1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317
|
/* Ergo, version 3.8.2, a program for linear scaling electronic structure
* calculations.
* Copyright (C) 2023 Elias Rudberg, Emanuel H. Rubensson, Pawel Salek,
* and Anastasia Kruchinina.
*
* This program is free software: you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation, either version 3 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program. If not, see <http://www.gnu.org/licenses/>.
*
* Primary academic reference:
* Ergo: An open-source program for linear-scaling electronic structure
* calculations,
* Elias Rudberg, Emanuel H. Rubensson, Pawel Salek, and Anastasia
* Kruchinina,
* SoftwareX 7, 107 (2018),
* <http://dx.doi.org/10.1016/j.softx.2018.03.005>
*
* For further information about Ergo, see <http://www.ergoscf.org>.
*/
/** @file boysfunction.cc
@brief Code for Boys function evaluation.
@author: Elias Rudberg <em>responsible</em>
*/
#include <stdlib.h>
#include <cmath>
#include <cstring>
#include <cassert>
#include "boysfunction.h"
#include "pi.h"
#include "output.h"
#include "utilities.h"
#include "mat_gblas.h"
#include "config.h" // Needed to get the PRECISION_SINGLE macro
static ergo_real
semiFactorial(int n)
{
assert(n >= -1);
if(n <= 1)
return 1;
return n * semiFactorial(n - 2);
}
static ergo_real
BoysFunction_raw_simpson(int n, ergo_real x, int noOfIntegrationIntervals)
{
const int N = noOfIntegrationIntervals;
ergo_real h = (ergo_real)0.5 / N;
ergo_real sum = 0;
for(int k = 0; k <= 2*N; k++)
{
ergo_real tk = (ergo_real)k / (2*N);
// Compute f(tk) = exp(-x*tk*tk) * pow(tk, 2*n)
ergo_real foftk = template_blas_exp(-x*tk*tk);
if(n != 0)
{
if(k != 0)
foftk *= template_blas_pow(tk, (ergo_real)(2*n));
else
foftk = 0;
}
// OK, foftk done, now add to sum.
if(k == 0 || k == 2*N)
{
sum += foftk;
continue;
}
if(k % 2 == 1)
{
sum += 4 * foftk;
continue;
}
sum += 2 * foftk;
}
return (h/3) * sum;
}
static ergo_real BoysFunctionIntegrand(int n, ergo_real x, ergo_real t) {
return template_blas_exp(-x*t*t) * template_blas_pow(t, (ergo_real)(2*n));
}
/* Numerical integration using Boole's rule */
static ergo_real
BoysFunction_raw_booles_rule(int n, ergo_real x, int noOfIntegrationIntervals)
{
const int N = noOfIntegrationIntervals;
ergo_real intervalWidth = (ergo_real)1 / N;
ergo_real h = intervalWidth / 4;
ergo_real sum = 0;
for(int k = 0; k < N; k++) {
ergo_real x1 = (ergo_real)k / N;
ergo_real x2 = x1 + h;
ergo_real x3 = x1 + 2*h;
ergo_real x4 = x1 + 3*h;
ergo_real x5 = x1 + 4*h;
ergo_real f_of_x1 = BoysFunctionIntegrand(n, x, x1);
ergo_real f_of_x2 = BoysFunctionIntegrand(n, x, x2);
ergo_real f_of_x3 = BoysFunctionIntegrand(n, x, x3);
ergo_real f_of_x4 = BoysFunctionIntegrand(n, x, x4);
ergo_real f_of_x5 = BoysFunctionIntegrand(n, x, x5);
sum += 7 * f_of_x1 +
32 * f_of_x2 +
12 * f_of_x3 +
32 * f_of_x4 +
7 * f_of_x5;
}
return (2*h/45)*sum;
}
/* Numerical integration using 7-point Gauss-Lobatto rule */
static ergo_real
BoysFunction_raw_GaussLobatto(int n, ergo_real x, int noOfIntegrationIntervals, ergo_real endPt = 1)
{
if(endPt == 0)
return 0;
// If integrand at endPt is almost zero we get better accuracy by integrating over a shorter interval.
