namd.git
2 years agoUpdate list of platforms. release-2-11
Jim Phillips [Mon, 21 Dec 2015 19:33:38 +0000 (13:33 -0600)]
Update list of platforms.

2 years agoUpdates from release notes.
Jim Phillips [Mon, 21 Dec 2015 19:33:20 +0000 (13:33 -0600)]
Updates from release notes.

2 years agoUpdate to 2.11.
Jim Phillips [Sun, 20 Dec 2015 22:30:25 +0000 (16:30 -0600)]
Update to 2.11.

2 years agoCleaner fix for issue when periodic patch grid is one in any dimension.
Jim Phillips [Sun, 20 Dec 2015 17:43:08 +0000 (11:43 -0600)]
Cleaner fix for issue when periodic patch grid is one in any dimension.

2 years agoFix issue when periodic patch grid is one in any dimension.
Jim Phillips [Sun, 20 Dec 2015 17:20:38 +0000 (11:20 -0600)]
Fix issue when periodic patch grid is one in any dimension.

2 years agoAdd optional third argument, tolerance, and list all greater differences if given.
Jim Phillips [Sun, 20 Dec 2015 17:06:40 +0000 (11:06 -0600)]
Add optional third argument, tolerance, and list all greater differences if given.

2 years agoAvoid NaN in Langevin piston when cellBasisVector3 z component is zero.
Jim Phillips [Thu, 17 Dec 2015 21:22:03 +0000 (15:22 -0600)]
Avoid NaN in Langevin piston when cellBasisVector3 z component is zero.

2 years agoVarious updates, including removing +commthread and adding ++mpiexec-no-n.
Jim Phillips [Wed, 16 Dec 2015 20:30:07 +0000 (14:30 -0600)]
Various updates, including removing +commthread and adding ++mpiexec-no-n.

2 years agoAdd LJ tail correction overflow fix and Mac CUDA binaries.
Jim Phillips [Wed, 16 Dec 2015 20:01:23 +0000 (14:01 -0600)]
Add LJ tail correction overflow fix and Mac CUDA binaries.

2 years agoUpdate Charm++ version 6.7.0-pre to 6.7.0.
Jim Phillips [Wed, 16 Dec 2015 18:49:05 +0000 (12:49 -0600)]
Update Charm++ version 6.7.0-pre to 6.7.0.

2 years agoDisable Xeon Phi asynchronous offload due to memory and performance leak.
Jim Phillips [Wed, 16 Dec 2015 18:42:44 +0000 (12:42 -0600)]
Disable Xeon Phi asynchronous offload due to memory and performance leak.

2 years agoMinor reordering.
Jim Phillips [Fri, 11 Dec 2015 21:00:29 +0000 (15:00 -0600)]
Minor reordering.

2 years agoFix memory leak.
Jim Phillips [Fri, 11 Dec 2015 20:43:31 +0000 (14:43 -0600)]
Fix memory leak.

2 years agoSplit very long initialize function to keep compiler from crashing.
Jim Phillips [Fri, 11 Dec 2015 15:58:02 +0000 (09:58 -0600)]
Split very long initialize function to keep compiler from crashing.

2 years agoFix memory leaks in reloadGridforceGrid.
Jim Phillips [Thu, 10 Dec 2015 22:39:42 +0000 (16:39 -0600)]
Fix memory leaks in reloadGridforceGrid.

2 years agoFix buffer overflow segfault on long topology file path in psf file.
Jim Phillips [Tue, 8 Dec 2015 22:39:54 +0000 (16:39 -0600)]
Fix buffer overflow segfault on long topology file path in psf file.

2 years agoFlag LJ tail corrections unimplemented in memopt builds.
Jim Phillips [Tue, 8 Dec 2015 15:41:37 +0000 (09:41 -0600)]
Flag LJ tail corrections unimplemented in memopt builds.

2 years agoExplicitly cast numAtoms to double in integer overflow fix to avoid accidental reversion.
Jim Phillips [Mon, 7 Dec 2015 22:17:19 +0000 (16:17 -0600)]
Explicitly cast numAtoms to double in integer overflow fix to avoid accidental reversion.

2 years agoFix integer overflow bug in VDW tail correction (LJcorrection), partially from Brian...
Jim Phillips [Mon, 7 Dec 2015 22:05:16 +0000 (16:05 -0600)]
Fix integer overflow bug in VDW tail correction (LJcorrection), partially from Brian Radak.

