2 * This file is part of the GROMACS molecular simulation package.
4 * Copyright (c) 2016,2017, 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.
38 * Implements internal utility functions for spline tables
40 * \author Erik Lindahl <erik.lindahl@gmail.com>
41 * \ingroup module_tables
45 #include "splineutil.h"
54 #include "gromacs/utility/alignedallocator.h"
55 #include "gromacs/utility/arrayref.h"
56 #include "gromacs/utility/exceptions.h"
57 #include "gromacs/utility/real.h"
58 #include "gromacs/utility/stringutil.h"
67 throwUnlessDerivativeIsConsistentWithFunction(const std::function<double(double)> &function,
68 const std::function<double(double)> &derivative,
69 const std::pair<real, real> &range)
71 // Since the numerical derivative will evaluate extra points
72 // we shrink the interval slightly to avoid calling the function with values
73 // outside the range specified.
74 double h = std::cbrt(GMX_DOUBLE_EPS); // ideal spacing
75 std::pair<double, double> newRange(range.first + h, range.second - h);
76 const int points = 1000; // arbitrary
77 double dx = (newRange.second - newRange.first) / points;
78 bool isConsistent = true;
79 double minFail = newRange.second;
80 double maxFail = newRange.first;
82 for (double x = newRange.first; x <= newRange.second; x += dx)
84 double analyticalDerivative = derivative(x);
85 double numericalDerivative = (function(x+h)-function(x-h))/(2*h);
86 double thirdDerivative = (derivative(x+h)-2.0*derivative(x)+derivative(x-h))/(h*h);
88 // We make two types of errors in numerical calculation of the derivative:
89 // - The truncation error: eps * |f| / h
90 // - The rounding error: h * h * |f'''| / 6.0
91 double expectedErr = GMX_DOUBLE_EPS*std::abs(function(x))/h + h*h*std::abs(thirdDerivative)/6.0;
93 // To avoid triggering false errors because of compiler optimization or numerical issues
94 // in the function evalulation we allow an extra factor of 10 in the expected error
95 if (std::abs(analyticalDerivative-numericalDerivative) > 10.0*expectedErr)
98 minFail = std::min(minFail, x);
99 maxFail = std::max(maxFail, x);
105 GMX_THROW(InconsistentInputError(formatString("Derivative inconsistent with analytical function in range [%d,%d]", minFail, maxFail)));
111 throwUnlessDerivativeIsConsistentWithFunction(ArrayRef<const double> function,
112 ArrayRef<const double> derivative,
114 const std::pair<real, real> &range)
116 std::size_t firstIndex = range.first / inputSpacing;
117 std::size_t lastIndex = range.second / inputSpacing;
118 bool isConsistent = true;
119 std::size_t minFail = lastIndex;
120 std::size_t maxFail = firstIndex;
122 // The derivative will access one extra point before/after each point, so reduce interval
123 for (std::size_t i = firstIndex + 1; (i + 1) < lastIndex; i++)
125 double inputDerivative = derivative[i];
126 double numericalDerivative = (function[i+1] - function[i-1]) / (2.0 * inputSpacing);
127 double thirdDerivative = (derivative[i+1] - 2.0*derivative[i] + derivative[i-1])/(inputSpacing * inputSpacing);
129 // We make two types of errors in numerical calculation of the derivative:
130 // - The truncation error: eps * |f| / h
131 // - The rounding error: h * h * |f'''| / 6.0
132 double expectedErr = GMX_DOUBLE_EPS*std::abs(function[i])/inputSpacing + inputSpacing*inputSpacing*std::abs(thirdDerivative)/6.0;
134 // To avoid triggering false errors because of compiler optimization or numerical issues
135 // in the function evalulation we allow an extra factor of 10 in the expected error
136 if (std::abs(inputDerivative-numericalDerivative) > 10.0*expectedErr)
138 isConsistent = false;
139 minFail = std::min(minFail, i);
140 maxFail = std::max(maxFail, i);
145 GMX_THROW(InconsistentInputError(formatString("Derivative inconsistent with numerical vector for elements %d-%d", minFail+1, maxFail+1)));
150 /*! \brief Update minQuotient if the ratio of this function value and its second derivative is smaller
152 * This is a utility function used in the functions to find the smallest quotient
155 * \param[in] previousPoint Value of function at x-h.
156 * \param[in] thisPoint Value of function at x.
157 * \param[in] nextPoint Value of function at x+h.
158 * \param[in] spacing Value of h.
159 * \param[inout] minQuotient Current minimum of such quotients, updated if this quotient is smaller.
162 updateMinQuotientOfFunctionAndSecondDerivative(double previousPoint,
166 double * minQuotient)
168 double value = std::abs( thisPoint );
169 double secondDerivative = std::abs( (previousPoint - 2.0 * thisPoint + nextPoint) / (spacing * spacing ) );
171 // Make sure we do not divide by zero. This limit is arbitrary,
172 // but it doesnt matter since this point will have a very large value,
173 // and the whole routine is searching for the smallest value.
174 secondDerivative = std::max(secondDerivative, static_cast<double>(std::sqrt(GMX_REAL_MIN)));
176 *minQuotient = std::min(*minQuotient, value / secondDerivative);
181 findSmallestQuotientOfFunctionAndSecondDerivative(const std::function<double(double)> &f,
182 const std::pair<real, real> &range)
184 // Since the numerical second derivative will evaluate extra points
185 // we shrink the interval slightly to avoid calling the function with values
186 // outside the range specified.
