Project

General

Profile

Activity

From 03/16/2018 to 04/14/2018

04/15/2018

08:26 PM Bug #1858: tests/ampi/privatization fails on mpi-linux-x86_64-smp autobuild
Notes:
The weird part is that printing out the PE # inside ampiCreateMain shows the call is from PE 0, but inside ...
Sam White
07:51 PM Feature #1866: Update Boost ULTs to v1.67.0
We could add a script to make updating this automatically easy. Sam White

04/14/2018

06:14 PM Feature #1866 (Merged): Update Boost ULTs to v1.67.0
Boost v1.67.0 was just released and it looks like there are some minor enhancements to the context library, which we ... Sam White
04:15 PM Feature #527: smp layer based on OpenMP 4.0 teams
Related: https://charm.cs.illinois.edu/gerrit/#/c/charm/+/2771/ Sam White
11:49 AM Feature #1693: Print template parameters when missing explicit instantiation
I added Ronak as a watcher since this is related to issue #1856. Sam White

04/13/2018

06:07 PM Bug #1500: Entry Methods Always Take lvalue References (feature/bug)
Should this be marked merged once we have perfect forwarding and r-value references supported? We already support r-v... Sam White
05:58 PM Feature #1864 (Implemented): Charm++ type traits for compile-time querying of types
https://charm.cs.illinois.edu/gerrit/#/c/charm/+/3988/ Sam White
02:52 PM Feature #1864: Charm++ type traits for compile-time querying of types
Looks like std::void_t is the only C++17 thing here, so we can just add a configure-time check for it and conditional... Sam White
01:35 PM Feature #1864: Charm++ type traits for compile-time querying of types
We can probably just take the code from Spectre ad stick it into charm++.h or some header. The only thing we need to ... Sam White
12:07 PM Feature #1864 (Merged): Charm++ type traits for compile-time querying of types
Spectre defines these here: https://github.com/nilsdeppe/spectre/blob/charm_simplifications/src/Parallel/TypeTraits.hpp Sam White
05:08 PM Bug #1854 (Implemented): Charmxi chokes on parameter packs
Evan Ramos
05:07 PM Bug #1853 (Implemented): Charmxi doesn't parse std::tuple
Evan Ramos
02:31 PM Bug #1859: Megampi hangs on multicore-linux-x86_64 with +CmiSleepOnIdle
This hang only happens when the test is run on charity with "+CmiSleepOnIdle". It doesn't reproduce on other lab mach... Sam White
10:26 AM Bug #1859 (New): Megampi hangs on multicore-linux-x86_64 with +CmiSleepOnIdle
Test case causing hang: ... Nitin Bhat
01:28 PM Feature #1863 (Implemented): PUP support for templated abstract base classes
https://charm.cs.illinois.edu/gerrit/#/c/charm/+/3985/ Sam White
12:05 PM Feature #1863 (Merged): PUP support for templated abstract base classes
https://github.com/sxs-collaboration/spectre/blob/develop/src/Parallel/CharmPupable.hpp
blocks at lines 17 & 29: htt...
Sam White
01:19 PM Bug #1858: tests/ampi/privatization fails on mpi-linux-x86_64-smp autobuild
Can also happen with netlrts-linux-x86_64-smp:... Evan Ramos
01:17 PM Bug #1861: Chare constructors cannot take noncopyable types as arguments
Wouldn't taking them by value defeat the purpose of std::move? Evan Ramos
11:48 AM Bug #1861 (Merged): Chare constructors cannot take noncopyable types as arguments
Currently chare constructors take their arguments by const l-value reference, which prohibits passing non-copyable ty... Sam White
12:35 PM Bug #1855 (Rejected): Zero copy send with std::move'ing the buffer to the runtime
This is a duplicate of: https://charm.cs.illinois.edu/redmine/issues/1865 Sam White
12:23 PM Feature #1865 (New): Implement zero copy translation for move semantics and rvalue refs
The buffer ownership semantics required by the zero copy approach can be derived from C++ syntax.
If the user pass...
Eric Bohm
12:00 PM Feature #1862: Simplify custom reducer registration
https://github.com/nilsdeppe/spectre/blob/charm_simplifications/src/Parallel/Reduction.hpp Sam White
11:59 AM Feature #1862 (New): Simplify custom reducer registration
Spectre does thi to allow free functions to be used and automatically registered: spectre/src/Parallel/Reduction.hpp Sam White
11:52 AM Bug #1692: Charmxi dies on Array Indices with Template Parameters
Spectre currently has to work around .ci file parsing limitations for this.
Changes in here to support ci file cha...
Sam White
11:43 AM Feature #1860 (New): Support HostBuffer shared memory allocation of one buffer per physical host ...
Some applications would like to have multiple processes per host (usually for NUMA and comm thread motivations) but s... Eric Bohm
11:02 AM Cleanup #1847: Eliminate sources of various compiler warnings
These should be fixed for 6.9.0 Sam White
11:01 AM Bug #1856: Keep track of entry method template arguments in traces for Projections
This would be great to have for 6.9.0, but I don't think we should block the release on it. Sam White
09:28 AM Bug #1849: mpi-linux-x86_64-syncft broken by zero copy API
Hmm, the same test passes for netlrts-darwin-x86_64-syncft. I fixed a recently introduced build error for syncft here... Sam White
09:04 PM Bug #1857 (Merged): Replace WIN32 macro checks with proper _WIN32
Sam White