if(BoysFunctionIntegrand(n, x, endPt) < template_blas_get_num_limit_min<ergo_real>())
return BoysFunction_raw_GaussLobatto(n, x, noOfIntegrationIntervals, endPt*0.5);
const ergo_real c_5_11 = (ergo_real)5/11;
const ergo_real c_2_11 = (ergo_real)2/11;
const ergo_real c_5_3 = (ergo_real)5/3 ;
const ergo_real sqrt15 = template_blas_sqrt((ergo_real)15);
// points xi
ergo_real x1 = 0;
ergo_real x2 = template_blas_sqrt(c_5_11 - c_2_11 * template_blas_sqrt(c_5_3));
ergo_real x3 = template_blas_sqrt(c_5_11 + c_2_11 * template_blas_sqrt(c_5_3));
ergo_real x4 = 1;
ergo_real x5 = -x2;
ergo_real x6 = -x3;
ergo_real x7 = -x4;
// weights wi
ergo_real w1 = (ergo_real)256/525;
ergo_real w2 = ((ergo_real)124 + (ergo_real)7*sqrt15) / 350;
ergo_real w3 = ((ergo_real)124 - (ergo_real)7*sqrt15) / 350;
ergo_real w4 = (ergo_real)1 / 21;
ergo_real w5 = w2;
ergo_real w6 = w3;
ergo_real w7 = w4;
const int N = noOfIntegrationIntervals;
ergo_real intervalWidth = (ergo_real)endPt / N;
ergo_real sum = 0;
for(int k = 0; k < N; k++) {
ergo_real a = (ergo_real)k*endPt / N;
ergo_real b = a + intervalWidth;
ergo_real b_minus_a_over_2 = (b-a)/2;
ergo_real a_plus_b_over_2 = (a+b)/2;
ergo_real f1 = BoysFunctionIntegrand(n, x, b_minus_a_over_2*x1+a_plus_b_over_2);
ergo_real f2 = BoysFunctionIntegrand(n, x, b_minus_a_over_2*x2+a_plus_b_over_2);
ergo_real f3 = BoysFunctionIntegrand(n, x, b_minus_a_over_2*x3+a_plus_b_over_2);
ergo_real f4 = BoysFunctionIntegrand(n, x, b_minus_a_over_2*x4+a_plus_b_over_2);
ergo_real f5 = BoysFunctionIntegrand(n, x, b_minus_a_over_2*x5+a_plus_b_over_2);
ergo_real f6 = BoysFunctionIntegrand(n, x, b_minus_a_over_2*x6+a_plus_b_over_2);
ergo_real f7 = BoysFunctionIntegrand(n, x, b_minus_a_over_2*x7+a_plus_b_over_2);
sum += b_minus_a_over_2 * (w1 * f1 + w2 * f2 + w3 * f3 + w4 * f4 + w5 * f5 + w6 * f6 + w7 * f7);
}
return sum;
}
void
BoysFunctionManager::init(void) {
if(Boys_init_flag == 1)
return;
Util::TimeMeter timeMeter;
ergo_real halfstep, kfactorial, BoysFuncRawResult, Ak, midx;
// Prepare Boys_list
halfstep = (ergo_real)BOYS_X_MAX / BOYS_NO_OF_INTERVALS * 0.5;
for(int j = 0; j < BOYS_NO_OF_INTERVALS; j++) {
midx = (ergo_real)BOYS_X_MAX * j / BOYS_NO_OF_INTERVALS + halfstep;
const int highest_N_needed = BOYS_N_MAX+BOYS_TAB_DEGREE-2;
ergo_real Boys_list_curr_midx[highest_N_needed+1];
// Use downward recursion
Boys_list_curr_midx[highest_N_needed] = BoysFunction_expensive(highest_N_needed, midx, 160); // FIXME DO NOT USE HARD-CODED VALUE HERE?
for(int n = highest_N_needed-1; n >= 0; n--)
Boys_list_curr_midx[n] = (2*midx*Boys_list_curr_midx[n+1] + template_blas_exp(-midx)) / (2*n+1);
// Now we have the Boys_list_curr_midx list so we can use that below
for(int n = 0; n < BOYS_N_MAX; n++) {
Boys_list[n].list[j].midx = midx;
kfactorial = 1;
int minusOneToPowk = 1;
for(int k = 0; k < BOYS_TAB_DEGREE; k++) {
BoysFuncRawResult = Boys_list_curr_midx[n+k];
Ak = minusOneToPowk * BoysFuncRawResult / kfactorial;
Boys_list[n].list[j].A[k] = Ak;
kfactorial *= k+1;
minusOneToPowk *= -1;
} /* END FOR k */
} /* END FOR j */
} /* END FOR n */
// Also prepare SavedPrefactor_list
for(int n = 0; n < BOYS_N_MAX; n++)
SavedPrefactor_list[n] = (semiFactorial(2*n-1) * template_blas_sqrt(pi) / template_blas_pow((ergo_real)2, (ergo_real)(n+1)));
// Set Boys_init_flag to indicate that initialization is done
Boys_init_flag = 1;
timeMeter.print(LOG_AREA_INTEGRALS, "BoysFunctionManager::init");
}
ergo_real
BoysFunctionManager::BoysFunction_pretabulated(int n, ergo_real x) const
{
const BoysFuncIntervalStruct* interval;
ergo_real intervalWidth, count, sum, deltax, deltaxtopowk;
int intervalIndex, k;
if(Boys_init_flag != 1)
throw std::runtime_error("Error in BoysFunctionManager::BoysFunction_pretabulated: (Boys_init_flag != 1).");
if(x < 0)
throw std::runtime_error("Error in BoysFunctionManager::BoysFunction_pretabulated: (x < 0).");
if(n >= BOYS_N_MAX)
throw std::runtime_error("Error in BoysFunctionManager::BoysFunction_pretabulated: (n >= BOYS_N_MAX).");
if(n < 0)
throw std::runtime_error("Error in BoysFunctionManager::BoysFunction_pretabulated: (n < 0).");
if(x >= BOYS_X_MAX) {
/* use "large x formula" */
return SavedPrefactor_list[n] / template_blas_pow((ergo_real)x, ((ergo_real)(2*n+1))/2);
}
/* choose which interval to use */
intervalWidth = (ergo_real)BOYS_X_MAX / BOYS_NO_OF_INTERVALS;
count = x / intervalWidth;
intervalIndex = (int)std::floor((long double)count); // FIXME: ACCURACY PROBLEM HERE FOR QUAD PRECISION?
