Mark Abraham [Fri, 16 Feb 2018 17:09:01 +0000 (18:09 +0100)]
Merge branch release-2016 into release-2018
Change-Id: If1c8c37bbfc44ae383999b828d1dfea4cf67d1ea
Mark Abraham [Fri, 16 Feb 2018 15:06:21 +0000 (16:06 +0100)]
Bumped patch version for next release
Change-Id: I174548b1be5261aea861ec29d550ee323ce3702e
Berk Hess [Wed, 7 Feb 2018 10:03:46 +0000 (11:03 +0100)]
Correct multisim MPI barrier
An MPI barrier used with multisim was called on incorrect ranks
and a barrier was missing.
Fixes #2403
Change-Id: I06a66167afc8dca9cdd4ca9b4f9a806984a6ec7a
Aleksei Iupinov [Thu, 15 Feb 2018 12:27:48 +0000 (13:27 +0100)]
Fix PME for large systems with Fermi GPUs
PME spread/gather CUDA kernel scheduling did not account for
compute capability limitations. Realistically this has only
caused it to fail on CC 2.x with input systems larger than
2^18 ~= 262k atoms. This is now fixed for all CUDA architectures.
Fixes #2409
Change-Id: I59295b5d53a341d08a221aebb52e1db9f1e80107
Mark Abraham [Fri, 16 Feb 2018 07:41:28 +0000 (08:41 +0100)]
Version 2016.5
Change-Id: I7dd31f2deea884f3075555547c4f3691bd4adae3
Aleksei Iupinov [Thu, 15 Feb 2018 14:19:33 +0000 (15:19 +0100)]
Add the missing comma in the quote
This was printed as:
GROMACS reminds you: "It's just B I O L O G Y, can't you see?Joe Jackson" ((null))
Change-Id: I69e86ecda8930fffeeab3a54f2f015339fd9b345
Mark Abraham [Wed, 14 Feb 2018 12:54:33 +0000 (13:54 +0100)]
Merge branch release-2016 into release-2018
This patch has already been incorporated into release-2018
Change-Id: I088c0d44b5620b3eb575e8a1d0a05adeefaee2cc
Berk Hess [Tue, 13 Feb 2018 10:40:28 +0000 (11:40 +0100)]
Fix range bug in add_selection_groups in tngio.cpp
If there were more molecule blocks than types in the topology,
atoms from the later blocks could be missing in selections written
to tng.
Change-Id: I8bf840b2f019ca9b3c2ccefcf48da866bf8ec559
Mark Abraham [Thu, 25 Jan 2018 08:34:45 +0000 (09:34 +0100)]
Fix bug in checkpoint error checking
Change-Id: I91a185913356a42949596d8f4ca34ec2394c24a5
Berk Hess [Tue, 13 Feb 2018 11:55:35 +0000 (12:55 +0100)]
Fix grompp net charge check
The grompp check for the net charge would ignore molecule blocks
at the end when molecule types are used in multiple, non consecutive
molecule blocks.
Fixes #2407
Change-Id: I4e7c6c609da99025174f180f9d4037ac9f9fd7fc
Berk Hess [Tue, 13 Feb 2018 11:55:35 +0000 (12:55 +0100)]
Fix grompp net charge check
The grompp check for the net charge would ignore molecule blocks
at the end when molecule types are used in multiple, non consecutive
molecule blocks.
Fixes #2407
Change-Id: I4e7c6c609da99025174f180f9d4037ac9f9fd7fc
Mark Abraham [Tue, 13 Feb 2018 07:11:35 +0000 (08:11 +0100)]
Prepare release notes for 2018.1
Change-Id: I342130c590bff57df1369d488fb98491e936f78f
Mark Abraham [Tue, 13 Feb 2018 16:11:19 +0000 (17:11 +0100)]
Use death test only when supported
In build farms for bioconda and DebiChem, some environments don't
support death tests, so hopefully if we use the GoogleTest machinery
to avoid using them when not supported we can avoid issues in such
environments.
Refs #2405
Change-Id: I17efe3b620b162d615b91a5f5cd4c8bd458143a5
Mark Abraham [Tue, 6 Feb 2018 16:22:16 +0000 (17:22 +0100)]
Use more portable python shebangs
Per https://www.python.org/dev/peps/pep-0394/#recommendation, we
should use env, and point it at python2. When we either make them 2/3
or just-3 compatible, this should change.
Some distros (notable Arch Linux) already point python at python3 so
we should choose to be explicit, and thus somewhat portable.
Fixes #2401
Change-Id: Icf98e04660c798d10b167f971b20698e0741994a
Mark Abraham [Thu, 8 Feb 2018 17:45:26 +0000 (18:45 +0100)]
Merge "Merge branch release-2016 into release-2018" into release-2018
Mark Abraham [Thu, 8 Feb 2018 16:55:11 +0000 (17:55 +0100)]
Merge branch release-2016 into release-2018
Resolved conflicts in commit matrices by implementing the same logical
change as in release-2016 (move msvc+icc config to post-submit matrix)
in release-2018.
