smith/caspt2/MSCASPT2_tasks12.cc

//
// BAGEL - Brilliantly Advanced General Electronic Structure Library
// Filename: MSCASPT2_tasks12.cc
// Copyright (C) 2014 Toru Shiozaki
//
// Author: Toru Shiozaki <shiozaki@northwestern.edu>
// Maintainer: Shiozaki group
//
// This file is part of the BAGEL package.
//
// 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/>.
//

#include <bagel_config.h>
#ifdef COMPILE_SMITH

#include <src/smith/caspt2/MSCASPT2_tasks12.h>

using namespace std;
using namespace bagel;
using namespace bagel::SMITH;
using namespace bagel::SMITH::MSCASPT2;

void Task550::Task_local::compute() {
  const Index x2 = b(0);
  const Index x3 = b(1);
  const Index x0 = b(2);
  const Index x1 = b(3);
  // tensor label: I832
  std::unique_ptr<double[]> odata(new double[out()->get_size(x2, x3, x0, x1)]);
  std::fill_n(odata.get(), out()->get_size(x2, x3, x0, x1), 0.0);
  std::unique_ptr<double[]> odata_sorted(new double[out()->get_size(x2, x3, x0, x1)]);
  std::fill_n(odata_sorted.get(), out()->get_size(x2, x3, x0, x1), 0.0);
  for (auto& c1 : *range_[0]) {
    for (auto& a2 : *range_[2]) {
      // tensor label: v2
      std::unique_ptr<double[]> i0data = in(0)->get_block(c1, a2, x0, x1);
      std::unique_ptr<double[]> i0data_sorted(new double[in(0)->get_size(c1, a2, x0, x1)]);
      sort_indices<0,1,2,3,0,1,1,1>(i0data, i0data_sorted, c1.size(), a2.size(), x0.size(), x1.size());
      // tensor label: l2
      std::unique_ptr<double[]> i1data = in(1)->get_block(c1, a2, x3, x2);
      std::unique_ptr<double[]> i1data_sorted(new double[in(1)->get_size(c1, a2, x3, x2)]);
      sort_indices<0,1,2,3,0,1,1,1>(i1data, i1data_sorted, c1.size(), a2.size(), x3.size(), x2.size());
      dgemm_("T", "N", x0.size()*x1.size(), x2.size()*x3.size(), a2.size()*c1.size(),
             1.0, i0data_sorted, a2.size()*c1.size(), i1data_sorted, a2.size()*c1.size(),
             1.0, odata_sorted, x0.size()*x1.size());
    }
  }
  sort_indices<3,2,0,1,1,1,1,1>(odata_sorted, odata, x0.size(), x1.size(), x3.size(), x2.size());
  out()->add_block(odata, x2, x3, x0, x1);
}

void Task551::Task_local::compute() {
  const Index x2 = b(0);
  const Index x3 = b(1);
  const Index x0 = b(2);
  const Index x1 = b(3);
  // tensor label: I832
  std::unique_ptr<double[]> odata(new double[out()->get_size(x2, x3, x0, x1)]);
  std::fill_n(odata.get(), out()->get_size(x2, x3, x0, x1), 0.0);
  std::unique_ptr<double[]> odata_sorted(new double[out()->get_size(x2, x3, x0, x1)]);
  std::fill_n(odata_sorted.get(), out()->get_size(x2, x3, x0, x1), 0.0);
  for (auto& a2 : *range_[2]) {
    for (auto& c1 : *range_[0]) {
      // tensor label: v2
      std::unique_ptr<double[]> i0data = in(0)->get_block(x0, a2, c1, x1);
      std::unique_ptr<double[]> i0data_sorted(new double[in(0)->get_size(x0, a2, c1, x1)]);
      sort_indices<1,2,0,3,0,1,1,1>(i0data, i0data_sorted, x0.size(), a2.size(), c1.size(), x1.size());
      // tensor label: l2
      std::unique_ptr<double[]> i1data = in(1)->get_block(c1, a2, x3, x2);
      std::unique_ptr<double[]> i1data_sorted(new double[in(1)->get_size(c1, a2, x3, x2)]);
      sort_indices<1,0,2,3,0,1,-1,2>(i1data, i1data_sorted, c1.size(), a2.size(), x3.size(), x2.size());
      dgemm_("T", "N", x0.size()*x1.size(), x2.size()*x3.size(), a2.size()*c1.size(),
             1.0, i0data_sorted, a2.size()*c1.size(), i1data_sorted, a2.size()*c1.size(),
             1.0, odata_sorted, x0.size()*x1.size());
    }
  }
  sort_indices<3,2,0,1,1,1,1,1>(odata_sorted, odata, x0.size(), x1.size(), x3.size(), x2.size());
  out()->add_block(odata, x2, x3, x0, x1);
}

