charm.git
6 years agoconfigure: Force XL C/C++ to complain about variadic templates during 44/544/5
YanhuaSun [Thu, 19 Feb 2015 19:59:43 +0000 (13:59 -0600)]
configure: Force XL C/C++ to complain about variadic templates during
testing

The offending template has to be instantiated for the compiler to
actually complain that it requires enable C++0x support.

Change-Id: Ib504ed4a5fead8c9736ac372b807b0c8afd8276b

6 years agoBug fixed in megatest: readonly only variable is never set in groupmulti.C 48/548/2
Xiang Ni [Thu, 19 Feb 2015 21:18:02 +0000 (15:18 -0600)]
Bug fixed in megatest: readonly only variable is never set in groupmulti.C

Change-Id: Ic11d70fea5aa5b6d0acaf9583eb0fa1fb0e9618d

6 years agoBug #447: Fix -memory verbose recursion in CmiPrintf/vfprintf 47/547/2
Michael Robson [Thu, 19 Feb 2015 21:04:57 +0000 (15:04 -0600)]
Bug #447: Fix -memory verbose recursion in CmiPrintf/vfprintf

Change-Id: Icb057a33c8b4f4c595f282572e3a4add25d5f40f

6 years agoMakefile: factor out repetitive bits about building libthreads-* objects 64/464/7
Phil Miller [Mon, 19 Jan 2015 22:26:30 +0000 (16:26 -0600)]
Makefile: factor out repetitive bits about building libthreads-* objects

Change-Id: I38032028c36720f7e839db28eae32cab50b28cc6

6 years agoMakefile: factor out repetitive bits about building libmemory-* objects 63/463/7
Phil Miller [Mon, 19 Jan 2015 22:00:34 +0000 (16:00 -0600)]
Makefile: factor out repetitive bits about building libmemory-* objects

Change-Id: I92aaf7b79ecf7d7b3c5a32c66590fcf744bfbfd9

6 years agoDrop the pmetest; Not an example, and does not work 04/504/5
Nikhil Jain [Thu, 19 Feb 2015 18:42:24 +0000 (12:42 -0600)]
Drop the pmetest; Not an example, and does not work

This code may be useful in creating a miniapplication for PME;
should be brought to life along with LeanMD.

Change-Id: I815e4a9681a2fa6759bcd0758e0789cf37f4ce07

6 years agoBug #655: fix data initialization and checking in varsizetest2 38/538/3
Eric Bohm [Mon, 16 Feb 2015 22:10:50 +0000 (16:10 -0600)]
Bug #655: fix data initialization and checking in varsizetest2

Message data was never initialized.  Consistency check was never enforced.
Both of these issues have been corrected.

Change-Id: I8c7e1d9a9603cb12760e228ce5c90ccb0d40db7a

6 years agoRefactor CmiHandler cast from CkRegisterHandler calls to definition 19/519/4
Phil Miller [Wed, 11 Feb 2015 18:00:44 +0000 (12:00 -0600)]
Refactor CmiHandler cast from CkRegisterHandler calls to definition

Change-Id: I17a6dd9798d5f111551aeb6136c36a9c3d14bb88

6 years agoBug fix: fix the bug that causes net-*-smp build to fail due to the pxshm patch 36/536/1
Xiang Ni [Mon, 16 Feb 2015 18:14:51 +0000 (12:14 -0600)]
Bug fix: fix the bug that causes net-*-smp build to fail due to the pxshm patch

The comm lock initialization got removed from machine-smp.c to machine-common-core.c
for the pxshm fix. However, the net build does not use the ConverseInit function
in machine-common-core.c. The lock initialization is added to the ConverseInit
function in net/machine.c to fix the bug.

Change-Id: I70b2c7a7ea8c7a6b5b073a7c3e8096ea7db91fe8

6 years agopxshm bugfix: changes to make pxshm work correctly in netlrts 88/388/8
Xiang Ni [Thu, 29 Jan 2015 17:04:04 +0000 (11:04 -0600)]
pxshm bugfix: changes to make pxshm work correctly in netlrts

Details of the changes:
1. conv-mach-pxshm: Use locks in SMP mode; default is fences. Locks
are needed because PCQueue do not work correctly in SMP mode without locks.
2. machine.c: Add pxshm scheduler at exit time.
3. machine-common-core: create comm_lock before initializing pxshm; needed in
barrier.
4. machine-pxshm: Update fence definitions using converse; use charmrun's id for
defining unique shm region; always wait for locks in SMP mode because in unlocked
version, PCQueue's access is not thread safe; always use LrtsBarrier before completing
initialization of pxshm;
5. converse.h: define CmiNetworkProgress differently if pxshm is used.

