2 * This file is part of the GROMACS molecular simulation package.
4 * Copyright (c) 2019, by the GROMACS development team, led by
5 * Mark Abraham, David van der Spoel, Berk Hess, and Erik Lindahl,
6 * and including many others, as listed in the AUTHORS file in the
7 * top-level source directory and at http://www.gromacs.org.
9 * GROMACS is free software; you can redistribute it and/or
10 * modify it under the terms of the GNU Lesser General Public License
11 * as published by the Free Software Foundation; either version 2.1
12 * of the License, or (at your option) any later version.
14 * GROMACS is distributed in the hope that it will be useful,
15 * but WITHOUT ANY WARRANTY; without even the implied warranty of
16 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
17 * Lesser General Public License for more details.
19 * You should have received a copy of the GNU Lesser General Public
20 * License along with GROMACS; if not, see
21 * http://www.gnu.org/licenses, or write to the Free Software Foundation,
22 * Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
24 * If you want to redistribute modifications to GROMACS, please
25 * consider that scientific software is very special. Version
26 * control is crucial - bugs must be traceable. We will be happy to
27 * consider code for inclusion in the official distribution, but
28 * derived work must not be called official GROMACS. Details are found
29 * in the README & COPYING files - if they are missing, get the
30 * official version at http://www.gromacs.org.
32 * To help us fund GROMACS development, we humbly ask that you cite
33 * the research papers on the package. Check out http://www.gromacs.org.
37 * \brief Implements update and constraints class using CUDA.
39 * The class combines Leap-Frog integrator with LINCS and SETTLE constraints.
41 * \todo The computational procedures in members should be integrated to improve
42 * computational performance.
44 * \author Artem Zhmurov <zhmurov@gmail.com>
46 * \ingroup module_mdlib
50 #include "update_constrain_cuda_impl.h"
59 #include "gromacs/gpu_utils/cudautils.cuh"
60 #include "gromacs/gpu_utils/devicebuffer.h"
61 #include "gromacs/gpu_utils/gputraits.cuh"
62 #include "gromacs/gpu_utils/vectype_ops.cuh"
63 #include "gromacs/mdlib/leapfrog_cuda.cuh"
64 #include "gromacs/mdlib/lincs_cuda.cuh"
65 #include "gromacs/mdlib/settle_cuda.cuh"
66 #include "gromacs/mdlib/update_constrain_cuda.h"
71 void UpdateConstrainCuda::Impl::integrate(GpuEventSynchronizer *fReadyOnDevice,
73 const bool updateVelocities,
74 const bool computeVirial,
76 const bool doTempCouple,
77 gmx::ArrayRef<const t_grp_tcstat> tcstat,
78 const bool doPressureCouple,
79 const float dtPressureCouple,
80 const matrix velocityScalingMatrix)
82 // Clearing virial matrix
83 // TODO There is no point in having separate virial matrix for constraints
86 // Make sure that the forces are ready on device before proceeding with the update.
87 fReadyOnDevice->enqueueWaitEvent(commandStream_);
89 // The integrate should save a copy of the current coordinates in d_xp_ and write updated once into d_x_.
90 // The d_xp_ is only needed by constraints.
91 integrator_->integrate(d_x_, d_xp_, d_v_, d_f_, dt,
93 doPressureCouple, dtPressureCouple, velocityScalingMatrix);
94 // Constraints need both coordinates before (d_x_) and after (d_xp_) update. However, after constraints
95 // are applied, the d_x_ can be discarded. So we intentionally swap the d_x_ and d_xp_ here to avoid the
96 // d_xp_ -> d_x_ copy after constraints. Note that the integrate saves them in the wrong order as well.
