Project

General

Profile

Activity

From 01/31/2013 to 03/01/2013

03/02/2013

11:38 PM Revision 29ac5c2f (charm): BGQ: Bug fix for limiting memory allocated on BGQ
Nikhil Jain
07:19 PM Cleanup #64 (Closed): remove archaic entries from FAQ
Eric Bohm
07:15 PM Revision 42fa749f (charm): A whole lot of FAQ updating
Eric Bohm
06:23 PM Feature #67: Reduce the impact of large messages sent to unlocated or migrated chare array elements
On systems that transmit even large messages directly, e.g. net-*, the explicit latency is only improved by buffering... Phil Miller
06:14 PM Feature #69 (New): Get large messages sent via the ZC API to be received directly using a GET fro...
On systems that communicate large messages via RDMA GET operations, it takes the latency of one fewer packet across t... Phil Miller

03/01/2013

05:50 PM Feature #68 (Closed): LRTS support for setting up a message to send and transmitting a GET handle
On platforms that use GET-based RDMA for large messages, there are various circumstances where it is useful to be abl... Phil Miller
05:39 PM Revision 10afaf6f (charm): use optimized memory allocator
Yanhua Sun
05:34 PM Feature #67 (Merged): Reduce the impact of large messages sent to unlocated or migrated chare arr...
Currently, any message sent to a chare array element (the *target*) whose location is uncached or has migrated is ref... Phil Miller
05:22 PM Feature #66 (New): Execution within a power budget
Given a power budget in terms of watts, the runtime should be able to adaptively adjust resource usage (e.g. DVFS) to... Phil Miller
05:15 PM Feature #65 (Merged): MetaLB: Separate cost/benefit analysis for running LB strategy and performi...
MetaLB potentially has two separate decisions to make as execution proceeds:
1. Is current load imbalance bad enough...
Phil Miller
05:06 PM Cleanup #64 (Closed): remove archaic entries from FAQ
Eric Bohm
04:40 PM Feature #28: Replica based FT
Make sure that the changes are reviewed by the core group before merge. Eric Bohm
04:25 PM Support #51: Regular performance benchmarks for AMPI
IMB
Mantevo
NAS
Eric Bohm
04:23 PM Bug #63 (Merged): megacon is leaking message memory
Clean up and correct megacon's handling of memory. where is the bug? Eric Bohm
04:07 PM Cleanup #62 (Closed): missing args to main
Add argument message to main in test and examples Eric Bohm
03:26 PM Bug #54: test/converse/machinetest fails in net smp
Ask senior person for help as needed. It is possible that the test itself is obsolete. Eric Bohm
03:17 PM Bug #61 (Closed): PAMI NAMD Exit issue
Test that PAMI NAMD exits cleanly on PAMI. Eric Bohm
02:27 PM Revision 294809fc (charm): Merge branch 'charm' of charmgit:charm into charm
Nikhil Jain
02:27 PM Revision bfd82d33 (charm): BGQ: Correction in queue creation
Nikhil Jain
02:18 PM Revision 083ee876 (charm): added support for obj ref and obj val type..
Bilge Acun
01:28 PM Bug #3: Provide full topology information in partititioned runs on BGQ
Nikhil, please document which platforms are affected by this and to what extent. It may make the most sense to open a... Phil Miller
10:58 AM Feature #60 (New): Exploit phase boundaries in iterative apps
https://wiki.engr.illinois.edu/x/TQDWCg
Many parallel programs exhibit phased behaviors, with differing patterns o...
Ramprasad Venkataraman
09:47 AM Revision 85ac4d97 (charm): locmgr: minor reindent
Ramprasad Venkataraman
09:47 AM Revision a84fc3a9 (charm): locmgr: remove unnecessary casts to CkArrayIndex
Ramprasad Venkataraman
03:27 AM Revision 0dbb02d1 (charm): BGQ: bug fix in L2Atomics
Nikhil Jain
03:12 AM Revision 51d6682f (charm): BGQ: bug fix in previous commit, order of parameters was reversed
Nikhil Jain
11:05 PM Revision 89ec6098 (charm): define another macro to protect intra-node migration optimiZation
Yanhua Sun

02/28/2013

10:50 AM Revision 625375c9 (charm): Fixing a bug. The bug was due to rec being defined again.
Harshitha Menon
09:07 PM Revision dd2a927b (charm): BGQ: modified converse exit for correctness
Sameer Kumar
08:52 PM Revision cce8c68c (charm): BGQ: modify charmrun for BGQ to accept exit status 1 as a successful e...
BG convention Nikhil Jain
08:34 PM Revision a314ffeb (charm): setting magic number in pup of chare for the intra node transfer case
Harshitha Menon
06:14 PM Revision f257459e (charm): Merge branch 'intra-node-balancing' of charmgit:charm into intra-node-...
Harshitha Menon
06:13 PM Revision 3e9b2b32 (charm): Fixing a Todo which was to use dummy PUP er only when the INTRA_NODE_B...
Harshitha Menon

02/27/2013

05:28 PM Revision 1275b7dd (charm): fix a bug to obtain scheduler queue length
Yanhua Sun
03:37 PM Revision 39cbb305 (charm): Using dummy PUP er for intra node transfer
Harshitha Menon
01:08 PM Revision ae864fd7 (charm): clear srcPe field which is a separate field when tracing comm thread.
Gengbin Zheng
10:20 PM Bug #59 (Merged): Support out-of-tree builds
--destination is buggy. --prefix should be supported. Nikhil Jain
10:19 PM Bug #58 (Closed): NAMD crashes if L2Atomics are enabled with PAMI LRTS
Nikhil Jain
10:17 PM Bug #56 (Closed): Shared build does not compile
Osman renamed the functions, and its resolved now. Nikhil Jain
09:41 PM Revision 009e998e (charm): converse: defining inter partition sends to intra if no support for re...
Nikhil Jain
06:54 PM Revision 002d1ad0 (charm): For the intra node load balancing scheme, avoid pup and instead just t...
Harshitha Menon