Change-Id: I9d58de8a0794cf8d08da9d2d9ac59874afd2da21

6 years agoBug #646: Remove now-duplicated $TESTOPTS 24/524/2
Phil Miller [Sun, 15 Feb 2015 06:53:51 +0000 (00:53 -0600)]
Bug #646: Remove now-duplicated $TESTOPTS

Change-Id: Ifc070590afdb93a2078591a4891286183a54091c

6 years agoDelete relic $CMK_F90OBJS from Fortran support for Origin2000 systems 83/483/5
Phil Miller [Wed, 28 Jan 2015 00:27:30 +0000 (18:27 -0600)]
Delete relic $CMK_F90OBJS from Fortran support for Origin2000 systems

Change-Id: I46cf5146c81ac1739e40b72b872ac8bb6f814d7d

6 years agoCorrect mis-matched argument in prototype for BgRegisterHandlerEx 23/523/2
Phil Miller [Fri, 13 Feb 2015 19:27:08 +0000 (13:27 -0600)]
Correct mis-matched argument in prototype for BgRegisterHandlerEx

Change-Id: I47bd36ec20f27ef50c37dfee40b3b79c12587674

6 years agoDisable stack-smashing protection based on relevant compilers, not OS 22/522/2
Phil Miller [Fri, 13 Feb 2015 18:44:44 +0000 (12:44 -0600)]
Disable stack-smashing protection based on relevant compilers, not OS

Stack-smashing protection is implemented on Mac OS and Windows, in addition to
Linux, by any or all of GCC, ICC, and Clang. It's been observed to cause
consistent crashes on Mac OS, which disppear when it's disabled. So, broaden
the efforts to disable it automatically, since it doesn't serve the purposes of
Charm++ code.

Change-Id: I9d47fde5e3ca3ae4241dc2233dc6816c000320da

6 years agoBug #646 Missing $(TESTOPTS) in lots of test 09/509/5
Ehsan Totoni [Mon, 9 Feb 2015 20:08:11 +0000 (14:08 -0600)]
Bug #646 Missing $(TESTOPTS) in lots of test

replaced all charmrun occurences with a run function
that add $(TESTOPTS) in all makefiles.

Change-Id: Icc37f934c8a8e57505e4c50a960129a8d7c8bd8b

6 years agoBug #598: Add topology test to test suite 20/520/1
Michael Robson [Thu, 12 Feb 2015 20:56:22 +0000 (14:56 -0600)]
Bug #598: Add topology test to test suite

Change-Id: Id8fe3d572098defe5950a9a36fe2ea4f7fb7723c

6 years agoBug #662: Added ifort -auto to gni-crayx* builds with intel compilers 17/517/2
Sam White [Wed, 11 Feb 2015 00:35:38 +0000 (18:35 -0600)]
Bug #662: Added ifort -auto to gni-crayx* builds with intel compilers

Change-Id: I2aa53861ebec19dd62c55fd4b18ccdb6d027c282

6 years agoBug #651: Mac OS X / Darwin defaults to clang, 32-bit builds fix 18/518/2
Sam White [Wed, 11 Feb 2015 02:03:27 +0000 (20:03 -0600)]
Bug #651: Mac OS X / Darwin defaults to clang, 32-bit builds fix

Change-Id: Ibfc372f78d8a15dfc561be6dc53f30dde4f26c0f

6 years agoAdd CkRegisterHandlerEx helper function 61/361/8
Ronak Buch [Thu, 28 Aug 2014 20:49:53 +0000 (15:49 -0500)]
Add CkRegisterHandlerEx helper function

Change-Id: Id4e59d168103678bb658b448271c24b4cc31db91

6 years agoBug #395 mlogft: Make message logging work with flag --with-production 16/516/4
Xiang Ni [Tue, 10 Feb 2015 20:00:13 +0000 (14:00 -0600)]
Bug #395 mlogft: Make message logging work with flag --with-production

The field "flags" was not set properly in message envelopes, thus
some messages that are supposed to be logged get freed.

Change-Id: I99aaf1a92e2e1975ee94cae08122a63f24799ad3

6 years agoMove the example for pencilFFT from test to examples 02/502/2
Nikhil Jain [Mon, 9 Feb 2015 17:10:03 +0000 (11:10 -0600)]
Move the example for pencilFFT from test to examples

Change-Id: I4b7615fa26c1e3ec2fca557eb80887676e1632ff

6 years agoBug #654: Throw warnings for Charm++ flags not supported by the current build 13/513/4
Eric Mikida [Mon, 9 Feb 2015 21:48:29 +0000 (15:48 -0600)]
Bug #654: Throw warnings for Charm++ flags not supported by the current build

Change-Id: Ic0b91e2437477e2112c378f9197ffe957e6fe1d1

6 years agoBug #608: shortened jacobi3d test execution times in tests/charm++/jacobi3d 10/510/3
Sam White [Mon, 9 Feb 2015 20:59:56 +0000 (14:59 -0600)]
Bug #608: shortened jacobi3d test execution times in tests/charm++/jacobi3d

Change-Id: I0c121128d6b327eff5f98529a8bcf406d9aba953

6 years agoCleanup #600: array4d test not run in nightly build. 79/479/3
Eric Bohm [Mon, 26 Jan 2015 19:24:27 +0000 (13:24 -0600)]
Cleanup #600: array4d test not run in nightly build.

Test contained several bugs.  The worst of which was improper use of
CkMulticast for marshalled messages, which caused a hang under some
conditions.  As there is no other example of 4D array use, I fixed the
bugs, removed the pointless and incorrect use of CkMulticast, revised
the CkArrayMap to map using all 4 dimensions with z innermost, and
added a README.  Given that it isn't a very an interesting test, but
is a decent example, I moved it to examples/charm++/hello and renamed
it to match the multi-dimensional array directory naming scheme there.

