charm.git
9 years agoMinor changes to ZoltanLB
Harshitha [Tue, 21 Feb 2012 22:50:36 +0000 (16:50 -0600)]
Minor changes to ZoltanLB

9 years agoMerge branch 'charm' of charmgit:charm into charm
Harshitha [Tue, 21 Feb 2012 22:40:53 +0000 (16:40 -0600)]
Merge branch 'charm' of charmgit:charm into charm

9 years agoZoltanLB which uses zoltan hyper-graph partitioner. This is a multicast aware load...
Harshitha [Tue, 21 Feb 2012 22:40:47 +0000 (16:40 -0600)]
ZoltanLB which uses zoltan hyper-graph partitioner. This is a multicast aware load balancer

9 years agouse signal with restart
Gengbin Zheng [Tue, 21 Feb 2012 19:09:48 +0000 (13:09 -0600)]
use signal with restart

9 years agono need for signal
Gengbin Zheng [Tue, 21 Feb 2012 19:08:15 +0000 (13:08 -0600)]
no need for signal

9 years agosignal is not needed
Gengbin Zheng [Tue, 21 Feb 2012 18:53:55 +0000 (12:53 -0600)]
signal is not needed

9 years agominor fix
Yanhua Sun [Tue, 21 Feb 2012 19:34:32 +0000 (13:34 -0600)]
minor fix

9 years agoMerge branch 'charm' of charmgit:charm into charm
Yanhua Sun [Tue, 21 Feb 2012 19:19:45 +0000 (13:19 -0600)]
Merge branch 'charm' of charmgit:charm into charm

9 years agofix a bug in worker thread sending messages in smp
Yanhua Sun [Tue, 21 Feb 2012 19:18:14 +0000 (13:18 -0600)]
fix a bug in worker thread sending messages in smp

9 years agominor changes
Gengbin Zheng [Tue, 21 Feb 2012 18:09:53 +0000 (12:09 -0600)]
minor changes

9 years ago--build-shared is not default
Gengbin Zheng [Tue, 21 Feb 2012 16:53:30 +0000 (10:53 -0600)]
--build-shared is not default

9 years agocharmxi: plug another memory leak in the AST nodes for chare classes
Phil Miller [Tue, 21 Feb 2012 08:20:24 +0000 (02:20 -0600)]
charmxi: plug another memory leak in the AST nodes for chare classes

9 years agocharmxi: close a simple memory leak
Phil Miller [Tue, 21 Feb 2012 07:47:59 +0000 (01:47 -0600)]
charmxi: close a simple memory leak

9 years agocharmxi: Factor out duplicated code calling `when' bodies, fixing bugs
Phil Miller [Tue, 21 Feb 2012 07:35:43 +0000 (01:35 -0600)]
charmxi: Factor out duplicated code calling `when' bodies, fixing bugs

The code that unpacked state variables to call the body of SDAG `when'
blocks was essentially duplicated across cases with 1 and more than 1
clause. One branch of the duplication had been modified in cfb58ab1ad
to fix a bug, but not the other one; the consolidation fixes
that. Also address a silly memory leak in the process.

9 years agocharmxi: Remove duplicate assignments
Phil Miller [Tue, 21 Feb 2012 06:01:55 +0000 (00:01 -0600)]
charmxi: Remove duplicate assignments

9 years agoFixed the PCQueueEmpty data-racing bug reported by Jim Philips, and the fix is verifi...
Chao Mei [Tue, 21 Feb 2012 03:49:53 +0000 (21:49 -0600)]
Fixed the PCQueueEmpty data-racing bug reported by Jim Philips, and the fix is verified by Jim Philips

9 years agoAdd support for an arbitrary number of triggers on an sdag statement
Aaron Becker [Tue, 21 Feb 2012 04:02:34 +0000 (22:02 -0600)]
Add support for an arbitrary number of triggers on an sdag statement

9 years agoMerge branch 'charm' of charmgit:charm into charm
Aaron Becker [Tue, 21 Feb 2012 03:33:09 +0000 (21:33 -0600)]
Merge branch 'charm' of charmgit:charm into charm

9 years agoMerge branch 'charm' of charmgit:charm into charm
Aaron Becker [Mon, 20 Feb 2012 22:49:10 +0000 (16:49 -0600)]
Merge branch 'charm' of charmgit:charm into charm

9 years agoleanmd charj example structural code
Aaron Becker [Mon, 20 Feb 2012 22:48:32 +0000 (16:48 -0600)]
leanmd charj example structural code