02/26/2013

04:49 PM Revision 7607a1a4 (charm): disable L2AtomicQueue
Yanhua Sun
01:37 PM Revision debc72f8 (charm): coordination library: useful in iterative codes where objects 'check-i...
This library ensures that progress is
made even when a certain PE doesn't have any resident objects. Used by ChaNGa.
Pritish Jetley
01:31 PM Revision e3549d4f (charm): cache: removed references to file CkCachePointerContainer.h
Pritish Jetley
01:28 PM Revision 7d0725bd (charm): cache: added data structure that wraps around osl's hashtable
Pritish Jetley
01:27 PM Revision 1c90aafc (charm): cache: added orion's threadsafe hashtable
Pritish Jetley
01:25 PM Revision 66ada090 (charm): code for smp cache
Pritish Jetley
11:56 AM Revision 0450b156 (charm): does not have to print log flushing warning when none happens.
Gengbin Zheng
10:32 AM Revision ebbab1e1 (charm): check whether object is ready to migrate
Yanhua Sun
10:00 AM Revision 844c5c5e (charm): check whether object is ready to migrate
Yanhua Sun
04:10 AM Revision d9b6dc20 (charm): BGQ: Remove the env to set fifo size if using charmrun
Nikhil Jain
03:54 AM Revision 232eefa6 (charm): BGQ: explorations to tackle the bug in RZV mode
Sameer Kumar
03:52 AM Revision 707e1dc8 (charm): BGQ: misc commits, sets some parameters
Sameer Kumar
03:50 AM Revision aac1a19e (charm): BGQ: Adding a slow spin wait barrier for exit
Nikhil Jain
03:49 AM Revision e385d78b (charm): BGQ: Adding blocking atomic queue
Sameer Kumar
03:48 AM Revision 0e2ca6d5 (charm): BGQ: Adding functions to be used when NodeQ is available
Sameer Kumar
03:46 AM Revision ab2f7fe6 (charm): BGQ: Adding L2Atomic mutex
Sameer Kumar
02:44 AM Revision 4dbd3329 (charm): Since this change did not fix the issues as expected, I am reverting it
Revert "BGQ: changing the way context is selected for send - I found this to be useful"
This reverts commit 57b64f4b...
Nikhil Jain
09:39 PM Revision d63d290f (charm): Minor problem fixed: variable and method name clash between TempAwareC...
Osman Sarood

02/25/2013

05:41 PM Revision a7e02263 (charm): doc: use macro \charm instead of explicitly spelling Charm++
Ramprasad Venkataraman
05:32 PM Revision b46cc8ee (charm): Revert "docs: Define all term macros with a trailing xspace, so that w...
This reverts commit 53ddc227a8d01ac5deae1cdaab739d26ff0740b9.
This change causes trouble with latex2html, because it ...
Ramprasad Venkataraman
05:32 PM Revision 723411e6 (charm): doc: manually insert xspace after each newcommand
Ramprasad Venkataraman
04:03 PM Bug #57 (In Progress): problem with mempool for AMPI programs
Nikhil Jain
12:03 PM Bug #57 (Closed): problem with mempool for AMPI programs
I believe that there is a problem with mempool in Charm++ and it's affecting AMPI programs in some cases.
For curr...
Celso Mendes
03:51 PM Bug #39 (Closed): PAMI FIFO Size bug for large messages
Added the warning to src code. Added the note in release notes. Nikhil Jain
03:44 PM Revision c07d8d90 (charm): add -print-prog-name= to mimic gcc.
Gengbin Zheng
11:26 AM Revision 6081cd84 (charm): add simple hello to test intra-node balance
Yanhua Sun
11:23 AM Revision 4102fefe (charm): modify for intra-node balancing
Yanhua Sun
12:04 AM Revision 5b0a82ea (charm): doc: correction in +sumonly option
Nikhil Jain
11:46 PM Revision c505fb23 (charm): doc: add documentation on partitioning
Nikhil Jain
11:45 PM Revision b7eb904b (charm): doc: adding pamilrts as an example to install
Nikhil Jain
06:50 PM Revision 57b64f4b (charm): BGQ: changing the way context is selected for send - I found this to b...
for stablizing pamilrts, hence adding it for PAMI untill the issue is resolved. Nikhil Jain
06:47 PM Revision efd0809b (charm): release: updating the CHANGES for 6.5 release with partitioning and
PAMI Nikhil Jain
06:39 PM Revision c05d34f4 (charm): BGQ: Adding detailed warning when we suspect that low level buffers ma...
Nikhil Jain

02/23/2013

10:22 PM Revision ff61f9e3 (charm): Scalable-start for SMP version - patch suggested by Jim
Abhishek Gupta

02/22/2013

05:21 PM Bug #56: Shared build does not compile
I mean addition of temperature aware LB has prevented successful compilation of shared library. Nikhil Jain
03:44 PM Bug #56 (Closed): Shared build does not compile
Adding a new load balancing prevents charm compilation. Nikhil Jain
04:28 PM Revision f4c3be29 (charm): bug fix due to return CmiTrue
Yanhua Sun
03:50 PM Revision d7366ee4 (charm): basic framwork without prefetch works
Xiang Ni
03:25 PM Revision 7a1c4c2c (charm): Charj: more of less emitting refs correctly, old syntax still needs to...
Bilge Acun
02:54 PM Feature #29 (In Progress): Reduction Starting messages
A first cut implementation of the scheme is implemented. Tested for the case with no local chares on any PE.
The ...
Harshitha Menon
02:30 PM Revision e25f57f3 (charm): Charj: more on ref types..
Bilge Acun
11:16 AM Revision a04d35f9 (charm): smp fix for cpv
Yanhua Sun
11:14 AM Revision 7ff399a8 (charm): minor fix for i
Yanhua Sun
08:57 PM Revision 6349e441 (charm): keep the original msg and push to other PEs
Yanhua Sun
07:18 PM Revision dc6fcad8 (charm): add work stealing for intra node load
Yanhua Sun
06:43 PM Revision 0291ab8f (charm): init for out of core
Xiang Ni