Change-Id: I167c462230b6e6abb60cf0e807301778d20527d3

6 years agoChange the entry method attribute in Metabalancer to expedited instead of immediate... 11/511/2
Harshitha [Mon, 9 Feb 2015 21:23:48 +0000 (15:23 -0600)]
Change the entry method attribute in Metabalancer to expedited instead of immediate. Metabalancer is a group and immediate is supported only in nodegroup

Change-Id: Ib8b165a9ec7d3fa89e3725068ba58dfeac965f78

6 years agoBug #553 charmrun: Switch default remote shell from rsh to ssh 08/508/2
Phil Miller [Mon, 9 Feb 2015 18:19:09 +0000 (12:19 -0600)]
Bug #553 charmrun: Switch default remote shell from rsh to ssh

Update all user-facing documentation and messages accordingly

Change-Id: Ib8646b56ef3db16a9af7f13e04392b42646a0dd9

6 years agoBug #553: Drop support for HP-UX relic 'remsh' 98/498/4
Phil Miller [Mon, 9 Feb 2015 16:19:42 +0000 (10:19 -0600)]
Bug #553: Drop support for HP-UX relic 'remsh'

Change-Id: I921b55a6091295d8068a053a742f183b86312c9c

6 years agoBug #323: Add additional example to "Marshalling User-Defined Structures and Classes... 99/499/3
Ronak Buch [Mon, 9 Feb 2015 16:09:54 +0000 (10:09 -0600)]
Bug #323: Add additional example to "Marshalling User-Defined Structures and Classes" section of manual

Change-Id: I8f9fd038549cb759a6390c6a3776d9b41e5a0b16

6 years agoBug #163 SMP ++local: Remove duplicate call to CmiDestroyLocks to fix double free... 03/503/4
YanhuaSun [Mon, 9 Feb 2015 17:23:26 +0000 (11:23 -0600)]
Bug #163 SMP ++local: Remove duplicate call to CmiDestroyLocks to fix double free error on signals

Change-Id: I08df3721e9e03336b81dac57811227ea11e41243

6 years agoBug #651: Mac OS X / Darwin builds should default to compiling with clang, not g++ 07/507/2
Sam White [Mon, 9 Feb 2015 18:48:59 +0000 (12:48 -0600)]
Bug #651: Mac OS X / Darwin builds should default to compiling with clang, not g++

Change-Id: Ibcfe27b08f2c619f6c0f197ab92c85f08c90b2e6

6 years agoBug#653: Fix reference to example for readonly 05/505/1
Nikhil Jain [Mon, 9 Feb 2015 18:15:03 +0000 (12:15 -0600)]
Bug#653: Fix reference to example for readonly

Change-Id: I7d8a6c36a8547f6108b4c8b44e452697cd59483b

6 years agoIssue #602: removing a redundant chare array 'hello' test 97/497/2
Lukasz Wesolowski [Sun, 8 Feb 2015 09:49:57 +0000 (17:49 +0800)]
Issue #602: removing a redundant chare array 'hello' test

Change-Id: I3eb01001e2599440413ca3bdef13f451f78ca6b1

6 years agoEnable out-of-tree builds 04/304/6
Phil Miller [Thu, 13 Jun 2013 19:29:17 +0000 (14:29 -0500)]
Enable out-of-tree builds

For instance, the following now works:

    alice:/tmp/alice/charm$ ~/path/to/charm/build charm++ net-linux-x86_64

Change-Id: I15567fee17fbd7f66458555e784c6f8b2abee34d

6 years agoParFUM: Rearrange code to deal with where generated template code lands 72/472/2
Phil Miller [Tue, 19 Mar 2013 22:23:34 +0000 (17:23 -0500)]
ParFUM: Rearrange code to deal with where generated template code lands

Change-Id: I793f4dde55a0c2be7a63a422e9794973e3bef26a

6 years agoexamples/collide: fix for modern style 71/471/3
Phil Miller [Wed, 21 Jan 2015 01:26:21 +0000 (19:26 -0600)]
examples/collide: fix for modern style

Change-Id: Ic68b8d7213ef9f8f18e384f251878c31520947c7

6 years agoAllow passing arguments to a chare's parent class constructor 68/468/3
Phil Miller [Thu, 11 Jul 2013 23:26:58 +0000 (18:26 -0500)]
Allow passing arguments to a chare's parent class constructor

Replace existing constructors of the CBaseT1 template with a variadic
template constructor to enable this.

Change-Id: I419b257eeef7d4953c3d6d40295d93eaaddfd752

6 years agoInitial support for automatically enabling C++11 support 57/457/4
Phil Miller [Tue, 20 Jan 2015 21:11:22 +0000 (15:11 -0600)]
Initial support for automatically enabling C++11 support

Change-Id: I76a323c44855e77d080bfe4ee6b9f2b7725fc658

6 years agofix bug and clean up jacobi examples 98/398/9
Yanhua Sun [Mon, 13 Oct 2014 16:43:23 +0000 (11:43 -0500)]
fix bug and clean up jacobi examples

The boundary condition constrain was wrong because it constrains the
boundary of each sub block. When using different number of subblocks,
the boundary condition changes. Fixed this by only constrains the
boundary of boundary subblock. Replace all tabs with 2 spaces.