Change-Id: Id5d731e37e54b8c3896763d0a4ca496f99612f0f
Berk Hess [Wed, 24 Jan 2018 15:26:08 +0000 (16:26 +0100)]
Use SIMD bondeds without perturbed interactions
With free-energy, but without perturbed bonded interactions,
we would use the SIMD accelerated bondeds functions.
Change-Id: I67bdb0cba9c6ec20b076576af940d2be51fa6c9f
Berk Hess [Wed, 7 Feb 2018 12:15:19 +0000 (13:15 +0100)]
Allow bonds with distance=0
We should allow overlapping atoms in harmonic bonds. But the code
would cause a floating point exception and incorrect free-energy
derivatives.
Note that using sqrt(x) instead of x*invsqrt(x) is faster on Intel.
Change-Id: I87dd1e84f8f526ac40e482bccc94d89e33a1d112
Aleksei Iupinov [Tue, 6 Feb 2018 15:29:43 +0000 (16:29 +0100)]
Zero unused values to silence the warning
Change-Id: I8419a823e9a94434c3c534da4bf98ca79ff6ead7
Mark Abraham [Wed, 31 Jan 2018 07:48:16 +0000 (08:48 +0100)]
Prevent log file energy average printing dividing by zero
If there's not enough data for averages, skip the average printing
entirely. In principle, pr_ebin could be called from elsewhere than
print_ebin, so now we print "N/A" rather than "nan" or segfaulting.
Added some const correctness. Added some tests. Unfortunately they
currently have to write a file to disk, but we should have the test
before refactoring it to run in memory. The tests needed to be able
to clean up their memory, so runner can now also call done_ebin
but for stability we'll leave that off for now.
TODO When merging to master branch enable the call to done_ebin
in runner().
Fixes #2394
Change-Id: Id2e151664dd4553e49880f96bdd90aacc945f992
Mark Abraham [Tue, 6 Feb 2018 07:52:13 +0000 (08:52 +0100)]
Use isfinite unambiguously
Patch provdied by Veselin Kolev
Fixes #2400
Change-Id: I1923998a2288e20cb339e7fc59210cbf9ac5bfa8
Aleksei Iupinov [Mon, 5 Feb 2018 11:31:57 +0000 (12:31 +0100)]
Only setup Ewald in the generic group kernel if it's used
Helps to not divide by garbage values.
Change-Id: I67999c56b2ce6105d154e7f90b137f67bafd4673
Berk Hess [Mon, 5 Feb 2018 15:05:23 +0000 (16:05 +0100)]
Set cutoff modifiers in forcerec
The cutoff modifiers were not copied from interaction_const_t
to forcerec_t which meant only the generic kernels was used with
the group scheme.
Fixes #2399
Change-Id: I2890cbd55e648e434ad3817edc4e5fe4c5697162
Berk Hess [Mon, 5 Feb 2018 09:38:22 +0000 (10:38 +0100)]
Fix assertion failure in TPI
Fixes #2398
Change-Id: I89ddde7bdb6c061411160009df1f87e7c2cef0a1
Mark Abraham [Wed, 24 Jan 2018 13:23:00 +0000 (14:23 +0100)]
Move Windows icc config to post-submit testing
The front end and standard library for this build are the same as the
msvc config, so there is not a great deal of coverage added by this
config. Post-submit does not use this slave at all. Only running one
job per matrix on this slave will permit us to use only one executor
on it, which might reduce the number of strange failures and hangups.
Fixes #2387
Change-Id: Iedc421b7f4ea6d74f0b4cbd5cc88ea958b427213
Roland Schulz [Fri, 19 Jan 2018 21:23:27 +0000 (14:23 -0700)]
AVX512 GCC 5.3 work-around
GCC 5.3 has bug in overload resolution causing the AVX512
and scalar function to become ambiguous.
Change-Id: I1db4696f19c05c9eb3a97320a8c6c5b41823cca9
Roland Schulz [Thu, 18 Jan 2018 22:13:03 +0000 (14:13 -0800)]
Remove incorrect release-note entry
Change was reverted before the release.
Change-Id: Id46ba251edc9a44cd62275d03ecfe15d5230cfd9
Berk Hess [Fri, 12 Jan 2018 14:23:40 +0000 (15:23 +0100)]
Fix COMM removal on part of the system
COMM removal requested for part of the system acted on the whole
system.
Fixes #2381
Change-Id: Ic36ce0d261dd9b5aeceef1803f6a97e6a9ca8716
Berk Hess [Thu, 18 Jan 2018 10:51:36 +0000 (11:51 +0100)]
Add check in grompp to avoid assertion failure
With an mdp file with a parameter present with both the current name
and the old name which automatically gets replaced, an assertion
would fail. Now a fatal error is issued.
