alexxy/gromacs.git
2 years agoAdd PP decomposition simluationWorkload flags
Szilárd Páll [Thu, 12 Aug 2021 16:57:42 +0000 (18:57 +0200)]
Add PP decomposition simluationWorkload flags

Add the complete set of flags that indicate whether there is PP
decomposition using CPU/GPU communication.

These flags allow simplifications in the force schedule.

The additional benefit is a reduced reliance on passing around commrec
for checks related to parallelization.

Refs #3913

2 years agoIntroduce ObservablesReducer
Mark Abraham [Wed, 18 Aug 2021 14:36:19 +0000 (14:36 +0000)]
Introduce ObservablesReducer

2 years agoFix compiler warnings introduced in 8f6ea9eb6c (!1865)
Andrey Alekseenko [Wed, 18 Aug 2021 09:42:11 +0000 (12:42 +0300)]
Fix compiler warnings introduced in 8f6ea9eb6c (!1865)

- Add ifdef around clang-specific pragmas.
- Add newline at the end of file.

2 years agoSimplify nblib listed forces type tests
Joe Jordan [Wed, 18 Aug 2021 09:31:18 +0000 (09:31 +0000)]
Simplify nblib listed forces type tests

2 years agoGpuEventSynchronizer: extract backend-specific functionality
Andrey Alekseenko [Wed, 18 Aug 2021 08:52:57 +0000 (08:52 +0000)]
GpuEventSynchronizer: extract backend-specific functionality

No new functionality or added checks.

Preparatory work for adding advanced event accounting logic to GpuEventSynchronizer.

DeviceEvent class contains only backend-specific functions and minimal sanity checks.
The more advanced logic (the one we're trying to fix in #3988) is kept in
GpuEventSynchronizer, and currently left unchanged (1:1 for OpenCL and SYCL,
very relaxed rules for CUDA).

Refs #2527, #3988.

2 years agoUpdate GPU CI configuration
Mark Abraham [Tue, 17 Aug 2021 17:01:37 +0000 (17:01 +0000)]
Update GPU CI configuration

CUDA 11.4.1 updates the most recent supported versions of gcc and
clang so we should update those also.

ROCM 4.3 is out, and is more likely than any older version to be the
one we will support in 2022, so we should update them.

Several recent bug fixes in hipSYCL are worth having to avoid needing
to work around them in other changes.

Previously we re-used the hipsycl container for clang-tidy with CUDA,
but as we wish that to remain at clang version 11, we need an alternative
solution. Simplest would be to add CUDA to the linting image, but that
also builds LLVM with TSAN. As that TSAN build didn't work out of the box
on a CUDA base image, I split the TSAN image off from the image that does
linting, and added CUDA to the latter.

2 years agoAdd workaround for CorrelationsTest with Intel ICPX
Andrey Alekseenko [Tue, 17 Aug 2021 16:39:44 +0000 (16:39 +0000)]
Add workaround for CorrelationsTest with Intel ICPX

2 years agoAdd basic tests for GpuEventSynchronizer
Andrey Alekseenko [Tue, 17 Aug 2021 15:29:14 +0000 (15:29 +0000)]
Add basic tests for GpuEventSynchronizer

2 years agoMove the ownership of the xUpdatedOnDevice event to update constraints
Artem Zhmurov [Tue, 17 Aug 2021 15:13:06 +0000 (15:13 +0000)]
Move the ownership of the xUpdatedOnDevice event to update constraints

2 years agoAdd Urey-Bradley listed interaction to nblib
Joe Jordan [Tue, 17 Aug 2021 09:03:09 +0000 (09:03 +0000)]
Add Urey-Bradley listed interaction to nblib

2 years agoRandom cosmetic fixes
Andrey Alekseenko [Tue, 17 Aug 2021 05:26:56 +0000 (05:26 +0000)]
Random cosmetic fixes

2 years agoAdd common header to GpuEventSynchronizer
Artem Zhmurov [Mon, 16 Aug 2021 16:04:30 +0000 (19:04 +0300)]
Add common header to GpuEventSynchronizer

Move the pre-processor conditional to a common header, that can be
included unconditionally. Will simplify code unification further on.