04/12/2018

11:39 AM Bug #1857 (Implemented): Replace WIN32 macro checks with proper _WIN32
Some of the "WIN32"s are from libraries that we didn't develop defining their own WIN32 macro based on _WIN32: https:... Sam White
09:23 AM Bug #1853 (In Progress): Charmxi doesn't parse std::tuple
https://charm.cs.illinois.edu/gerrit/#/c/charm/+/3980/ Sam White
09:23 AM Bug #1854 (In Progress): Charmxi chokes on parameter packs
https://charm.cs.illinois.edu/gerrit/#/c/charm/+/3980/ Sam White
09:21 AM Bug #1858: tests/ampi/privatization fails on mpi-linux-x86_64-smp autobuild
Also hanging on multicore-linux-x86_64 Sam White

04/11/2018

01:35 AM Bug #1858 (Merged): tests/ampi/privatization fails on mpi-linux-x86_64-smp autobuild
Autobuild runs this test on the 'courtesy' lab machine:... Sam White
10:49 PM Bug #1857 (Merged): Replace WIN32 macro checks with proper _WIN32
See comments on this patch: https://charm.cs.illinois.edu/gerrit/#/c/charm/+/3939/... Sam White

04/10/2018

12:14 PM Bug #1849: mpi-linux-x86_64-syncft broken by zero copy API
Of course we also need to make the test skip this case when run with +p1 to avoid failing out of the tests.
With s...
Sam White
07:12 PM Bug #1849 (In Progress): mpi-linux-x86_64-syncft broken by zero copy API
Yeah, in-memory checkpointing can't work in the +p1 case, since there's no remote memory to checkpoint and restart fr... Sam White
10:42 AM Documentation #1611 (Merged): Document network dependent rdma thresholds, above which benefits of...
Nitin Bhat
09:35 AM Bug #1692: Charmxi dies on Array Indices with Template Parameters
Nils pointed out that they have a templated array index type in Spectre that that allows users to define custom indic... Sam White