9 years agoNDMeshStreamer: refactored MeshStreamer into an abstract class;
Lukasz Wesolowski [Mon, 20 Feb 2012 02:21:38 +0000 (20:21 -0600)]
NDMeshStreamer: refactored MeshStreamer into an abstract class;
implemented GroupMeshStreamer - a concrete implementation of MeshStreamer

concrete implementations correspond to the different

9 years agocode cleanup
Gengbin Zheng [Sun, 19 Feb 2012 06:33:19 +0000 (00:33 -0600)]
code cleanup

9 years agocomplete urgent send
Gengbin Zheng [Sun, 19 Feb 2012 06:29:15 +0000 (00:29 -0600)]
complete urgent send

9 years agoadd another sanity check in alloc_mempool_block to prevent a block is allocated too...
Gengbin Zheng [Sat, 18 Feb 2012 19:16:20 +0000 (11:16 -0800)]
add another sanity check in alloc_mempool_block to prevent a block is allocated too big for max mempool size.

9 years agoreduced BIG_MSG threshold to 8MB. If it is 16MB as before, mempool can expand up...
Gengbin Zheng [Sat, 18 Feb 2012 19:01:51 +0000 (11:01 -0800)]
reduced BIG_MSG threshold to 8MB. If it is 16MB as before, mempool can expand up to 32MB, which makes it possible to register for fit in the max mempool threshold.
Add more sanity check about the values and change names of runtime parameters and environment variables.

9 years agoincrease periodicCallInterval aray size
Gengbin Zheng [Sat, 18 Feb 2012 16:38:42 +0000 (08:38 -0800)]
increase periodicCallInterval aray size

9 years agofixed potential bugs with mem register
Yanhua Sun [Sat, 18 Feb 2012 08:30:20 +0000 (02:30 -0600)]
fixed potential bugs with mem register

9 years agoMerge branch 'charm' of charmgit:charm into charm
Yanhua Sun [Sat, 18 Feb 2012 00:10:32 +0000 (18:10 -0600)]
Merge branch 'charm' of charmgit:charm into charm

9 years agomerge first
Yanhua Sun [Sat, 18 Feb 2012 00:10:01 +0000 (18:10 -0600)]
merge first

9 years agoadd urgentSend flag.
Gengbin Zheng [Sat, 18 Feb 2012 00:07:47 +0000 (18:07 -0600)]
add urgentSend flag.

9 years agoMerge branch 'charm' of charmgit:charm into charm
Gengbin Zheng [Sat, 18 Feb 2012 00:02:08 +0000 (18:02 -0600)]
Merge branch 'charm' of charmgit:charm into charm

9 years agoadded a deadlock detection
Gengbin Zheng [Sat, 18 Feb 2012 00:01:39 +0000 (18:01 -0600)]
added a deadlock detection
expot CHARM_UGNI_NO_DEADLOCK_CHECK=yes to disable it

9 years agoskip this sequential test when it is on bgp
Gengbin Zheng [Fri, 17 Feb 2012 23:04:06 +0000 (17:04 -0600)]
skip this sequential test when it is on bgp

9 years agoMerge branch 'charm' of charmgit:charm into charm
Gengbin Zheng [Fri, 17 Feb 2012 22:53:23 +0000 (16:53 -0600)]
Merge branch 'charm' of charmgit:charm into charm

9 years agosome change from Nikhil
Gengbin Zheng [Fri, 17 Feb 2012 22:52:41 +0000 (16:52 -0600)]
some change from Nikhil

9 years agoMerge branch 'charm' of charmgit:charm into charm
Aaron Becker [Fri, 17 Feb 2012 22:07:36 +0000 (16:07 -0600)]
Merge branch 'charm' of charmgit:charm into charm

9 years agopatch to support mpi-bluegenep-xlc-smp mpi-bluegenep-mpcc-smp
Eric Bohm [Fri, 17 Feb 2012 20:12:59 +0000 (20:12 +0000)]
patch to support mpi-bluegenep-xlc-smp mpi-bluegenep-mpcc-smp

9 years agofix numPes possibly not defined with tracing is off
Gengbin Zheng [Fri, 17 Feb 2012 20:10:03 +0000 (14:10 -0600)]
fix numPes possibly not defined with tracing is off

9 years agofix symbol resolving code left in CFGBuilder
Aaron Becker [Fri, 17 Feb 2012 19:56:39 +0000 (13:56 -0600)]
fix symbol resolving code left in CFGBuilder