2 years agoUpdates for updated Python version requirement.
M. Eric Irrgang [Mon, 16 Aug 2021 16:01:38 +0000 (19:01 +0300)]
Updates for updated Python version requirement.

* Update documented required Python version.
* Remove some references to Python 3.6 and 3.7.
* Cite issues where appropriate.

Refs #3716

2 years agoAdd `breathe` package to CI Python venvs.
M. Eric Irrgang [Mon, 16 Aug 2021 10:08:01 +0000 (13:08 +0300)]
Add `breathe` package to CI Python venvs.

It seems likely that at least one of the projects documented in the
GROMACS manual will want to incorporate documentation from C++ headers
(Doxygen) into the Sphinx pages. We can update the container build
script in advance so that the package will be available for
experimentation and testing after the next container rebuild.

See also #2698

2 years agoRemove support for Python 3.6
Mark Abraham [Mon, 16 Aug 2021 12:05:49 +0000 (14:05 +0200)]
Remove support for Python 3.6

This version goes end-of-life before we will release GROAMCS 2022, so
we should stop testing it. We decided this a long time ago, but didn't
actually do it.

Refs #3708

2 years agogrompp no longer modifies nstcomm
Berk Hess [Tue, 29 Jun 2021 09:43:07 +0000 (11:43 +0200)]
grompp no longer modifies nstcomm

2 years agoUpdates to nblib listed forces kernels
Joe Jordan [Mon, 16 Aug 2021 14:18:22 +0000 (14:18 +0000)]
Updates to nblib listed forces kernels

2 years agoRework nblib kernel checks and setup code
Joe Jordan [Mon, 16 Aug 2021 13:49:41 +0000 (13:49 +0000)]
Rework nblib kernel checks and setup code

2 years agoUpdate regular CI images to OneAPI 2021.3.0
Andrey Alekseenko [Fri, 13 Aug 2021 17:59:44 +0000 (20:59 +0300)]
Update regular CI images to OneAPI 2021.3.0

2 years agoDo not allow AtomLocality::All for coordinates and velocities transfer.
Artem Zhmurov [Sun, 15 Aug 2021 13:40:55 +0000 (13:40 +0000)]
Do not allow AtomLocality::All for coordinates and velocities transfer.

1. Use AtomLocality:Local instead of AtomLocality::All in PME-only ranks
   and in PME tests.
2. Assert on using AtomLocality::All for positions and velocities transfers,
   print verbose message on assert.

This reduces the number of possible streams used in StatePropagatoerData, hence
reduces its complexity. In preparation for buffer state tracking.

Refs. #3323

Refs. #3020

2 years agoFix MSVC error in toppush.cpp
Dmitry Morozov [Fri, 13 Aug 2021 15:47:49 +0000 (15:47 +0000)]
Fix MSVC error in toppush.cpp

2 years agoProvide better defaults for GMX_GPU_NB_CLUSTER_SIZE with SYCL
Mark Abraham [Fri, 13 Aug 2021 08:06:18 +0000 (08:06 +0000)]
Provide better defaults for GMX_GPU_NB_CLUSTER_SIZE with SYCL

Now that we know whether a SYCL configuration targets hipSYCL and if
so for which hardware, we can provide a default non-bonded cluster
size that works well in all SYCL cases.

Refs #3965

2 years agoRemove duplicate waiting for coordinates
Andrey Alekseenko [Fri, 13 Aug 2021 07:39:35 +0000 (07:39 +0000)]
Remove duplicate waiting for coordinates

2 years agoAdded class for generating standard CP2K input
Dmitry Morozov [Thu, 12 Aug 2021 09:40:38 +0000 (09:40 +0000)]
Added class for generating standard CP2K input

2 years agoNblib doc fixes
ejjordan [Tue, 10 Aug 2021 14:08:38 +0000 (16:08 +0200)]
Nblib doc fixes

Update some doc strings to make future merges more concise.