04/09/2018

06:40 PM Bug #1849: mpi-linux-x86_64-syncft broken by zero copy API
The first expression is true. The while loop seems to become an infinite loop when both pe and budpe are 0.
Print...
Nitin Bhat
03:28 PM Bug #1849: mpi-linux-x86_64-syncft broken by zero copy API
Which of the while expressions is always true? Sam White
02:02 PM Bug #1849: mpi-linux-x86_64-syncft broken by zero copy API
Launching a serial version using gdb ./pgm causes the hang at the following location:
*CkMemCheckPT::BuddyPE (pe=...
Nitin Bhat
01:41 PM Bug #1849: mpi-linux-x86_64-syncft broken by zero copy API
I see the hang as well, when I add the fix. However, I am intermittently seeing the hang on earlier commits too. I tr... Nitin Bhat
08:44 AM Bug #1849: mpi-linux-x86_64-syncft broken by zero copy API
This is a release blocker Sam White
06:34 PM Bug #1852 (Merged): verbs-linux-x86_64 failure in tests/charm++/pingpong with +p2 after direct AP...
Sam White
05:50 PM Bug #1852 (Implemented): verbs-linux-x86_64 failure in tests/charm++/pingpong with +p2 after dire...
Fix: https://charm.cs.illinois.edu/gerrit/#/c/charm/+/3971/ Nitin Bhat
08:44 AM Bug #1852: verbs-linux-x86_64 failure in tests/charm++/pingpong with +p2 after direct API merge
This is a release blocker Sam White
06:04 PM Bug #1853: Charmxi doesn't parse std::tuple
Evan, can you take a stab at this and parameter packs (issue #1854) in the next week or two to see if we can get it i... Sam White
03:22 PM Bug #1853 (Merged): Charmxi doesn't parse std::tuple
Nils pointed this out, Spectre currently patches charmxi's generated code to get around this. Sam White
04:53 PM Feature #559: Generic (un)marshalling code to replace code generation per entry-method
From Nils, we can do this with an initializer_list that uses std::forward on the variadic template arguments. Sam White
04:47 PM Bug #1856 (Merged): Keep track of entry method template arguments in traces for Projections
We currently don't record the template arguments of entry methods in traces for Projections, so if you have two 'entr... Sam White
04:22 PM Bug #1855 (Rejected): Zero copy send with std::move'ing the buffer to the runtime
This would allow the application to tell the runtime to send a buffer without copying it and remove the need for the ... Sam White
03:25 PM Bug #1679: Do Not Require Default Constructors for Serializable Classes
For direct documentation here, you can now specify a constructor that takes "PUP::reconstruct" as its argument Sam White
03:23 PM Bug #1854 (Merged): Charmxi chokes on parameter packs
Nils is currently working around this in Spectre. charmxi doesn't parse "..." properly. Sam White
08:51 PM Support #1681: Document Exception Handling in Charm++
Make the RTS use std::set_terminate: https://charm.cs.illinois.edu/redmine/issues/1730
Mark the entire RTS as 'noe...
Sam White
07:59 PM Bug #1737: tests/charm++/pingpong and examples/charm++/zerocopy/pingpong fail when run on 2 proce...
Just curious: which system do we want pamilrts-linux-x86_64 (not ppc64le?) to run on? Sam White

04/08/2018

10:36 AM Bug #1852 (Merged): verbs-linux-x86_64 failure in tests/charm++/pingpong with +p2 after direct AP...
It seems to be failing autobuild consistently here since the direct API verbs patch was merged. The next case in this... Sam White

04/07/2018

04:36 PM Bug #1850 (Merged): multicore builds failing to parse CLAs correctly in tests/charm++/load_balanc...
Sam White
12:22 PM Bug #1850 (Implemented): multicore builds failing to parse CLAs correctly in tests/charm++/load_b...
https://charm.cs.illinois.edu/gerrit/#/c/charm/+/3964/ Sam White
10:41 AM Bug #1850 (Merged): multicore builds failing to parse CLAs correctly in tests/charm++/load_balanc...
Since the change to command line argument parsing in charmrun on multicore builds, they are all failing to parse lb_t... Sam White
02:11 PM Bug #1849: mpi-linux-x86_64-syncft broken by zero copy API
The issue looks to be that mpi/conv-mach-syncft.h was not updated with the additional field 'msgType' in the MPI dire... Sam White
10:39 AM Bug #1849 (Merged): mpi-linux-x86_64-syncft broken by zero copy API
... Sam White
01:02 PM Feature #1851 (New): Ensure PPCAtomicQueue has proper alignment/padding on ppc64le
PPCAtomicQueue is used by pamilrts on both BGQ and POWER8/9 systems. POWER8/9 has 128 Bytes cache line size, while BG... Sam White