2 years agoFix an interger overflow bug that gave the wrong tail corrections for moderate-large...
Peter Freddolino [Sun, 6 Dec 2015 04:02:19 +0000 (22:02 -0600)]
Fix an interger overflow bug that gave the wrong tail corrections for moderate-large systems

2 years agoFix instability when running minimize command multiple times. release-2-11-beta-2
Jim Phillips [Fri, 4 Dec 2015 21:43:18 +0000 (15:43 -0600)]
Fix instability when running minimize command multiple times.

2 years agoUpdate to 2.11b2.
Jim Phillips [Thu, 3 Dec 2015 21:48:42 +0000 (15:48 -0600)]
Update to 2.11b2.

2 years agoRemove abnormal exits used to avoid hanging threads in multicore since Charm++ now...
Jim Phillips [Thu, 3 Dec 2015 21:38:00 +0000 (15:38 -0600)]
Remove abnormal exits used to avoid hanging threads in multicore since Charm++ now calls exit().

2 years agoExit abnormally to avoid hanging CUDA runtime threads.
Jim Phillips [Thu, 3 Dec 2015 17:48:15 +0000 (11:48 -0600)]
Exit abnormally to avoid hanging CUDA runtime threads.

2 years agoMake CUDA full exclusion checksum work across multiple devices and follow same except...
Jim Phillips [Thu, 3 Dec 2015 17:02:21 +0000 (11:02 -0600)]
Make CUDA full exclusion checksum work across multiple devices and follow same exceptions as CPU.

2 years agoFix CUDA on Mac, in particular adding -rpath /usr/local/cuda/lib to avoid "CUDA drive...
Jim Phillips [Thu, 3 Dec 2015 14:44:40 +0000 (08:44 -0600)]
Fix CUDA on Mac, in particular adding -rpath /usr/local/cuda/lib to avoid "CUDA driver version is insufficient for CUDA runtime version" error.  (CUDA driver installs /usr/local/cuda/lib/libcuda.dylib stub.)

2 years agoUpdate CUDA toolkit versions and gencode flags.
Jim Phillips [Wed, 2 Dec 2015 17:52:45 +0000 (11:52 -0600)]
Update CUDA toolkit versions and gencode flags.

2 years agoPrevent calling run when minimization is enabled (as it won't actually minimize).
Jim Phillips [Mon, 30 Nov 2015 22:50:39 +0000 (16:50 -0600)]
Prevent calling run when minimization is enabled (as it won't actually minimize).

2 years agoMultiplexed bias-exchange umbrella sampling script.
Jim Phillips [Tue, 24 Nov 2015 05:47:33 +0000 (23:47 -0600)]
Multiplexed bias-exchange umbrella sampling script.

2 years agoAdd EXCHANGE_ACCEPT output on stderr to match non-multiplexed version.
Jim Phillips [Tue, 24 Nov 2015 05:42:16 +0000 (23:42 -0600)]
Add EXCHANGE_ACCEPT output on stderr to match non-multiplexed version.

2 years agoFrom Antti-Pekka Hynninen, fix issue when periodic patch grid is one in any dimension.
Jim Phillips [Tue, 24 Nov 2015 04:42:35 +0000 (22:42 -0600)]
From Antti-Pekka Hynninen, fix issue when periodic patch grid is one in any dimension.

2 years agoFrom Brian Radak, revert vdwForceSwitching with LJcorrection changes.
Jim Phillips [Thu, 19 Nov 2015 22:29:35 +0000 (16:29 -0600)]
From Brian Radak, revert vdwForceSwitching with LJcorrection changes.

2 years agoChange +ppn to ++ppn when parsed by charmrun.
Jim Phillips [Tue, 17 Nov 2015 18:49:55 +0000 (12:49 -0600)]
Change +ppn to ++ppn when parsed by charmrun.

2 years agoEliminate duplicate PME processor flag implementations.
Jim Phillips [Fri, 13 Nov 2015 04:33:35 +0000 (22:33 -0600)]
Eliminate duplicate PME processor flag implementations.

2 years agoAdd multiplexed replica exchange script. release-2-11-beta-1
Jim Phillips [Sun, 8 Nov 2015 22:20:05 +0000 (16:20 -0600)]
Add multiplexed replica exchange script.