02/21/2013

11:59 AM Revision e1888a43 (charm): doc: stop building ps versions of manuals
Ramprasad Venkataraman
11:58 AM Revision 1717cbe1 (charm): Ref example update
Bilge Acun
11:51 AM Revision 00a0522f (charm): doc: retain backup manual copies only if make web fails
Ramprasad Venkataraman
09:47 AM Support #55: update webpage /help to reflect new urls for fem, ifem, netfem and mblock manuals
Ramprasad Venkataraman wrote:
> These dormant manuals are now under /www/manuals/old/*
* The charisma manual has ...
Ramprasad Venkataraman

02/20/2013

05:02 PM Revision c57305de (charm): ref type parsing more..
Bilge Acun
04:24 PM Support #55 (Closed): update webpage /help to reflect new urls for fem, ifem, netfem and mblock m...
These dormant manuals are now under /www/manuals/old/* Ramprasad Venkataraman
04:21 PM Revision c627dffc (charm): doc: make web shouldn't look for nonexistent manual.css
Ramprasad Venkataraman
04:10 PM Support #36 (Resolved): Nightly manual rebuild
Old, dormant manuals like fem, ifem, netfem, mblock will not be built everyday. Instead they're served out of a persi... Ramprasad Venkataraman
03:57 PM Revision 81250e8d (charm): doc: dont build charisma manual until its updated
Ramprasad Venkataraman
03:57 PM Revision 42acacfb (charm): doc: do not build fem, ifem, netfem, multiblock manuals every time
Ramprasad Venkataraman
03:57 PM Revision 2b656a39 (charm): doc: nuke manuals for dead libraries - irecv, tempo, ddt, barrier
Ramprasad Venkataraman
03:57 PM Revision ab3b3547 (charm): doc: nuke some ifpdf clauses from dormant f90 bindings manual
Ramprasad Venkataraman
03:57 PM Revision 86be07c8 (charm): doc: nuke empty __ck__ manual
Ramprasad Venkataraman
02:57 PM Revision 6d8ae3ac (charm): Removing an echo that get left in the previous commit
Nikhil Jain
03:49 AM Revision c3db1abc (charm): Fix the bug in Makefile that prevented building ROM-IO
Nikhil Jain
12:41 AM Bug #54 (Merged): test/converse/machinetest fails in net smp
tests/converse/machinetest example gives segmentation violation for smp version in net build. In non smp build, it is... Bilge Acun

02/19/2013

01:39 PM Support #35: Release notes for 6.5.0
Release notes look fine to me. Eric Bohm
08:05 PM Support #35 (Resolved): Release notes for 6.5.0
Eric, please check that I haven't misrepresented anything, or left out some entry that should have been in the wiki b... Phil Miller
11:37 AM Revision 4bf6a728 (charm): cleanup
Bilge Acun
11:15 AM Revision f2a80f31 (charm): Fix old typo, from CmiDestoryLocks to CmiDestroyLocks
Bilge Acun
08:21 PM Bug #1 (Closed): Document template entities better
Phil Miller
08:04 PM Revision 4e32aa67 (charm): Add release notes for version 6.5.0
Phil Miller
08:04 PM Revision 66b97370 (charm): megatest: Use newer CkIndex syntax
Phil Miller

02/18/2013

05:53 PM Bug #53 (New): SDAG case constructs and Bigsim traces
Inspired by some other discussion, I went to look at the implementation of SDAG case clauses to see how they'd intera... Phil Miller
01:22 PM Cleanup #48: Factor out SdagConstruct types into separate classes
To give a little more guidance on this, there are a number of methods of SdagConstruct that switch on what type of co... Phil Miller
01:13 PM Feature #47: Accurate file/line numbers in emitted SDAG atomic blocks
This will probably be a bit easier once at least the 'atomic' portion of #48 is done. Phil Miller
01:04 PM Revision d4fb59dd (charm): Revert "Sameer's patch for PAMI issues and improvements"
This reverts commit 6f23cc04b0ba98d78b3e584f14395a091b3ec174. Nikhil Jain
09:15 PM Feature #49: Define and document semantics of dynamic Insertion in Chare arrays wrt broadcasts an...
So, the meeting on this happened, and it seemed that the consensus settled on "application developers are responsible... Phil Miller
06:15 PM Revision 70055be2 (charm): Charj: new Ref example added
Bilge Acun
06:14 PM Revision 5821fa2e (charm): Charj: pointer syntax inital parsing done
Bilge Acun

02/17/2013

12:29 AM Revision 6f23cc04 (charm): Sameer's patch for PAMI issues and improvements
Sameer Kumar
06:35 PM Revision 32975f66 (charm): A possible fix for the error seen in which program crashes during CmiS...
Nikhil Jain

02/16/2013

01:39 PM Bug #32 (Closed): PAMI LRTS SMP with L2Atomic Queue crashed
Phil Miller
06:04 PM Revision 0c695a3d (charm): some mofidicatioins
Bilge Acun

02/15/2013

05:49 PM Bug #50 (Closed): isomalloc problem under Cray compiler
Nikhil Jain
05:07 PM Bug #1: Document template entities better
Please review the revised template documentation and close the issue if it looks good. Phil Miller
05:07 PM Bug #1 (Resolved): Document template entities better
Rewritten and pushed. Phil Miller
04:46 PM Revision b488003e (charm): Manual: update template documentation to describe entry method templat...
Phil Miller
01:44 PM Revision bb5bd02f (charm): make these functions static inline to avoid multiple definition collis...
Eric Bohm
11:30 AM Revision 5b4909ee (charm): lrts exit function fix
Bilge Acun