04/06/2018

06:41 PM Bug #1846 (Merged): mpi-win64 machine-rdma.h 'void *': unknown size error
Sam White
04:58 PM Bug #1846 (Implemented): mpi-win64 machine-rdma.h 'void *': unknown size error
https://charm.cs.illinois.edu/gerrit/#/c/charm/+/3960/ Sam White
11:46 AM Bug #1846: mpi-win64 machine-rdma.h 'void *': unknown size error
Offending commit: https://charm.cs.illinois.edu/gerrit/#/c/charm/+/3120/ Sam White
06:41 PM Bug #1848 (Merged): Make multicore charmrun parse all arguments
Sam White
05:31 PM Bug #1848 (Implemented): Make multicore charmrun parse all arguments
https://charm.cs.illinois.edu/gerrit/3961 Evan Ramos
05:17 PM Bug #1848 (Merged): Make multicore charmrun parse all arguments
Currently the multicore version of charmrun only parses arguments until it hits one that doesn't start with a '+'. Th... Sam White
05:18 PM Cleanup #1847 (In Progress): Eliminate sources of various compiler warnings
Sam White
04:38 PM Cleanup #1847: Eliminate sources of various compiler warnings
I've updated the patch on gerrit. These are the only warnings left now from what I originally posted above:... Sam White
02:35 PM Cleanup #1847: Eliminate sources of various compiler warnings
> The if condition is only to avoid warnings that would happen if we define _GNU_SOURCE when it is already defined. E... Sam White
02:31 PM Cleanup #1847: Eliminate sources of various compiler warnings
Sam White wrote:
> I think the noreturn warnings can be solved by adding "while(1);" to the end of those functions: ...
Evan Ramos
12:41 PM Cleanup #1847: Eliminate sources of various compiler warnings
I think the noreturn warnings can be solved by adding "while(1);" to the end of those functions: https://stackoverflo... Sam White
10:16 AM Cleanup #1847: Eliminate sources of various compiler warnings
I got rid of most of these here: https://charm.cs.illinois.edu/gerrit/#/c/charm/+/3955/
The remaining ones are:
...
Sam White
09:30 AM Cleanup #1847 (Merged): Eliminate sources of various compiler warnings
Looking over autobuild I see a lot compiler warnings, some of which were added recently, that should be easily remove... Sam White
02:22 PM Bug #1801: charmrun on multicore builds doesn't consume ++local
Yikes, that's correct, it works if ++local comes before the program's name. Sam White
02:17 PM Bug #1801: charmrun on multicore builds doesn't consume ++local
It looks like the multicore charmrun shell script is only set up to parse arguments until it reaches one that does no... Evan Ramos
02:14 PM Bug #1801: charmrun on multicore builds doesn't consume ++local
I am still seeing ++local not consumed on multicore-darwin-x86_64. It looks like it gets passed through to the RTS wh... Sam White
02:03 PM Bug #1789: ++oneWthPerSocket doesn't work on Darwin
https://charm.cs.illinois.edu/gerrit/#/c/charm/+/3958/ Sam White
07:51 PM Feature #1742: Send large readonly variables using the Nocopy API to reduce memory footprint and ...
Moving this after 6.9 Nitin Bhat
07:16 PM Bug #937: [local] entry methods don't set tracing event dependencies
Bump Sam White
07:11 PM Feature #921: Entry tag [inline] is unable to optimize away most of the overhead
tests/charm++/pingpong has some of them, though I don't think it passes its arguments by const& Sam White
07:09 PM Feature #921: Entry tag [inline] is unable to optimize away most of the overhead
Is there any existing code to profile/time these? Evan Ramos
07:01 PM Bug #1699: [inline] entry methods should use perfect forwarding in C++11
I think so. Evan Ramos