Change-Id: Ib6d7979e6ababf070daf4c3a990cd3b3d866bcf1

6 years agoEnable node-queue when using multicore build 86/486/2
Nikhil Jain [Thu, 29 Jan 2015 07:08:25 +0000 (01:08 -0600)]
Enable node-queue when using multicore build

Change-Id: I92250e6e6ea24f67aa111b8d3d8a53b52770099d

6 years agoBug #649 Blue Gene/Q Fortran: Configure paths, flags, and libraries for GCC/XL instea... 76/476/3
Phil Miller [Fri, 23 Jan 2015 20:45:10 +0000 (14:45 -0600)]
Bug #649 Blue Gene/Q Fortran: Configure paths, flags, and libraries for GCC/XL instead of Absoft

The values of $CMK_CF90 (Fortran 90 compiler), $CMK_F90LIBS (Fortran 90 runtime
libraries to link), and $CMK_F90_MODINC (the command-line flag to prefix paths
to search for compiled Fortran module files) were all set by successively
copying and pasting values from previous conv-mach.sh and cc-gcc.sh files. They
were set for Absoft Fortran, originating with commit
64787aef49f939fc4082871d91fce365f338f90c on net-linux/conv-mach.sh in 2001.

They're wrong everywhere, but cause special pain on Blue Gene/Q. That's because
Absoft used `-p' for $CMK_F90_MODINC. On GCC and XL Fortran, that activates
gprof-style instrumentation. However, the flag isn't passed on the linker
command line, because module files are no longer needed.

Without linking in gprof support, a variable that the instrumentation code uses
for a spin lock never gets initialized. The code tries to take this lock in the
first Fortran code that the system starts to execute. Because that variable
isn't set to the `unlocked' value, the code spins forever. Thus, the program
hangs at startup in gprof instrumentation code that no one expected to see
there:

  #0  0x000000000167e07c in __mcount_internal (frompc=17248776, selfpc=16783752) at mcount.c:78
  78              while (atomic_compare_and_exchange_bool_acq (&p->mcount_hwthd, hwthd, -1));
  (gdb) bt
  #0  0x000000000167e07c in __mcount_internal (frompc=17248776, selfpc=16783752) at mcount.c:78
  #1  0x000000000165ce78 in ._mcount ()
  #2  0x0000000001001988 in mpi_main () at fort.f90:1
  #3  0x0000000001073208 in AMPI_Fallback_Main (argc=1, argv=0x19c6696ca0) at ampi.C:487
  #4  0x0000000001072cac in MPI_threadstart_t::start (this=0x19c6918e40) at ampi.C:663
  #5  0x0000000001072e8c in AMPI_threadstart (data=0x19c678c820) at ampi.C:680
  #6  0x000000000100514c in startTCharmThread (msg=0x19c678c800) at tcharm.C:162
  #7  0x000000000109885c in CthStartThread (fn=@0x1934390: 0x10050f0 <startTCharmThread(TCharmInitMsg*)>, arg=0x19c678c800) at threads.c:1694
  #8  0x0000000001098820 in ?? () at threads.c:663

Change-Id: I3b78d18294ede29546343e2308361ed5b6c61597

6 years agonetlrts & verbs: remove unused variable declaration _charmEpoch 28/428/3
Bilge Acun [Mon, 10 Nov 2014 22:21:18 +0000 (16:21 -0600)]
netlrts & verbs: remove unused variable declaration _charmEpoch

Left here during the intial lrts transformation, not needed anymore.

Change-Id: Iaffc8bd607cb321a6f60239963906ad6023cc642

6 years agoMention additional example code in TRAM documentation 80/480/2
Lukasz Wesolowski [Tue, 27 Jan 2015 04:48:45 +0000 (12:48 +0800)]
Mention additional example code in TRAM documentation

Change-Id: I64e8ca4587486a3696838399e9fe96fd2206bb2f

6 years agocharmc: Add a -touch-on-failure option to replace ugly stuff in the makefile 62/462/2
Phil Miller [Mon, 19 Jan 2015 21:25:11 +0000 (15:25 -0600)]
charmc: Add a -touch-on-failure option to replace ugly stuff in the makefile

Change-Id: I131ab38819574901f49b3ebcf89f2c4dba349665

6 years agoCleanup #538: TSAN Data races in global variable init, cmd-line arguments. 29/429/6
Eric Bohm [Mon, 5 Jan 2015 23:02:48 +0000 (17:02 -0600)]
Cleanup #538: TSAN Data races in global variable init, cmd-line arguments.

Most of these solved by limiting to CmiRank==0.  A few required CmiLock.
One required conversion from static to CpvStatic.
The locks are only during initialization, with negligible performance impact.

Initialize only once semantics required split of arg list processing from variable initialization for global statics.

Change-Id: I92ce897fd5759cb83e9fd973e947886d75743fcb

6 years agoCleaning up charm++ reduction examples 21/421/3
Bilge Acun [Tue, 4 Nov 2014 20:07:47 +0000 (14:07 -0600)]
Cleaning up charm++ reduction examples

Put the reduction examples in one place: examples/charm++/reductions/
Update the manual to point new place of the examples.
Remove unnecessary copy AllReduce/bench folder.

Change-Id: I622279389fb2cfc3308850aa24a274ea3605f39a

6 years agoBug #612: delete unused and soon to redundant commlib tests 77/477/2
Eric Bohm [Fri, 23 Jan 2015 23:12:03 +0000 (17:12 -0600)]
Bug #612: delete unused and soon to redundant commlib tests

Change-Id: Ib91702205d9fb6f41a9cd3022f0a3125cb358138

6 years agoAdd missing include <functional> 78/478/1
Phil Miller [Mon, 26 Jan 2015 00:31:34 +0000 (18:31 -0600)]
Add missing include <functional>

On Visual Studio 2013, this isn't included by any of the other headers
here, and so the build breaks.

