 1.13 | wilmarth | src/libs/ck-libs/ParFUM/ParFUM_SA.C |
| Latest fixes to edge_bisect. Serial looks to be working fine now. | |||
| 2008-04-04 00:38:56 | 1.3 | bohm | src/conv-core/cmidirect.h |
| 2008-04-04 00:38:56 | 1.8 | bohm | src/arch/bluegenep/machine.c |
| improve robustness in cmidirect implementation | |||
| 2008-04-03 21:18:29 | 1.41 | abecker | src/libs/ck-libs/ParFUM/adapt_adj.C |
| fixed silly bug in copyEdgeAdjacency | |||
| 2008-04-03 17:37:57 | 1.39 | wilmarth | src/libs/ck-libs/ParFUM/bulk_adapt_ops.C |
| *** empty log message *** | |||
| 2008-04-03 17:27:09 | 1.4 | gzheng | tests/charm++/chkpt/hello.C |
| comment | |||
| 2008-04-03 13:54:23 | 1.38 | wilmarth | src/libs/ck-libs/ParFUM/bulk_adapt_ops.C |
| *** empty log message *** | |||
| 2008-04-03 12:32:28 | 1.15 | jetley | tests/charm++/ckdirect-stencil/stencil3d.C |
|
corrected array access bug, count is now a double instead of an int (haven't thought about machine precision
issues yet). per iteration time is printed now. | |||
| 2008-04-03 02:55:35 | 2.67 | gzheng | src/ck-core/ckreduction.C |
| two new cknodereduction class variables need to be pup'ed so that checkpoint/restart will work. | |||
| 2008-04-03 02:25:27 | 2.52 | gzheng | src/ck-core/ckcheckpoint.C |
| added a node barrier to avoid a race condition of overwriting readonlies. | |||
| 2008-04-02 19:30:17 | 1.2 | bohm | src/conv-core/cmidirect.h |
| bg/p ckdirect update | |||
| 2008-04-02 17:38:01 | 1.37 | wilmarth | src/libs/ck-libs/ParFUM/bulk_adapt_ops.C |
| Fixed a memory bug... | |||
| 2008-04-02 15:29:59 | 1.5 | bohm | src/arch/net-linux-amd64/conv-mach-ibverbs.h |
| Adds CMK_DIRECT as a conv-mach symbol which can be tested for CmiDirect support. | |||
| 2008-04-02 15:28:33 | 1.7 | bohm | src/arch/bluegenep/machine.c |
| 2008-04-02 15:28:33 | 1.4 | bohm | src/arch/bluegenep/conv-mach.h |
|
Implements CmiDirect interface on Blue Gene/P.
Adds CMK_DIRECT as a conv-mach symbol which can be tested for CmiDirect support. | |||
| 2008-04-02 13:53:30 | 2.329 | schak | src/arch/net/machine.c |
| 2008-04-02 13:53:30 | 2.44 | schak | src/arch/net/machine-ibverbs.c |
|
Removed mallocs and frees in cmidirect, while calling cmidirect_ready
and after a message has been received | |||
| 2008-04-02 11:43:56 | 1.36 | wilmarth | src/libs/ck-libs/ParFUM/bulk_adapt_ops.C |
| *** empty log message *** | |||
| 2008-04-01 21:06:45 | 1.2 | bhatele | examples/charm++/ckdirect/matmul3d/Makefile |
| *** empty log message *** | |||
| 2008-04-01 20:44:21 | 1.85 | cheelee | tools/projections/projections/gui/MainWindow.java |
| 2008-04-01 20:44:12 | 1.305 | cheelee | tools/projections/bin/projections.jar |
| Added the index number to the summary data dump. | |||
| 2008-04-01 18:55:39 | 1.3 | cheelee | tools/projections/projections/gui/graph/SummaryXAxis.java |
| 2008-04-01 18:55:38 | 1.84 | cheelee | tools/projections/projections/gui/MainWindow.java |
| 2008-04-01 18:55:38 | 1.12 | cheelee | tools/projections/projections/gui/GraphXAxisCanvas.java |
| 2008-04-01 18:55:38 | 1.71 | cheelee | tools/projections/projections/gui/Analysis.java |