void Task552::Task_local::compute() {
  const Index x2 = b(0);
  const Index x3 = b(1);
  const Index x0 = b(2);
  const Index x1 = b(3);
  // tensor label: I832
  std::unique_ptr<double[]> odata(new double[out()->get_size(x2, x3, x0, x1)]);
  std::fill_n(odata.get(), out()->get_size(x2, x3, x0, x1), 0.0);
  std::unique_ptr<double[]> odata_sorted(new double[out()->get_size(x2, x3, x0, x1)]);
  std::fill_n(odata_sorted.get(), out()->get_size(x2, x3, x0, x1), 0.0);
  for (auto& c1 : *range_[0]) {
    for (auto& a2 : *range_[2]) {
      // tensor label: v2
      std::unique_ptr<double[]> i0data = in(0)->get_block(x0, x1, c1, a2);
      std::unique_ptr<double[]> i0data_sorted(new double[in(0)->get_size(x0, x1, c1, a2)]);
      sort_indices<2,3,0,1,0,1,1,1>(i0data, i0data_sorted, x0.size(), x1.size(), c1.size(), a2.size());
      // tensor label: l2
      std::unique_ptr<double[]> i1data = in(1)->get_block(c1, a2, x3, x2);
      std::unique_ptr<double[]> i1data_sorted(new double[in(1)->get_size(c1, a2, x3, x2)]);
      sort_indices<0,1,2,3,0,1,1,1>(i1data, i1data_sorted, c1.size(), a2.size(), x3.size(), x2.size());
      dgemm_("T", "N", x0.size()*x1.size(), x2.size()*x3.size(), a2.size()*c1.size(),
             1.0, i0data_sorted, a2.size()*c1.size(), i1data_sorted, a2.size()*c1.size(),
             1.0, odata_sorted, x0.size()*x1.size());
    }
  }
  sort_indices<3,2,0,1,1,1,1,1>(odata_sorted, odata, x0.size(), x1.size(), x3.size(), x2.size());
  out()->add_block(odata, x2, x3, x0, x1);
}

void Task553::Task_local::compute() {
  const Index x3 = b(0);
  const Index x1 = b(1);
  const Index x0 = b(2);
  const Index x2 = b(3);
  // tensor label: I835
  std::unique_ptr<double[]> i0data = in(0)->get_block(x2, x3, x0, x1);
  std::unique_ptr<double[]> i0data_sorted(new double[in(0)->get_size(x2, x3, x0, x1)]);
  sort_indices<1,3,2,0,0,1,1,1>(i0data, i0data_sorted, x2.size(), x3.size(), x0.size(), x1.size());
  // tensor label (calculated on-the-fly): Gamma132
  {
    if (x0 == x1) {
      std::unique_ptr<double[]> o1data(new double[out(1)->get_size(x3, x2)]);
      std::fill_n(o1data.get(), out(1)->get_size(x3, x2), 0.0);
      for (int ix2 = 0; ix2 != x2.size(); ++ix2) {
        for (int ix1 = 0; ix1 != x1.size(); ++ix1) {
          for (int ix3 = 0; ix3 != x3.size(); ++ix3) {
            o1data[ix3+x3.size()*(ix2)] +=
              1.0 * i0data_sorted[ix3+x3.size()*(ix1+x1.size()*(ix1+x0.size()*(ix2)))];
          }
        }
      }
      out(1)->add_block(o1data, x3, x2);
    }
  }
  {
    if (x0 == x2) {
      std::unique_ptr<double[]> o1data(new double[out(1)->get_size(x3, x1)]);
      std::fill_n(o1data.get(), out(1)->get_size(x3, x1), 0.0);
      for (int ix2 = 0; ix2 != x2.size(); ++ix2) {
        for (int ix1 = 0; ix1 != x1.size(); ++ix1) {
          for (int ix3 = 0; ix3 != x3.size(); ++ix3) {
            o1data[ix3+x3.size()*(ix1)] +=
              -2.0 * i0data_sorted[ix3+x3.size()*(ix1+x1.size()*(ix2+x0.size()*(ix2)))];
          }
        }
      }
      out(1)->add_block(o1data, x3, x1);
    }
  }
  {
    std::unique_ptr<double[]> o2data(new double[out(2)->get_size(x3, x1, x0, x2)]);
    std::fill_n(o2data.get(), out(2)->get_size(x3, x1, x0, x2), 0.0);
    sort_indices<0,1,2,3,1,1,1,1>(i0data_sorted, o2data, x3.size(), x1.size(), x0.size(), x2.size());
    out(2)->add_block(o2data, x3, x1, x0, x2);
  }
}