187 double h = std::pow( GMX_DOUBLE_EPS, 0.25 );
188 std::pair<double, double> newRange(range.first + h, range.second - h);
189 const int points = 1000; // arbitrary
190 double dx = (newRange.second - newRange.first) / points;
191 double minQuotient = GMX_REAL_MAX;
193 for (double x = newRange.first; x <= newRange.second; x += dx)
195 updateMinQuotientOfFunctionAndSecondDerivative(f(x-h), f(x), f(x+h), h, &minQuotient);
197 return static_cast<real>(minQuotient);
206 findSmallestQuotientOfFunctionAndSecondDerivative(ArrayRef<const double> function,
208 const std::pair<real, real> &range)
211 std::size_t firstIndex = range.first / inputSpacing;
212 std::size_t lastIndex = range.second / inputSpacing;
213 double minQuotient = GMX_REAL_MAX;
215 for (std::size_t i = firstIndex + 1; (i + 1) < lastIndex; i++)
217 updateMinQuotientOfFunctionAndSecondDerivative(function[i-1], function[i], function[i+1], inputSpacing, &minQuotient);
219 return static_cast<real>(minQuotient);
224 /*! \brief Update minQuotient if the ratio of this function value and its third derivative is smaller
226 * This is a utility function used in the functions to find the smallest quotient
229 * \param[in] previousPreviousPoint Value of function at x-2h.
230 * \param[in] previousPoint Value of function at x-h.
231 * \param[in] thisPoint Value of function at x.
232 * \param[in] nextPoint Value of function at x+h.
233 * \param[in] nextNextPoint Value of function at x+2h.
234 * \param[in] spacing Value of h.
235 * \param[inout] minQuotient Current minimum of such quotients, updated if this quotient is smaller.
238 updateMinQuotientOfFunctionAndThirdDerivative(double previousPreviousPoint,
239 double previousPoint,
242 double nextNextPoint,
244 double * minQuotient)
246 double value = std::abs( thisPoint );
247 double thirdDerivative = std::abs((nextNextPoint - 2 * nextPoint + 2 * previousPoint - previousPreviousPoint) / (2 * spacing * spacing * spacing));
249 // Make sure we do not divide by zero. This limit is arbitrary,
250 // but it doesnt matter since this point will have a very large value,
251 // and the whole routine is searching for the smallest value.
252 thirdDerivative = std::max(thirdDerivative, static_cast<double>(std::sqrt(GMX_REAL_MIN)));
254 *minQuotient = std::min(*minQuotient, value / thirdDerivative);
259 findSmallestQuotientOfFunctionAndThirdDerivative(const std::function<double(double)> &f,
260 const std::pair<real, real> &range)
262 // Since the numerical second derivative will evaluate extra points
263 // we shrink the interval slightly to avoid calling the function with values
264 // outside the range specified.
265 double h = std::pow( GMX_DOUBLE_EPS, 0.2 ); // optimal spacing for 3rd derivative
266 std::pair<double, double> newRange(range.first + 2*h, range.second - 2*h);
267 const int points = 1000; // arbitrary
268 double dx = (newRange.second - newRange.first) / points;
269 double minQuotient = GMX_REAL_MAX;
271 for (double x = newRange.first; x <= newRange.second; x += dx)
273 updateMinQuotientOfFunctionAndThirdDerivative(f(x-2*h), f(x-h), f(x), f(x+h), f(x+2*h), h, &minQuotient);
275 return static_cast<real>(minQuotient);
280 findSmallestQuotientOfFunctionAndThirdDerivative(ArrayRef<const double> function,
282 const std::pair<real, real> &range)
285 std::size_t firstIndex = range.first / inputSpacing;
286 std::size_t lastIndex = range.second / inputSpacing;
287 double minQuotient = GMX_REAL_MAX;
289 for (std::size_t i = firstIndex + 2; (i + 2) < lastIndex; i++)
291 updateMinQuotientOfFunctionAndThirdDerivative(function[i-2], function[i-1], function[i], function[i+1], function[i+2], inputSpacing, &minQuotient);
293 return static_cast<real>(minQuotient);
299 vectorSecondDerivative(ArrayRef<const double> f, double spacing)
303 GMX_THROW(APIError("Too few points in vector for 5-point differentiation"));
306 std::vector<double> d(f.size());
309 // 5-point formula evaluated for points 0,1
311 d[i] = (11 * f[i+4] - 56 * f[i+3] + 114 * f[i+2] - 104 * f[i+1] + 35 * f[i]) / ( 12 * spacing * spacing);
313 d[i] = (-f[i+3] + 4 * f[i+2] + 6 * f[i+1] - 20 * f[i] + 11 * f[i-1]) / ( 12 * spacing * spacing);
315 for (std::size_t i = 2; i < d.size() - 2; i++)
317 // 5-point formula evaluated for central point (2)
318 d[i] = (-f[i+2] + 16 * f[i+1] - 30 * f[i] + 16 * f[i-1] - f[i-2]) / (12 * spacing * spacing);
321 // 5-point formula evaluated for points 3,4
323 d[i] = (11 * f[i+1] - 20 * f[i] + 6 * f[i-1] + 4 * f[i-2] - f[i-3]) / ( 12 * spacing * spacing);
325 d[i] = (35 * f[i] - 104 * f[i-1] + 114 * f[i-2] - 56 * f[i-3] + 11 * f[i-4]) / ( 12 * spacing * spacing);
332 } // namespace internal