Fixes #2386
Change-Id: Idc66145afaca6a1e0cb868dd6aba2405e2dd7171
Aleksei Iupinov [Tue, 16 Jan 2018 17:08:57 +0000 (18:08 +0100)]
Fix box scaling in PME mixed mode
A TODO is left to test and handle this better in master.
Fixes #2385
Change-Id: I34d01d3ea6802f624888a1fd85eef72711a82f2f
Berk Hess [Tue, 16 Jan 2018 14:13:08 +0000 (15:13 +0100)]
Re-enable GPU with walls and 1 energy group
With a single non-bonded energy groups and walls we can use a GPU
for non-bonded calculations.
Change-Id: I95e72d6acc8852091bb269b350cdbf67f56e19fc
Berk Hess [Tue, 16 Jan 2018 13:02:37 +0000 (14:02 +0100)]
Remove tumbling ice-cube warning with SD
With SD there is friction, so ice cubes will not tumble.
Change-Id: Ibe91856529c26386e3f3683f95b574a9c6e7a875
Berk Hess [Fri, 12 Jan 2018 10:50:33 +0000 (11:50 +0100)]
Avoid mdrun echoing "No option -multi"
mdrun would print as many messages "No option -multi" as there
are MPI ranks to stderr.
Also updated -multi to -multidir in an error message.
Fixes #2377
Change-Id: I697956e28f904b148974183ad85ed7ed4ec82221
Szilárd Páll [Thu, 11 Jan 2018 16:37:21 +0000 (17:37 +0100)]
Call GPU profiler reset on separate PME ranks
As separate PME ranks can now offlaod to GPU, to be able to collect
profiling data, the reset call was added.
Fixes #2378
Change-Id: I48ad62ab51165347471efc9c5710c49f0001b2e0
Mark Abraham [Thu, 11 Jan 2018 02:35:05 +0000 (03:35 +0100)]
Bumped version for next patch release
Change-Id: I264bca65d7d327387196453faeccbedb87e4fcb8
Mark Abraham [Tue, 9 Jan 2018 21:25:31 +0000 (22:25 +0100)]
Version 2018
Removed "-rc1" per policy. Bumped regressiontests hash.
Updated release notes for final release, including release date, and
removing beta-phase fix content. Sorted tools page by tool name/topic.
Fixed some text to refer to changes in the past tense. Updated
some line wrapping.
Change-Id: Ia82a1cef3270dbdddd173c58248fc9bf6ea420c5
Mark Abraham [Wed, 10 Jan 2018 12:22:27 +0000 (13:22 +0100)]
Announce in user log files that features are deprecated.
These are merely informational notes, not warnings or errors.
Refs #1781, #1971, #2136
Change-Id: I96e19acb0e15d3f42b0929f555b451299a2882e4
Roland Schulz [Wed, 10 Jan 2018 18:26:18 +0000 (10:26 -0800)]
Fix GMX_ENABLE_AVX512_TESTS being internal
Internal implies FORCE and thus variable couldn't be used to disable test.
Change-Id: I07521af7e998b4b032106f020a918ed2a0e7f747
Mark Abraham [Tue, 9 Jan 2018 17:52:37 +0000 (18:52 +0100)]
Fix automated GPU ID assignment
When we permitted separate PME-GPU ranks, we should have relaxed
this logic also.
However, the performance in such cases is not very predictable, so if
there's a distribution of tasks with more than one task to a GPU that
is uneven, then we should require the user to specify exactly what
they want. This also reinstates the 2016-era behaviour where, if
running multiple PP ranks on GPUs, that mdrun will not by default
produce an unbalanced mapping with more than one task per GPU.
Change-Id: I5b2fad317ecbb4e5e02fccd68e15350b678df34c
Szilárd Páll [Fri, 5 Jan 2018 02:54:44 +0000 (03:54 +0100)]
User guide updates to GPU section
Refs #2359
Change-Id: Ied4945f83dc19f93a51aa78f51b7cea18021f5b8
Szilárd Páll [Tue, 9 Jan 2018 19:33:24 +0000 (20:33 +0100)]
Release notes tweaks
* Move dynamic pruning to features.
* Add new CPU arch support to the perf highlights.
Change-Id: Ia6228cb779ebcebd4ba81505c8545b85f00ffc90
Mark Abraham [Wed, 6 Dec 2017 01:13:35 +0000 (12:13 +1100)]
Add release notes to source repo
Release notes content is current up to 2018-rc1, and will need to be
updated for the final release. Also several TODOs to remove the
content of beta-phase fixes, and text about them.