02/14/2013

04:16 PM Revision 40e6aef6 (charm): update minor version number
Eric Bohm
04:15 PM Revision c961676d (charm): configure time detection of Cray BIConfig API
Eric Bohm
03:43 PM Support #52 (Closed): Build and test Charm++ applications as part of regular testing
Phil Miller
03:42 PM Support #51 (New): Regular performance benchmarks for AMPI
Phil Miller
12:47 PM Bug #50: isomalloc problem under Cray compiler
As suggested by Gengbin, adding "-memory os-isomalloc -dynamic" to the build line causes the code to build and run fi... Celso Mendes
10:40 AM Bug #50 (Closed): isomalloc problem under Cray compiler
I'm facing a problem when using isomalloc under an AMPI build with the Cray compiler on Blue Waters.
After buildin...
Celso Mendes
12:37 AM Feature #49 (New): Define and document semantics of dynamic Insertion in Chare arrays wrt broadca...
Define the semantics of dynamic insertion, make modifications in Charm to adhere to those semantics, and document it. Nikhil Jain
09:35 PM Cleanup #48 (Merged): Factor out SdagConstruct types into separate classes
The interface translator charmxi has a class SdagConstruct that is used to represent all kinds of entities in the SDA... Phil Miller

02/13/2013

03:21 PM Revision 2870b5d7 (charm): srcpe is charm envelope is doubled used when tracing communication thr...
Gengbin Zheng
12:02 PM Feature #47 (Merged): Accurate file/line numbers in emitted SDAG atomic blocks
We should emit #line directives with our generated SDAG code when it contains bits written by application developers,... Phil Miller
11:29 AM Bug #39: PAMI FIFO Size bug for large messages
Descheduling from 6.5 release because the pamilrts build seems to be sufficiently stable Phil Miller
10:51 AM Revision fa9f3cd4 (charm): add notification when objects safely migrated for proactive fault tole...
Xiang Ni
09:11 PM Revision e47531a5 (charm): Using 5s as failure detection timeout.
Esteban Meneses

02/12/2013

05:13 PM Revision 328e2b7e (charm): Adding padding to envelope in causal message-logging to make xlc work....
Esteban Meneses
12:06 PM Revision a8f03381 (charm): adding missing test files
Bilge Acun
09:12 PM Revision a965abb9 (charm): Adding print to indicate message-logging protocol.
Esteban Meneses

02/11/2013

04:38 PM Revision 22d74fd8 (charm): clean up
Yanhua Sun
04:27 PM Revision 23d330ea (charm): Merge branch 'net-lrts' of charmgit:charm into net-lrts
Bilge Acun
04:26 PM Revision a4185bf9 (charm): bug fix in CmiMachineProgress function
Bilge Acun
04:24 PM Feature #23: Optimization of MPI layer
Quoting Ralf's Email on this-
As for MPI, we've found something that apparently makes it be as fast as gemini_gni-...
Nikhil Jain
04:24 PM Support #36: Nightly manual rebuild
Moved the cron job to panache. It appears to have run as scheduled, but does not complete due to various errors when ... Lukasz Wesolowski
03:54 PM Bug #45 (Closed): examples/bigsim/emulator: Parallel build copies charmrun multiple times
Phil Miller
09:44 AM Bug #45 (Resolved): examples/bigsim/emulator: Parallel build copies charmrun multiple times
Phil Miller
01:01 AM Bug #45 (Closed): examples/bigsim/emulator: Parallel build copies charmrun multiple times
When multiple linking steps run in parallel, they will each default to copying charmrun into the build directory. In ... Phil Miller
03:50 PM Bug #46 (Closed): Cray builds try to compile 32-bit versions of GOT global swapping
The test in global-elfgot.C for 32/64 bit type definitions was based on the CMK_AMD64 macro that was set for several ... Phil Miller
03:35 PM Revision de6114a1 (charm): Apply to mpi-crayxe same fix from Phil in gemini_gni-crayxe, to fix co...
Celso Mendes
03:26 PM Revision ba31ec6c (charm): remove pxshm vshm from the net specific code since it is already in co...
Yanhua Sun
03:16 PM Bug #43 (Closed): Incorrect result on BGQ
Nikhil Jain
03:15 PM Bug #43: Incorrect result on BGQ
instead_test.sh used on BGQ had a bug - after _make test_ was called, _status = $?_ was used to store the status, and... Nikhil Jain
02:41 PM Revision 46cc541f (charm): split persistent functions from machine.c
Yanhua Sun
02:35 PM Revision a508d808 (charm): tcharm: fix Windows linking error from missing threadpool extern decla...
Phil Miller
02:34 PM Revision cf43f5e7 (charm): send function cleanup
Bilge Acun
01:47 PM Revision 942a026a (charm): minor changes
Bilge Acun
12:17 PM Revision 7e6c72ce (charm): gemini_gni-crayxe: define CMK_AMD64 to fix compilation error in global...
Phil Miller
11:30 AM Revision d583a5fd (charm): gemini_gni: typo fix limite -> limit
Phil Miller
10:51 AM Revision fef4064f (charm): MSA Histogram example: Distribute number of worker threads to allow an...
Phil Miller
10:48 AM Revision e887c3b9 (charm): MSA Histogram example: End output with a newline
Phil Miller
10:48 AM Revision 965cc6a0 (charm): MSA Histogram example: insert missing variable declaration
Phil Miller
12:34 AM Revision b1b8f51d (charm): examples/bigsim/emulator/Makefile: Don't be so fancy with target-speci...
Phil Miller

02/10/2013

09:48 AM Support #44 (New): Macports / Homebrew installation scripts
Write a Macport script to ease installation for users of Mac OS X. Phil Miller
10:21 PM Bug #43: Incorrect result on BGQ
Nikhil added return code from runjob command and fixed a bug in charmrun for cases in which a job is submitted outsid... Nikhil Jain
10:18 PM Bug #43 (Closed): Incorrect result on BGQ
Autobuild shows result as good even when the tests are failing on BGQ. Nikhil Jain
10:05 PM Bug #15 (Closed): Non-isomalloc threads
Nikhil Jain

02/09/2013

08:19 AM Revision dafc4aa5 (charm): minor changes
Bilge Acun
03:23 AM Revision 2c852028 (charm): Correcting charmrun for all BGQ target for cases where job is submitte...
a script Nikhil Jain
02:05 AM Revision c7c8ab9b (charm): Charmrun in mpi-bluegeneq was outdated (from BGL time), using the new one
Nikhil Jain
02:03 AM Revision 2a0fd776 (charm): Adding return status for charmrun of BGQ when using runjob
Nikhil Jain