void Task554::Task_local::compute() {
  const Index x2 = b(0);
  const Index x3 = b(1);
  const Index x0 = b(2);
  const Index x1 = b(3);
  // tensor label: I835
  std::unique_ptr<double[]> odata(new double[out()->get_size(x2, x3, x0, x1)]);
  std::fill_n(odata.get(), out()->get_size(x2, x3, x0, x1), 0.0);
  std::unique_ptr<double[]> odata_sorted(new double[out()->get_size(x2, x3, x0, x1)]);
  std::fill_n(odata_sorted.get(), out()->get_size(x2, x3, x0, x1), 0.0);
  for (auto& c2 : *range_[0]) {
    for (auto& a1 : *range_[2]) {
      // tensor label: v2
      std::unique_ptr<double[]> i0data = in(0)->get_block(c2, x0, x1, a1);
      std::unique_ptr<double[]> i0data_sorted(new double[in(0)->get_size(c2, x0, x1, a1)]);
      sort_indices<0,3,1,2,0,1,1,1>(i0data, i0data_sorted, c2.size(), x0.size(), x1.size(), a1.size());
      // tensor label: l2
      std::unique_ptr<double[]> i1data = in(1)->get_block(x3, a1, c2, x2);
      std::unique_ptr<double[]> i1data_sorted(new double[in(1)->get_size(x3, a1, c2, x2)]);
      sort_indices<2,1,0,3,0,1,-1,2>(i1data, i1data_sorted, x3.size(), a1.size(), c2.size(), x2.size());
      dgemm_("T", "N", x0.size()*x1.size(), x2.size()*x3.size(), c2.size()*a1.size(),
             1.0, i0data_sorted, c2.size()*a1.size(), i1data_sorted, c2.size()*a1.size(),
             1.0, odata_sorted, x0.size()*x1.size());
    }
  }
  sort_indices<3,2,0,1,1,1,1,1>(odata_sorted, odata, x0.size(), x1.size(), x3.size(), x2.size());
  out()->add_block(odata, x2, x3, x0, x1);
}

void Task555::Task_local::compute() {
  const Index x3 = b(0);
  const Index x0 = b(1);
  const Index x1 = b(2);
  const Index x2 = b(3);
  // tensor label: I838
  std::unique_ptr<double[]> i0data = in(0)->get_block(x2, x3, x0, x1);
  std::unique_ptr<double[]> i0data_sorted(new double[in(0)->get_size(x2, x3, x0, x1)]);
  sort_indices<1,2,3,0,0,1,1,1>(i0data, i0data_sorted, x2.size(), x3.size(), x0.size(), x1.size());
  // tensor label (calculated on-the-fly): Gamma142
  {
    if (x1 == x2) {
      std::unique_ptr<double[]> o1data(new double[out(1)->get_size(x3, x0)]);
      std::fill_n(o1data.get(), out(1)->get_size(x3, x0), 0.0);
      for (int ix2 = 0; ix2 != x2.size(); ++ix2) {
        for (int ix0 = 0; ix0 != x0.size(); ++ix0) {
          for (int ix3 = 0; ix3 != x3.size(); ++ix3) {
            o1data[ix3+x3.size()*(ix0)] +=
              2.0 * i0data_sorted[ix3+x3.size()*(ix0+x0.size()*(ix2+x1.size()*(ix2)))];
          }
        }
      }
      out(1)->add_block(o1data, x3, x0);
    }
  }
  {
    std::unique_ptr<double[]> o2data(new double[out(2)->get_size(x3, x0, x1, x2)]);
    std::fill_n(o2data.get(), out(2)->get_size(x3, x0, x1, x2), 0.0);
    sort_indices<0,1,2,3,1,1,-1,1>(i0data_sorted, o2data, x3.size(), x0.size(), x1.size(), x2.size());
    out(2)->add_block(o2data, x3, x0, x1, x2);
  }
}

