charm.git
11 years agoUpdate NQueens program. This program can be a benchmark test for seed balancer
Yanhua Yanhua [Wed, 28 Jul 2010 17:25:32 +0000 (12:25 -0500)]
Update NQueens program. This program can be a benchmark test for seed balancer

11 years agoMerge branch 'charm' of charmgit:charm into charm
Yanhua Yanhua [Wed, 28 Jul 2010 17:09:54 +0000 (12:09 -0500)]
Merge branch 'charm' of charmgit:charm into charm

11 years agoMerge branch 'charm' of charmgit:charm into charm
Gengbin Zheng [Tue, 27 Jul 2010 19:37:15 +0000 (15:37 -0400)]
Merge branch 'charm' of charmgit:charm into charm

11 years agouse more robust test from configure to compile asm code of switching tls.
Gengbin Zheng [Tue, 27 Jul 2010 19:35:57 +0000 (15:35 -0400)]
use more robust test from configure to compile asm code of switching tls.

11 years agore-generated
Gengbin Zheng [Tue, 27 Jul 2010 19:35:09 +0000 (14:35 -0500)]
re-generated

11 years agoadded test for asm code of switching TLS.
Gengbin Zheng [Tue, 27 Jul 2010 19:33:18 +0000 (15:33 -0400)]
added test for asm code of switching  TLS.

11 years agocasting type to make xlc compiler happen.
Gengbin Zheng [Tue, 27 Jul 2010 03:29:03 +0000 (22:29 -0500)]
casting type to make xlc compiler happen.

11 years agoAdding an abort whenever users compile charm++ without control point support, but...
Isaac Dooley [Mon, 26 Jul 2010 19:55:14 +0000 (14:55 -0500)]
Adding an abort whenever users compile charm++ without control point support, but try to use it via command line argument to their program.

11 years agoa generic KLevelTree which builds a tree with same branching factor at all levels...
Gengbin Zheng [Mon, 26 Jul 2010 17:04:29 +0000 (12:04 -0500)]
a generic KLevelTree which builds a tree with same branching factor at all levels (the numbers of processors has to be just enough to build the perfect tree)

11 years agominor change in debugging output
Gengbin Zheng [Mon, 26 Jul 2010 16:51:24 +0000 (11:51 -0500)]
minor change in debugging output

11 years agocontrol point seems to break bigsim, disable it for now.
Gengbin Zheng [Mon, 26 Jul 2010 16:50:01 +0000 (11:50 -0500)]
control point seems to break bigsim, disable it for now.

11 years agomodifyScheduler.C and tracec.C all depends on charm++, move them from conv_core to...
Gengbin Zheng [Mon, 26 Jul 2010 04:20:33 +0000 (23:20 -0500)]
modifyScheduler.C and tracec.C all depends on charm++, move them from conv_core to ck_core library.

11 years agoadd -tlsglobal for tls globals
Gengbin Zheng [Sun, 25 Jul 2010 02:33:15 +0000 (21:33 -0500)]
add -tlsglobal for tls globals

11 years agoMerge branch 'charm' of charmgit:charm into charm
Gengbin Zheng [Sat, 24 Jul 2010 15:08:13 +0000 (10:08 -0500)]
Merge branch 'charm' of charmgit:charm into charm

11 years agofixed fortran compiler settings
Gengbin Zheng [Sat, 24 Jul 2010 15:06:31 +0000 (10:06 -0500)]
fixed fortran compiler settings
use g++ instead bgxlC as linker command.

11 years agodon't force compiling the fortran code.
Gengbin Zheng [Sat, 24 Jul 2010 15:02:46 +0000 (10:02 -0500)]
don't force compiling the fortran code.

11 years agoold sun compiler does not support CkCallbackResumeThread.
Gengbin Zheng [Sat, 24 Jul 2010 14:38:57 +0000 (09:38 -0500)]
old sun compiler does not support CkCallbackResumeThread.
The problem is that CkCallbackResumeThread object's destructor is not called for some reason in order to suspect the current thread.

11 years agosuppress error when building with tls globals.
Gengbin Zheng [Sat, 24 Jul 2010 14:38:38 +0000 (09:38 -0500)]
suppress error when building with tls globals.