2 years agoFix few issues with open-source Intel LLVM (Clang 13)
Andrey Alekseenko [Wed, 11 Aug 2021 16:22:05 +0000 (19:22 +0300)]
Fix few issues with open-source Intel LLVM (Clang 13)

- Add -Wno-reserved-identifier to avoid false-positive with _real
  user-defined literal (https://bugs.llvm.org/show_bug.cgi?id=50644).
- Change the alignment-checking code to avoid "warning: performing
  pointer subtraction with a null pointer may have undefined behavior".
- Update some comments about SYCL version quirks.

There are some deprecation warnings related to SYCL naming being brought
into agreement with the final SYCL2020 standard, but those will be fixed
once they appear in any official release.

2 years agoAdd hipSYCL sanity check
Andrey Alekseenko [Wed, 11 Aug 2021 10:12:21 +0000 (10:12 +0000)]
Add hipSYCL sanity check

2 years agoImprove testutils infrastructure
Mark Abraham [Tue, 10 Aug 2021 19:52:49 +0000 (19:52 +0000)]
Improve testutils infrastructure

2 years agoAdd SYCL SETTLE kernel
Artem Zhmurov [Mon, 9 Aug 2021 14:52:34 +0000 (14:52 +0000)]
Add SYCL SETTLE kernel

Refs #3931

2 years agoFix GMX_PYTHON_PACKAGE option.
M. Eric Irrgang [Mon, 9 Aug 2021 12:56:48 +0000 (12:56 +0000)]
Fix GMX_PYTHON_PACKAGE option.

2 years agoOnly enable GPU FFT test for build supporting it
Andrey Alekseenko [Mon, 9 Aug 2021 10:47:36 +0000 (10:47 +0000)]
Only enable GPU FFT test for build supporting it

2 years agoMinor clean-up to sample_restraint tests.
M. Eric Irrgang [Mon, 9 Aug 2021 09:20:01 +0000 (09:20 +0000)]
Minor clean-up to sample_restraint tests.

2 years agoAvoid expensive atom-type lookups in grompp
Mark Abraham [Sat, 7 Aug 2021 14:38:20 +0000 (14:38 +0000)]
Avoid expensive atom-type lookups in grompp

For each dihedral interaction for which parameters needed to be looked
up, many non-inline function calls were made to see if dihedral types
matched. Instead they are now compared as two ranges with std::equal.

2 years agoAdd a helpful dev docs link.
M. Eric Irrgang [Fri, 6 Aug 2021 08:22:48 +0000 (11:22 +0300)]
Add a helpful dev docs link.

Try to connect the contribution guidelines better
to the requirements on acceptable commits.

2 years agoLift atom type lookup out of inner loops
Mark Abraham [Mon, 26 Jul 2021 08:31:26 +0000 (10:31 +0200)]
Lift atom type lookup out of inner loops

Grompp loops over molecule types, looking up force parameters for each
interaction from the associated bond types for the system (e.g. from
the force field). The atom types for that interaction have to be
looked up from the atoms for the molecule type that contains it, but
this should be done only once, before considering each bond type as a
possible match. The lookups for both A- and B-state parameters are now
lifted out of the loops over bond types, simplifying the logic and
significantly improving performance.

Once that is done, one custom function could be replaced by
std::equal.

Improved some variable naming

Apply 1 suggestion(s) to 1 file(s)

2 years agoRemove leftover licenses for optional and string_view
Andrey Alekseenko [Wed, 4 Aug 2021 09:53:39 +0000 (12:53 +0300)]
Remove leftover licenses for optional and string_view

After 175d3bb6 (MR !98), we don't bundle std::optional and
std::string_view, so no need to mention licenses of their reference
implementations.