9 years agofix pup for ckmulticast
Xiang Ni [Fri, 17 Feb 2012 07:19:43 +0000 (01:19 -0600)]
fix pup for ckmulticast

9 years agotake big msg registration into register size
Yanhua Sun [Fri, 17 Feb 2012 07:18:02 +0000 (01:18 -0600)]
take big msg registration into register size

9 years agoMerge branch 'charm' of charmgit:charm into charm
Yanhua Sun [Fri, 17 Feb 2012 06:19:36 +0000 (00:19 -0600)]
Merge branch 'charm' of charmgit:charm into charm

9 years agoin SMP, deregister mem block also search mempool of other ranks.
Gengbin Zheng [Fri, 17 Feb 2012 06:17:55 +0000 (00:17 -0600)]
in SMP, deregister mem block also search mempool of other ranks.

9 years agoMerge branch 'charm' of charmgit:charm into charm
Yanhua Sun [Fri, 17 Feb 2012 06:14:14 +0000 (00:14 -0600)]
Merge branch 'charm' of charmgit:charm into charm

9 years agoMerge branch 'charm' of charmgit:charm into charm
Gengbin Zheng [Fri, 17 Feb 2012 02:35:55 +0000 (20:35 -0600)]
Merge branch 'charm' of charmgit:charm into charm

9 years agomake conservation ugni settting
Yanhua Sun [Fri, 17 Feb 2012 01:18:19 +0000 (19:18 -0600)]
make conservation ugni settting

9 years agoMerge branch 'charm' of charmgit:charm into charm
Aaron Becker [Fri, 17 Feb 2012 00:40:57 +0000 (18:40 -0600)]
Merge branch 'charm' of charmgit:charm into charm

9 years agoNDMeshStreamer: committing the version of the code which performs best
Lukasz Wesolowski [Thu, 16 Feb 2012 22:54:36 +0000 (22:54 +0000)]
NDMeshStreamer: committing the version of the code which performs best
for 16k nodes of RandomAccess on BG/P; caching locations is disabled
by default and flushing sends largest buffer in each dimension

9 years agoadd a placeholder for CMK_SMP_TRACE_COMMTHREAD
Gengbin Zheng [Thu, 16 Feb 2012 17:52:25 +0000 (11:52 -0600)]
add a placeholder for CMK_SMP_TRACE_COMMTHREAD

9 years agocode clean up, most indentation.
Gengbin Zheng [Thu, 16 Feb 2012 04:33:35 +0000 (22:33 -0600)]
code clean up, most indentation.

9 years agochange of data structure in gni-rc-status
Yanhua Sun [Wed, 15 Feb 2012 21:54:17 +0000 (13:54 -0800)]
change of data structure in gni-rc-status

9 years agospeedup sendbuffersmsg sendrdma
Yanhua Sun [Wed, 15 Feb 2012 21:26:50 +0000 (13:26 -0800)]
speedup sendbuffersmsg sendrdma

9 years agoMerge branch 'charm' of charmgit:charm into charm
Yanhua Sun [Wed, 15 Feb 2012 19:39:14 +0000 (11:39 -0800)]
Merge branch 'charm' of charmgit:charm into charm

9 years agocode clean up
Yanhua Sun [Wed, 15 Feb 2012 19:39:03 +0000 (11:39 -0800)]
code clean up

9 years agoMerge branch 'charm' of charmgit:charm into charm
Gengbin Zheng [Wed, 15 Feb 2012 15:33:12 +0000 (07:33 -0800)]
Merge branch 'charm' of charmgit:charm into charm

9 years agoconfig files for inmem checkpointing
Gengbin Zheng [Wed, 15 Feb 2012 15:32:10 +0000 (07:32 -0800)]
config files for inmem checkpointing

9 years agofixed a smp mempool register bug
Yanhua Sun [Wed, 15 Feb 2012 06:06:01 +0000 (22:06 -0800)]
fixed a smp mempool register bug

9 years agoNODE_CHECKPOINT now works for SMP
Gengbin Zheng [Wed, 15 Feb 2012 05:27:16 +0000 (21:27 -0800)]
NODE_CHECKPOINT now works for SMP

9 years agoblock_header adds a pointer to the very first block (i.e. the mempool ptr)
Gengbin Zheng [Wed, 15 Feb 2012 05:10:58 +0000 (21:10 -0800)]
block_header adds a pointer to the very first block (i.e. the mempool ptr)

9 years agoadd dependence to machine.h
Gengbin Zheng [Wed, 15 Feb 2012 05:10:30 +0000 (21:10 -0800)]
add dependence to machine.h