if((intervalIndex < 0) || (intervalIndex >= BOYS_NO_OF_INTERVALS))
throw std::runtime_error("Error in BoysFunctionManager::BoysFunction_pretabulated: bad intervalIndex.");
interval = &Boys_list[n].list[intervalIndex];
sum = 0;
deltax = x - interval->midx;
deltaxtopowk = 1;
for(k = 0; k < BOYS_TAB_DEGREE; k++)
{
ergo_real Ak = interval->A[k];
sum += Ak * deltaxtopowk;
deltaxtopowk *= deltax;
}
return sum;
}
ergo_real
BoysFunctionManager::BoysFunction(int n, ergo_real x) const {
return BoysFunction_pretabulated(n, x);
}
ergo_real BoysFunctionManager::BoysFunction_expensive(int n, ergo_real x, int noOfIntegrationIntervals, int method) const {
if(method == 0) {
// default case
#ifdef PRECISION_SINGLE
return BoysFunction_raw_simpson(n, x, noOfIntegrationIntervals);
#else
return BoysFunction_raw_GaussLobatto(n, x, noOfIntegrationIntervals);
#endif
}
else if(method == 1)
return BoysFunction_raw_simpson(n, x, noOfIntegrationIntervals);
else if(method == 2)
return BoysFunction_raw_booles_rule(n, x, noOfIntegrationIntervals);
else if(method == 3)
return BoysFunction_raw_GaussLobatto(n, x, noOfIntegrationIntervals);
else
throw std::runtime_error("Error in BoysFunctionManager::BoysFunction_expensive: bad mthod value.");
}
BoysFunctionManager::BoysFunctionManager() : Boys_list(BOYS_N_MAX), Boys_init_flag(0) {
for(int i = 0; i < BOYS_N_MAX; i++)
SavedPrefactor_list[i] = 0;
}
/** Function needed for Chunks&Tasks usage. */
void BoysFunctionManager::write_to_buffer ( char * dataBuffer, size_t const bufferSize ) const {
char* p = dataBuffer;
if(bufferSize < get_size())
throw std::runtime_error("Error in BoysFunctionManager::write_to_buffer: bufferSize too small.");
// Boys_list
memcpy(p, &Boys_list[0], BOYS_N_MAX*sizeof(BoysFuncIntervalSetStruct));
p += BOYS_N_MAX*sizeof(BoysFuncIntervalSetStruct);
// SavedPrefactor_list
memcpy(p, SavedPrefactor_list, sizeof(SavedPrefactor_list));
p += sizeof(SavedPrefactor_list);
// Boys_init_flag
memcpy(p, &Boys_init_flag, sizeof(int));
p += sizeof(int);
// DONE!
}
/** Function needed for Chunks&Tasks usage. */
size_t BoysFunctionManager::get_size() const {
return BOYS_N_MAX*sizeof(BoysFuncIntervalSetStruct) + sizeof(SavedPrefactor_list) + sizeof(int);
}
/** Function needed for Chunks&Tasks usage. */
void BoysFunctionManager::assign_from_buffer ( char const * dataBuffer, size_t const bufferSize) {
const char* p = dataBuffer;
// Boys_list
memcpy(&Boys_list[0], p, BOYS_N_MAX*sizeof(BoysFuncIntervalSetStruct));
p += BOYS_N_MAX*sizeof(BoysFuncIntervalSetStruct);
// SavedPrefactor_list
memcpy(SavedPrefactor_list, p, sizeof(SavedPrefactor_list));
p += sizeof(SavedPrefactor_list);
// Boys_init_flag
memcpy(&Boys_init_flag, p, sizeof(int));
p += sizeof(int);
// DONE!
if(static_cast<size_t>(p-dataBuffer) > bufferSize)
throw std::runtime_error("Error: (p > bufferSize).");
}
|