2 years agoAdd GPU 3D FFT tests
Mark Abraham [Tue, 3 Aug 2021 13:42:42 +0000 (13:42 +0000)]
Add GPU 3D FFT tests

2 years agoUpdate bundled GoogleTest to current HEAD
Mark Abraham [Tue, 3 Aug 2021 09:30:45 +0000 (09:30 +0000)]
Update bundled GoogleTest to current HEAD

2 years agoReorganize COPYING file for auto-detection of licenses
Erik Lindahl [Mon, 2 Aug 2021 14:27:57 +0000 (16:27 +0200)]
Reorganize COPYING file for auto-detection of licenses

It seems our previous complex COPYING file that did not start with
the LGPL license causes GitLab to detect our license as "other", rather
than an OSI-approved one. By clearly putting the LGPL license first
in the file and then add the more permissible exceptions we allow
at the end we avoid this.

2 years agoRemove StatePropagatorDataGpu::getVelocitiesReadyOnDeviceEvent
Andrey Alekseenko [Tue, 3 Aug 2021 07:01:34 +0000 (07:01 +0000)]
Remove StatePropagatorDataGpu::getVelocitiesReadyOnDeviceEvent

It was unused, since we were using the copied data only in the stream in
which we did the copy. So, the synchronization was performed implicitly,
with no need to wait for the event.

Also added a comment about a rare redundant copy.

Refs #3988

2 years agoUpdate to support CUDA 11.4
Mark Abraham [Wed, 28 Jul 2021 06:33:06 +0000 (08:33 +0200)]
Update to support CUDA 11.4

Update to test the newest supported CUDA now that it can support GROMACS

Refs #3912

2 years agoAdd LICENSE file to make GitLab detect LGPL v2.1
Erik Lindahl [Mon, 2 Aug 2021 12:22:25 +0000 (14:22 +0200)]
Add LICENSE file to make GitLab detect LGPL v2.1

2 years agoFix style violations
Mark Abraham [Mon, 2 Aug 2021 08:59:07 +0000 (08:59 +0000)]
Fix style violations

2 years agoRemove MPI comm from GPU PME-PP force transfer initiation
Alan Gray [Sun, 1 Aug 2021 08:24:19 +0000 (08:24 +0000)]
Remove MPI comm from GPU PME-PP force transfer initiation

2 years agoOnly define GMX_SYCL_DPCPP macro in SYLC builds
Szilárd Páll [Fri, 30 Jul 2021 15:38:56 +0000 (17:38 +0200)]
Only define GMX_SYCL_DPCPP macro in SYLC builds

Previously the macro could get defined not only when the build was not a
hipSYCL build, but also when the build was not a SYCL build at all.

2 years agoSimplify short-circuit logic in grompp
Mark Abraham [Mon, 26 Jul 2021 08:47:19 +0000 (10:47 +0200)]
Simplify short-circuit logic in grompp

This is simpler to understand, and also prepares for a future
change that removes bB while improving performance.

2 years agoCheck nvcc accepts flags before using them
Mark Abraham [Thu, 29 Jul 2021 05:57:14 +0000 (05:57 +0000)]
Check nvcc accepts flags before using them

2 years agoOptimize grompp VDW parameter-generation allocations
Mark Abraham [Mon, 26 Jul 2021 02:50:13 +0000 (04:50 +0200)]
Optimize grompp VDW parameter-generation allocations

2 years agoRemove duplicate waiting for coordinates
Andrey Alekseenko [Mon, 26 Jul 2021 11:14:30 +0000 (11:14 +0000)]
Remove duplicate waiting for coordinates

2 years agoFix Doxygen issues introduced in 997d9e02 (MR !1783)
Andrey Alekseenko [Fri, 9 Jul 2021 21:30:19 +0000 (00:30 +0300)]
Fix Doxygen issues introduced in 997d9e02 (MR !1783)

2 years agoFix modular simulator MTTK
Pascal Merz [Mon, 28 Jun 2021 18:00:58 +0000 (14:00 -0400)]
Fix modular simulator MTTK

The modular simulator MTTK pressure coupling was slightly off due to
different reasons:
* The PV term was taken at the time of the update of the extended
  degree of freedom rather than at the time of reporting. This leads to
  slight differences when using nsttcoupling and nstpcoupling > 1.
* The NH chain coupled to the thermostat would alter the extended
  variable and hence the integral before it is reported. This NHC update
  already belongs to the next step, so shouldn't change the reported
  integral.
* The NH chain couple to the thermostat would not update the propagator
  scaling. Again, this leads to very slight differences when using
  nsttcoupling > 1.