04/05/2018

06:56 PM Bug #1699: [inline] entry methods should use perfect forwarding in C++11
Would that mean we'd basically have to make every single entry method a templated method? Well, I guess just every si... Eric Mikida
06:41 PM Bug #1699: [inline] entry methods should use perfect forwarding in C++11
I believe the build failures we are seeing with Eric's patch are a result of the generated code not meeting the requi... Evan Ramos
04:59 PM Bug #1846 (Merged): mpi-win64 machine-rdma.h 'void *': unknown size error
... Jim Phillips
04:45 PM Feature #921: Entry tag [inline] is unable to optimize away most of the overhead
Yeah, I am unsure what is left to do here that isn't covered by Bug #1699. It might be good to have another round of ... Sam White
04:30 PM Feature #921: Entry tag [inline] is unable to optimize away most of the overhead
I'm having trouble understanding what problem(s) still need fixing for this issue to be considered complete. It appea... Evan Ramos
02:57 PM Support #1725 (In Progress): Improve pup_stl testing
Samarth Kulshreshtha
02:42 PM Feature #1809 (Merged): Direct API for the OFI layer
Nitin Bhat
02:42 PM Feature #1807 (Merged): Direct API for the Verbs layer
Nitin Bhat
12:25 PM Feature #1808 (Merged): Direct API for the MPI layer
Nitin Bhat
11:55 AM Feature #1805 (Merged): Direct API for the GNI layer
Nitin Bhat
10:15 AM Feature #1810 (Merged): Direct API - Support multiple user operation modes
Nitin Bhat

04/04/2018

05:03 PM Support #1842 (Closed): Make Jenkins SMP builds run faster
Adding +CmiSleepOnIdle seems to have fixed the issue in Jenkins SMP builds. I also increased the parallel builds from... Sam White
02:33 PM Support #1842: Make Jenkins SMP builds run faster
Actually it is a much more drastic improvement in some cases: from 64511.965036 us to 36.540985 us for 1D array pingp... Sam White
02:27 PM Support #1842: Make Jenkins SMP builds run faster
On my lab machine (beauty), adding +CmiSleepOnIdle to TESTOPTS cuts down the time taken for SMP pingpong with +p2 by ... Sam White
09:36 AM Support #1842: Make Jenkins SMP builds run faster
The netlrts-linux-x86_64-smp autobuild target should I think have +CmiSleepOnIdle added to its TESTOPTS as well. netl... Sam White
08:44 PM Support #1842 (Closed): Make Jenkins SMP builds run faster
We are seeing a lot of timeouts in SMP builds for Jenkins commit-triggered builds, because tests/charm++/pingpong/ ta... Sam White
05:02 PM Feature #984: AMPI support for missing MPI-3 non-blocking collectives
Implemented MPI_I{(ex)scan,reduce_scatter(_block)} using calls to their blocking variants: https://charm.cs.illinois.... Sam White
04:57 PM Bug #1843 (Merged): Undefined references to Converse QD
Sam White
11:52 AM Bug #1843 (Implemented): Undefined references to Converse QD
Nitin Bhat
11:30 AM Bug #1843: Undefined references to Converse QD
Fix is here: https://charm.cs.illinois.edu/gerrit/#/c/charm/+/3944/ Sam White
11:27 AM Bug #1843: Undefined references to Converse QD
The error I am seeing is:... Nitin Bhat
11:23 AM Bug #1843 (Merged): Undefined references to Converse QD
Builds with -j1, -j2, -j4 seems to be going fine. (For this reason, this was probably not caught in autobuild).
I'...
Nitin Bhat
03:03 PM Feature #65: MetaLB: Separate cost/benefit analysis for running LB strategy and performing migration
Estimation of migration costs is already included in the model based method to select Load balancing strategy. So thi... Kavitha Chandrasekar
02:26 PM Feature #1667: Direct API for nocopy operations on sender-side and receiver-side
Except for documentation for the Direct API, which is targeted for 6.9, all the other tasks are for later releases. Nitin Bhat
02:23 PM Documentation #1845 (Merged): Documentation for the Zerocopy Direct API
Nitin Bhat
02:19 PM Feature #1844 (New): SMP/non-SMP agnostic job launching arguments
Currently a user has to know whether Charm++ was built in SMP mode or not when they are running their application. We... Sam White
12:14 PM Feature #921: Entry tag [inline] is unable to optimize away most of the overhead
Reassigning to Evan in the hopes that he will have the right mix of time and expertise. Eric Bohm
11:45 AM Support #1841 (Closed): Add -Werror=vla to Jenkins per-commit builds
Nitin Bhat
11:45 AM Support #1841: Add -Werror=vla to Jenkins per-commit builds
Added it to non-smp build - charm-commit-triggered and smp build - test-charm-smp.
Nitin Bhat
08:38 PM Support #1841 (Closed): Add -Werror=vla to Jenkins per-commit builds
This was done a year ago, but has apparently been undone sometime since then: https://charm.cs.illinois.edu/redmine/i... Sam White
10:51 AM Feature #1579: Mark CkAbort [[noreturn]]
I'm seeing compiler warnings about this now:... Sam White
07:57 PM Bug #1699: [inline] entry methods should use perfect forwarding in C++11
It is certainly possible that my implementation is correct, this is really my first time doing anything specifically ... Eric Mikida
07:21 PM Feature #1803 (Merged): Direct API for the LRTS layer
Sam White