Change-Id: I316a5a4cd0add0acc23ce78b5df0ea6ee5163c82

6 years agoAdding a test target for kNeighbor test 67/467/5
Bilge Acun [Tue, 20 Jan 2015 21:33:07 +0000 (15:33 -0600)]
Adding a test target for kNeighbor test

Change-Id: I80f054a763a90620ce94d3afd0e1ad87227de8d7

6 years agoFixing an interface issue of collision library 74/474/2
Ehsan Totoni [Wed, 21 Jan 2015 20:26:21 +0000 (14:26 -0600)]
Fixing an interface issue of collision library

Charm collision library provides an interface
for a "client" group to be defined but it has
to extend a client group it provides. Therefore,
when the library is built, its .dec.h header
needs to be copied to /include. This change addresses this.
A mpi-interop example was hacking around by
copying that declaration file, which is fixed here as well.

Change-Id: I72c80ca8048f98fafc1e3941c60c620eb7214c04

6 years agonetlrts & verbs: remove unused CmiIdleState declaration 27/427/2
Bilge Acun [Mon, 10 Nov 2014 20:13:35 +0000 (14:13 -0600)]
netlrts & verbs: remove unused CmiIdleState declaration

Left here during the intial lrts transformation, not needed anymore.

Change-Id: Ibce866034027cba2942f74c7b0e7ac26bf27cf88

6 years ago#556 multicore: Migrate from 'net' underpinnings to 'netlrts' 79/379/6
Phil Miller [Mon, 29 Sep 2014 23:29:35 +0000 (18:29 -0500)]
#556 multicore: Migrate from 'net' underpinnings to 'netlrts'

Change-Id: I65109d2c6b3f1083799d9f6fb89833a998041ef6

6 years agotests/charm++/streamingAllToAll: Add missing $(TESTOPTS) 69/469/2
Phil Miller [Tue, 20 Jan 2015 23:27:50 +0000 (17:27 -0600)]
tests/charm++/streamingAllToAll: Add missing $(TESTOPTS)

Change-Id: I3b06b1f67f33e9edd2efd0d01ef95a8f3f339dce

6 years agoBug #609: Delete unused and not very useful tests/charm++/commSpeed 61/461/2
Phil Miller [Sat, 17 Jan 2015 21:50:47 +0000 (15:50 -0600)]
Bug #609: Delete unused and not very useful tests/charm++/commSpeed

This has been superseded by more intensive and targeted test programs.

Change-Id: Ib9219f95ff682cb5e8eb6f0adbdf3ee8cee6c77e

6 years agodoc/markupSanitizer.py #328: Support Python 2 and 3 through Unicode string literal... 60/460/2
Nicholas Bock [Sat, 17 Jan 2015 21:40:04 +0000 (15:40 -0600)]
doc/markupSanitizer.py #328: Support Python 2 and 3 through Unicode string literal agnosticism

Change-Id: Ia902814d8921dfd04ab06725f7a5bb1d2a792612

6 years agoclean code for converting integer to memory address for gni 40/340/7
YanhuaSun [Mon, 4 Aug 2014 03:42:39 +0000 (22:42 -0500)]
clean code for converting integer to memory address for gni

Change-Id: Id4d7daf0e98dc645f4ce1d441fc6975c1d449497

6 years agoBug #615: startupTest not being tested in nightly build 70/470/1
Eric Bohm [Wed, 21 Jan 2015 00:47:57 +0000 (18:47 -0600)]
Bug #615: startupTest not being tested in nightly build

Updated to current CkArrayIndex API.  Fixed bug in Makefile.
Added to top level test Makefile.

Change-Id: Iec520eec87a0e0cd6dc22f61571c5ddc3867d79a

6 years agoFix rca error with Cray compiler 58/458/3
YanhuaSun [Thu, 15 Jan 2015 21:59:08 +0000 (15:59 -0600)]
Fix rca error with Cray compiler

Change-Id: I8f40cf5736dea3324c15465d82deefbcb144cecc

6 years ago323 - Improve documentation and examples regarding marshalling of user-defined types... 59/459/1
Ronak Buch [Fri, 16 Jan 2015 23:41:03 +0000 (17:41 -0600)]
323 - Improve documentation and examples regarding marshalling of user-defined types using PUP

Change-Id: Iaae55f2d62ca392d4d6014d3b0236e6565d93ef1

6 years ago#192: Remove deprecated CmiBool typedef and CmiTrue/CmiFalse macros 04/404/2
Phil Miller [Wed, 15 Oct 2014 17:27:28 +0000 (12:27 -0500)]
#192: Remove deprecated CmiBool typedef and CmiTrue/CmiFalse macros

Change-Id: I2a9c426d1905acb30b3677efad7a3081e3a4671c

6 years agockpt: Include readonly message variables in checkpoints and restore them on restart 53/453/2
Esteban Meneses-Rojas [Thu, 11 Dec 2014 22:09:21 +0000 (16:09 -0600)]
ckpt: Include readonly message variables in checkpoints and restore them on restart

Change-Id: Ibe679f748c9fc6ad5a1f8b1982e3faf71b9ee4a5

6 years agouiuc.edu to illinois.edu 55/455/1
Shanna DeSouza [Thu, 11 Dec 2014 20:19:10 +0000 (14:19 -0600)]
uiuc.edu to illinois.edu

Change-Id: I96c04f2979494c4860c9844996d378a92c890b05

6 years agoRevert "Change "cd $$d; $(MAKE)" to "cd $$d && $(MAKE)" to catch errors in cd command" 49/449/2
Lukasz Wesolowski [Tue, 9 Dec 2014 21:55:43 +0000 (15:55 -0600)]
Revert "Change "cd $$d; $(MAKE)" to "cd $$d && $(MAKE)" to catch errors in cd command"

The change was causing an immediate error when running make in doc

This reverts commit 0d277d6b045b48bea9ca94d50727118f33a8d126.