These slight inaccuracies went undetected because they are minimal in
normal usage. To emphasize them in testing, the compressibility in the
simulator equivalence tests was strongly increased. This shows the
differences and would make the tests fail without the fixes in the
current change.

To implement these changes, the responsibility to calculate the
integral and set the scaling factor was moved from the elements to the
MTTK data structure which is accessed by all elements changing the
external degree of freedom and the box.

2 years agoAvoid passing a temporary object to emplace_back
Andrey Alekseenko [Fri, 2 Jul 2021 06:45:08 +0000 (06:45 +0000)]
Avoid passing a temporary object to emplace_back

2 years agoUse MDP database for simulator equivalence tests
Pascal Merz [Tue, 29 Jun 2021 22:00:20 +0000 (18:00 -0400)]
Use MDP database for simulator equivalence tests

Changes the simulator equivalence tests to use the additional MDP
database introduced recently. This will allow to test algorithms
requiring additional MDP parameters more cleanly.

This also changes the MDP database behavior to overwrite default
parameters. This does not change any current behavior, as in the
current usage, no parameters are defined in both the default and
the additional parameter sets. Moving forward, it's more likely
that we will want to explicitly overwrite default parameters with
the additional database rather than have parameter redefinitions
silently fail.

2 years agoTest with newer compilers in CI
Mark Abraham [Wed, 30 Jun 2021 11:14:21 +0000 (11:14 +0000)]
Test with newer compilers in CI

Now compiles with clang 12 and gcc 11 without warnings in various
configurations. Did not update santizer configurations as they're
already on clang 11.

Adjusted various CI job intents to be either accurate or placed on
more suitable configurations.

Eliminated one build container, since we target only gcc versions that
are most recent supported, oldest supported, and if needed also those
required for e.g. support of most recent and oldest supported CUDA
versions.

Code changes mostly related to false positives.

Suppresses a false-positive warning in GoogleTest

Avoids warnings about loop variables referring to temporaries. Our
usage was safe, but the warning is potentially valuable elsewhere, so
we should respect it.

Rewrote a test with idiomatic value-parametrization to avoid a warning
about a reference to a temporary

Refs #3912

2 years agoPort basic pull test to gtest framework
Berk Hess [Wed, 30 Jun 2021 10:50:18 +0000 (10:50 +0000)]
Port basic pull test to gtest framework

2 years agoFix GCC warning
Andrey Alekseenko [Wed, 30 Jun 2021 08:33:20 +0000 (11:33 +0300)]
Fix GCC warning

2 years agoMove GPU 3D FFT code to fft module
Mark Abraham [Tue, 29 Jun 2021 06:36:52 +0000 (08:36 +0200)]
Move GPU 3D FFT code to fft module

Then we can have tests of all the 3D FFT implementations together

Refs #3965

2 years agoTest gmxapi and clients through Py 3.9.
M. Eric Irrgang [Tue, 29 Jun 2021 14:36:44 +0000 (14:36 +0000)]
Test gmxapi and clients through Py 3.9.

* Begin moving to the new CI matrix config conventions.
* Clarify test intentions.
* Add testing for sample_restraint in master branch.
* Use the clang-8 build artifacts instead of gcc-10
  for faster gmxapi build.
* Adjust CMake options to get installed headers
  needed by consumers of the build artifact.

2 years agoFix python_packaging Dockerfiles.
M. Eric Irrgang [Fri, 25 Jun 2021 09:50:35 +0000 (12:50 +0300)]
Fix python_packaging Dockerfiles.

* Don't try to copy a directory that was removed in a recent commit.
* Add CMake option to install needed headers.
* Reduce image size.