Change-Id: I6f5999327c9ac4ab56d8057b5eb2b4f8351eda87
Erik Lindahl [Tue, 2 Jan 2018 10:55:31 +0000 (11:55 +0100)]
Improve pdb2gmx for nonstandard residue types
If explicit non-blank chain identifiers are set, it
will now be a hard error if the residue types in each
chain do not match. For blank chain ID we still need
to allow detection of non-chain parts, but this case
too now provides more explicit output information.
Fixes #2370.
Change-Id: Ic70f4d91fddd80a22e91eb1a0672a15838966a75
Szilárd Páll [Fri, 5 Jan 2018 02:51:26 +0000 (03:51 +0100)]
Install guide updates
Updated install guide to reflect chnges in for the 2018 release.
Refs #2359
Change-Id: Ic679e7d7e96392c2c28185db1f71928146cd742d
Szilárd Páll [Tue, 19 Dec 2017 18:20:42 +0000 (19:20 +0100)]
Environment variable docs maintenance
Removed deprecated variables (some ancient, some recent), corrected some
misspelled ones.
Added undocumented environment variables.
Left a comment with handy one-liner to discover leftovers as well as
undocumented variables.
Refs #2359
Change-Id: Icb80cf661e2a53c5049fe18d687b987fb2c025f1
Aleksei Iupinov [Fri, 5 Jan 2018 11:25:28 +0000 (12:25 +0100)]
Do not print PME counter breakdown if it's empty
Fixes #2361
Change-Id: I4272cec58d837d2f1038d23ceb1eaf52736e3d52
Mark Abraham [Wed, 3 Jan 2018 10:35:38 +0000 (11:35 +0100)]
Handle erroneous command line args better
Some gmx modules need to be able to accept non-option arguments, and
some should not. Introduced enough functionality to support such
behaviour, while giving useful error messages in cases where the
command line is merely missing hyphens (which can happen e.g. when
people copy-paste from inconveniently built PDF files for tutorials).
Increased test coverage of relevant cases.
Removed some useless command-line argument strings from test cases
that never needed them.
Also tested some behaviours of handling string options, and renamed
some test input strings to reflect the intent.
Fixes #2153
Change-Id: Icec02ea86a57a8b129a621121518036b9f032078
Erik Lindahl [Mon, 8 Jan 2018 23:14:25 +0000 (00:14 +0100)]
Clarified gmx rmsf documentation
Made it clearer what input files coordinates
are taken from when writing B-factors.
Fixes #2320.
Change-Id: Ic47cbb7d9b1036c9e87a77b244dde3cd3da5df36
Berk Hess [Tue, 9 Jan 2018 20:35:20 +0000 (21:35 +0100)]
Reword CPU/GPU imbalance notes
Changes text in CPU/GPU imbalance from "performance loss" to "wasting
resources", since in some cases one can not get higher performance.
Replaced "GPU has less load" by "CPU has more load".
Removed hint to reduce the cut-off, since one often can not do this.
Note that with CUDA all theses notes are never printed, since we no
longer have timings on (by default), unlike with OpenCL.
Fixes #2253
Change-Id: Ib4a9752ad27c1cd2a3cd751a217249694a56d3b7
Mark Abraham [Tue, 9 Jan 2018 16:50:46 +0000 (17:50 +0100)]
Deactivate autoselect of PME on GPU on separate rank
Although the code does work if such a setup is specified manually, we
intended to require the user to specify mdrun -nb gpu -pme -gpu -npme
1 to get this setup.
Change-Id: I27e0e517435e0640f3a2bab1c62fc61034ec7789
Mark Abraham [Tue, 9 Jan 2018 15:02:34 +0000 (16:02 +0100)]
Clean up user guide for release
Various work in progress remains on master branch, but we need
something decent looking for this release, even if incomplete.
Reordered the list of files within the user guide roughly along the
lines of what might be likely to interest a user needing information,
but that's not exactly an exact science.
Change-Id: I2c13f318844e42f796804430b4519e18e62b4857
Mark Abraham [Tue, 9 Jan 2018 07:42:07 +0000 (08:42 +0100)]
Merge branch release-2016 into release-2018
Change-Id: I22d5b28794a4bbb3b122af339a282695aa64f266
Berk Hess [Mon, 8 Jan 2018 21:45:46 +0000 (22:45 +0100)]
Do not turn on DLB at replica exchange
Turning on DLB right after exchanging replicas caused an assertion
failure and is also useless.
Fixes #2298
Change-Id: I20c3cb6ef3d74907d53d447fa9b0c9168f03c769
Berk Hess [Mon, 8 Jan 2018 18:37:34 +0000 (19:37 +0100)]
Fix triclinic domain decomposition bug
With triclinic unit-cells with vectors a,b,c, the domain decomposition
would communicate an incorrect halo along dimension x when b[x]!=0
and vector c not parallel to the z-axis. The halo cut-off bound plane
was tilted incorrect along x/z with an error approximately
proportional to b[x]*(c[x] - b[x]*c[y]/b[y]).