04/03/2018

10:12 PM Cleanup #1840 (Merged): Direct API - Pass CkNcpyDestination and CkNcpySource by reference (instea...
Sam White

04/02/2018

05:43 PM Cleanup #1840 (Implemented): Direct API - Pass CkNcpyDestination and CkNcpySource by reference (i...
Fix: https://charm.cs.illinois.edu/gerrit/#/c/charm/+/3935/ Nitin Bhat
05:39 PM Cleanup #1840 (Merged): Direct API - Pass CkNcpyDestination and CkNcpySource by reference (instea...
Nitin Bhat
04:14 PM Feature #1804 (Merged): Direct API - Add CMA support for netlrts, multicore and pami builds
Nitin Bhat
03:47 PM Bug #1832 (Merged): SMP hangs in megatest/multisectiontest
Nitin Bhat
02:51 PM Bug #1832 (Implemented): SMP hangs in megatest/multisectiontest
Fix: https://charm.cs.illinois.edu/gerrit/#/c/charm/+/3934/ Nitin Bhat
02:53 PM Bug #738: prevent direct calls to sdag entry functions
Openatom changes corresponding to this bug fix have been merged: https://charm.cs.illinois.edu/gerrit/#/c/openatom/+/... Nitin Bhat
08:12 AM Cleanup #1839 (Merged): tests/charm++/pmetest/ doesn't have any source files in it?
Sam White

03/30/2018

12:23 PM Bug #1220 (Merged): AMPI: Support tlsglobals with dynamically linked objects
Sam White
10:03 AM Bug #1811 (In Progress): Cross-array section reduction and multicast with callback set at contrib...
Raghavendra Kanakagiri
09:29 PM Bug #1832: SMP hangs in megatest/multisectiontest
This is a QD bug introduced by the group dependence patch.
It hangs multisection test only because that test uses ...
Juan Galvez

03/29/2018

05:10 PM Feature #1579 (Merged): Mark CkAbort [[noreturn]]
Sam White
10:51 AM Feature #1579: Mark CkAbort [[noreturn]]
The patch above has been updated to only mark CkAbort() as noreturn. We can't safely mark CkExit() as noreturn becaus... Sam White
04:04 PM Cleanup #1839 (Implemented): tests/charm++/pmetest/ doesn't have any source files in it?
Nikhil removed this test in 2015, but forgot to get rid of the Makefile: https://charm.cs.illinois.edu/gerrit/#/c/cha... Sam White
03:38 PM Cleanup #1839 (Merged): tests/charm++/pmetest/ doesn't have any source files in it?
I don't know what's going on with this test, but there's nothing besides a Makefile in its directory Sam White
03:46 PM Bug #1789 (Merged): ++oneWthPerSocket doesn't work on Darwin
Sam White
03:42 PM Bug #1797 (Merged): ARM64 build and stability issues
Evan confirmed that making context threads the default has fixed the ARM 64-bit builds. Sam White
03:41 PM Bug #1798 (Merged): uFcontext thread support for 64-bit ARM systems
Marking this complete for now since we've worked around it by using context threads on ARM8.
There's another issue...
Sam White
03:13 PM Feature #1417 (Merged): Reduce CkReductionMsg envelope size
Sam White
01:33 PM Bug #1699: [inline] entry methods should use perfect forwarding in C++11
If it would be useful at all, it is possible for lvalue and rvalue reference overloads to exist side by side.
<pre...
Evan Ramos
10:46 AM Bug #1699: [inline] entry methods should use perfect forwarding in C++11
Here's the patch: https://charm.cs.illinois.edu/gerrit/#/c/charm/+/3788/ Sam White
09:20 AM Bug #1699: [inline] entry methods should use perfect forwarding in C++11
This should not be an API breaking change. If implemented correctly perfect forwarding does not break existing code, ... Nils Deppe
10:52 AM Cleanup #980: Cleanup examples/charm++/cuda/hello
Not a release blocker Sam White
10:48 AM Feature #1695: Allow "type aliases" for explicit instantiations of member function templates
This seems unlikely to get done before 6.9.0 Sam White