02/08/2013

05:24 PM Revision 629d6993 (charm): initial commit
Bilge Acun
04:57 PM Bug #15 (Resolved): Non-isomalloc threads
Added fix to the files; works with memory-aliasing on simple jacobi example. Will wait for autobuild to get it mark c... Nikhil Jain
04:48 PM Revision 58feeaaf (charm): Bug fix for mempool based isomalloc - correction for cases that use a ...
mechanism for stack Nikhil Jain
04:43 PM Feature #30 (Closed): MPI Interoperability
Phil Miller
04:31 PM Feature #38: Verify that compress works with envelope reduction branch
Could we get a test against (your suitably modified version of) NAMD, too, since that actually intends to use this? Phil Miller
04:13 PM Feature #38 (Resolved): Verify that compress works with envelope reduction branch
I tested persistent pingpong with message reduction and compression. It works fine Yanhua Sun
03:18 PM Feature #38 (Closed): Verify that compress works with envelope reduction branch
Consult Lukasz regarding envelope reduction branch.
test merge with compression
g/envelope-reduction
Eric Bohm
03:40 PM Bug #42 (Closed): Large TLS segments cause crashes on Cray XE (exposed by AMPI using TLS for priv...
Memory incompatibility with large TLS > 64k.
Bug reported to Cray and being worked on by them.
From email:
"U...
Eric Bohm
03:31 PM Feature #41 (New): LRTS Machine layer flow control
Design general scheme to manage flow control for messages. Eric Bohm
03:30 PM Bug #40 (Closed): Report PAMI large/many msgs issue to IBM

Investigate possibility of a check with appropriate error condition to be emitted from SPI or PAMI.
Submit small e...
Eric Bohm
03:29 PM Bug #39 (Closed): PAMI FIFO Size bug for large messages
Connected to rendezvous issues, but can occur independently.
Workaround is to increase environment variable,
f...
Eric Bohm
03:24 PM Revision 456bfa67 (charm): Merge branch 'charm' into gzheng/envelope-reduction
Lukasz Wesolowski
03:14 PM Bug #37 (Merged): Metabalancer's use of immediate messages causes hangs with QD
Replicate this bug with current charm, or show that its fixed and close. Eric Bohm
03:10 PM Feature #31 (Closed): AMPI
Eric Bohm
03:09 PM Support #36: Nightly manual rebuild
Make cron job to automate this.
Consult Ram about process of building as needed.
Eric Bohm
02:52 PM Support #36 (Closed): Nightly manual rebuild
Make sure the Charm++ documentation is being built regularly and automatically. Phil Miller
03:06 PM Feature #6 (Closed): Disk checkpoints with partitions should use a separate directory per partition
Eric Bohm
11:12 PM Feature #6 (Resolved): Disk checkpoints with partitions should use a separate directory per parti...
Phil Miller
02:50 PM Support #35 (Closed): Release notes for 6.5.0
Edit the list of note-worthy changes from the wiki into a coherent set of release notes.
https://wiki.engr.illinois....
Phil Miller
02:49 PM Bug #1 (In Progress): Document template entities better
Phil Miller
11:59 AM Revision 92515ffd (charm): examples/bigsim/emulator/Makefile: Move TESTOPTS back to the end of th...
Phil Miller
11:10 AM Bug #32 (Resolved): PAMI LRTS SMP with L2Atomic Queue crashed
The bug is that LrtsQueueCreate() did not return the queue pointer. Yanhua Sun
11:29 PM Revision 87737a26 (charm): Merge branch 'charm' of charmgit:charm into charm
Yanhua Sun
11:29 PM Revision 39d1e429 (charm): remove SPECIFIC_IDLE to make lrtsNotifyIdle general
Yanhua Sun
11:11 PM Revision d07dc310 (charm): ckpt: drop some now-unnecessary std:: qualification
Phil Miller
11:10 PM Revision 5969567e (charm): ckpt: split separate partitions files into separate directories
Closes #6 Phil Miller
10:29 PM Revision d6a7a2aa (charm): move init persistent to persisnt.c file
Yanhua Sun
09:15 PM Revision ff4a4947 (charm): (1) fix L2Atomic Queue bug (2) add spin L2AtomicQueue on idle
Yanhua Sun
09:12 PM Revision e46c99a9 (charm): Merge branch 'charm' of charmgit:charm into charm
Yanhua Sun
09:12 PM Revision 928d279f (charm): add debugLog for pami
Yanhua Sun
06:08 PM Revision 4cbe0394 (charm): Reducer type "random" should produce a message with reduction type "ra...
Lukasz Wesolowski

02/07/2013

04:51 PM Feature #34 (New): Reduce Charm Message Send Overhead for Marshalled Messages
How many lines of code does it take to send a Chare Array message down into the Converse layer?
LOC Experiment
...
Eric Bohm
04:47 PM Support #33 (In Progress): Jenkins for Autobuild
It would be very useful to have Jenkins integrated in with Autobuild.
# Firing off automatic tests on a limited nu...
Eric Bohm
03:30 PM Feature #29: Reduction Starting messages
Phil Miller wrote:
> @Eric: Is this actually critical to shipping the next release? It hadn't been talked about that...
Eric Bohm
03:16 PM Feature #29: Reduction Starting messages
@Eric: Is this actually critical to shipping the next release? It hadn't been talked about that way in previous discu... Phil Miller
03:28 PM Revision 0d8b4435 (charm): examples/bigsim/emulator/Makefile: Finish fixing parallel build by onl...
Phil Miller
02:55 PM Bug #32 (Closed): PAMI LRTS SMP with L2Atomic Queue crashed
Extended LRTS to use L2AtomicQueue to replace PCQueue in SMP for PAMI. However, it crashed in megatest Yanhua Sun
11:23 AM Revision 062e55b8 (charm): examples/bigsim/emulator/Makefile: Fix parallel compilation and clean ...
Phil Miller
10:49 AM Revision 232c5fb2 (charm): Adding support for disk checkpoint to fast message-logging.
Esteban Meneses
01:26 AM Revision f39a2ddc (charm): Protecting cmiqueue.h with #if
Nikhil Jain
12:59 AM Revision e9532fec (charm): Moving SPECIFIC_QUEUE out of machine.c where it should not occur.
Nikhil Jain
12:07 AM Bug #15 (In Progress): Non-isomalloc threads
Nikhil Jain
12:02 AM Revision 2adfd5f1 (charm): add in overlap for now
Jonathan Lifflander
11:59 PM Feature #31 (Closed): AMPI
This task is the parent task for most of AMPI related work. Nikhil Jain
11:57 PM Revision a288cfe5 (charm): sdag: walk through the tree and output the relevant pieces
Jonathan Lifflander
11:31 PM Feature #30 (Closed): MPI Interoperability
Task for bug fixes and features related to interoperability. Nikhil Jain