void Task556::Task_local::compute() {
  const Index x2 = b(0);
  const Index x3 = b(1);
  const Index x0 = b(2);
  const Index x1 = b(3);
  // tensor label: I838
  std::unique_ptr<double[]> odata(new double[out()->get_size(x2, x3, x0, x1)]);
  std::fill_n(odata.get(), out()->get_size(x2, x3, x0, x1), 0.0);
  std::unique_ptr<double[]> odata_sorted(new double[out()->get_size(x2, x3, x0, x1)]);
  std::fill_n(odata_sorted.get(), out()->get_size(x2, x3, x0, x1), 0.0);
  for (auto& a1 : *range_[2]) {
    for (auto& c2 : *range_[0]) {
      // tensor label: v2
      std::unique_ptr<double[]> i0data = in(0)->get_block(x0, a1, c2, x1);
      std::unique_ptr<double[]> i0data_sorted(new double[in(0)->get_size(x0, a1, c2, x1)]);
      sort_indices<1,2,0,3,0,1,1,1>(i0data, i0data_sorted, x0.size(), a1.size(), c2.size(), x1.size());
      // tensor label: l2
      std::unique_ptr<double[]> i1data = in(1)->get_block(x3, a1, c2, x2);
      std::unique_ptr<double[]> i1data_sorted(new double[in(1)->get_size(x3, a1, c2, x2)]);
      sort_indices<1,2,0,3,0,1,1,2>(i1data, i1data_sorted, x3.size(), a1.size(), c2.size(), x2.size());
      dgemm_("T", "N", x0.size()*x1.size(), x2.size()*x3.size(), c2.size()*a1.size(),
             1.0, i0data_sorted, c2.size()*a1.size(), i1data_sorted, c2.size()*a1.size(),
             1.0, odata_sorted, x0.size()*x1.size());
    }
  }
  sort_indices<3,2,0,1,1,1,1,1>(odata_sorted, odata, x0.size(), x1.size(), x3.size(), x2.size());
  out()->add_block(odata, x2, x3, x0, x1);
}

void Task557::Task_local::compute() {
  const Index x3 = b(0);
  const Index x2 = b(1);
  const Index x0 = b(2);
  const Index x1 = b(3);
  // tensor label: I847
  std::unique_ptr<double[]> i0data = in(0)->get_block(x2, x3, x0, x1);
  std::unique_ptr<double[]> i0data_sorted(new double[in(0)->get_size(x2, x3, x0, x1)]);
  sort_indices<1,0,2,3,0,1,1,1>(i0data, i0data_sorted, x2.size(), x3.size(), x0.size(), x1.size());
  // tensor label (calculated on-the-fly): Gamma122
  {
    if (x0 == x1) {
      std::unique_ptr<double[]> o1data(new double[out(1)->get_size(x3, x2)]);
      std::fill_n(o1data.get(), out(1)->get_size(x3, x2), 0.0);
      for (int ix1 = 0; ix1 != x1.size(); ++ix1) {
        for (int ix2 = 0; ix2 != x2.size(); ++ix2) {
          for (int ix3 = 0; ix3 != x3.size(); ++ix3) {
            o1data[ix3+x3.size()*(ix2)] +=
              2.0 * i0data_sorted[ix3+x3.size()*(ix2+x2.size()*(ix1+x0.size()*(ix1)))];
          }
        }
      }
      out(1)->add_block(o1data, x3, x2);
    }
  }
  {
    if (x0 == x2) {
      std::unique_ptr<double[]> o1data(new double[out(1)->get_size(x3, x1)]);
      std::fill_n(o1data.get(), out(1)->get_size(x3, x1), 0.0);
      for (int ix1 = 0; ix1 != x1.size(); ++ix1) {
        for (int ix2 = 0; ix2 != x2.size(); ++ix2) {
          for (int ix3 = 0; ix3 != x3.size(); ++ix3) {
            o1data[ix3+x3.size()*(ix1)] +=
              -1.0 * i0data_sorted[ix3+x3.size()*(ix2+x2.size()*(ix2+x0.size()*(ix1)))];
          }
        }
      }
      out(1)->add_block(o1data, x3, x1);
    }
  }
  {
    std::unique_ptr<double[]> o2data(new double[out(2)->get_size(x3, x2, x0, x1)]);
    std::fill_n(o2data.get(), out(2)->get_size(x3, x2, x0, x1), 0.0);
    sort_indices<0,1,2,3,1,1,-1,1>(i0data_sorted, o2data, x3.size(), x2.size(), x0.size(), x1.size());
    out(2)->add_block(o2data, x3, x2, x0, x1);
  }
}