11 years agofixed compilation error.
Gengbin Zheng [Sat, 24 Jul 2010 05:19:26 +0000 (00:19 -0500)]
fixed compilation error.

11 years agoMerge branch 'charm' of charmgit:charm into charm
Isaac Dooley [Sat, 24 Jul 2010 05:12:11 +0000 (00:12 -0500)]
Merge branch 'charm' of charmgit:charm into charm

11 years agoFixing linking issues on BG/P for control point framework.
Isaac Dooley [Sat, 24 Jul 2010 05:11:40 +0000 (00:11 -0500)]
Fixing linking issues on BG/P for control point framework.

11 years agoremove an assert which may sometimes fail (maybe due to immediate msg)
Gengbin Zheng [Sat, 24 Jul 2010 04:59:31 +0000 (23:59 -0500)]
remove an assert which may sometimes fail (maybe due to immediate msg)

11 years agointegrated Eduardo's TLS based globals scheme.
Gengbin Zheng [Sat, 24 Jul 2010 04:42:58 +0000 (23:42 -0500)]
integrated Eduardo's TLS based globals scheme.
Two new thread libraries are geenrated, one for context:  context-tls, the other for qt:  qt-tls.

11 years agoremoved a statement for debugging.
Gengbin Zheng [Sat, 24 Jul 2010 04:34:32 +0000 (23:34 -0500)]
removed a statement for debugging.

11 years agoRevert "fix some brain damaged multiple definition of pe"
Gengbin Zheng [Sat, 24 Jul 2010 04:32:25 +0000 (23:32 -0500)]
Revert "fix some brain damaged multiple definition of pe"

This reverts commit 7d8af457038c7ea9d845d8e03ea1e5ea47c3cd62.

11 years agouse charmxi to generate decl.def to .ci rule.
Gengbin Zheng [Sat, 24 Jul 2010 04:09:35 +0000 (23:09 -0500)]
use charmxi to generate decl.def to .ci rule.

11 years agodependence to charmxi use its real full path (argv[0])
Gengbin Zheng [Sat, 24 Jul 2010 04:08:23 +0000 (23:08 -0500)]
dependence to charmxi use its real full path (argv[0])

11 years agosplit some tls definition into a new header file.
Gengbin Zheng [Sat, 24 Jul 2010 04:06:54 +0000 (23:06 -0500)]
split some tls definition into a new header file.

11 years agofix some brain damaged multiple definition of pe
Eric Bohm [Fri, 23 Jul 2010 23:35:12 +0000 (18:35 -0500)]
fix some brain damaged multiple definition of pe

11 years agoMerge branch 'charm' of charmgit:charm into charm
Eric Bohm [Fri, 23 Jul 2010 23:24:09 +0000 (18:24 -0500)]
Merge branch 'charm' of charmgit:charm into charm

11 years agoCray Node identification: make variable names match their meaningf
Phil Miller [Fri, 23 Jul 2010 19:34:44 +0000 (14:34 -0500)]
Cray Node identification: make variable names match their meaningf

11 years agomake getXTNodeID calls smp safe
Abhinav Bhatele [Fri, 23 Jul 2010 19:06:12 +0000 (15:06 -0400)]
make getXTNodeID calls smp safe

11 years agofix for cputopology for smp version of mpi-crayxt
Abhinav Bhatele [Fri, 23 Jul 2010 17:23:53 +0000 (13:23 -0400)]
fix for cputopology for smp version of mpi-crayxt

11 years agotest CkCallbackResumeThread (an intended usage than exposing thread_delay)
Gengbin Zheng [Fri, 23 Jul 2010 21:39:25 +0000 (16:39 -0500)]
test CkCallbackResumeThread (an intended usage than exposing thread_delay)

11 years agoMerge branch 'charm' of charmgit:charm into charm
Gengbin Zheng [Fri, 23 Jul 2010 16:37:04 +0000 (11:37 -0500)]
Merge branch 'charm' of charmgit:charm into charm

11 years agoremove an NULL exit function.
Gengbin Zheng [Fri, 23 Jul 2010 21:20:18 +0000 (16:20 -0500)]
remove an NULL exit function.