Change-Id: I54b4ba127ceb0f105779883e380d43f9ff08e924

6 years agoUpdate the version 48/448/1
Nikhil Jain [Mon, 8 Dec 2014 16:00:30 +0000 (10:00 -0600)]
Update the version

Change-Id: I8a0f6a5fa00c069e33eddd527384d680b02e0aef

6 years agoModify the test target to match our convention 47/447/2
Nikhil Jain [Sat, 6 Dec 2014 19:40:36 +0000 (13:40 -0600)]
Modify the test target to match our convention

Do not pass ++local explicitly, let it come from
TESTOPTS if needed.

Change-Id: I28793dd9497a1131207614675276717cf6e3c1b9

6 years agobug #555: Enable restart using different numNodes 46/446/2
Xiang Ni [Sat, 6 Dec 2014 16:03:37 +0000 (10:03 -0600)]
bug #555: Enable restart using different numNodes

If restart is done using different number of nodes, use
node 0's state for all nodes; similar scheme as we do
when numPEs is different

Change-Id: I9af292c10ebc773852334ab9153cd513917d296a

6 years agoTRAM user manual update and cleanup 45/445/2
Lukasz Wesolowski [Sat, 6 Dec 2014 05:08:19 +0000 (23:08 -0600)]
TRAM user manual update and cleanup

Change-Id: I73236099e158a163c03e10c3592d636cdfb35ada

6 years agoAdded MPI_STATUS(ES)_IGNORE declarations to the Fortran AMPI header 38/438/2
Sam White [Tue, 25 Nov 2014 20:04:48 +0000 (14:04 -0600)]
Added MPI_STATUS(ES)_IGNORE declarations to the Fortran AMPI header

Change-Id: I1aa8e3a704715936554ca05ee94879c1974fcf0c

6 years agoAdding test of fine-grained all-to-all using TRAM 40/440/5
Lukasz Wesolowski [Thu, 4 Dec 2014 23:10:15 +0000 (17:10 -0600)]
Adding test of fine-grained all-to-all using TRAM

Change-Id: If8cf3e9f90f97ad5841f8e8fa786951050fe3575

6 years agoAdd HPC Challenge Random Access to examples directory 39/439/4
Lukasz Wesolowski [Thu, 4 Dec 2014 00:21:20 +0000 (18:21 -0600)]
Add HPC Challenge Random Access to examples directory

Change-Id: I7851ca4592e7cd4caaba01cbccd66a4a4181f9ff

6 years ago#628 ckio: Check for EINTR when syncing data 44/444/2
Ronak Buch [Fri, 5 Dec 2014 22:17:27 +0000 (16:17 -0600)]
#628 ckio: Check for EINTR when syncing data

Change-Id: If94c237fa65cce12e20608e4f50bb13dfa04bf57

6 years agoDisable IMMEDIATE msgs in multicore builds, no comm thread 42/442/4
Nikhil Jain [Fri, 5 Dec 2014 21:22:30 +0000 (15:22 -0600)]
Disable IMMEDIATE msgs in multicore builds, no comm thread

Change-Id: I2caf3acf903b8e4119efdca4d43d7d390f68282b

6 years agoexamples/collide: Fix Makefiles to test binaries with explicit relative path 24/424/6
Phil Miller [Fri, 7 Nov 2014 17:10:51 +0000 (11:10 -0600)]
examples/collide: Fix Makefiles to test binaries with explicit relative path

Change-Id: I3cf05f0cd473a1b6f0d61330ad2b0473938907c0

6 years agoFix a bug in Charm++ pingpong - missing break statement 43/443/3
Lukasz Wesolowski [Fri, 5 Dec 2014 21:38:27 +0000 (15:38 -0600)]
Fix a bug in Charm++ pingpong - missing break statement

Change-Id: I526929bf537bfacaf35be7ac8a6d2ba8b48abd30

6 years agoadd test to detect macro conflict arising due to C++ MPI bindings 34/434/3
Anirudh Jayakumar [Tue, 25 Nov 2014 08:27:42 +0000 (02:27 -0600)]
add test to detect macro conflict arising due to C++ MPI bindings

Change-Id: I9eab5e121133013a3cc9863ad35e73c6e57f6e35

6 years agonet layer time: do not subtract Cmi_clock 41/441/1
Jim Phillips [Fri, 5 Dec 2014 17:31:09 +0000 (11:31 -0600)]
net layer time: do not subtract Cmi_clock

Subtracting Cmi_clock leads to incorrect output in cases
where Charm's scheduler is restarted multiple times, for
example as done in NAMD.