| 2008-04-01 18:55:38 | 1.4 | cheelee | tools/projections/projections/analysis/ProjMain.java |
| 2008-04-01 18:55:31 | 1.304 | cheelee | tools/projections/bin/projections.jar |
|
Added a feature to provide parameters to Projections at the start to control
the summary output if available. | |||
| 2008-04-01 16:34:25 | 1.14 | wilmarth | src/libs/ck-libs/ParFUM/bulk_adapt_ops.h |
| 2008-04-01 16:34:25 | 1.35 | wilmarth | src/libs/ck-libs/ParFUM/bulk_adapt_ops.C |
| 2008-04-01 16:34:25 | 1.31 | wilmarth | src/libs/ck-libs/ParFUM/adapt_adj.h |
| 2008-04-01 16:34:25 | 1.11 | wilmarth | src/libs/ck-libs/ParFUM/ParFUM_SA.h |
| 2008-04-01 16:34:25 | 1.7 | wilmarth | src/libs/ck-libs/ParFUM/ParFUM_SA.ci |
| 2008-04-01 16:34:25 | 1.12 | wilmarth | src/libs/ck-libs/ParFUM/ParFUM_SA.C |
|
Edge bisection code. Works on one serial element, but locking breaks on the second
bisection. | |||
| 2008-04-01 14:34:38 | 1.83 | cheelee | tools/projections/projections/gui/MainWindow.java |
| 2008-04-01 14:34:28 | 1.303 | cheelee | tools/projections/bin/projections.jar |
|
Fixed silly bug where the summary file dump is only written out to disk if
a re-binning of it's interval is required. Now the summary dump is written out for all cases. | |||
| 2008-04-01 14:33:16 | 1.1 | bhatele | examples/charm++/ckdirect/matmul3d/matmul3d.h |
| 2008-04-01 14:33:16 | 1.1 | bhatele | examples/charm++/ckdirect/matmul3d/matmul3d.ci |
| 2008-04-01 14:33:16 | 1.1 | bhatele | examples/charm++/ckdirect/matmul3d/matmul3d.C |
| 2008-04-01 14:33:15 | 1.1 | bhatele | examples/charm++/ckdirect/matmul3d/Makefile |
| first version after move | |||
| 2008-04-01 12:55:51 | 1.30 | abecker | src/libs/ck-libs/ParFUM/adapt_adj.h |
| fixed declaration for replaceadaptadjonedge | |||
| 2008-03-31 21:40:28 | 1.14 | jetley | tests/charm++/ckdirect-stencil/stencil3d.C |
| have an ARR_CHECK version | |||
| 2008-03-31 20:31:01 | 1.7 | jetley | tests/charm++/ckdirect-stencil/stencil3d.ci |
| 2008-03-31 20:31:01 | 1.13 | jetley | tests/charm++/ckdirect-stencil/stencil3d.C |
| fixed bug | |||
| 2008-03-31 18:52:57 | 1.48 | gioachin | src/conv-core/memory.c |
|
fixed compilation problem when compiling charm with CMK_OPTIMIZE and the
user program without CMK_OPTIMIZE | |||
| 2008-03-31 17:33:49 | 1.1 | bohm | src/arch/common/cc-icc10amd.sh |
| 2008-03-31 17:33:48 | 1.1 | bohm | src/arch/common/cc-icc10amd.h |
| icc on ranger's opterons hates -xT using -xO instead | |||
| 2008-03-31 16:08:03 | 1.29 | abecker | src/libs/ck-libs/ParFUM/adapt_adj.h |
| 2008-03-31 16:08:03 | 1.40 | abecker | src/libs/ck-libs/ParFUM/adapt_adj.C |
| Added functions to add to and clear edge adjacencies | |||
| 2008-03-31 15:31:54 | 2.43 | schak | src/arch/net/machine-ibverbs.c |
| Poll cmidirect q for 1 processor runs as well | |||
| 2008-03-31 12:53:20 | 1.32 | abecker | src/libs/ck-libs/ParFUM/mesh_modify.C |
| 2008-03-31 12:53:20 | 1.34 | abecker | src/libs/ck-libs/ParFUM/bulk_adapt_ops.C |
| 2008-03-31 12:53:20 | 1.28 | abecker | src/libs/ck-libs/ParFUM/adapt_adj.h |