03/28/2018

05:41 PM Feature #1731: Complete spack installation script
I just asked Edward and he just hadn't seen the comments on the pull request for whatever reason. So he'll keep worki... Sam White
05:22 PM Feature #1731: Complete spack installation script
@Sam, do you know what the status is? I'd be happy to take this over, but I don't have permissions to modify the pull... Matthias Diener
08:16 PM Feature #1731: Complete spack installation script
Spack pull request here: https://github.com/spack/spack/pull/7291
I'm not sure if Edward Kim is still working for ...
Sam White
03:32 PM Bug #1838: Reevaluate LBCommTable initialization
LBCommData is 88 Bytes, and TableState is 4 Bytes.
This commit cuts TableState from 4 Bytes to 1 Byte: https://cha...
Sam White
02:17 PM Bug #1838 (New): Reevaluate LBCommTable initialization
This is what LBCommTable currently looks like:... Sam White
03:13 PM Bug #1831 (Merged): Shrink/expand broken by charmrun refactoring
Sam White
02:33 PM Feature #177 (In Progress): objid_t: load balancing infrastructure should use objid_t
https://charm.cs.illinois.edu/gerrit/#/c/3742/ Kavitha Chandrasekar
10:32 AM Bug #1832: SMP hangs in megatest/multisectiontest
Most of the hangs or crashes are in megatest/multisection test, but today SMP seemed to fail in tests/converse/commbe... Juan Galvez
09:34 AM Bug #1832: SMP hangs in megatest/multisectiontest
Yeah, I thought it had been fixed by the above patch.
So what we are getting now is a random hang, not always in t...
Juan Galvez
08:37 AM Bug #1832: SMP hangs in megatest/multisectiontest
Originally [1] caused failures in this test. That initial failure was then fixed by [2], but now we are seeing this h... Sam White
08:27 AM Bug #1832: SMP hangs in megatest/multisectiontest
Do we know the commit that triggered this bug? Juan Galvez
08:26 PM Bug #1832: SMP hangs in megatest/multisectiontest
This is one of the few bugs left that would block the release of 6.9.0 Sam White
11:19 PM Bug #1837 (New): uFcontext thread issue on ARM 64 bit systems
Currently, uFcontext based ULT doesn't work on ARM64 bit systems.
For ARM 64bit systems, context is set by defaul...
Seonmyeong Bak
08:21 PM Feature #641: protect load balancer from variable cpu clock
Please close this issue if it is done Sam White
08:19 PM Documentation #1656: Update manual entries on Load Balancing strategies
This would be good to have for 6.9.0 Sam White