When c[x] > b[x]*c[y]/b[y], the communicated halo was too small, which
could cause instabilities or silent errors.
When c[x] < b[x]*c[y]/b[y], the communicated halo was too large, which
could cause some communication overhead.
Fixes #2125
Change-Id: I2109542292beca5be26eddc262e0974c4ae825ea
Carsten Kutzner [Mon, 8 Jan 2018 10:46:47 +0000 (11:46 +0100)]
Fix a few typos in SIMD documentation
Change-Id: Ia64f936098172eb4e13d3bacf1caa824c43779fc
Erik Lindahl [Fri, 15 Dec 2017 15:16:02 +0000 (16:16 +0100)]
Test that nvcc/host compiler combo works
Compile a trivial CUDA program during CMake time
to catch both unsupported nvcc/host compiler
version combinations and other unknown errors.
Fixes #1616.
Change-Id: I3cc55e4d0db9d6eb01e8a7cd8916cc7a7a1e21fd
Erik Lindahl [Fri, 29 Dec 2017 09:52:28 +0000 (10:52 +0100)]
Make TNG writing work with multiple identical steps
Introduce a wrapper structure around TNG so we detect
and correct for cases when writing multiple frames with
the same step, or non-zero initial steps to TNG files.
This will avoid frames overwriting each other, and make
sure the time per frame is correct.
Fixes #2189.
Change-Id: I3f44a2681138351d89183e3e196269f2b2f258fc
Erik Lindahl [Fri, 5 Jan 2018 10:49:45 +0000 (11:49 +0100)]
Add AVX_512 and KNC symbols to FFTW SIMD test
Otherwise the CMake code might complain loudly about
FFTW not being accelerated on KNC or KNL hosts.
Change-Id: I888936e9cec18436d775cb3c6eedc21335ab0e95
Mark Abraham [Tue, 19 Dec 2017 07:25:06 +0000 (18:25 +1100)]
Remove FindCUDA from source tarball
When we decided to bundle FindCUDA.cmake and supporting files to keep
our developer and Jenkins builds free of warnings, we did not take
care to ensure that the user gets the behaviour from their cmake,
since the combination of their cmake and its FindCUDA.cmake is much
more likely to work correctly than their cmake and FindCUDA.cmake from
cmake 3.4.3.
The testing of the release tarball as part of the release workflow
will ensure that we don't silently ship a broken CUDA build.
Refs #2276
Change-Id: I9ed6fa97921eb5c3ce3b4a6ecc401b4dd26e1251
Erik Lindahl [Thu, 28 Dec 2017 13:51:35 +0000 (14:51 +0100)]
Improve frame time/step handling in trjconv
Store the exact step in PDB/GRO file headers,
and be more careful about not claiming to have
time or step information when it was not available.
This change will avoid some of the problems
described in #2189, but it does not yet properly
fix the issue in the TNG library.
Refs #2189.
Change-Id: I44bb59fbca83da53f6e8d4e494ae6476a82eb7cd
Erik Lindahl [Wed, 27 Dec 2017 11:50:51 +0000 (12:50 +0100)]
Implement changes for CMake policy 0068
CMake-3.9 introduced a changed behavior for RPATH
vs. install_name options on OS X. This avoids relying
on functionality that will be removed in future
CMake versions.
Change-Id: Ic3697615bc372d09d45dfd007ba5a088624de051
Mark Abraham [Mon, 4 Dec 2017 02:57:49 +0000 (13:57 +1100)]
Add test config of GPU build on slave without GPU
This will test that e.g. GPU detection and cleanup code does not give
unexpected fatal errors.
Change-Id: Ic3314777b6f530fe02aa55e05ca3d352eb786d7d
Paul Bauer [Wed, 3 Jan 2018 12:31:24 +0000 (13:31 +0100)]
Fix gmx msd when using COM removal and molecules
Changed order of code to actually assign correct coordinates before
copying the data, and modified data structure size when using COM
removal and individual molecules.
Fixes #2043
Change-Id: Ic16f05a589609a43f14fd75753ca8589cf3d8c42
Erik Lindahl [Wed, 3 Jan 2018 10:48:34 +0000 (11:48 +0100)]
Fix trjconv for tdump<frame timestep
Set frame timestep before starting the loop by
reading first two frames and rewinding, and make
sure we always write something to the dump output
based on best-guess (if there is at least one
input frame present).
Fixes #1832.
Change-Id: Ic1db912c966448e68b307cfc1fd492f4677e7ab4
Erik Lindahl [Wed, 3 Jan 2018 14:44:01 +0000 (15:44 +0100)]
Fix index error in gmx chi
An error in the index construction could lead
to segfaults. However, the actual indices were
correct, so it should not have produced any
incorrect results.