void Task558::Task_local::compute() {
  const Index x2 = b(0);
  const Index x3 = b(1);
  const Index x0 = b(2);
  const Index x1 = b(3);
  // tensor label: I847
  std::unique_ptr<double[]> odata(new double[out()->get_size(x2, x3, x0, x1)]);
  std::fill_n(odata.get(), out()->get_size(x2, x3, x0, x1), 0.0);
  std::unique_ptr<double[]> odata_sorted(new double[out()->get_size(x2, x3, x0, x1)]);
  std::fill_n(odata_sorted.get(), out()->get_size(x2, x3, x0, x1), 0.0);
  for (auto& c1 : *range_[0]) {
    for (auto& a2 : *range_[2]) {
      // tensor label: v2
      std::unique_ptr<double[]> i0data = in(0)->get_block(c1, x0, x1, a2);
      std::unique_ptr<double[]> i0data_sorted(new double[in(0)->get_size(c1, x0, x1, a2)]);
      sort_indices<0,3,1,2,0,1,1,1>(i0data, i0data_sorted, c1.size(), x0.size(), x1.size(), a2.size());
      // tensor label: l2
      std::unique_ptr<double[]> i1data = in(1)->get_block(c1, a2, x3, x2);
      std::unique_ptr<double[]> i1data_sorted(new double[in(1)->get_size(c1, a2, x3, x2)]);
      sort_indices<0,1,2,3,0,1,1,2>(i1data, i1data_sorted, c1.size(), a2.size(), x3.size(), x2.size());
      dgemm_("T", "N", x0.size()*x1.size(), x2.size()*x3.size(), a2.size()*c1.size(),
             1.0, i0data_sorted, a2.size()*c1.size(), i1data_sorted, a2.size()*c1.size(),
             1.0, odata_sorted, x0.size()*x1.size());
    }
  }
  sort_indices<3,2,0,1,1,1,1,1>(odata_sorted, odata, x0.size(), x1.size(), x3.size(), x2.size());
  out()->add_block(odata, x2, x3, x0, x1);
}

void Task559::Task_local::compute() {
  const Index x2 = b(0);
  const Index x3 = b(1);
  const Index x0 = b(2);
  const Index x1 = b(3);
  // tensor label: I847
  std::unique_ptr<double[]> odata(new double[out()->get_size(x2, x3, x0, x1)]);
  std::fill_n(odata.get(), out()->get_size(x2, x3, x0, x1), 0.0);
  std::unique_ptr<double[]> odata_sorted(new double[out()->get_size(x2, x3, x0, x1)]);
  std::fill_n(odata_sorted.get(), out()->get_size(x2, x3, x0, x1), 0.0);
  for (auto& c1 : *range_[0]) {
    // tensor label: h1
    std::unique_ptr<double[]> i0data = in(0)->get_block(c1, x0);
    std::unique_ptr<double[]> i0data_sorted(new double[in(0)->get_size(c1, x0)]);
    sort_indices<0,1,0,1,1,1>(i0data, i0data_sorted, c1.size(), x0.size());
    // tensor label: l2
    std::unique_ptr<double[]> i1data = in(1)->get_block(x3, x2, c1, x1);
    std::unique_ptr<double[]> i1data_sorted(new double[in(1)->get_size(x3, x2, c1, x1)]);
    sort_indices<2,0,1,3,0,1,1,1>(i1data, i1data_sorted, x3.size(), x2.size(), c1.size(), x1.size());
    dgemm_("T", "N", x0.size(), x1.size()*x2.size()*x3.size(), c1.size(),
           1.0, i0data_sorted, c1.size(), i1data_sorted, c1.size(),
           1.0, odata_sorted, x0.size());
  }
  sort_indices<2,1,0,3,1,1,1,1>(odata_sorted, odata, x0.size(), x3.size(), x2.size(), x1.size());
  out()->add_block(odata, x2, x3, x0, x1);
}