2 years agoUpdate to 2.11b1.
Jim Phillips [Sat, 7 Nov 2015 17:11:20 +0000 (11:11 -0600)]
Update to 2.11b1.

2 years agoAdd "replicaDcdFile off" command to return to writing local trajectory file.
Jim Phillips [Fri, 6 Nov 2015 15:29:26 +0000 (09:29 -0600)]
Add "replicaDcdFile off" command to return to writing local trajectory file.

2 years agoFix BlueGeneQ compile error.
Jim Phillips [Thu, 5 Nov 2015 14:17:10 +0000 (08:17 -0600)]
Fix BlueGeneQ compile error.

2 years agoAdd "replicaDcdFile index ?filename?" command to allow writing to trajectory files...
Jim Phillips [Wed, 4 Nov 2015 17:20:22 +0000 (11:20 -0600)]
Add "replicaDcdFile index ?filename?" command to allow writing to trajectory files from multiple replica paritions.  The named file is opened once and written to by any replica using that index.

2 years agoChanges from Brian Radak:
Jim Phillips [Mon, 2 Nov 2015 15:29:17 +0000 (09:29 -0600)]
Changes from Brian Radak:
1) Properly implement vdwForceSwitching involving alchemical atoms and substantially cleanup the code in ComputeNonbondedTI.C and ComputeNonbondedFEP.C; the implementation and variable names now more accurately match those in ComputeNonbondedUtil.C. The appropriate TI derivatives have also been implemented.
2) Properly implement a long-range isotropic correction to the LJ energy and virial when vdwForceSwitching is on. The relevant sections in Molecule.C have also been cleaned up so that the expressions for other switching scenarios are more regularized.
3) All of the relevant output warnings for the above two changes have been changed (e.g. a warning no longer appears stating the vdwForceSwitching will be turned off for alchemical atoms, bc it won't be)
4) A noteworthy change for vdwForceSwitching is that the standard potential defined by that method is exactly zero beyond cutoff so that no LJcorrection exists. When LJcorrection is off, this is still the default behavior. However, when LJcorrection is on, the potential shifts are defined differently so that a non-zero tail correction can be derived in line with a true Lennard-Jones potential. A warning to this effect is displayed to the user.

2 years agoAdd "startup" command to trigger startup without energy evaluation.
Jim Phillips [Fri, 30 Oct 2015 03:58:50 +0000 (22:58 -0500)]
Add "startup" command to trigger startup without energy evaluation.

2 years agoAdd note about commands not allowed within replicaEval.
Jim Phillips [Fri, 30 Oct 2015 03:40:57 +0000 (22:40 -0500)]
Add note about commands not allowed within replicaEval.

2 years agoSupport PrgEnv-intel.
Jim Phillips [Wed, 28 Oct 2015 18:11:06 +0000 (13:11 -0500)]
Support PrgEnv-intel.

2 years agoAllow mpi-smp CUDA builds on Windows only.
Jim Phillips [Wed, 28 Oct 2015 17:01:39 +0000 (12:01 -0500)]
Allow mpi-smp CUDA builds on Windows only.

2 years agoMake replicaSend asynchronous, but potentially overtaking, when called from inside...
Jim Phillips [Tue, 27 Oct 2015 04:01:09 +0000 (23:01 -0500)]
Make replicaSend asynchronous, but potentially overtaking, when called from inside replicaEval.

2 years agoExit abnormally to avoid hanging MIC offload threads.
Jim Phillips [Fri, 23 Oct 2015 20:18:42 +0000 (15:18 -0500)]
Exit abnormally to avoid hanging MIC offload threads.

2 years agoCUDA-enabled NAMD no longer requires at least one patch per thread.
Jim Phillips [Fri, 23 Oct 2015 19:18:34 +0000 (14:18 -0500)]
CUDA-enabled NAMD no longer requires at least one patch per thread.

2 years agoFix CUDA/MIC for single patch.
Jim Phillips [Fri, 23 Oct 2015 19:01:20 +0000 (14:01 -0500)]
Fix CUDA/MIC for single patch.

2 years agoAdd enabletotalforces.
Jim Phillips [Fri, 23 Oct 2015 16:15:27 +0000 (11:15 -0500)]
Add enabletotalforces.