9 years agosome sanity check of env variables settings
Gengbin Zheng [Tue, 14 Feb 2012 21:10:22 +0000 (13:10 -0800)]
some sanity check of env variables settings

9 years agouse -d option which is needed for SMP
Gengbin Zheng [Tue, 14 Feb 2012 19:27:23 +0000 (11:27 -0800)]
use -d option which is needed for SMP

9 years agoimplement urgent send queue, default off
Gengbin Zheng [Tue, 14 Feb 2012 07:00:11 +0000 (01:00 -0600)]
implement urgent send queue, default off

9 years agoremove some broken MACHSTATE
Gengbin Zheng [Tue, 14 Feb 2012 06:23:52 +0000 (00:23 -0600)]
remove some broken MACHSTATE

9 years agofix a bug in SendRdmaMsg that double pop the queue
Gengbin Zheng [Tue, 14 Feb 2012 06:13:35 +0000 (00:13 -0600)]
fix a bug in SendRdmaMsg that double pop the queue

9 years agoin SMP case make sure SendBufferMsg loop all buffered message
Gengbin Zheng [Tue, 14 Feb 2012 06:09:30 +0000 (00:09 -0600)]
in SMP case make sure SendBufferMsg loop all buffered message

9 years agoYanhua's change to impose send flow control and max mempool cap
Gengbin Zheng [Tue, 14 Feb 2012 05:55:44 +0000 (23:55 -0600)]
Yanhua's change to impose send flow control and max mempool cap

9 years agochange CmiReadSize to return long type
Gengbin Zheng [Tue, 14 Feb 2012 05:47:26 +0000 (23:47 -0600)]
change CmiReadSize to return long type

9 years agoadd dummy functions used for tracing comm thread so that converse programs can compile.
Gengbin Zheng [Tue, 14 Feb 2012 05:39:35 +0000 (23:39 -0600)]
add dummy functions used for tracing comm thread so that converse programs can compile.

9 years agoat restart, restore broadcast spring cleaner
Gengbin Zheng [Mon, 13 Feb 2012 03:39:49 +0000 (21:39 -0600)]
at restart, restore broadcast spring cleaner

9 years agoadded RECV_MAX to limit the maximum register memory for recv
Yanhua Sun [Mon, 13 Feb 2012 02:06:19 +0000 (20:06 -0600)]
added RECV_MAX to limit the maximum register memory for recv

9 years agochange in mempool for send/recv msgs in transactions
Yanhua Sun [Mon, 13 Feb 2012 00:31:41 +0000 (18:31 -0600)]
change in mempool for send/recv msgs in transactions

9 years agoMerge branch 'charm' of charmgit:charm into charm
Yanhua Sun [Mon, 13 Feb 2012 00:18:29 +0000 (18:18 -0600)]
Merge branch 'charm' of charmgit:charm into charm

9 years agocombine two functions, fix a bug of missing rdma msg
Yanhua Sun [Mon, 13 Feb 2012 00:18:23 +0000 (18:18 -0600)]
combine two functions, fix a bug of missing rdma msg

9 years agotrace functions for comm threads always defined now for consistence.
Gengbin Zheng [Sun, 12 Feb 2012 23:18:36 +0000 (17:18 -0600)]
trace functions for comm threads always defined now for consistence.

9 years agofixed a crash bug when npes == 1 in SMP
Gengbin Zheng [Sun, 12 Feb 2012 18:01:19 +0000 (10:01 -0800)]
fixed a crash bug when npes == 1 in SMP

9 years agoNDMeshStreamer performance optimizations and bug fixes:
Lukasz Wesolowski [Fri, 10 Feb 2012 02:00:53 +0000 (20:00 -0600)]
NDMeshStreamer performance optimizations and bug fixes:
(1) only compute the index information necessary for sending
in the current step
(2) process items immediately at intermediate steps when intermediate
node turns out to be the final destination
(3) fixed a bug in the location caching scheme
(4) code cleanup

9 years agosolve the memory leak after restart
Xiang Ni [Thu, 9 Feb 2012 17:08:31 +0000 (11:08 -0600)]
solve the memory leak after restart

9 years agominor
Yanhua Sun [Thu, 9 Feb 2012 05:47:22 +0000 (23:47 -0600)]
minor

9 years agoMerge branch 'charm' of charmgit:charm into charm
Eric Bohm [Wed, 8 Feb 2012 18:54:01 +0000 (12:54 -0600)]
Merge branch 'charm' of charmgit:charm into charm