2 years agoMerge branch release-2021 into master
Paul Bauer [Tue, 29 Jun 2021 07:54:40 +0000 (09:54 +0200)]
Merge branch release-2021 into master

Resolved Conflicts:
api/gmxapi/CMakeLists.txt
cmake/gmxGenerateVersionInfoWithoutGit.cmake
cmake/gmxManageClangCudaConfig.cmake
python_packaging/sample_restraint/src/cpp/CMakeLists.txt
python_packaging/src/CMakeLists.txt
python_packaging/src/gmxapi/export_system.cpp
python_packaging/src/gmxapi/version.py
python_packaging/src/setup.py
src/gromacs/applied_forces/awh/awh.cpp
src/gromacs/ewald/pme_gpu_program_impl.cu
src/gromacs/gmxana/gmx_dipoles.cpp
src/gromacs/tools/convert_tpr.cpp

2 years agoAllow gmxapi.commandline.cli to check GMXBIN.
M. Eric Irrgang [Fri, 27 Nov 2020 17:39:38 +0000 (20:39 +0300)]
Allow gmxapi.commandline.cli to check GMXBIN.

If `shutil.which` initially fails, try again using
the equivalent of `PATH=$GMXBIN`.

Refs #2961

2 years agoRevert "Enable back floating-point exceptions for SYCL build"
Andrey Alekseenko [Mon, 28 Jun 2021 07:28:53 +0000 (10:28 +0300)]
Revert "Enable back floating-point exceptions for SYCL build"

This reverts commit c502562354b1de48aa4e8e102a04d7586ce345e3.

Turns out, the FPEs are still occasionally triggered, at least for DG1.

2 years agoFix leak in copy_t_atoms
Dmitry Morozov [Tue, 29 Jun 2021 06:43:08 +0000 (06:43 +0000)]
Fix leak in copy_t_atoms

2 years agoAdd nblib virials function with test
Joe Jordan [Mon, 28 Jun 2021 09:18:17 +0000 (09:18 +0000)]
Add nblib virials function with test

2 years agoRestore gmxapi._gmxapi.add_mdmodule() Python functionality.
M. Eric Irrgang [Mon, 28 Jun 2021 07:35:36 +0000 (07:35 +0000)]
Restore gmxapi._gmxapi.add_mdmodule() Python functionality.

2 years agoDecouple PME GPU 3DFFT from PME GPU module
Mark Abraham [Sat, 26 Jun 2021 16:48:46 +0000 (16:48 +0000)]
Decouple PME GPU 3DFFT from PME GPU module

2 years agoFix wallcycle counting
Paul Bauer [Fri, 25 Jun 2021 16:09:56 +0000 (16:09 +0000)]
Fix wallcycle counting

2 years agoapply reviewer suggestions
M. Eric Irrgang [Thu, 24 Jun 2021 14:50:08 +0000 (14:50 +0000)]
apply reviewer suggestions

2 years agoDon't run full CI when stages fail.
M. Eric Irrgang [Thu, 24 Jun 2021 13:43:51 +0000 (16:43 +0300)]
Don't run full CI when stages fail.

See https://docs.gitlab.com/ee/ci/yaml/#when

Fixes #3482

2 years agoUpdate hipSYCL CI container
Mark Abraham [Thu, 24 Jun 2021 14:49:45 +0000 (14:49 +0000)]
Update hipSYCL CI container

2 years agoLet gmxapi.commandline access GROMACS installation details.
M. Eric Irrgang [Thu, 24 Jun 2021 14:15:29 +0000 (14:15 +0000)]
Let gmxapi.commandline access GROMACS installation details.

2 years agoRequire Python 3.7 for gmxapi in GROMACS 2022
M. Eric Irrgang [Thu, 24 Jun 2021 11:32:18 +0000 (11:32 +0000)]
Require Python 3.7 for gmxapi in GROMACS 2022

2 years agoIntroduce AngleInteractionType
ejjordan [Mon, 21 Jun 2021 14:34:29 +0000 (16:34 +0200)]
Introduce AngleInteractionType

This adds a template from which subsequent angle types can be generated.