02/06/2013

05:36 PM Feature #29 (Closed): Reduction Starting messages
Due to migratability, the current Chare Array reduction mechanic triggers partition wide broadcasts. This is problem... Eric Bohm
01:11 AM Revision 3be62353 (charm): Inserting missing #if to protect checkpointing code
Nikhil Jain
12:31 AM Feature #28 (In Progress): Replica based FT
Integrate replication based fault tolerance to mainline. Nikhil Jain
12:29 AM Feature #23 (In Progress): Optimization of MPI layer
I've heard through the grapevine (not to be confused with Harshitha's LB strategy) that Ralf has achieved near-parity... Phil Miller
12:16 AM Feature #23 (New): Optimization of MPI layer
Revisit MPI layer and possibly rewrite to improve performance. Nikhil Jain
12:28 AM Feature #19 (In Progress): Extend TopoManager to work on an n-dimensional torus
Nikhil Jain
11:04 PM Feature #19 (New): Extend TopoManager to work on an n-dimensional torus
Extend the TopoManager API to work on any n-dimensional torus. Nikhil Jain
12:28 AM Feature #17 (In Progress): Native MPI Interoperability with PAMI machine layer
Nikhil Jain
11:02 PM Feature #17 (Closed): Native MPI Interoperability with PAMI machine layer
Suport interoperation of Charm and MPI on PAMI. Nikhil Jain
12:28 AM Feature #18 (In Progress): Native MPI Interoperability with UGNI
Nikhil Jain
11:03 PM Feature #18 (Closed): Native MPI Interoperability with UGNI
Implement interoperability of Charm and MPI on UGNI. Nikhil Jain
12:27 AM Feature #26 (In Progress): Eliminate need to call CBase_Foo::pup in Foo::pup
Phil Miller
12:26 AM Feature #26 (Merged): Eliminate need to call CBase_Foo::pup in Foo::pup
https://wiki.engr.illinois.edu/display/ppl/Incantationectomy
Have a look at branch phil/base_pup for work in progr...
Phil Miller
12:26 AM Feature #27 (Closed): Make CkIO safe to use in real application code with proper flush and close ...
Phil Miller
12:18 AM Feature #25 (Closed): Verbs in LRTS
Separate verbs from NET and port it to LRTS. Nikhil Jain
12:17 AM Feature #24 (Closed): LRTS based Net layer
Port net layer to LRTS. Nikhil Jain
12:14 AM Feature #22 (Closed): Stack aliasing on BGQ
Add support for stack aliasing on BGQ. Nikhil Jain
12:11 AM Projections Feature #21 (New): 3D view in Projections
Java based 3D view in projections is buggy - fix it. Nikhil Jain
12:04 AM Cleanup #20 (Rejected): Warnings about ignored return values from scanf
Quoth Nikhil:
> The following the warning I see very often while compiling codes (that I mentioned in the core meeti...
Phil Miller
10:38 PM Feature #16 (New): Projections tracing with MPI Interoperability
Make programs that use multiple modules in interoperation with MPI dump correct projections files. Nikhil Jain
10:33 PM Bug #15 (Closed): Non-isomalloc threads
If threads are used in non-isomalloc mode for stacks such as stack copy, they may crash since isomalloc pup is always... Nikhil Jain
10:31 PM Bug #14 (Closed): PAMI-SMP (both with and w/o ASYNC) Autobuild failure
PAMI-SMP-ASYNC does not pass auto build. Nikhil Jain
10:19 PM Feature #13 (New): CkCallback to section breaks
This code fails to run (runtime error on CkCallback::pup), but compiles fine.
CProxy_ArrayA proxy = CProxy_ArrayA:...
Jonathan Lifflander
06:57 PM Feature #2: Reduce memory usage of reductions
This will be much easier when #12 is done, since it will only need to be done once. Phil Miller
06:56 PM Cleanup #12 (New): Factor out massive duplication in reductions
The code that implements reductions is duplicated to varying extents across groups, nodegroups, and arrays. Refactor ... Phil Miller
06:11 PM Feature #11 (New): Make CCS work with partitions
Phil Miller
06:10 PM Feature #10 (Closed): Ensure that Bigsim logging works with partitions
Phil Miller
06:10 PM Feature #9 (New): Ensure that record-replay works with partitions
Phil Miller
06:10 PM Feature #8 (New): Ensure that various features work with partitions
Phil Miller
06:07 PM Cleanup #7 (Closed): cppcheck cleanup
Removing errors detailed by cppcheck for proposal Michael Robson
06:05 PM Feature #6 (Closed): Disk checkpoints with partitions should use a separate directory per partition
Phil Miller
06:04 PM Bug #5 (Closed): Many-to-many primitive broken on DCMF
The DCMF and PAMI machine layers offer an optimized many-to-many primitive that has been useful in NAMD, but it's cur... Phil Miller
06:03 PM Feature #4 (Rejected): CmiOneSided on Cray GNI (Gemini, Aries/Cascade)
Make CmiDirect work on Gemini/uGNI. Phil Miller
06:03 PM Bug #3 (Closed): Provide full topology information in partititioned runs on BGQ
On some platforms that have meaningful network topology, running a partitioned job results in getting linearized topo... Phil Miller