Change-Id: Ia52d7715b03f2899367af9950d15b06b6ee0d839

6 years agockio: remove entire array (including groups and elements) after session completes. 33/433/5
Gengbin Zheng [Tue, 18 Nov 2014 22:52:34 +0000 (16:52 -0600)]
ckio: remove entire array (including groups and elements) after session completes.

The intermediate reduction added in 03ec7085a3945d174ed260aa14c52a0cede3a241 undid
the whole-array destruction.

Change-Id: I2c31ff3ae19d812109706032dd65e874dfe7557f

6 years agoArray reductions: When deleting the whole array, stop counting contributors 37/437/1
Phil Miller [Tue, 25 Nov 2014 21:36:09 +0000 (15:36 -0600)]
Array reductions: When deleting the whole array, stop counting contributors

When a chare array is being destroyed, the infrastructure to count contributors
on each PE, and restructure the reduction tree when one becomes barren, is
just extra work.

This extra work causes a problem because it sends messages to group elements
that are being deleted. These messages represent leaked memory each time an
array is destroyed. They also prevent QD from ever triggering thereafter.

Discovered and characterized by Tom Quinn in discussion on
http://charm.cs.illinois.edu/gerrit/433/

Change-Id: I04a7c994a08fdc0485c344157ac872287e5e1f21

6 years agoTypo fix 35/435/1
Phil Miller [Tue, 25 Nov 2014 18:42:22 +0000 (12:42 -0600)]
Typo fix

Change-Id: Ib375eb8e97e7273f0116ee8097e90a04686c25a8

6 years agofix remote_event bug on CrayXC 32/432/2
Yanhua Sun [Tue, 18 Nov 2014 04:55:28 +0000 (20:55 -0800)]
fix remote_event bug on CrayXC

GNI_CQ_GET_INST_ID() allows a user to retrieve the instance ID (inst_id)
field from an event dequeued from a source completion queue (attached to a GNI endpoint).
GNI_CQ_GET_REM_INST_ID() allows a user to retrieve the instance ID field
from an event dequeued from a destination completion queue (attached to
GNI registered memory).

Change-Id: I91d5f8e000d162f87f7bc9a7f8d8f37ed89e7a95

6 years agoFeature #574 win: Copy binaries to <build>/bin inside Makefile 31/431/2
Michael Robson [Wed, 12 Nov 2014 07:08:02 +0000 (01:08 -0600)]
Feature #574 win: Copy binaries to <build>/bin inside Makefile

Change-Id: Ic6e4499959c58f3eb2a516b7c36b07519d63cea0

6 years agoPartitions: Print output redirection message on partition 0 node 0 30/430/1
Phil Miller [Tue, 11 Nov 2014 19:25:24 +0000 (13:25 -0600)]
Partitions: Print output redirection message on partition 0 node 0

Change-Id: I0cb9f29d690d206e0dcbe47601e1992a9d0a4e7e

6 years agoDisable remote_event macro for CMK_CRAYXC 25/425/6
Yanhua Sun [Fri, 7 Nov 2014 18:02:10 +0000 (10:02 -0800)]
Disable remote_event macro for CMK_CRAYXC

Change-Id: I691154bb9352ff4790df3b28451c4f8529d8a97c

6 years agofix uGNI machine.c send_smsg_message function call error 26/426/3
Yanhua Sun [Sat, 8 Nov 2014 04:16:28 +0000 (20:16 -0800)]
fix uGNI machine.c send_smsg_message function call error

Change-Id: I6162d576f18acaaee18fee67d45ecb7ddbe97546

6 years agoIssue 374, BGQ Mapping: use kernel call to find the mapping 22/422/2
Nikhil Jain [Tue, 4 Nov 2014 21:49:20 +0000 (21:49 +0000)]
Issue 374, BGQ Mapping: use kernel call to find the mapping

Change-Id: I19391f3662f82e978899a48cfb45ee4ed194e822

6 years agoAdd pxshm conv-mach in verbs 70/370/5
Nikhil Jain [Mon, 15 Sep 2014 20:29:17 +0000 (15:29 -0500)]
Add pxshm conv-mach in verbs

This should enable use of pxshm with verbs layer. Use of
lock is needed for correctness.

Change-Id: Ia89893f8dbff3ff7d68d88ee0b55f8a22df40fe6

6 years agoFeature #574 win: Copy binaries to <build>/bin not /usr/bin 23/423/7
Michael Robson [Wed, 5 Nov 2014 19:00:40 +0000 (13:00 -0600)]
Feature #574 win: Copy binaries to <build>/bin not /usr/bin

Change-Id: I0e383dbc7b5426bb4ebd1df80923845b5b320deb

6 years agoChanges to handle failures during checkpointing 15/415/4
Xiang Ni [Tue, 28 Oct 2014 04:24:02 +0000 (23:24 -0500)]
Changes to handle failures during checkpointing

This commit enables Charm++ RTS to handle failures that occur during checkpointing: this is achieved by keeping the previous checkpoint around till the new checkpoint is generated on all processors. CmiReduce is used to find the last safe checkpoint on all processors when failures happen.

