VRR_GTG_1d_xx_xx.h

/*
 *  Copyright (C) 2004-2021 Edward F. Valeev
 *
 *  This file is part of Libint.
 *
 *  Libint is free software: you can redistribute it and/or modify
 *  it under the terms of the GNU Lesser General Public License as published by
 *  the Free Software Foundation, either version 3 of the License, or
 *  (at your option) any later version.
 *
 *  Libint 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 Lesser General Public License for more details.
 *
 *  You should have received a copy of the GNU Lesser General Public License
 *  along with Libint.  If not, see <http://www.gnu.org/licenses/>.
 *
 */
 
#ifndef _libint2_src_lib_libint_vrrgtg1dxxxx_h_
#define _libint2_src_lib_libint_vrrgtg1dxxxx_h_
 
#include <cstdlib>
#include <cassert>
#include <libint2.h>
#include <util_types.h>
 
namespace libint2 {
 
  template <unsigned int CartesianAxis, int La, int Lb, int Lc, int Ld, bool vectorize>
  struct VRR_GTG_1d_xx_xx {
 
    static void compute(const Libint_t* inteval,
                        LIBINT2_REALTYPE* target,
                        const LIBINT2_REALTYPE* src0) {
 
      enum XYZ {x=0, y=1, z=2};
      assert(CartesianAxis == x || CartesianAxis == y || CartesianAxis == z);
      assert(!vectorize);
 
      const unsigned int veclen = vectorize ? inteval->veclen : 1;
 
      // corner case: (00|00)
      if (La == 0 && Lb == 0 && Lc == 0 && Ld == 0) {
        for (unsigned int v=0; v!=veclen; ++v)
          target[v] = src0[v];
        return;
      }
 
      //---------------------------------------------
      // Part (1): build (a+b 0|c+d 0)
      //---------------------------------------------
 
      LIBINT2_REALTYPE apb_0_GTG_cpd_0[La+Lb+1][Lc+Ld+1];
      apb_0_GTG_cpd_0[0][0] = src0[0];
 
      const LIBINT2_REALTYPE *pfac0_0, *pfac0_1;
      const LIBINT2_REALTYPE *pfac1_0 = inteval->R12kG12_pfac1_0;
      const LIBINT2_REALTYPE *pfac1_1 = inteval->R12kG12_pfac1_1;
      const LIBINT2_REALTYPE *pfac2 = inteval->R12kG12_pfac2;
      switch (CartesianAxis) {
        case x:
        pfac0_0 = inteval->R12kG12_pfac0_0_x;
        pfac0_1 = inteval->R12kG12_pfac0_1_x;
        break;
        case y:
        pfac0_0 = inteval->R12kG12_pfac0_0_y;
        pfac0_1 = inteval->R12kG12_pfac0_1_y;
        break;
        case z:
        pfac0_0 = inteval->R12kG12_pfac0_0_z;
        pfac0_1 = inteval->R12kG12_pfac0_1_z;
        break;
        default: assert(false);
      }
 
      // build (0 0|1 0)
      if (Lc+Ld > 0) {
        apb_0_GTG_cpd_0[0][1] = pfac0_1[0] * apb_0_GTG_cpd_0[0][0];
#if LIBINT2_FLOP_COUNT
        inteval->nflops[0] += 1;
#endif
      }
 
      // build (0 0|c+d 0)
      if (Lc+Ld > 1) {
        for(int c_plus_d=1; c_plus_d!=Lc+Ld; ++c_plus_d) {
          apb_0_GTG_cpd_0[0][c_plus_d+1] = pfac0_1[0] * apb_0_GTG_cpd_0[0][c_plus_d] +
                                c_plus_d * pfac1_1[0] * apb_0_GTG_cpd_0[0][c_plus_d-1];
        }
#if LIBINT2_FLOP_COUNT
        inteval->nflops[0] += 4*(Lc+Ld-1);
#endif
      }
 
      // build (1 0|0 0)
      if (La+Lb > 0) {
        apb_0_GTG_cpd_0[1][0] = pfac0_0[0] * apb_0_GTG_cpd_0[0][0];
#if LIBINT2_FLOP_COUNT
        inteval->nflops[0] += 1;
#endif
      }
 
      // build (a+b 0|0 0)
      if (La+Lb > 1) {
        for(int a_plus_b=1; a_plus_b!=La+Lb; ++a_plus_b) {
          apb_0_GTG_cpd_0[a_plus_b+1][0] = pfac0_0[0] * apb_0_GTG_cpd_0[a_plus_b][0] +
                                a_plus_b * pfac1_0[0] * apb_0_GTG_cpd_0[a_plus_b-1][0];
        }
#if LIBINT2_FLOP_COUNT
        inteval->nflops[0] += 4*(La+Lb-1);
#endif
      }
 
