2 * This file is part of the GROMACS molecular simulation package.
4 * Copyright (c) 2017,2018,2019,2020, 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.
43 #include <gmock/gmock.h>
44 #include <gtest/gtest.h>
46 #include "gromacs/applied_forces/awh/bias.h"
47 #include "gromacs/applied_forces/awh/correlationgrid.h"
48 #include "gromacs/applied_forces/awh/pointstate.h"
49 #include "gromacs/mdtypes/awh_params.h"
50 #include "gromacs/utility/stringutil.h"
52 #include "testutils/refdata.h"
53 #include "testutils/testasserts.h"
61 //! The number of lambda states to use in the tests.
62 const int numLambdaStates = 16;
65 * Struct that gathers all input for setting up and using a Bias
67 struct AwhFepLambdaStateTestParameters
69 AwhFepLambdaStateTestParameters() = default;
71 AwhFepLambdaStateTestParameters(AwhFepLambdaStateTestParameters&& o) noexcept :
73 awhDimParams(o.awhDimParams),
74 awhBiasParams(o.awhBiasParams),
75 awhParams(o.awhParams),
76 dimParams(std::move(o.dimParams))
78 awhBiasParams.dimParams = &awhDimParams;
79 awhParams.awhBiasParams = &awhBiasParams;
81 double beta; //!< 1/(kB*T)
83 AwhDimParams awhDimParams; //!< Dimension parameters pointed to by \p awhBiasParams
84 AwhBiasParams awhBiasParams; //!< Bias parameters pointed to by \[ awhParams
85 AwhParams awhParams; //!< AWH parameters, this is the struct to actually use
87 std::vector<DimParams> dimParams; //!< Dimension parameters for setting up Bias
90 //! Helper function to set up the C-style AWH parameters for the test
91 static AwhFepLambdaStateTestParameters getAwhFepLambdaTestParameters(int eawhgrowth, int eawhpotential)
93 AwhFepLambdaStateTestParameters params;
97 AwhDimParams& awhDimParams = params.awhDimParams;
99 awhDimParams.period = 0;
100 // Correction for removal of GaussianGeometryFactor/2 in histogram size
101 awhDimParams.diffusion = 1e-4 / (0.12927243028700 * 2);
102 awhDimParams.origin = 0;
103 awhDimParams.end = numLambdaStates - 1;
104 awhDimParams.coordValueInit = awhDimParams.origin;
105 awhDimParams.coverDiameter = 0;
106 awhDimParams.eCoordProvider = eawhcoordproviderFREE_ENERGY_LAMBDA;
108 AwhBiasParams& awhBiasParams = params.awhBiasParams;
110 awhBiasParams.ndim = 1;
111 awhBiasParams.dimParams = &awhDimParams;
112 awhBiasParams.eTarget = eawhtargetCONSTANT;
113 awhBiasParams.targetBetaScaling = 0;
114 awhBiasParams.targetCutoff = 0;
115 awhBiasParams.eGrowth = eawhgrowth;
116 awhBiasParams.bUserData = FALSE;
117 awhBiasParams.errorInitial = 1.0 / params.beta;
118 awhBiasParams.shareGroup = 0;
119 awhBiasParams.equilibrateHistogram = FALSE;
121 int64_t seed = 93471803;
123 params.dimParams.push_back(DimParams::fepLambdaDimParams(numLambdaStates, params.beta));
125 AwhParams& awhParams = params.awhParams;
127 awhParams.numBias = 1;
128 awhParams.awhBiasParams = &awhBiasParams;
129 awhParams.seed = seed;
130 awhParams.nstOut = 0;
131 awhParams.nstSampleCoord = 1;
132 awhParams.numSamplesUpdateFreeEnergy = 10;
133 awhParams.ePotential = eawhpotential;
134 awhParams.shareBiasMultisim = FALSE;
139 //! Convenience typedef: growth type enum, potential type enum, disable update skips
140 typedef std::tuple<int, int, BiasParams::DisableUpdateSkips> BiasTestParameters;
142 /*! \brief Test fixture for testing Bias updates
144 class BiasFepLambdaStateTest : public ::testing::TestWithParam<BiasTestParameters>
147 //! Random seed for AWH MC sampling