Fixes #1814.
Change-Id: I191de7248615994d5c0dfb39eccafd53ebbedb67
Paul Bauer [Wed, 3 Jan 2018 10:39:00 +0000 (11:39 +0100)]
Correct dihedral restraint section in Manual
Fixes #2038
Change-Id: I37bfaa3083efe5721b2d30f4b1038062fc4dbe2b
Erik Lindahl [Wed, 3 Jan 2018 09:25:09 +0000 (10:25 +0100)]
Fix complexity for large exclusion orders
To avoid exploding computational complexity
for highly connected molecules with large
values for excluded neighbors, avoid
adding a neighbor to the temporary nnb
structure if it is already present as a
lower-order neighbor.
Fixes #2260.
Change-Id: I30c27a281ddcbbac175f79ab9d5bba49f2f988ae
Mark Abraham [Wed, 3 Jan 2018 09:17:00 +0000 (10:17 +0100)]
Fix duplicate warning
Thread-MPI, GPUs and energy groups was issuing this warning
twice. Note that no-MPI and real-MPI are still handled correctly -
they warn on the second call, which is the only one made in those
cases.
Fixes #2372
Change-Id: I68eb0505056b8edd5233a6abf1f2e30b09bbacee
Erik Lindahl [Sun, 31 Dec 2017 16:24:23 +0000 (17:24 +0100)]
Don't allow multiple energy groups for GPU runs
Exit with a fatal error instead of only warning, since the
latter leads to writing data for energy groups that
is incorrect to the energy file.
Fixes #1822.
Change-Id: I34ccb10bba6d6e1350283e34ebc908c6f830baab
Erik Lindahl [Sun, 31 Dec 2017 11:11:10 +0000 (12:11 +0100)]
Remove duplications in GMXLIB search paths
Remove entries that are duplicated, or identical
to the default search path, to avoid e.g.
listing identical force fields multiple times.
Fixes #1928.
Change-Id: Ib3caa116b4a6e1ce50424408832a96da52eba774
Erik Lindahl [Mon, 25 Dec 2017 21:34:14 +0000 (22:34 +0100)]
Avoid FTZ triggering simd test failures
For very small arguments on platforms without
FMA support, the Intel compiler's default
usage of flush-to-zero for denormal values can
lead to slight deviations. Since this is a
range we really don't care about, and non-FMA
platforms are anyway a thing of the past, just
avoid testing a very small range around that
threshold for non-FMA SIMD platforms.
Fixes #2335.
Change-Id: I7cfb14284e79534f114891ba4dcd1f1c7223faf3
Erik Lindahl [Tue, 2 Jan 2018 23:41:17 +0000 (00:41 +0100)]
Document required position restraint files
Explain properly to users why the files are
suddenly required when it has worked fine without
them for a decade.
Fixes #2237.
Change-Id: I7990994d2da1f1af95b07bb0110b9bfdd89f3e25
Erik Lindahl [Sun, 31 Dec 2017 13:20:10 +0000 (14:20 +0100)]
Allow empty lines in hdb files
Skip lines that consist only of whitespace. Not a universal
solution for fixing hdb files, but better than the user
getting very strange error messages that don't say anything
about whitespace.
Fixes #2028.
Change-Id: I0962c89d0251f83da7fe0440efd31fb7ca468690
Erik Lindahl [Fri, 29 Dec 2017 22:18:38 +0000 (23:18 +0100)]
Fix OpenCL compiles on Mac OS
Confirmed to work on Mac OS 10.13.2 running
on a Macbook Pro with Radeon Pro 560.
Fixes #2369.
Change-Id: I0b0056075ecbaf7c11e0022ed13e43cfe2e7484c
Mark Abraham [Tue, 2 Jan 2018 07:55:09 +0000 (07:55 +0000)]
Fix membed by calling the proper mtop finalization routine
Fixes #2364
Refs #2371
Change-Id: I27166acec903e4c0cf3b9a54aaf4b9db7d6974b0
Erik Lindahl [Sat, 30 Dec 2017 21:28:47 +0000 (22:28 +0100)]
Improve manual from comments by Gang Liu
Fixes #2078.
Change-Id: If1267822be4963a59bac4dc07f9731c9489cc3e3
Erik Lindahl [Sat, 30 Dec 2017 16:55:36 +0000 (17:55 +0100)]
Make it explicit that time units apply to begin/end times
The help documentation could be interpreted as if the
begin/end frame selection was always specified in ps, no
matter what the time unit was set to. This makes it
explicit that the time unit selection applies, and that ps
is just the default unit.
Fixes #2235.
Change-Id: If010a8d6f8e3e1949039ff4f32a5cc00058466e0
Erik Lindahl [Sun, 31 Dec 2017 14:03:05 +0000 (15:03 +0100)]
Don't require matching names between rtp and tdb files
This was only documented in the source. It's a remnant
from the days when all force fields were in the same
directory, and no longer necessary. With this change we
will properly match all termini to all amino acids.