| Moved bulk adaptivity attr handling code from FEM_add_element_local to BulkAdapt::add_element | |||
| 2008-03-31 08:58:52 | 1.13 | wilmarth | src/libs/ck-libs/ParFUM/bulk_adapt_ops.h |
| 2008-03-31 08:58:52 | 1.33 | wilmarth | src/libs/ck-libs/ParFUM/bulk_adapt_ops.C |
| 2008-03-31 08:58:52 | 1.27 | wilmarth | src/libs/ck-libs/ParFUM/adapt_adj.h |
| 2008-03-31 08:58:52 | 1.10 | wilmarth | src/libs/ck-libs/ParFUM/ParFUM_SA.h |
| 2008-03-31 08:58:52 | 1.6 | wilmarth | src/libs/ck-libs/ParFUM/ParFUM_SA.ci |
| 2008-03-31 08:58:52 | 1.11 | wilmarth | src/libs/ck-libs/ParFUM/ParFUM_SA.C |
| Needed newline at EOF. | |||
| 2008-03-30 02:10:16 | 1.5 | bhatele | examples/charm++/topology/matmul3d/matmul3d.h |
| 2008-03-30 02:10:16 | 1.5 | bhatele | examples/charm++/topology/matmul3d/matmul3d.ci |
| 2008-03-30 02:10:16 | 1.10 | bhatele | examples/charm++/topology/matmul3d/matmul3d.C |
| fixed a reduction bug | |||
| 2008-03-30 00:39:19 | 1.100 | gzheng | src/ck-perf/trace-summary.C |
| open sts file only at the traceClose time instead of fopen at the beginning of the execution. | |||
| 2008-03-30 00:23:37 | 1.9 | bhatele | examples/charm++/topology/matmul3d/matmul3d.C |
| look for bugs | |||
| 2008-03-29 16:24:26 | 1.6 | jetley | tests/charm++/ckdirect-stencil/stencil3d.ci |
| 2008-03-29 16:24:26 | 1.12 | jetley | tests/charm++/ckdirect-stencil/stencil3d.C |
| 1d instead of 3d chare array | |||
| 2008-03-29 13:56:51 | 1.4 | schak | tests/charm++/ckdirect-stencil/Makefile |
| Makefile with 2 targets, ckdirect and msg based | |||
| 2008-03-29 04:06:33 | 1.5 | jetley | tests/charm++/ckdirect-stencil/stencil3d.ci |
| 2008-03-29 04:06:33 | 1.11 | jetley | tests/charm++/ckdirect-stencil/stencil3d.C |
| re-enabled eltsComp | |||
| 2008-03-28 16:11:11 | 1.9 | gzheng | src/arch/mpi-linux-amd64/conv-mach.h |
| use os malloc as default now | |||
| 2008-03-28 15:18:21 | 1.39 | abecker | src/libs/ck-libs/ParFUM/adapt_adj.C |
| code cleanup | |||
| 2008-03-28 15:02:47 | 2.301 | gzheng | src/conv-core/convcore.c |
| remove c++ style comments in c. | |||
| 2008-03-28 14:19:24 | 1.8 | bhatele | examples/charm++/topology/matmul3d/matmul3d.C |
|
unnecessary copy removed and use of local pointers for chares on the same
processor | |||
| 2008-03-28 13:48:59 | 1.15 | gzheng | src/util/pup_c.h |
| include conv-config.h for the int8 type definition. | |||
| 2008-03-28 12:48:02 | 1.14 | gzheng | src/util/pup_c.h |
| 2008-03-28 12:48:02 | 1.13 | gzheng | src/util/pup_c.C |
| define a new pup_int8. | |||
| 2008-03-28 12:47:22 | 2.36 | gzheng | src/conv-core/isomalloc.c |
| use a new puper to pup int8 type. | |||
| 2008-03-28 11:23:41 | 2.35 | gzheng | src/conv-core/isomalloc.c |
| a complete change to promot int to long to overcome the intmax limit. | |||
| 2008-03-28 10:19:28 | 1.4 | jetley | tests/charm++/ckdirect-stencil/stencil3d.ci |
| 2008-03-28 10:19:27 | 1.10 | jetley | tests/charm++/ckdirect-stencil/stencil3d.C |
| no eltsComp | |||
| 2008-03-28 09:08:00 | 2.307 | gzheng | src/conv-core/converse.h |
|
changed the return type of CmiIsomallocLength to long.