2 years agoFor TclForces and colvars only send total forces when needed, add new enabletotalforc...
Jim Phillips [Thu, 22 Oct 2015 05:03:12 +0000 (00:03 -0500)]
For TclForces and colvars only send total forces when needed, add new enabletotalforces/disabletotalforces TclForces commands.

2 years agoTcl forces loadtotalforces return data for groups as well as atoms.
Jim Phillips [Tue, 20 Oct 2015 22:42:32 +0000 (17:42 -0500)]
Tcl forces loadtotalforces return data for groups as well as atoms.

2 years agoTcl forces loadtotalforces return data for groups as well as atoms; fix various crashes.
Jim Phillips [Tue, 20 Oct 2015 22:42:21 +0000 (17:42 -0500)]
Tcl forces loadtotalforces return data for groups as well as atoms; fix various crashes.

2 years agoPrevent multicore builds from hanging waiting for CUDA runtime threads to exit.
Jim Phillips [Tue, 20 Oct 2015 21:24:08 +0000 (16:24 -0500)]
Prevent multicore builds from hanging waiting for CUDA runtime threads to exit.

2 years agoIncrease Charm++ version to 6.7.0-pre (pre-release).
Jim Phillips [Thu, 15 Oct 2015 19:41:08 +0000 (14:41 -0500)]
Increase Charm++ version to 6.7.0-pre (pre-release).

2 years agoRefactor bonded compute loops to simplify hoisting invariants.
Jim Phillips [Mon, 12 Oct 2015 20:09:27 +0000 (15:09 -0500)]
Refactor bonded compute loops to simplify hoisting invariants.

2 years agoOnly print initial and benchmark times once, not after every run command.
Jim Phillips [Mon, 12 Oct 2015 15:48:41 +0000 (10:48 -0500)]
Only print initial and benchmark times once, not after every run command.

2 years agoEliminate spurious "Original numsteps XXX will be ignored." messages.
Jim Phillips [Mon, 12 Oct 2015 15:14:13 +0000 (10:14 -0500)]
Eliminate spurious "Original numsteps XXX will be ignored." messages.

2 years agoLink psfgen module statically in NAMD - use "package require psfgen" to enable.
Jim Phillips [Fri, 9 Oct 2015 20:47:16 +0000 (15:47 -0500)]
Link psfgen module statically in NAMD - use "package require psfgen" to enable.

2 years agoEnable linking psfgen module statically in NAMD.
Jim Phillips [Fri, 9 Oct 2015 20:46:20 +0000 (15:46 -0500)]
Enable linking psfgen module statically in NAMD.

2 years agoAdd new command "topology alias" to provide alternate (e.g., shorter) residue names...
Jim Phillips [Fri, 9 Oct 2015 19:14:05 +0000 (14:14 -0500)]
Add new command "topology alias" to provide alternate (e.g., shorter) residue names that are used in output files, unlike "pdbalias residue" that only translates residue names on input during segment-building.  Add binary coordinate option to coordpdb and binary velocity options to readpsf and writenamdbin.  Increase version to 1.6.3.

2 years agoFix non-Tcl build.
Jim Phillips [Wed, 30 Sep 2015 13:17:13 +0000 (08:17 -0500)]
Fix non-Tcl build.

2 years agoDisable CkLoop for Win64 builds.
Jim Phillips [Fri, 25 Sep 2015 21:10:43 +0000 (16:10 -0500)]
Disable CkLoop for Win64 builds.

2 years agoUpdate to CUDA 6.5, which is needed for Visual Studio 2013.
Jim Phillips [Fri, 25 Sep 2015 20:52:32 +0000 (15:52 -0500)]
Update to CUDA 6.5, which is needed for Visual Studio 2013.

2 years agoSwitch default build to multicore-win64.
Jim Phillips [Fri, 25 Sep 2015 20:51:41 +0000 (15:51 -0500)]
Switch default build to multicore-win64.

2 years agorebuild dependencies
Jim Phillips [Fri, 25 Sep 2015 20:47:34 +0000 (15:47 -0500)]
rebuild dependencies

2 years agoReplace CkExit() with NAMD_bug() or BackEnd::exit() as appropriate.
Jim Phillips [Fri, 25 Sep 2015 20:07:00 +0000 (15:07 -0500)]
Replace CkExit() with NAMD_bug() or BackEnd::exit() as appropriate.