97 lincsCuda_->apply(d_xp_, d_x_,
98 updateVelocities, d_v_, 1.0/dt,
99 computeVirial, virial);
100 settleCuda_->apply(d_xp_, d_x_,
101 updateVelocities, d_v_, 1.0/dt,
102 computeVirial, virial);
104 // scaledVirial -> virial (methods above returns scaled values)
105 float scaleFactor = 0.5f/(dt*dt);
106 for (int i = 0; i < DIM; i++)
108 for (int j = 0; j < DIM; j++)
110 virial[i][j] = scaleFactor*virial[i][j];
114 coordinatesReady_->markEvent(commandStream_);
119 UpdateConstrainCuda::Impl::Impl(const t_inputrec &ir,
120 const gmx_mtop_t &mtop,
121 const void *commandStream,
122 GpuEventSynchronizer *xUpdatedOnDevice) :
123 coordinatesReady_(xUpdatedOnDevice)
125 GMX_ASSERT(xUpdatedOnDevice != nullptr, "The event synchronizer can not be nullptr.");
126 commandStream != nullptr ? commandStream_ = *static_cast<const CommandStream*>(commandStream) : commandStream_ = nullptr;
129 integrator_ = std::make_unique<LeapFrogCuda>(commandStream_);
130 lincsCuda_ = std::make_unique<LincsCuda>(ir.nLincsIter, ir.nProjOrder, commandStream_);
131 settleCuda_ = std::make_unique<SettleCuda>(mtop, commandStream_);
135 UpdateConstrainCuda::Impl::~Impl()
139 void UpdateConstrainCuda::Impl::set(DeviceBuffer<float> d_x,
140 DeviceBuffer<float> d_v,
141 const DeviceBuffer<float> d_f,
144 const int numTempScaleValues)
146 GMX_ASSERT(d_x != nullptr, "Coordinates device buffer should not be null.");
147 GMX_ASSERT(d_v != nullptr, "Velocities device buffer should not be null.");
148 GMX_ASSERT(d_f != nullptr, "Forces device buffer should not be null.");
150 d_x_ = reinterpret_cast<float3*>(d_x);
151 d_v_ = reinterpret_cast<float3*>(d_v);
152 d_f_ = reinterpret_cast<float3*>(d_f);
156 reallocateDeviceBuffer(&d_xp_, numAtoms_, &numXp_, &numXpAlloc_, nullptr);
158 reallocateDeviceBuffer(&d_inverseMasses_, numAtoms_,
159 &numInverseMasses_, &numInverseMassesAlloc_, nullptr);
161 // Integrator should also update something, but it does not even have a method yet
162 integrator_->set(md, numTempScaleValues, md.cTC);
163 lincsCuda_->set(idef, md);
164 settleCuda_->set(idef, md);
167 void UpdateConstrainCuda::Impl::setPbc(const t_pbc *pbc)
169 setPbcAiuc(pbc->ndim_ePBC, pbc->box, &pbcAiuc_);
170 integrator_->setPbc(pbc);
171 lincsCuda_->setPbc(pbc);
172 settleCuda_->setPbc(pbc);
175 void UpdateConstrainCuda::Impl::waitCoordinatesReadyOnDevice()
177 coordinatesReady_->waitForEvent();
180 GpuEventSynchronizer* UpdateConstrainCuda::Impl::getCoordinatesReadySync()
182 return coordinatesReady_;
185 UpdateConstrainCuda::UpdateConstrainCuda(const t_inputrec &ir,
186 const gmx_mtop_t &mtop,
187 const void *commandStream,
188 GpuEventSynchronizer *xUpdatedOnDevice)
189 : impl_(new Impl(ir, mtop, commandStream, xUpdatedOnDevice))
193 UpdateConstrainCuda::~UpdateConstrainCuda() = default;
195 void UpdateConstrainCuda::integrate(GpuEventSynchronizer *fReadyOnDevice,
197 const bool updateVelocities,
198 const bool computeVirial,
200 const bool doTempCouple,
201 gmx::ArrayRef<const t_grp_tcstat> tcstat,
202 const bool doPressureCouple,
203 const float dtPressureCouple,
204 const matrix velocityScalingMatrix)
206 impl_->integrate(fReadyOnDevice,
207 dt, updateVelocities, computeVirial, virialScaled,
208 doTempCouple, tcstat,
209 doPressureCouple, dtPressureCouple, velocityScalingMatrix);
212 void UpdateConstrainCuda::set(DeviceBuffer<float> d_x,
213 DeviceBuffer<float> d_v,
214 const DeviceBuffer<float> d_f,
217 const int numTempScaleValues)
219 impl_->set(d_x, d_v, d_f, idef, md, numTempScaleValues);
222 void UpdateConstrainCuda::setPbc(const t_pbc *pbc)
227 void UpdateConstrainCuda::waitCoordinatesReadyOnDevice()
229 impl_->waitCoordinatesReadyOnDevice();
232 GpuEventSynchronizer* UpdateConstrainCuda::getCoordinatesReadySync()
234 return impl_->getCoordinatesReadySync();