Fixes #2026, #2027.
Change-Id: Ie7bf8e65892281cc2744146a5525be0c4afdcecf
Erik Lindahl [Mon, 1 Jan 2018 23:28:28 +0000 (00:28 +0100)]
Don't write reference pull group 0 to log
This is an internal group used for absolute
references, which cannot be set by users, so
printing it just leads to confusion.
Fixes #2143.
Change-Id: I89ee3a0e3a65855735b745f4ed7615ca47bfd36a
Erik Lindahl [Mon, 1 Jan 2018 16:53:00 +0000 (17:53 +0100)]
Fix gmx density for non-mass calculations
Implemented fix proposed by Klark Chen and Reid
Van Lehn so that we always use mass and never
charge/electron density to center systems.
Fixes #2230.
Change-Id: Id49741bd44349d43a27b5f20f4e498d2fd4ba1f9
Erik Lindahl [Mon, 1 Jan 2018 20:14:35 +0000 (21:14 +0100)]
Clarify XTC error message
Explain that the fatal error can also be
triggered by coordinates that are NaN, or
if the scaled values cannot be represented
with integers.
Fixes #2152.
Change-Id: I6eec60df48fe3994258bf8dce00ff33328c93f13
Mark Abraham [Tue, 26 Dec 2017 07:59:24 +0000 (18:59 +1100)]
First release candidate for version 2018
Change-Id: I893e4a9c0f2684955554ad537e6d9609e5f83102
Mark Abraham [Thu, 21 Dec 2017 10:56:59 +0000 (21:56 +1100)]
Separate canDetectGpus and findGpus futher, and fix tests
Renamed detect_gpus to findGpus so that no code can silently call
detect_gpus while forgetting to call the required canDetectGpus first.
Some test code is updated accordingly, which should have happened
earlier. The function with the new name now needs no return value, so
the formerly confusing return value of zero for success is no longer
present.
Shifted some more responsibilities from findGpus to canDetectGpus, so
that the latter now has responsibility for ensuring that when it
returns true, the former will always succeed.
Fixed tests that compile with CUDA, but cannot run unless there
are visible comatible devices and a valid context.
Refs #2347, #2322, #2321
Change-Id: I34acf8be4c0f0dcc29e931d83c970ba945865ca7
Roland Schulz [Wed, 27 Dec 2017 18:59:04 +0000 (10:59 -0800)]
Replace GMX_ALIGNED with alignas and SIMD alignment define
This allows each platform to define its required or
preferred SIMD alignment, and we avoid using alignments
larger than what might be supported on each architecture.
Also fix overalignment in pairs.
Fixes #2365
Change-Id: I4793adf31d186eade8a1fd8c920ab75c685ad53f
Mark Abraham [Tue, 26 Dec 2017 13:01:43 +0000 (14:01 +0100)]
Silence POWER8 compiler warnings
Change-Id: Id116b048d775d36ce8ac942380cfd64deed3bb7e
Roland Schulz [Tue, 26 Dec 2017 18:51:23 +0000 (10:51 -0800)]
Fix quote
Change-Id: Ic763145c231951cd8e61cf36f3486de7a3d30e30
Erik Lindahl [Tue, 12 Dec 2017 19:29:41 +0000 (20:29 +0100)]
Fix builds on ARM & clarify (ARM) GPU support
Fixed a typo in architecture.h that prevented
the Neon Asimd instructions from being selected,
and updated the CPU brand detection to also look
for a new label with Tegra X1 on Ubuntu 16.04
Clarified in error messages and documentation that
Gromacs in fact does not build all supported GPU
architectures by default, explain the common cases
when things might fail, exactly what the user
should do to enable the support, and how the
support strings should be formatted.
Fixes #2287.
Change-Id: I87a2eb81ee11b78f072e3ef359a00c75eb7ec24b
Mark Abraham [Fri, 22 Dec 2017 14:11:52 +0000 (01:11 +1100)]
Merge branch release-2016 into release-2018
Change-Id: I909443584916c574f94b38d0aab2701163faecad
Mark Abraham [Thu, 21 Dec 2017 14:00:56 +0000 (01:00 +1100)]
Merge branch release-5-1 into release-2016
Change-Id: Ic4aa87c01dfe281c9c8aa31fbc0f7f9fcbe9752d
Mark Abraham [Thu, 21 Dec 2017 13:56:25 +0000 (00:56 +1100)]
Bump minor version for theoretical new release
Change-Id: I54152717d87ad84e7fa5549887b587aa10fe43e4
Pascal Merz [Thu, 14 Dec 2017 22:55:34 +0000 (15:55 -0700)]
Documentation and fixes for physical validation
Addresses #2349
Adds documentation for the physical validation suite in
docs/dev-manual/physical_validation.rst
As this was misunderstandable, changed the default behavior of
`make check-phys` and `make check-all` to actually run the simulations.