9 years agoinit maxfiles to number of physnodes
Eric Bohm [Wed, 8 Feb 2012 18:53:37 +0000 (12:53 -0600)]
init maxfiles to number of physnodes

9 years agoMerge branch 'charm' of charmgit:charm into charm
Aaron Becker [Wed, 8 Feb 2012 17:35:04 +0000 (11:35 -0600)]
Merge branch 'charm' of charmgit:charm into charm

9 years agoreverted to its original state before the latest wave of changes
Gengbin Zheng [Wed, 8 Feb 2012 06:51:54 +0000 (00:51 -0600)]
reverted to its original state before the latest wave of changes

9 years agoKeep old interface of Cmidirect and added gemini cmidirect
Yanhua Sun [Wed, 8 Feb 2012 06:18:31 +0000 (00:18 -0600)]
Keep old interface of Cmidirect and added gemini cmidirect

9 years agorevert the pingpong using cmi_direct
Yanhua Sun [Wed, 8 Feb 2012 05:34:09 +0000 (23:34 -0600)]
revert the pingpong using cmi_direct

9 years agoMerge remote-tracking branch 'origin/pario' into charm
Phil Miller [Wed, 8 Feb 2012 01:22:55 +0000 (19:22 -0600)]
Merge remote-tracking branch 'origin/pario' into charm

9 years agodescribes what the example is for
Eric Bohm [Wed, 8 Feb 2012 00:28:34 +0000 (18:28 -0600)]
describes what the example is for

9 years agoMerge branch 'charm' of charmgit:charm into charm
Eric Bohm [Wed, 8 Feb 2012 00:24:33 +0000 (18:24 -0600)]
Merge branch 'charm' of charmgit:charm into charm

9 years agoAdd map object for pupdisk to distribute them to one per node when possible
Eric Bohm [Wed, 8 Feb 2012 00:23:55 +0000 (18:23 -0600)]
Add map object for pupdisk to distribute them to one per node when possible

9 years agoMeshStreamer and NDMeshStreamer: default to a single buffer per destination
Lukasz Wesolowski [Tue, 7 Feb 2012 23:00:41 +0000 (17:00 -0600)]
MeshStreamer and NDMeshStreamer: default to a single buffer per destination
when first argument to constructor is too small or invalid.

9 years agoadded missing cmidirect.c
Yanhua Sun [Tue, 7 Feb 2012 22:41:38 +0000 (14:41 -0800)]
added missing cmidirect.c

9 years agoMerge branch 'charm' of charmgit:charm into charm
Yanhua Sun [Tue, 7 Feb 2012 21:20:39 +0000 (13:20 -0800)]
Merge branch 'charm' of charmgit:charm into charm

9 years agoadded CmiDirect in Gemini
Yanhua Sun [Tue, 7 Feb 2012 21:20:22 +0000 (13:20 -0800)]
added CmiDirect in Gemini

9 years agominor tweaks on print messages
Gengbin Zheng [Tue, 7 Feb 2012 20:23:21 +0000 (14:23 -0600)]
minor tweaks on print messages

9 years agoadded a new pattern for repeating: e.g 2X0-4
Gengbin Zheng [Tue, 7 Feb 2012 05:35:28 +0000 (23:35 -0600)]
added a new pattern for repeating:  e.g 2X0-4
For Cray, with +setcpuaffinity and no pemap and commap, it will compute a map for comm threads

9 years agominor tweak on print
Gengbin Zheng [Tue, 7 Feb 2012 04:49:40 +0000 (22:49 -0600)]
minor tweak on print

9 years agoworking version. needs small amount of placing tweaking to put
Eric Bohm [Tue, 7 Feb 2012 01:58:02 +0000 (19:58 -0600)]
working version.  needs small amount of placing tweaking to put
one element per node

9 years agofix an invalid pointer reference error whern npes = 1
Gengbin Zheng [Mon, 6 Feb 2012 15:57:31 +0000 (09:57 -0600)]
fix an invalid pointer reference error whern npes = 1

9 years agominor celan up code, allow using a link list instead of looping sendQ.
Gengbin Zheng [Mon, 6 Feb 2012 06:53:09 +0000 (00:53 -0600)]
minor celan up code, allow using a link list instead of looping sendQ.

9 years agonew env "CHARM_UGNI_MEMPOOL_SIZE"
Gengbin Zheng [Sat, 4 Feb 2012 19:33:30 +0000 (13:33 -0600)]
new env "CHARM_UGNI_MEMPOOL_SIZE"