11 years agokeep up to date (for blueprint)
Gengbin Zheng [Fri, 23 Jul 2010 16:36:09 +0000 (11:36 -0500)]
keep up to date (for blueprint)

11 years agoCharmrun ++mpiexec: Switch to standard-described -n vs -np
Phil Miller [Thu, 1 Jul 2010 19:47:10 +0000 (14:47 -0500)]
Charmrun ++mpiexec: Switch to standard-described -n vs -np

11 years agoCharmrun: Don't require a nodelist file under ++mpiexec
Phil Miller [Sat, 26 Jun 2010 17:00:23 +0000 (12:00 -0500)]
Charmrun: Don't require a nodelist file under ++mpiexec

11 years agoCharmrun: Support another MPI variant under ++mpiexec
Phil Miller [Fri, 25 Jun 2010 22:33:45 +0000 (17:33 -0500)]
Charmrun: Support another MPI variant under ++mpiexec

11 years agoCharmrun: don't background mpiexec like rsh
Phil Miller [Fri, 25 Jun 2010 21:26:21 +0000 (16:26 -0500)]
Charmrun: don't background mpiexec like rsh

11 years agoCharmrun: quote environment variables in arguments
Phil Miller [Mon, 14 Jun 2010 16:07:04 +0000 (11:07 -0500)]
Charmrun: quote environment variables in arguments

11 years agoCharmrun: support MPIRUN_RANK/NPROCS as well as OMPI_foo
Phil Miller [Thu, 10 Jun 2010 02:21:18 +0000 (21:21 -0500)]
Charmrun: support MPIRUN_RANK/NPROCS as well as OMPI_foo

11 years agoa minimal working pup for control point manager to allow exit function to work (when...
Gengbin Zheng [Fri, 23 Jul 2010 15:38:01 +0000 (10:38 -0500)]
a minimal working pup for control point manager to allow exit function to work (when controlpoint is not used). Some flags which the exit function checks are otherwise uninitialized.

11 years agoDelete another old CVS header
Phil Miller [Fri, 23 Jul 2010 15:09:59 +0000 (10:09 -0500)]
Delete another old CVS header

11 years agoSupport multiple inheritance of chares from more than 2 base classes
Phil Miller [Fri, 23 Jul 2010 15:06:48 +0000 (10:06 -0500)]
Support multiple inheritance of chares from more than 2 base classes

11 years agoRemove useless and misleading method
Phil Miller [Fri, 23 Jul 2010 15:05:19 +0000 (10:05 -0500)]
Remove useless and misleading method

11 years agoDelete long commented-out code
Phil Miller [Fri, 23 Jul 2010 15:01:51 +0000 (10:01 -0500)]
Delete long commented-out code

11 years agocharmxi: generically generate code for up to 9 base classes
Phil Miller [Fri, 23 Jul 2010 14:58:52 +0000 (09:58 -0500)]
charmxi: generically generate code for up to 9 base classes

11 years agoRename templated base classes consistently to ease translator's job
Phil Miller [Fri, 23 Jul 2010 14:43:34 +0000 (09:43 -0500)]
Rename templated base classes consistently to ease translator's job

11 years agoAdd templates to inherit from up to 9 base classes
Phil Miller [Fri, 23 Jul 2010 00:40:17 +0000 (19:40 -0500)]
Add templates to inherit from up to 9 base classes

11 years agoRemove ancient CVS headers
Phil Miller [Thu, 22 Jul 2010 23:08:07 +0000 (18:08 -0500)]
Remove ancient CVS headers

11 years agoCPU Affinity: Remove +coremap, obsoleted by +pemap and +commap
Phil Miller [Fri, 25 Jun 2010 17:55:08 +0000 (12:55 -0500)]
CPU Affinity: Remove +coremap, obsoleted by +pemap and +commap

11 years agoTLS register swapping code (from Edurado) for __thread based global variable handling.
Gengbin Zheng [Fri, 23 Jul 2010 13:59:21 +0000 (08:59 -0500)]
TLS register swapping code (from Edurado) for __thread based global variable handling.

11 years agoadded -Wl,-bbigtoc to linker to increase TOC --- was getting TOC overflow error when...
Gengbin Zheng [Fri, 23 Jul 2010 13:36:40 +0000 (08:36 -0500)]
added -Wl,-bbigtoc to linker to increase TOC --- was getting TOC overflow error when building shared lib. This however may increase binary size.