2 years agoPrevent multicore builds from hanging waiting for Tcl notifier thread to exit.
Jim Phillips [Fri, 25 Sep 2015 18:17:35 +0000 (13:17 -0500)]
Prevent multicore builds from hanging waiting for Tcl notifier thread to exit.

2 years agoMove preprocessor code outside of macros.
Jim Phillips [Fri, 25 Sep 2015 17:04:00 +0000 (12:04 -0500)]
Move preprocessor code outside of macros.

2 years agoInclude algorithm header for std::min.
Jim Phillips [Fri, 25 Sep 2015 16:45:00 +0000 (11:45 -0500)]
Include algorithm header for std::min.

2 years agoDefine __thread for WIN32.
Jim Phillips [Fri, 25 Sep 2015 16:42:27 +0000 (11:42 -0500)]
Define __thread for WIN32.

2 years agoChange "and" to "&&" in preprocessor macro.
Jim Phillips [Fri, 25 Sep 2015 16:36:47 +0000 (11:36 -0500)]
Change "and" to "&&" in preprocessor macro.

2 years agoActually make colvarsConfig optional.
Jim Phillips [Wed, 23 Sep 2015 12:57:58 +0000 (07:57 -0500)]
Actually make colvarsConfig optional.

2 years agoMajor update to collective variables module from Giacomo Fiorin and Jerome Henin...
Jim Phillips [Tue, 22 Sep 2015 21:29:06 +0000 (16:29 -0500)]
Major update to collective variables module from Giacomo Fiorin and Jerome Henin: Tcl scripts to expand existing functionality, improvements to histogram bias, colvarsConfig is now optional ("cv" command), and bug fixes.

2 years agoAlias Mac g++ build scripts to use clang.
Jim Phillips [Fri, 18 Sep 2015 21:05:39 +0000 (16:05 -0500)]
Alias Mac g++ build scripts to use clang.

2 years agoUpdate compiler options, require C++ 11 to support Charm++ 6.7.0.
Jim Phillips [Fri, 18 Sep 2015 20:58:43 +0000 (15:58 -0500)]
Update compiler options, require C++ 11 to support Charm++ 6.7.0.

2 years agoSilence warning about using the result of an assignment as a condition without parent...
Jim Phillips [Fri, 18 Sep 2015 18:57:09 +0000 (13:57 -0500)]
Silence warning about using the result of an assignment as a condition without parentheses.

2 years agoUpdate compiler options, require C++ 11 to support Charm++ 6.7.0.
Jim Phillips [Fri, 18 Sep 2015 18:36:15 +0000 (13:36 -0500)]
Update compiler options, require C++ 11 to support Charm++ 6.7.0.

2 years agoUpdate compiler options, require C++ 11 to support Charm++ 6.7.0.
Jim Phillips [Fri, 18 Sep 2015 18:29:23 +0000 (13:29 -0500)]
Update compiler options, require C++ 11 to support Charm++ 6.7.0.

2 years agoUpdate compiler options, require C++ 11 to support Charm++ 6.7.0.
Jim Phillips [Fri, 18 Sep 2015 17:03:15 +0000 (12:03 -0500)]
Update compiler options, require C++ 11 to support Charm++ 6.7.0.

2 years agoUpdated Drude parameter text based on recommendation from
David Hardy [Fri, 18 Sep 2015 15:07:16 +0000 (10:07 -0500)]
Updated Drude parameter text based on recommendation from
Hui Li (force field developer in Benoit Roux's group).

2 years agoFor CUDA version with fixed atoms, remove warning about incorrect pressure and do...
Jim Phillips [Fri, 4 Sep 2015 22:26:30 +0000 (17:26 -0500)]
For CUDA version with fixed atoms, remove warning about incorrect pressure and do not force fixedAtomsForces on when pressure control is used.

2 years agoMake pressure calculation in CUDA version with fixed atoms present but fixedAtomsForc...
Jim Phillips [Fri, 4 Sep 2015 22:20:02 +0000 (17:20 -0500)]
Make pressure calculation in CUDA version with fixed atoms present but fixedAtomsForces off match CPU version, and add fixedAtomsForceOutput option to not zero forces on fixed atoms on output, both from Antti-Pekka Hynninen.

2 years agoFix potential uninitialized data at startup for replicaBarrier.
Jim Phillips [Fri, 28 Aug 2015 22:00:18 +0000 (17:00 -0500)]
Fix potential uninitialized data at startup for replicaBarrier.