151 std::unique_ptr<Bias> bias_;
153 BiasFepLambdaStateTest()
155 /* We test all combinations of:
157 * eawhgrowthLINEAR: final, normal update phase
158 * ewahgrowthEXP_LINEAR: intial phase, updated size is constant
159 * eawhpotential (test both, but for the FEP lambda state dimension MC will in practice be used,
160 * except that eawhpotentialCONVOLVED also gives a potential output):
161 * eawhpotentialUMBRELLA: MC on lambda state
162 * eawhpotentialCONVOLVED: MD on a convolved potential landscape (falling back to MC on lambda state)
163 * disableUpdateSkips (should not affect the results):
164 * BiasParams::DisableUpdateSkips::yes: update the point state for every sample
165 * BiasParams::DisableUpdateSkips::no: update the point state at an interval > 1 sample
167 * Note: It would be nice to explicitly check that eawhpotential
168 * and disableUpdateSkips do not affect the point state.
169 * But the reference data will also ensure this.
173 BiasParams::DisableUpdateSkips disableUpdateSkips;
174 std::tie(eawhgrowth, eawhpotential, disableUpdateSkips) = GetParam();
176 /* Set up a basic AWH setup with a single, 1D bias with parameters
177 * such that we can measure the effects of different parameters.
179 const AwhFepLambdaStateTestParameters params =
180 getAwhFepLambdaTestParameters(eawhgrowth, eawhpotential);
182 seed_ = params.awhParams.seed;
184 double mdTimeStep = 0.1;
186 bias_ = std::make_unique<Bias>(-1, params.awhParams, params.awhBiasParams, params.dimParams,
187 params.beta, mdTimeStep, 1, "", Bias::ThisRankWillDoIO::No,
192 TEST_P(BiasFepLambdaStateTest, ForcesBiasPmf)
194 gmx::test::TestReferenceData data;
195 gmx::test::TestReferenceChecker checker(data.rootChecker());
199 /* Make strings with the properties we expect to be different in the tests.
200 * These also helps to interpret the reference data.
202 std::vector<std::string> props;
203 props.push_back(formatString("stage: %s", bias.state().inInitialStage() ? "initial" : "final"));
204 props.push_back(formatString("convolve forces: %s", bias.params().convolveForce ? "yes" : "no"));
205 props.push_back(formatString("skip updates: %s", bias.params().skipUpdates() ? "yes" : "no"));
207 SCOPED_TRACE(gmx::formatString("%s, %s, %s", props[0].c_str(), props[1].c_str(), props[2].c_str()));
209 std::vector<double> force, pot;
211 double potentialJump = 0;
212 double mdTimeStep = 0.1;
215 /* Some energies to use as base values (to which some noise is added later on). */
216 std::vector<double> neighborLambdaEnergies(numLambdaStates);
217 std::vector<double> neighborLambdaDhdl(numLambdaStates);
218 const double magnitude = 12.0;
219 for (int i = 0; i < numLambdaStates; i++)
221 neighborLambdaEnergies[i] = magnitude * std::sin(i * 0.1);
222 neighborLambdaDhdl[i] = magnitude * std::cos(i * 0.1);
225 for (int step = 0; step < nSteps; step++)
227 int umbrellaGridpointIndex = bias.state().coordState().umbrellaGridpoint();
228 awh_dvec coordValue = { bias.getGridCoordValue(umbrellaGridpointIndex)[0], 0, 0, 0 };
229 double potential = 0;
230 gmx::ArrayRef<const double> biasForce = bias.calcForceAndUpdateBias(
231 coordValue, neighborLambdaEnergies, neighborLambdaDhdl, &potential, &potentialJump,
232 nullptr, nullptr, step * mdTimeStep, step, seed_, nullptr);
234 force.push_back(biasForce[0]);
235 pot.push_back(potential);
238 /* When skipping updates, ensure all skipped updates are performed here.