2 years agoSYCL NBNXM: Use hand-crafted energy reduction
Andrey Alekseenko [Thu, 24 Jun 2021 08:43:43 +0000 (08:43 +0000)]
SYCL NBNXM: Use hand-crafted energy reduction

2 years agoEnable back floating-point exceptions for SYCL build
Andrey Alekseenko [Wed, 23 Jun 2021 19:46:05 +0000 (22:46 +0300)]
Enable back floating-point exceptions for SYCL build

Originally, they were disabled in
48fb95de1e456569deed5096924e617d511cf109 because Intel IGC compiler was
triggering them while doing JIT compilation of NBNXM kernels from the
intermediate bytecode.

Now:

- The problem was fixed in IGC 1.0.7683 (compute-runtime 21.24.20098).

- Either NBNXM kernels or the DPC++ compiler changed, and the generated
  SPIR-V code does not trigger FPEs even for earlier IGC versions.

2 years agoActivate the existing 20.04 CI containers
Mark Abraham [Tue, 22 Jun 2021 06:02:59 +0000 (06:02 +0000)]
Activate the existing 20.04 CI containers

2 years agoReworked convert-tpr runtime extension
Eliane Briand [Fri, 28 May 2021 07:11:10 +0000 (07:11 +0000)]
Reworked convert-tpr runtime extension

Fixes extension of run using -until flag.

Fixes #4056

2 years agoContinue fixing warnings with clang as host compiler
Mark Abraham [Mon, 21 Jun 2021 07:57:13 +0000 (07:57 +0000)]
Continue fixing warnings with clang as host compiler

2 years agoMinor docs fixes
Andrey Alekseenko [Fri, 18 Jun 2021 21:12:59 +0000 (00:12 +0300)]
Minor docs fixes

- Imbalanced quoting in a math formula causing incorrect rendering.
- Unnecessary escaping of an underscore in a code block.

2 years agoSpeed up grompp LJ parameter generation
Berk Hess [Fri, 18 Jun 2021 13:12:44 +0000 (13:12 +0000)]
Speed up grompp LJ parameter generation

2 years agoSuppressed -Wold-style-cast for clang as nvcc host compiler
Mark Abraham [Fri, 18 Jun 2021 04:56:39 +0000 (06:56 +0200)]
Suppressed -Wold-style-cast for clang as nvcc host compiler

Also updated some cache variable names to fit the style used for other
warning suppressions, in case we get to re-use the configure-time
result in future.

2 years agoUpdate device detection code for hipSYCL
Andrey Alekseenko [Fri, 18 Jun 2021 09:16:12 +0000 (09:16 +0000)]
Update device detection code for hipSYCL

2 years agoAllow to build and install nblib and gmxapi without GMX TESTS enabled
Alexey Shvetsov [Fri, 18 Jun 2021 09:06:58 +0000 (09:06 +0000)]
Allow to build and install nblib and gmxapi  without GMX TESTS enabled

2 years agoIntroduce ReferenceTemperatureManager
Pascal Merz [Fri, 18 Jun 2021 08:33:18 +0000 (08:33 +0000)]
Introduce ReferenceTemperatureManager

2 years agoFixed docs build
Mark Abraham [Thu, 17 Jun 2021 20:51:23 +0000 (20:51 +0000)]
Fixed docs build

2 years agoFix warnings when clang is the CUDA host compiler
Mark Abraham [Thu, 17 Jun 2021 20:19:13 +0000 (20:19 +0000)]
Fix warnings when clang is the CUDA host compiler

2 years agoFix clang-format in src/gromacs/fft/fft_mkl.cpp
Andrey Alekseenko [Thu, 17 Jun 2021 12:19:23 +0000 (15:19 +0300)]
Fix clang-format in src/gromacs/fft/fft_mkl.cpp