11 years agoCkExit() always call user exit function on Pe 0 now, which is more consistent.
Gengbin Zheng [Fri, 23 Jul 2010 03:47:02 +0000 (22:47 -0500)]
CkExit() always call user exit function on Pe 0 now, which is more consistent.
The pitfall of current implementation (same as before) has the problem that
when multiple CkExit() are called, the user register function may overlap.

11 years ago+sumonly mode apparantly is broken, fixed.
Gengbin Zheng [Fri, 23 Jul 2010 03:44:42 +0000 (22:44 -0500)]
+sumonly mode apparantly is broken, fixed.
Call CkExit() instead of _CkExit.

11 years agoChange timestep filename format to work with paraview 3.x
Aaron Becker [Thu, 22 Jul 2010 21:21:15 +0000 (16:21 -0500)]
Change timestep filename format to work with paraview 3.x

11 years agoAdded migration constructors to control points groups.
Isaac Dooley [Thu, 22 Jul 2010 19:48:20 +0000 (14:48 -0500)]
Added migration constructors to control points groups.

11 years agocharmxi/Make.depends: put the charmxi dep in the right place
Phil Miller [Thu, 22 Jul 2010 02:28:05 +0000 (21:28 -0500)]
charmxi/Make.depends: put the charmxi dep in the right place

Only generate a dependence on charmxi for ci files internal to Charm++,
as indicated by the -intrinsic option. This dependence needs to be on the
foo.ci.stamp: foo.ci commands, so that charmc actually gets invoked.

11 years agodisable control point in with-production.
Gengbin Zheng [Wed, 21 Jul 2010 23:52:15 +0000 (18:52 -0500)]
disable control point in with-production.

11 years agofixed dependence of all ci files to charmxi.
Gengbin Zheng [Wed, 21 Jul 2010 23:24:17 +0000 (18:24 -0500)]
fixed dependence of all ci files to charmxi.
In old setting, changing charmxi does not recompile ci files.

11 years agofix a buffer overfow bug.
Gengbin Zheng [Wed, 21 Jul 2010 22:00:03 +0000 (17:00 -0500)]
fix a buffer overfow bug.

11 years agoMerge branch 'charm' of charmgit:charm into charm
Gengbin Zheng [Wed, 21 Jul 2010 21:45:12 +0000 (16:45 -0500)]
Merge branch 'charm' of charmgit:charm into charm

11 years agoadded --enable/disable-controlpoint.
Gengbin Zheng [Wed, 21 Jul 2010 21:44:50 +0000 (16:44 -0500)]
added --enable/disable-controlpoint.

11 years agoFixing problem when USE_CRITICAL_PATH_HEADER_ARRAY is not defined, PathHistoryEnvelop...
Isaac Dooley [Wed, 21 Jul 2010 21:43:32 +0000 (16:43 -0500)]
Fixing problem when USE_CRITICAL_PATH_HEADER_ARRAY is not defined, PathHistoryEnvelope::reset() is still called in PathHistoryEnvelope::PathHistoryEnvelope(), but it is not defined in pathHistory.C.

11 years agoMerge branch 'charm' of charmgit:charm into charm
Isaac Dooley [Wed, 21 Jul 2010 21:15:06 +0000 (16:15 -0500)]
Merge branch 'charm' of charmgit:charm into charm

11 years agoAdding CMK_WITH_CONTROLPOINT macros around control point code.
Isaac Dooley [Wed, 21 Jul 2010 21:14:30 +0000 (16:14 -0500)]
Adding CMK_WITH_CONTROLPOINT macros around control point code.

11 years agoDocs: note slowness in CPU topology gathering from DNS issues
Phil Miller [Wed, 21 Jul 2010 20:56:36 +0000 (15:56 -0500)]
Docs: note slowness in CPU topology gathering from DNS issues

11 years agoxlc: don't pass archaic -qstaticinline
Phil Miller [Wed, 21 Jul 2010 20:53:41 +0000 (15:53 -0500)]
xlc: don't pass archaic -qstaticinline

After seeing compilation failures on Blue Drop, we heard the following
via NCSA:

From IBM compiler team ....
> I see the customer is using -qstaticinline. Tell them to stop using
  that option. The -qstaticinline option was intended to help help
  port non-standard conforming code (eg. v3) over to the first
  standard compliant compilers (eg. v4 & v5).