239 * This should result in the same bias state as at output in a normal run.
241 if (bias.params().skipUpdates())
243 bias.doSkippedUpdatesForAllPoints();
246 std::vector<double> pointBias, logPmfsum;
247 for (auto& point : bias.state().points())
249 pointBias.push_back(point.bias());
250 logPmfsum.push_back(point.logPmfSum());
253 constexpr int ulpTol = 10;
255 checker.checkSequence(props.begin(), props.end(), "Properties");
256 checker.setDefaultTolerance(absoluteTolerance(magnitude * GMX_DOUBLE_EPS * ulpTol));
257 checker.checkSequence(force.begin(), force.end(), "Force");
258 checker.checkSequence(pot.begin(), pot.end(), "Potential");
259 checker.setDefaultTolerance(relativeToleranceAsUlp(1.0, ulpTol));
260 checker.checkSequence(pointBias.begin(), pointBias.end(), "PointBias");
261 checker.checkSequence(logPmfsum.begin(), logPmfsum.end(), "PointLogPmfsum");
264 /* Scan initial/final phase, MC/convolved force and update skip (not) allowed
265 * Both the convolving and skipping should not affect the bias and PMF.
266 * It would be nice if the test would explicitly check for this.
267 * Currently this is tested through identical reference data.
269 INSTANTIATE_TEST_CASE_P(WithParameters,
270 BiasFepLambdaStateTest,
271 ::testing::Combine(::testing::Values(eawhgrowthLINEAR, eawhgrowthEXP_LINEAR),
272 ::testing::Values(eawhpotentialUMBRELLA, eawhpotentialCONVOLVED),
273 ::testing::Values(BiasParams::DisableUpdateSkips::yes,
274 BiasParams::DisableUpdateSkips::no)));
276 // Test that we detect coverings and exit the initial stage at the correct step
277 TEST(BiasFepLambdaStateTest, DetectsCovering)
279 const AwhFepLambdaStateTestParameters params =
280 getAwhFepLambdaTestParameters(eawhgrowthEXP_LINEAR, eawhpotentialCONVOLVED);
282 const double mdTimeStep = 0.1;
284 Bias bias(-1, params.awhParams, params.awhBiasParams, params.dimParams, params.beta, mdTimeStep,
285 1, "", Bias::ThisRankWillDoIO::No);
287 const int64_t exitStepRef = 320;
289 bool inInitialStage = bias.state().inInitialStage();
291 /* Some energies to use as base values (to which some noise is added later on). */
292 std::vector<double> neighborLambdaEnergies(numLambdaStates);
293 std::vector<double> neighborLambdaDhdl(numLambdaStates);
294 const double magnitude = 12.0;
295 for (int i = 0; i < numLambdaStates; i++)
297 neighborLambdaEnergies[i] = magnitude * std::sin(i * 0.1);
298 neighborLambdaDhdl[i] = magnitude * std::cos(i * 0.1);
302 /* Normally this loop exits at exitStepRef, but we extend with failure */
303 for (step = 0; step <= 2 * exitStepRef; step++)
305 int umbrellaGridpointIndex = bias.state().coordState().umbrellaGridpoint();
306 awh_dvec coordValue = { bias.getGridCoordValue(umbrellaGridpointIndex)[0], 0, 0, 0 };
308 double potential = 0;
309 double potentialJump = 0;
310 bias.calcForceAndUpdateBias(coordValue, neighborLambdaEnergies, neighborLambdaDhdl,
311 &potential, &potentialJump, nullptr, nullptr, step, step,
312 params.awhParams.seed, nullptr);
314 inInitialStage = bias.state().inInitialStage();
321 EXPECT_EQ(false, inInitialStage);
324 EXPECT_EQ(exitStepRef, step);