2 years agoFix startup race condition in replicaBarrier (modified patch from Nikhil Jain).
Jim Phillips [Fri, 28 Aug 2015 20:03:05 +0000 (15:03 -0500)]
Fix startup race condition in replicaBarrier (modified patch from Nikhil Jain).

2 years agoAutomatically enable fixedAtomsForces for fixed atoms with new CUDA kernel.
Jim Phillips [Fri, 7 Aug 2015 20:52:11 +0000 (15:52 -0500)]
Automatically enable fixedAtomsForces for fixed atoms with new CUDA kernel.

2 years agoCUDA optimization and refactoring from Antti-Pekka Hynninen.
Jim Phillips [Fri, 7 Aug 2015 20:34:10 +0000 (15:34 -0500)]
CUDA optimization and refactoring from Antti-Pekka Hynninen.

2 years agoCray XC persistent feature is no longer broken (since before 2.10 release).
Jim Phillips [Wed, 29 Jul 2015 18:35:04 +0000 (13:35 -0500)]
Cray XC persistent feature is no longer broken (since before 2.10 release).

3 years agoDon't use SETTLE if any atoms in water are fixed.
Jim Phillips [Wed, 10 Jun 2015 21:48:07 +0000 (16:48 -0500)]
Don't use SETTLE if any atoms in water are fixed.

3 years agoRead NBFIX modified 1-4 terms from CHARMM parameter files.
Jim Phillips [Tue, 2 Jun 2015 14:29:53 +0000 (09:29 -0500)]
Read NBFIX modified 1-4 terms from CHARMM parameter files.

3 years agoProvide useful error message when water molecule has wrong number of atoms.
Jim Phillips [Fri, 22 May 2015 21:20:49 +0000 (16:20 -0500)]
Provide useful error message when water molecule has wrong number of atoms.

3 years agoFlush FEP file at end of run.
Jim Phillips [Wed, 18 Mar 2015 23:14:39 +0000 (18:14 -0500)]
Flush FEP file at end of run.

3 years agoLimit CkLoop parallelism on X-pencil FFT due to false sharing.
Jim Phillips [Sun, 15 Mar 2015 17:30:13 +0000 (12:30 -0500)]
Limit CkLoop parallelism on X-pencil FFT due to false sharing.

3 years agoFix race conditions in node-parallel receive and disable except for CUDA-memopt.
Jim Phillips [Sun, 15 Mar 2015 17:04:19 +0000 (12:04 -0500)]
Fix race conditions in node-parallel receive and disable except for CUDA-memopt.

3 years agoIncrease X and Z pencil CkLoop parallelism.
Jim Phillips [Sat, 14 Mar 2015 19:54:20 +0000 (14:54 -0500)]
Increase X and Z pencil CkLoop parallelism.

3 years agoDistribute iteration over flags array in chargeGridReady().
Jim Phillips [Sat, 14 Mar 2015 19:29:46 +0000 (14:29 -0500)]
Distribute iteration over flags array in chargeGridReady().

3 years agoEnable use of CkLoop library.
Jim Phillips [Fri, 13 Mar 2015 20:20:02 +0000 (15:20 -0500)]
Enable use of CkLoop library.

3 years agoChange "#if USE_CKLOOP" to "#if CMK_SMP && USE_CKLOOP"
Jim Phillips [Fri, 13 Mar 2015 20:16:09 +0000 (15:16 -0500)]
Change "#if USE_CKLOOP" to "#if CMK_SMP && USE_CKLOOP"

3 years agoEnable node-parallel receives and remove unnecessary locks.
Jim Phillips [Fri, 13 Mar 2015 16:38:30 +0000 (11:38 -0500)]
Enable node-parallel receives and remove unnecessary locks.

3 years agoDistribute wrapping and unwrapping of PME charge and potential grids across threads...
Jim Phillips [Fri, 13 Mar 2015 04:36:50 +0000 (23:36 -0500)]
Distribute wrapping and unwrapping of PME charge and potential grids across threads by merging into message-processing code.

3 years agoAllow setting proxyTreeBranchFactor without explicity enabling spanning tree.
Jim Phillips [Wed, 11 Mar 2015 23:04:37 +0000 (18:04 -0500)]
Allow setting proxyTreeBranchFactor without explicity enabling spanning tree.