Something forgot to check in yesterday. | |||
| 2008-03-28 03:39:52 | 1.9 | jetley | tests/charm++/ckdirect-stencil/stencil3d.C |
| no redundant computation | |||
| 2008-03-28 02:59:14 | 1.3 | jetley | tests/charm++/ckdirect-stencil/stencil3d.ci |
| 2008-03-28 02:59:13 | 1.8 | jetley | tests/charm++/ckdirect-stencil/stencil3d.C |
| 2008-03-28 02:59:13 | 1.3 | jetley | tests/charm++/ckdirect-stencil/Makefile |
|
fixed bug. all elements being touched now.
next item - remove redundant computations and replace with copies. | |||
| 2008-03-28 00:50:35 | 1.7 | jetley | tests/charm++/ckdirect-stencil/stencil3d.C |
| *** empty log message *** | |||
| 2008-03-28 00:49:06 | 1.6 | jetley | tests/charm++/ckdirect-stencil/stencil3d.C |
| *** empty log message *** | |||
| 2008-03-28 00:46:31 | 1.5 | jetley | tests/charm++/ckdirect-stencil/stencil3d.C |
| added eltsComp | |||
| 2008-03-28 00:28:37 | 1.4 | jetley | tests/charm++/ckdirect-stencil/stencil3d.C |
| iterations++ was in the wrong place | |||
| 2008-03-27 21:52:06 | 2.34 | gzheng | src/conv-core/isomalloc.c |
| remove the constraint that the number of slots on each processor can not exceed the intMax. This is done by prompting relevant int type to long. | |||
| 2008-03-27 18:28:32 | 1.2 | jetley | tests/charm++/ckdirect-stencil/stencil3d.ci |
| 2008-03-27 18:28:32 | 1.3 | jetley | tests/charm++/ckdirect-stencil/stencil3d.C |
| 2008-03-27 18:28:32 | 1.2 | jetley | tests/charm++/ckdirect-stencil/Makefile |
| *** empty log message *** | |||
| 2008-03-27 18:04:17 | 1.31 | abecker | src/libs/ck-libs/ParFUM/mesh_modify.C |
| 2008-03-27 18:04:17 | 1.26 | abecker | src/libs/ck-libs/ParFUM/adapt_adj.h |
|
Added initialization and cleanup for bulk adjacency attributes in add/remove
element | |||
| 2008-03-27 17:22:54 | 1.2 | jetley | tests/charm++/ckdirect-stencil/stencil3d.C |
| *** empty log message *** | |||
| 2008-03-27 17:00:08 | 2.33 | gzheng | src/conv-core/isomalloc.c |
| fixed a bug that when calling an operation on a remote processor (grab or free), it passed wrong nslots parameter. | |||
| 2008-03-27 11:31:05 | 1.199 | gzheng | src/scripts/charmc |
| fixed Filippo's check in. Get rid of preloading of conv-config.sh, which changed the value of $@ on some platform for example IBM SP. | |||
| 2008-03-26 19:46:28 | 2.9 | gioachin | src/conv-core/debug-conv.h |
| 2008-03-26 19:46:28 | 2.16 | gioachin | src/conv-core/debug-conv.c |
| 2008-03-26 19:46:28 | 1.221 | gioachin | src/xlat-i/xi-symbol.C |
| 2008-03-26 19:46:28 | 2.10 | gioachin | src/ck-core/debug-charm.h |
| 2008-03-26 19:46:28 | 2.24 | gioachin | src/ck-core/debug-charm.C |
| fixed previous checkin errors | |||
| 2008-03-26 18:20:12 | 1.449 | schak | src/scripts/Makefile |
| Include cmidirect.h and ckdirect.h in include directory of build | |||
| 2008-03-26 17:44:13 | 1.10 | jetley | tests/charm++/pingpong/pingpong.C |
| 2008-03-26 17:44:12 | 1.1 | jetley | src/conv-core/cmidirect.h |