03/27/2018

06:09 PM Bug #1699: [inline] entry methods should use perfect forwarding in C++11
It's not clear what to do here for 6.9.0, any thoughts Nils? Sam White
06:08 PM Bug #1798 (In Progress): uFcontext thread support for 64-bit ARM systems
As noted on the gerrit issue, this still isn't passing any tests that use ULTs. Maybe for now just make the default o... Sam White
06:06 PM Bug #1665 (Merged): DDT needs to reference count and free its type objects
Sam White
03:26 PM Feature #1657: CMA support for nocopy sends using the Entry Method API across processes on the sa...
Moving it out of 6.9 release as the Broadcast API has higher priority. Nitin Bhat
03:26 PM Feature #1468: Enable pre-pinning memory for the zero-copy message sends through the Entry Method...
Moving it out of 6.9 release as the Nocopy Broadcast API has higher priority. Nitin Bhat

03/26/2018

05:35 PM Feature #1378: 64-bit Charm message sizes
I don't think we want to expand to support 64bit sizes for all messages.
That would pointlessly blow up the size f...
Eric Bohm
03:44 PM Feature #1378 (In Progress): 64-bit Charm message sizes
Evan Ramos
03:42 PM Bug #1220 (Implemented): AMPI: Support tlsglobals with dynamically linked objects
Evan Ramos
03:29 PM Feature #1113 (Merged): smart-build.pl should detect supercomputers with specialized software env...
Evan Ramos
03:28 PM Bug #522 (Feedback): static linking breaks on multicore builds with 'undefined reference to `get_...
Proposing closure unless the issue can be reproduced. Evan Ramos
02:55 PM Bug #1831 (Implemented): Shrink/expand broken by charmrun refactoring
Fixed in: https://charm.cs.illinois.edu/gerrit/3900
Tested with netlrts-darwin-x86_64 and netlrts-linux-x86_64.
Evan Ramos
09:28 AM Cleanup #1836 (New): Cleanup warnings in charmdebug-python
Compiling multicore-linux-x86_64 with ICC on Stampede2 gives the following warnings, some in the generated code:
<pr...
Sam White

03/20/2018

03:57 PM Bug #738: prevent direct calls to sdag entry functions
I have made changes to Openatom to convert direct sdag calls to proxy calls and have this change on a branch called `... Nitin Bhat
03:38 PM Bug #522 (In Progress): static linking breaks on multicore builds with 'undefined reference to `g...
I am unable to reproduce this. My command was:
./build AMPI multicore-linux-x86_64 -j4 --no-build-shared -optimize...
Evan Ramos
02:38 PM Bug #756: CUDA build does not correctly find cuda location
Sam White wrote:
> What does that documentation patch have to do with this issue?
From the Description:
> tl;dr ...
Michael Robson

03/19/2018

10:47 AM Bug #1832 (In Progress): SMP hangs in megatest/multisectiontest
Nitin Bhat
10:38 AM Feature #541 (Implemented): SMP mesage passing must enforce memory ordering consistency
Gerrit: https://charm.cs.illinois.edu/gerrit/#/c/3877/ Nitin Bhat
10:00 AM Feature #1834 (Merged): Functionality to Broadcast large buffers using the nocopy API
Nitin Bhat

03/17/2018

04:42 PM Cleanup #1833 (New): Cleanup SDAG Closure refnum handling with enable_if
The SDAG code has this in a comment:... Sam White
04:21 PM Bug #1832 (Merged): SMP hangs in megatest/multisectiontest
Jenkins builds and the netlrts-linux-x86_64-smp autobuild target are hanging pretty consistently in tests/charm++/meg... Sam White
12:02 PM Bug #1822 (Merged): megatest/multisection test failures caused by changes to group dependencies
Evan created a patch to prevent similar issues in user code: https://charm.cs.illinois.edu/gerrit/#/c/3871/ Sam White
12:00 PM Bug #1819 (Merged): bigsim failing lb_test inside CmiMemoryUsage()
Sam White
12:00 PM Bug #1828 (Merged): Infinite recursion inside malloc_info in CmiMemoryUsage
Sam White

03/16/2018

01:08 PM Bug #1828: Infinite recursion inside malloc_info in CmiMemoryUsage
We reverted the malloc_info patch for now Sam White
01:07 PM Bug #1819: bigsim failing lb_test inside CmiMemoryUsage()
We reverted the malloc_info patch for now Sam White
 

Also available in: Atom