This might take very long, but since the physical validation tests need to
be turned on explicitly via cmake option, the chances of somebody using the
tests by mistake are low. The `check` targets are:
* `make check`: Run unit and regression tests (unchanged)
* `make check-phys`: Run simulations needed for physical validation, then
run physical validation tests
* `make check-phys-analyze`: Only run physical validation tests, assuming
that simulations were run previously and are available.
* `make check-all`: Combination of `make check` and `make check-phys`
Additionally, `make check-phys-prepare` can be used to prepare GROMACS
input files and a script to run the simulations needed for the physical
validation tests.
Bugfixed dependecy issue that attempted to run validation test before
target `gmx`.
Fix in physical_validation package to avoid including matplotlib by default
(and hence requiring X and Gtk).
Improved output of physical validation script (update while running
simulation, info about script after preparing input files, etc).
Additional minor changes not mentioned in #2349:
* Allowing name of test to be different of directory of input files, which
allows to define several tests on the same input files.
* Adapted some tests - more thorough checking of ensembles (check
temperature- and pressure-dependence separately *and* collectively), and
updated tolerance levels.
* Bumped physical_validation package to newest version to include bugfixes
and stability improvements.
Change-Id: I7219f5392ed068ba1b2e13b30efe5aa4e36ff586
Mark Abraham [Thu, 7 Dec 2017 04:36:49 +0000 (15:36 +1100)]
Update double-precision test configurations
These changes improve coverage of double precision, using more release
mode, particularly with latest gcc and icc, and using 128-bit SIMD,
which have been cases that were buggy recently. The other aspects of
the configurations that have been modified have been
non-critical. Where appropriate, brief rationales are recorded. This
resolves an old TODO item in the post-submit matrix.
Fixed a sign mismatch in initializing an OpenCL variable that didn't
need to be initialized.
Noted relevant new TODOs.
Refs #2300, #2325, #2326, #2334, #2335, #2336, #2337, #2338
Change-Id: I131fa1a6776d1e7809799c3f931a1fc8100fcdc9
Mark Abraham [Thu, 21 Dec 2017 08:47:01 +0000 (19:47 +1100)]
Version 5.1.5
Bumped SOVERSION_MINOR and REGRESSIONTEST_HASH
Change-Id: Ie62ecacbe53f4f39e4592a7593ed53ffe47fd642
Roland Schulz [Wed, 20 Dec 2017 19:42:17 +0000 (11:42 -0800)]
Update checks for BuildOwnFFT
- Allow on Windows (e.g. WSL, Mingw)
- Disallow with Ninja (broken)
Fixes #2356
Change-Id: I8ac5dd520f92b882dcaeb009792fae2d6e9f0062
Aleksei Iupinov [Wed, 20 Dec 2017 13:16:38 +0000 (14:16 +0100)]
Set GMX_GPU_AUTO to FALSE with GMX_GPU defined
Refs #1985, #2357
Change-Id: I5cada97015ee94717ea6eb988b3a84a351f11293
Mark Abraham [Wed, 20 Dec 2017 11:13:00 +0000 (22:13 +1100)]
Use new defaultRealTolerance() to stabilise some tests
The single-precision values in these tests sometimes failed to compare
equal with string-ified versions when the test was run from a
double-precision build on Windows. That difference may have originated
in how a rounding mode was implemented in different libraries, but
ultimately doesn't matter because the value should not be compared as
if it was computed in double precision.
Change-Id: I9b7c2b9409145cc579229f0866935754e3a9dcac
Mark Abraham [Wed, 20 Dec 2017 12:33:56 +0000 (13:33 +0100)]
Merge "Merge branch release-2016 into release-2018" into release-2018
Mark Abraham [Tue, 19 Dec 2017 13:23:59 +0000 (00:23 +1100)]
Merge branch release-2016 into release-2018
Change-Id: I2c3aa754de1b8ff971854740da9815fff8a41f0d
Aleksei Iupinov [Tue, 19 Dec 2017 15:40:23 +0000 (16:40 +0100)]
Remove potentially wrong "per user request" note from npme reporting
Refs #2204
Change-Id: Idd2127737b7b1af9cec9f7b547478b8b82cdb59a
Berk Hess [Tue, 19 Dec 2017 10:55:25 +0000 (11:55 +0100)]
Make acceleration correction VCM mode work
The new acceleration correction VCM mode did not actually correct
the coordinate for the acceleration, since a null pointer was passed.
Introduced an extra CGLO flag to allow for correction of the
coordinates, but leave the initial coordinates unaffected.
Change-Id: I673793902df7241085fff20c63cf3ce88ef60313