| 2008-03-26 17:44:12 | 2.1 | jetley | src/ck-core/ckdirect.h |
| 2008-03-26 17:44:12 | 2.94 | jetley | src/ck-core/charm.h |
| 2008-03-26 17:44:12 | 1.4 | jetley | tests/charm++/pingpong/Makefile |
| 2008-03-26 17:44:11 | 2.42 | jetley | src/arch/net/machine-ibverbs.c |
| CmiDirect/CkDirect changes | |||
| 2008-03-26 17:30:14 | 1.1 | jetley | tests/charm++/ckdirect-stencil/Makefile |
| *** empty log message *** | |||
| 2008-03-26 17:23:00 | 1.1 | jetley | tests/charm++/ckdirect-stencil/stencil3d.ci |
| 2008-03-26 17:23:00 | 1.1 | jetley | tests/charm++/ckdirect-stencil/stencil3d.C |
| 3d versions, with and without CkDirect | |||
| 2008-03-26 16:15:01 | 1.198 | gioachin | src/scripts/charmc |
| forgot check for machines without python | |||
| 2008-03-26 13:31:26 | 1.448 | gioachin | src/scripts/Makefile |
| *** empty log message *** | |||
| 2008-03-25 22:33:20 | 2.24 | gioachin | src/ck-core/register.h |
| 2008-03-25 22:33:20 | 2.38 | gioachin | src/ck-core/register.C |
| 2008-03-25 22:33:20 | 2.93 | gioachin | src/ck-core/charm.h |
| 2008-03-25 22:33:20 | 1.105 | gioachin | src/scripts/configure |
| 2008-03-25 22:33:19 | 1.7 | gioachin | src/libs/ck-libs/pythonCCS/PythonCCS-client.h |
| 2008-03-25 22:33:19 | 1.6 | gioachin | src/libs/ck-libs/pythonCCS/PythonCCS-client.C |
|
Added flag "-debug" to charmc for usage with charmdebug.
Some work done on CmiReduce to make it possible to have concurrent reductions. Added memory tagging by chare ID in charmdebug. Fixed breakpointing with charmdebug: now each entry point is defined by its sequential index in the EP table (before a string match was performed). Added stepping: delivery of a single message in the queue. Also, all the system entry methods are currently always executed, and never get buffered by charmdebug (they can still breakpoint if the user so chooses). Fixed race condition in python interface. To do that the interface itself has been modified, in particular with regard to high-level python. Python is now more flexible and it allows any chare collection to be "python" and receive requests. The replies are reduced to a PythonCCS mainchare which will handle the socket. The python directory has now another module (charmdebug-python) for usage by charmdebug introspection code. | |||
| 2008-03-25 22:21:34 | 2.42 | gioachin | src/conv-ccs/conv-ccs.c |
| 2008-03-25 22:21:34 | 2.300 | gioachin | src/conv-core/convcore.c |
| 2008-03-25 22:21:34 | 1.220 | gioachin | src/xlat-i/xi-symbol.C |
| 2008-03-25 22:21:34 | 1.14 | gioachin | src/xlat-i/xi-main.C |
| 2008-03-25 22:21:33 | 1.47 | gioachin | src/conv-core/memory.c |
| 2008-03-25 22:21:33 | 1.12 | gioachin | src/conv-core/memory-charmdebug.c |
| 2008-03-25 22:21:33 | 2.8 | gioachin | src/conv-core/debug-conv.h |
| 2008-03-25 22:21:33 | 2.15 | gioachin | src/conv-core/debug-conv.c |
| 2008-03-25 22:21:33 | 2.306 | gioachin | src/conv-core/converse.h |