We should be close enough to compliant not to need it.

11 years agouse getifaddrs (when available) for getting local IP.
Gengbin Zheng [Wed, 21 Jul 2010 20:08:22 +0000 (15:08 -0500)]
use getifaddrs (when available) for getting local IP.

11 years agoa commented debug print
Gengbin Zheng [Wed, 21 Jul 2010 20:06:35 +0000 (15:06 -0500)]
a commented debug print

11 years agocheck if cp -p works (cp on some machines like the new upcrc does not take -p)
Gengbin Zheng [Wed, 21 Jul 2010 17:51:57 +0000 (12:51 -0500)]
check if cp -p works  (cp on some machines like the new upcrc does not take -p)

11 years agoMerge commit 'bad1e25ac34e5b0c0cba41a71d136e9c667990a1' into charm reviewed-6.2.2
Isaac Dooley [Wed, 21 Jul 2010 16:34:01 +0000 (11:34 -0500)]
Merge commit 'bad1e25ac34e5b0c0cba41a71d136e9c667990a1' into charm

11 years agoreset reduction_started flag to 0.
Gengbin Zheng [Tue, 20 Jul 2010 16:31:27 +0000 (11:31 -0500)]
reset reduction_started flag to 0.

11 years agofor FT, when load balancing happens, processor with 0 objects may start sending stats...
Gengbin Zheng [Tue, 20 Jul 2010 05:06:14 +0000 (00:06 -0500)]
for FT, when load balancing happens, processor with 0 objects may start sending stats immediately, leading to inconsistent state at the time of doing checkpointing. Explicitly handle this scenario with a state flag.

11 years agoMerge branch 'charm' of charmgit:charm into charm
Chao Mei [Fri, 16 Jul 2010 00:01:54 +0000 (19:01 -0500)]
Merge branch 'charm' of charmgit:charm into charm

11 years agoOnly copy the argv on the proc whose rank is not 0. This fixed the subtle bug in
Chao Mei [Thu, 15 Jul 2010 23:55:56 +0000 (18:55 -0500)]
Only copy the argv on the proc whose rank is not 0. This fixed the subtle bug in
NAMD in parsing application arguments together with Charm arguments. NAMD will call converseInit and initCharm seperately, so the copied argv which will be processed by converseInit will not be passed to the initCharm, and the argv to initCharm is the original argv passed by NAMD binary. In this case, NAMD's arguments are not correctly set.

11 years agoCharj: add some infrastructure for extern types and pre-define a few
Aaron Becker [Thu, 15 Jul 2010 19:40:58 +0000 (14:40 -0500)]
Charj: add some infrastructure for extern types and pre-define a few

11 years agoa fix for typo
Gengbin Zheng [Wed, 14 Jul 2010 21:30:27 +0000 (16:30 -0500)]
a fix for typo

11 years agoadded a barrier when receiving migration msg, before actually migrating objects.This...
Gengbin Zheng [Wed, 14 Jul 2010 18:59:15 +0000 (13:59 -0500)]
added a barrier when receiving migration msg, before actually migrating objects.This seems to help when migrating objects interferences broadcast tree operations.

11 years agoMerge branch 'charm' of charmgit:charm into charm
Yanhua Yanhua [Tue, 13 Jul 2010 20:24:19 +0000 (15:24 -0500)]
Merge branch 'charm' of charmgit:charm into charm

11 years agoFIX for commit 26b17fb9.
Filippo Gioachin [Tue, 13 Jul 2010 10:47:42 +0000 (05:47 -0500)]
FIX for commit 26b17fb9.
That commit broke the reading of processors lists

11 years agoPrinting that we are recording a processor
Filippo Gioachin [Tue, 13 Jul 2010 10:47:31 +0000 (05:47 -0500)]
Printing that we are recording a processor

11 years agoCharj: modify mainchare detection to avoid runtime crash
Aaron Becker [Mon, 12 Jul 2010 21:21:01 +0000 (16:21 -0500)]
Charj: modify mainchare detection to avoid runtime crash