Change-Id: Ib83de2c5d66b6257809d365b1c617298bb89f9c3

6 years ago#161 verbs, netlrts: Complete rename of 'noprocforcommthread' to 'sleepOnIdle' 77/377/3
Phil Miller [Mon, 29 Sep 2014 22:32:51 +0000 (17:32 -0500)]
#161 verbs, netlrts: Complete rename of 'noprocforcommthread' to 'sleepOnIdle'

Change-Id: I96aeca113f1dcbdfb81264a7e33d151b738c3bbd

6 years agonet & multicore: Print overall execution time at shutdown 76/376/3
Phil Miller [Mon, 29 Sep 2014 16:40:35 +0000 (11:40 -0500)]
net & multicore: Print overall execution time at shutdown

Change-Id: I01b187ac9a242537d518c1fa9f4e3acf0dc8b27b

6 years agoMake examples/collide workable 19/419/1
Phil Miller [Fri, 31 Oct 2014 19:14:32 +0000 (14:14 -0500)]
Make examples/collide workable

Add a top-level Makefile, and use $TESTOPTS in `make test' in the
subdirectory Makefiles.

Change-Id: I6016fa8081b3f251fdcba29a2c23a45f1c0f1d3f

6 years agoWindows unix2nt_cc: Add support for un-defining preprocessor macros 10/410/4
Phil Miller [Tue, 21 Oct 2014 19:07:31 +0000 (14:07 -0500)]
Windows unix2nt_cc: Add support for un-defining preprocessor macros

Change-Id: Id85176dd7c0c2ad9118c53fffa7b31cb44288c7b

6 years agonetlrts & verbs: eliminate usage of CmiState in the machine layer 03/403/6
Bilge Acun [Wed, 15 Oct 2014 15:26:30 +0000 (10:26 -0500)]
netlrts & verbs: eliminate usage of CmiState in the machine layer

CmiState usage is abstracted away from the machine layer by replacing
CmiGetPeGlobal with CmiMyPeGlobal and adding a new function,
CmiIdleLock_hasMessage to check if the state has messages

Change-Id: I772849f6228c95b94cd8f6d8baa5488b99cf11d7

6 years agoSet stack size for QT threads 14/414/3
Ehsan Totoni [Sun, 26 Oct 2014 16:51:19 +0000 (11:51 -0500)]
Set stack size for QT threads

QT code allocates the stack but does not
set the stack size in the base thread data structure.
This was causing a crash in net-linux builds for
thread stack check routine since it uses the stack size.

Change-Id: Ibccaf7b2629fc4150ba3076b975202bc64fe1be7

6 years agosolving issue of thread check code on win c 13/413/2
Ehsan Totoni [Fri, 24 Oct 2014 20:48:46 +0000 (15:48 -0500)]
solving issue of thread check code on win c

Win C compiler follows C90 standard that
prohibits mixing declarations and code.
The code is rearranged to follow those rules.

Change-Id: Ica7c9aa5f4c23619eea6e2138162139ec3332a4d

6 years agoBug #547 charmrun: exit(1) if mixed localhost and non-localhost detected in a nodelist 12/412/5
Michael Robson [Wed, 22 Oct 2014 02:42:06 +0000 (21:42 -0500)]
Bug #547 charmrun: exit(1) if mixed localhost and non-localhost detected in a nodelist

If localhost is used with other hostnames execution of the code will
hang unless the only other hostname used is the current host. This is
because nodes will try to connect to themselves instead of the intended node
due to using the loopback IP discovered from localhost.

Change-Id: Ia3d533eec16489741d58f5712a23c4df4e282ae3

6 years agoAdding sanity check (e.g. stack) in CthSuspend 97/397/4
Ehsan Totoni [Sun, 12 Oct 2014 20:29:32 +0000 (15:29 -0500)]
Adding sanity check (e.g. stack) in CthSuspend

CthSuspend is called after user functions so it is a good place
for checking for data corruption and stack overflow.
This code checks the magic number for corruption,
and checks the current stack pointer to see if it is still
between start and end of the thread's stack.

Change-Id: I22a39436e979085ecb692bdee9e15c1eb14b68ef

6 years agochanges for the new PE on Blue Waters. 11/411/4
Gengbin Zheng [Tue, 21 Oct 2014 19:43:36 +0000 (14:43 -0500)]
changes for the new PE on Blue Waters.

Cray-pmi package does not seem to be configured correctly by pkg-config.
However, pmi issue from the last PE (PMI env vars were not taken by CC
wrapper) is fixed. So pretty much we can discard the fix in conv-mach.sh
for the last PE.

Change-Id: I55b8fa4f20e41cc09770b136940cbaf197229fb5

6 years agonetlrts - verbs: initializing numNodes and myNodeID in LrtsInit 92/392/7
Bilge Acun [Tue, 7 Oct 2014 23:06:56 +0000 (18:06 -0500)]
netlrts - verbs: initializing numNodes and myNodeID in LrtsInit

eliminate usage of _Cmi_numnodes and _Cmi_mynode, replace with machine specific variables

Change-Id: I10ae3e7552a9d770d62f182877c2570c0200f325

6 years agoAdd Testopts to bgtest targets 09/409/2
Nikhil Jain [Tue, 21 Oct 2014 18:53:07 +0000 (13:53 -0500)]
Add Testopts to bgtest targets

Change-Id: Ic396f6f16945680aa52d17968ddafba20e4963eb

7 years agoMove TESTOPTS after binary name, so charmrun doesn't get confused with RTS flags 07/407/1
Phil Miller [Fri, 17 Oct 2014 21:54:00 +0000 (16:54 -0500)]
Move TESTOPTS after binary name, so charmrun doesn't get confused with RTS flags

Change-Id: Id20dace44164d333995160ad7972cd248551fe29