02/05/2013

05:59 PM Feature #2 (Merged): Reduce memory usage of reductions
The current system for reductions buffers input messages at each node of the tree until every input has been received... Phil Miller
05:54 PM Bug #1 (Closed): Document template entities better
Rewrite the section of the manual dealing with templated chares, messages, entry methods, and so forth. Phil Miller
03:47 PM Revision a9b56dd1 (charm): Add simple wrappers for runtime control of Cray Balanced Injection
Eric Bohm
02:48 AM Revision 5396158b (charm): Modified Charm++ reductions to accept dummy contributions with no data
from elements that do not have anything useful to contribute.
The dummy messages, submitted by calling contribute wit...
Lukasz Wesolowski
12:23 AM Revision 4f6798e5 (charm): Merge branch 'charm' of charmgit:charm into charm
Xiang Ni
12:23 AM Revision fb08657d (charm): protect the codes in net build
Xiang Ni
11:22 PM Revision 6f23c8a7 (charm): Compilation with GCC does not like inline declarations of variables an...
warning on some of the things; fixing them Nikhil Jain
11:03 PM Revision 3d0d81e9 (charm): Revert "Begining work on interoperatin in non-mpi layer"
This reverts commit 33b8aa64328a71627e09c9900f42f20847bb766c. Nikhil Jain
10:48 PM Revision d3a409eb (charm): Replace the #ifdef with #if in pingpong example
Nikhil Jain
10:22 PM Revision e6c40daa (charm): Merge branch 'charm' of charmgit:charm into charm
Nikhil Jain
08:43 PM Revision d41f2a53 (charm): fix the CmiDirect compilation error
Yanhua Sun

02/04/2013

04:06 AM Revision 740cc6d6 (charm): Bug Fix - Correct variable name for BGP TopoManager
Nikhil Jain
03:10 AM Revision ea0ad83c (charm): Bug Fix - intialize the subdir to null if num partitions is 1
Nikhil Jain
01:35 AM Revision 1e9fd8e3 (charm): Use defensive names for variables defined in macros, to avoid conflict...
Phil Miller
01:12 AM Revision 66a7aca3 (charm): make sure the consistence of the spare pe pool when there multiple par...
Xiang Ni
01:12 AM Revision bf65e421 (charm): remove the unnecessary conditional
Xiang Ni
01:12 AM Revision ecab1d99 (charm): protection for the previous changes
Xiang Ni
01:12 AM Revision 1d662332 (charm): Reverting addition of barrier to MPI layer - not all threads come to t...
and hence a CmiBarrier is bound to hang Nikhil Jain
01:12 AM Revision a7706431 (charm): Making changes to projection for handling partitions - earlier scheme ...
dump directories in all cases (even if the code is not compiled in tracemode
projections); also skipping directory if...
Nikhil Jain
01:12 AM Revision 1b426723 (charm): Making correction in CmiDirect of Gemini to work correctly with partit...
CmiDirect's implementation is fixed, this should work correctly Nikhil Jain
01:12 AM Revision e911b9cb (charm): Fixing bug in BGQTorus.h - incorrect variable name, and setting one more
variable Nikhil Jain
01:12 AM Revision fd817d80 (charm): Fixing ManytoMany and ListSend for correctness when used with partiton...
Nikhil Jain
01:12 AM Revision c879867a (charm): Fixing a bug in BGPTorus - private variables names were incorrect
Nikhil Jain
01:12 AM Revision 68fe5174 (charm): Adding a new flag for partition support, and making the change accordi...
converse.h Nikhil Jain
01:12 AM Revision 13987375 (charm): chkpt: Factor out file naming and opening to a separate function
This makes partition-handling easier, makes error checking and reporting
uniform, and will make later changes in path...
Phil Miller
01:12 AM Revision bf7636b8 (charm): Clubbing partition related function into a pound define protected by L...
providing dummy implementations to be used in non-LRTS cases Nikhil Jain
01:12 AM Revision a157ae04 (charm): Modifying the TopoManager related files for BG machines - if more than...
partitions are present, return a linear topology for now Nikhil Jain
01:12 AM Revision acdd4f88 (charm): For Cray machine > XT3, providing correct information even when multiple
partitions are present is straightforward, and hence I am using it. For XT3, I
leave it as it - after getting opinion...
Nikhil Jain
01:12 AM Revision 7b8dd26f (charm): Adding fix for DCMF many to many for multi partition run; I was not ab...
the correctness because the only use case (in NAMD) seems to be broken Nikhil Jain
01:12 AM Revision 6a8a4214 (charm): changes for disk checkpointing to write into different files in each p...
Xiang Ni
01:12 AM Revision 776b383d (charm): Adding a barrier before exit in MPI layer; this has been done to unavo...
incase some partitions wants to exit before others Nikhil Jain
01:12 AM Revision 064ffadc (charm): Adding a CMK pound define for layer that make use of LRTS interface;
Nikhil Jain
01:12 AM Revision 280b7c6f (charm): Forgot a colon in the previous commit
Nikhil Jain
01:12 AM Revision 6137c4bd (charm): Adding option for using replica as an alternate to partition
Nikhil Jain
01:12 AM Revision 5dd6bc0f (charm): Replacing replica with partition
Nikhil Jain
01:12 AM Revision 4f9d7bd6 (charm): This instance got missed last time Yanhua and I made replica compatibl...
persistent - fixing it now Nikhil Jain
01:12 AM Revision c93eaee8 (charm): creat different directories for different replicas
Yanhua Sun
01:12 AM Revision b4dd4170 (charm): Add support for +stdout; Change nomenclature of functions used by part...
Conflicts:
src/arch/util/machine-common-core.c
Nikhil Jain
01:12 AM Revision 1eaef590 (charm): Handle persistent comm in Gemini correctly:
Nikhil Jain
01:12 AM Revision 0ed4dfe4 (charm): Fix for VC++ - it does not like variable declarations after other code
Nikhil Jain
01:12 AM Revision b57f61ca (charm): Removing the local usage of CmiAssert which is bound to fail in replic...
environment
Conflicts:
src/arch/gemini_gni/machine-persistent.c
Nikhil Jain
01:12 AM Revision bf5c698c (charm): Remove usage of functions not meant to be exposed
Nikhil Jain
01:12 AM Revision ffcb4b3b (charm): Adding converse level functions for sending messages to remote replicas
Nikhil Jain
01:12 AM Revision 9c398c19 (charm): adding extern
Xiang Ni
01:12 AM Revision ae25fd55 (charm): First commmit to enable partition logic - added the concepts of localr...
global ranks; modifying LRTS to receive destNode; modifying converse to pass
global dest rank; trying to take care of...
Nikhil Jain
11:54 PM Revision 776ede5b (charm): Bug fix - in an earlier commit, the name of variable got changed from ...
CMK_SMK and this caused issues while running in SMP mode Nikhil Jain