      // build (1 0|c+d 0)
      if (La+Lb > 0 && Lc+Ld > 0) {
        for(int c_plus_d=1; c_plus_d<=Lc+Ld; ++c_plus_d) {
          apb_0_GTG_cpd_0[1][c_plus_d] = pfac0_0[0] * apb_0_GTG_cpd_0[0][c_plus_d] +
                                c_plus_d * pfac2[0] * apb_0_GTG_cpd_0[0][c_plus_d-1];
        }
#if LIBINT2_FLOP_COUNT
        inteval->nflops[0] += 4*(Lc+Ld-1);
#endif
      }
 
      // build (a+b 0|c+d 0)
      if (La+Lb > 1 && Lc+Ld > 0) {
        for(int a_plus_b=1; a_plus_b!=La+Lb; ++a_plus_b) {
          for(int c_plus_d=1; c_plus_d<=Lc+Ld; ++c_plus_d) {
            apb_0_GTG_cpd_0[a_plus_b+1][c_plus_d] = pfac0_0[0] * apb_0_GTG_cpd_0[a_plus_b][c_plus_d] +
                                         a_plus_b * pfac1_0[0] * apb_0_GTG_cpd_0[a_plus_b-1][c_plus_d] +
                                         c_plus_d *   pfac2[0] * apb_0_GTG_cpd_0[a_plus_b][c_plus_d-1];
          }
        }
#if LIBINT2_FLOP_COUNT
        inteval->nflops[0] += 7*(La+Lb-1)*(Lc+Ld-1);
#endif
      }
 
      //---------------------------------------------
      // Part (2): build (a b|c+d 0)
      //---------------------------------------------
 
      const LIBINT2_REALTYPE* AB;
      switch (CartesianAxis) {
        case x:
        AB = inteval->AB_x;
        break;
        case y:
        AB = inteval->AB_y;
        break;
        case z:
        AB = inteval->AB_z;
        break;
        default: assert(false);
      }
 
      LIBINT2_REALTYPE a_b_GTG_cpd_0[La+1][Lb+1][Lc+Ld+1];
      for(int c_plus_d=0; c_plus_d<=Lc+Ld; ++c_plus_d) {
        // copy (a+b 0| to a local 0,a+b buffer
        LIBINT2_REALTYPE b_a_GTG[La+Lb+1][La+Lb+1];
        for(int a_plus_b=0; a_plus_b<=La+Lb; ++a_plus_b) {
          b_a_GTG[0][a_plus_b] = apb_0_GTG_cpd_0[a_plus_b][c_plus_d];
        }
        // use HRR to compute b,a
        for(int b=1; b<=Lb; ++b) {
          for(int a=0; a<=La+Lb-b; ++a) {
            b_a_GTG[b][a] = b_a_GTG[b-1][a+1] + AB[0] * b_a_GTG[b-1][a];
          }
#if LIBINT2_FLOP_COUNT
            inteval->nflops[0] += 2 * (La+Lb-b+1);
#endif
        }
        // copy b,a to (a b|
        for(int b=0; b<=Lb; ++b) {
          for(int a=0; a<=La; ++a) {
            a_b_GTG_cpd_0[a][b][c_plus_d] = b_a_GTG[b][a];
          }
        }
      }
 
      //---------------------------------------------
      // Part (3): build (a b|c d)
      //---------------------------------------------
 
      const LIBINT2_REALTYPE* CD;
      switch (CartesianAxis) {
        case x:
        CD = inteval->CD_x;
        break;
        case y:
        CD = inteval->CD_y;
        break;
        case z:
        CD = inteval->CD_z;
        break;
        default: assert(false);
      }
 
      LIBINT2_REALTYPE* target_a_b_blk_ptr = target;
      const int Nd = (Ld+1);
      const int Ncd = (Lc+1)*Nd;
      for(int a=0; a<=La; ++a) {
        for(int b=0; b<=Lb; ++b, target_a_b_blk_ptr+=Ncd) {
          // copy |c+d 0) to a local 0,c+d buffer
          LIBINT2_REALTYPE d_c_GTG[Lc+Ld+1][Lc+Ld+1];
          for(int c_plus_d=0; c_plus_d<=Lc+Ld; ++c_plus_d) {
            d_c_GTG[0][c_plus_d] = a_b_GTG_cpd_0[a][b][c_plus_d];
          }
          // use HRR to compute d,c
          for(int d=1; d<=Ld; ++d) {
            for(int c=0; c<=Lc+Ld-d; ++c) {
              d_c_GTG[d][c] = d_c_GTG[d-1][c+1] + CD[0] * d_c_GTG[d-1][c];
            }
#if LIBINT2_FLOP_COUNT
              inteval->nflops[0] += 2 * (Lc+Ld-d+1);
#endif
          }
          // copy d,c to |c d)
          for(int d=0; d<=Ld; ++d) {
            for(int c=0, cd=d; c<=Lc; ++c, cd+=Nd) {
              target_a_b_blk_ptr[cd] = d_c_GTG[d][c];
            }
          }
        }
      }
 
      // done
    }
 
  };
 
};
 
#endif // header guard