2 years agoSYCL: Reduce the number of atomic ops in NBNXM fShift calculation
Andrey Alekseenko [Thu, 17 Jun 2021 13:34:35 +0000 (13:34 +0000)]
SYCL: Reduce the number of atomic ops in NBNXM fShift calculation

2 years agoImprove hipSYCL GpuEventSynchronizer implementation
Andrey Alekseenko [Thu, 3 Jun 2021 13:08:47 +0000 (16:08 +0300)]
Improve hipSYCL GpuEventSynchronizer implementation

Previously, an inefficient stub was added. Here, we do proper event
tracking using the HIPSYCL_EXT_QUEUE_WAIT_LIST extension.

Since now we can have to wait on multiple events, replaced the
std::optional with std::vector.

Closes #4064

2 years agoRework -Weverything
Mark Abraham [Thu, 17 Jun 2021 07:34:08 +0000 (07:34 +0000)]
Rework -Weverything

2 years agoMake doxygen assertion more helpful
Mark Abraham [Fri, 11 Jun 2021 12:25:41 +0000 (14:25 +0200)]
Make doxygen assertion more helpful

Knowing the name of the clashing symbol helps understand where the
problem lies.

2 years agoUse doxygen properly
Mark Abraham [Wed, 16 Jun 2021 09:00:19 +0000 (11:00 +0200)]
Use doxygen properly

2 years agoRemove workaround for old clang
Paul Bauer [Fri, 11 Jun 2021 09:37:38 +0000 (11:37 +0200)]
Remove workaround for old clang

No longer needed as we require a more recent clang than version 6.

2 years agoUse 20.04 for docs build
Mark Abraham [Tue, 15 Jun 2021 14:46:41 +0000 (16:46 +0200)]
Use 20.04 for docs build

Updated image with new tex font package needed for 20.04

2 years agoImprove robustness of StopSignalClient test.
M. Eric Irrgang [Fri, 11 Jun 2021 12:30:01 +0000 (15:30 +0300)]
Improve robustness of StopSignalClient test.

A previous commit inadvertently rendered this test ineffective.

This change clarifies the test and restores its utility.
Rearranged commentary and additional assertions should help to prevent
another regression in the future.

Additionally, the two simulations run in the test have been
further decoupled.

Fixes #4076

2 years agoPrepare free energy element for external FEP state setting
Pascal Merz [Wed, 16 Jun 2021 10:14:11 +0000 (10:14 +0000)]
Prepare free energy element for external FEP state setting

2 years agoUse unique_ptr<ListedForcesGpu> in forcerec
Joe Jordan [Wed, 16 Jun 2021 09:42:22 +0000 (09:42 +0000)]
Use unique_ptr<ListedForcesGpu> in forcerec

2 years agoFix reporting of quadrupole moment
Mark Abraham [Wed, 16 Jun 2021 06:47:26 +0000 (06:47 +0000)]
Fix reporting of quadrupole moment

2 years agoUpdate interface of gmx_stats_t
Mark Abraham [Tue, 15 Jun 2021 13:10:48 +0000 (15:10 +0200)]
Update interface of gmx_stats_t

The only error code returned was one that checked that there was data
points to work with. Given our style, that makes sense to implement
with an internal assertion at the point where it is necessary, and to
throw when the user input is observed to be inconsistent (ie.
requesting statistical work without providing a valid dataset).

At least one routines always returned the error code for success,
creating complexity which is now eliminated.

Used a template to permit one routine to be implemented in two
slightly different ways.

This change helps with warning-free compilation with gcc 12 by
eliminating the assertion that was only used in debug mode. It is now
always a throw, since it makes no sense to continue to report garbage
statistics in the case being checked. Now the enum class is also no
longer useful and is eliminated.

Uncovered bug #4080 which is fixed here, and backported to
release-2021 separately.

2 years agoSupport testing CUDA-aware MPI in CI
Mark Abraham [Tue, 15 Jun 2021 12:30:38 +0000 (12:30 +0000)]
Support testing CUDA-aware MPI in CI