02/03/2013

04:28 PM Revision f53218b7 (charm): ckcallback: write assigment operator, fix a few bugs with threads
Jonathan Lifflander
10:24 PM Revision 5230ac1b (charm): Making changes to enable time sharing on BGQ;
Nikhil Jain
10:08 PM Revision 326d4895 (charm): ckcallback: move methods to cpp file, move CallbackType into CkCallbac...
Jonathan Lifflander
09:41 PM Revision e65bad37 (charm): qd: fix compilation problem: pointer to member vs. member
Jonathan Lifflander
09:40 PM Revision cf33453d (charm): qd: make prepare message function static
Jonathan Lifflander
09:40 PM Revision e692afd2 (charm): qd: reformat code, remove old "register" declarations, remove todos
Jonathan Lifflander
09:39 PM Revision b66ecc68 (charm): qd: unpack message properly on the qd timeout branch
Jonathan Lifflander
09:38 PM Revision 8eef7cbd (charm): qd: some reformatting and change access protection for QdMsg
Jonathan Lifflander
09:32 PM Revision 33b8aa64 (charm): Begining work on interoperatin in non-mpi layer
Nikhil Jain
08:44 PM Revision a158756c (charm): Adding a time sharing example for Interoperation
Nikhil Jain
06:45 PM Revision 8e100ff5 (charm): qd: actually pack message properly (so the callback is transmitted cor...
Jonathan Lifflander
06:44 PM Revision 0b06b072 (charm): change to conform to the slight change in interface
Jonathan Lifflander
06:43 PM Revision 87eaf362 (charm): ckcallback: implement threads correctly, fix pups, polymorphic clone o...
Jonathan Lifflander

02/02/2013

06:10 PM Revision 8d8576a6 (charm): ckcallback: change to conform to the new interface
(unfinished: broke some things temporarily) Jonathan Lifflander
06:10 PM Revision db1519ac (charm): ckcallback: rewrite to use class hierarchy and virtual methods instead...
Jonathan Lifflander

02/01/2013

01:43 AM Revision 7aa4ac8b (charm): Adding async as a build target for pami build on BGQ
Nikhil Jain
09:29 PM Revision 77de81b5 (charm): Merge branch 'pami-lrts-ready' of charmgit:charm into charm
Nikhil Jain
08:30 PM Revision e279b858 (charm): fix a bug of message size alignment in persistent messages
Yanhua Sun
07:52 PM Revision 90cd8075 (charm): fix the size of memory for L2Atomic queue
Nikhil Jain
07:52 PM Revision 181a856f (charm): reserve one slot in L2Atomic Queue for node queue
Nikhil Jain
07:52 PM Revision ba935b9a (charm): Disabling LRTS specific List send till I can find a test case to verfi...
correctness Nikhil Jain
07:52 PM Revision 46a8ee5f (charm): Remove the partition related code;
Nikhil Jain
07:52 PM Revision 0aa4bbc1 (charm): Adding PAMI specific List Send functions
Nikhil Jain
07:52 PM Revision a754e87b (charm): Adding async progress based implementation a build target
Nikhil Jain
07:52 PM Revision d27bb376 (charm): Reverting a change made by Yanhua in the previous commit - one cannot ...
CmiBarrier after the context has been destroyed. Nikhil Jain
07:52 PM Revision 6463711c (charm): add L2Atomic Support, it compiles
Nikhil Jain
07:52 PM Revision 78895082 (charm): Bug Fix - I think targeting context to be random was a bad idea. Multiple
threads can be pushing to the same PE if the context is randomly chosen Nikhil Jain
07:52 PM Revision 1a385752 (charm): Some code clean up; bug fix in LrtsSendFunc - it is passed PE, not ran...
there is still a bug somewhere, looking for it Nikhil Jain
07:52 PM Revision bb61bdfa (charm): add CMIQueue macro to expand LRTS queue to use common pcqueue or not
Yanhua Sun
07:51 PM Revision 588d87a1 (charm): Fix the way exit happens
Nikhil Jain
07:51 PM Revision afccb02f (charm): add persistent message files and changes for PAMI BGQ
Yanhua Sun
07:51 PM Revision 0fde0171 (charm): Bug fixes, changes to make pamilrts work - two version work now...
non smp and smp without xlc - as of now xlc is not supported because cldb is not
being supported using xlc
Nikhil Jain
07:51 PM Revision 6e2a4796 (charm): Change from a pointer to actual data - bug fix
Nikhil Jain
07:51 PM Revision 2530521d (charm): First commit for porting PAMI to lrts
Nikhil Jain
07:51 PM Revision eaa77c45 (charm): Adding a new target pamilrts-bluegeneq
Nikhil Jain
07:51 PM Revision 211ab1c0 (charm): Modifying charmrun to behave corectly if launched from within a script
Nikhil Jain
06:19 PM Revision 089fb746 (charm): bug fix in get-based persistent scheme on Gemini
Yanhua Sun

01/31/2013

01:44 PM Revision 7fa8bd5f (charm): Revert "minor"
This reverts commit 638c8e6afb9093b4c223e1518da235db023a38de. Yanhua Sun
01:39 PM Revision 638c8e6a (charm): minor
Yanhua Sun
12:12 AM Revision 969375a0 (charm): bluegenq: Correct const error on revised CmiMemcpy macro
Phil Miller
12:05 AM Revision 7b868abc (charm): bluegenq: Conform CmiMemcpy macro to best practices
Phil Miller
 

Also available in: Atom