11 years agoCharj: fix primitive type lookup at package scope; fix documentation
Aaron Becker [Mon, 12 Jul 2010 21:20:17 +0000 (16:20 -0500)]
Charj: fix primitive type lookup at package scope; fix documentation

11 years agoMerge branch 'charm' of charmgit:charm into charm
Yanhua Yanhua [Fri, 9 Jul 2010 03:53:25 +0000 (22:53 -0500)]
Merge branch 'charm' of charmgit:charm into charm

11 years agowhen migration decision message is very big, split messages instead of broadcasting it.
Gengbin Zheng [Thu, 8 Jul 2010 21:18:50 +0000 (16:18 -0500)]
when migration decision message is very big, split messages instead of broadcasting it.

11 years agoCharj : making sure a mainchare has an appropriate default constructor
Minas Charalambides [Thu, 8 Jul 2010 12:51:22 +0000 (15:51 +0300)]
Charj   :   making sure a mainchare has an appropriate default constructor
            (including the CkArgMsg argument)

            got rid of redundant "hasDefaultCtor" arguments in Charj.stg
            as that information is already included in "sym" (the symbol
            argument corresponding to the class symbol being dealt with)

11 years agoCharj: fix trace macros in generated code
Aaron Becker [Wed, 7 Jul 2010 16:43:42 +0000 (11:43 -0500)]
Charj: fix trace macros in generated code

11 years agoCharj : Added the "if(p.isUnpacking())" check before calling
Minas Charalambides [Wed, 7 Jul 2010 15:19:15 +0000 (18:19 +0300)]
Charj   :   Added the "if(p.isUnpacking())" check before calling
            pupInitMethod() in the pup routine.

11 years agoMerge branch 'charm' of charmgit:charm into charm
Filippo Gioachin [Wed, 7 Jul 2010 09:10:36 +0000 (04:10 -0500)]
Merge branch 'charm' of charmgit:charm into charm

11 years agoWhen checking cross-corruption via memory backup, the allocation history from last...
Filippo Gioachin [Wed, 7 Jul 2010 09:09:59 +0000 (04:09 -0500)]
When checking cross-corruption via memory backup, the allocation history from last saved point must be kept

11 years agoHandling correctly the malloc and free of a memory block when mprotect is active...
Filippo Gioachin [Wed, 7 Jul 2010 09:09:02 +0000 (04:09 -0500)]
Handling correctly the malloc and free of a memory block when mprotect is active in the case of linked-list bookkeeping

11 years agoCorrectly protecting the memory segment (in particular when the slot is in the separa...
Filippo Gioachin [Wed, 7 Jul 2010 09:08:00 +0000 (04:08 -0500)]
Correctly protecting the memory segment (in particular when the slot is in the separate hashtable)

11 years agoAdded flag to indicate that a block has been protected by mprotect.
Filippo Gioachin [Wed, 7 Jul 2010 09:04:47 +0000 (04:04 -0500)]
Added flag to indicate that a block has been protected by mprotect.
This works around the lack of information provided by mprotect (the only way to check if a block has been protected would otherwise be write to it and generate a signal)

11 years agoInstead of being in "user code" most of the time, and entering the "system code"...
Filippo Gioachin [Wed, 7 Jul 2010 08:57:07 +0000 (03:57 -0500)]
Instead of being in "user code" most of the time, and entering the "system code" when certain function are called, start in "system code" and exit from it when user entry methods are called.
This is needed for +memory_mprotect so that the system code runs with access to blocks with chareID=0 (i.e it does not generate SIGFAULT)

11 years agoAdded some group definitions for command-line flags
Filippo Gioachin [Wed, 7 Jul 2010 08:55:21 +0000 (03:55 -0500)]
Added some group definitions for command-line flags

11 years agoOnly ignore built targets at the top level
Phil Miller [Tue, 6 Jul 2010 19:58:27 +0000 (14:58 -0500)]
Only ignore built targets at the top level

11 years agoCUDA: Build hybrid API as part of Charm++, not before
Phil Miller [Tue, 6 Jul 2010 19:58:04 +0000 (14:58 -0500)]
CUDA: Build hybrid API as part of Charm++, not before