void Task560::Task_local::compute() {
  const Index x5 = b(0);
  const Index x0 = b(1);
  const Index x4 = b(2);
  const Index x3 = b(3);
  const Index x2 = b(4);
  const Index x1 = b(5);
  // tensor label: I856
  std::unique_ptr<double[]> i0data = in(0)->get_block(x3, x4, x5, x0, x1, x2);
  std::unique_ptr<double[]> i0data_sorted(new double[in(0)->get_size(x3, x4, x5, x0, x1, x2)]);
  sort_indices<2,3,1,0,5,4,0,1,1,1>(i0data, i0data_sorted, x3.size(), x4.size(), x5.size(), x0.size(), x1.size(), x2.size());
  // tensor label (calculated on-the-fly): Gamma169
  {
    if (x2 == x3) {
      std::unique_ptr<double[]> o2data(new double[out(2)->get_size(x5, x0, x4, x1)]);
      std::fill_n(o2data.get(), out(2)->get_size(x5, x0, x4, x1), 0.0);
      for (int ix1 = 0; ix1 != x1.size(); ++ix1) {
        for (int ix3 = 0; ix3 != x3.size(); ++ix3) {
          for (int ix4 = 0; ix4 != x4.size(); ++ix4) {
            for (int ix0 = 0; ix0 != x0.size(); ++ix0) {
              for (int ix5 = 0; ix5 != x5.size(); ++ix5) {
                o2data[ix5+x5.size()*(ix0+x0.size()*(ix4+x4.size()*(ix1)))] +=
                  1.0 * i0data_sorted[ix5+x5.size()*(ix0+x0.size()*(ix4+x4.size()*(ix3+x3.size()*(ix3+x2.size()*(ix1)))))];
              }
            }
          }
        }
      }
      out(2)->add_block(o2data, x5, x0, x4, x1);
    }
  }
  {
    std::unique_ptr<double[]> o3data(new double[out(3)->get_size(x5, x0, x4, x3, x2, x1)]);
    std::fill_n(o3data.get(), out(3)->get_size(x5, x0, x4, x3, x2, x1), 0.0);
    sort_indices<0,1,2,3,4,5,1,1,1,1>(i0data_sorted, o3data, x5.size(), x0.size(), x4.size(), x3.size(), x2.size(), x1.size());
    out(3)->add_block(o3data, x5, x0, x4, x3, x2, x1);
  }
}

void Task561::Task_local::compute() {
  const Index x3 = b(0);
  const Index x4 = b(1);
  const Index x5 = b(2);
  const Index x0 = b(3);
  const Index x1 = b(4);
  const Index x2 = b(5);
  // tensor label: I856
  std::unique_ptr<double[]> odata(new double[out()->get_size(x3, x4, x5, x0, x1, x2)]);
  std::fill_n(odata.get(), out()->get_size(x3, x4, x5, x0, x1, x2), 0.0);
  std::unique_ptr<double[]> odata_sorted(new double[out()->get_size(x3, x4, x5, x0, x1, x2)]);
  std::fill_n(odata_sorted.get(), out()->get_size(x3, x4, x5, x0, x1, x2), 0.0);
  for (auto& a1 : *range_[2]) {
    // tensor label: v2
    std::unique_ptr<double[]> i0data = in(0)->get_block(x0, a1, x1, x2);
    std::unique_ptr<double[]> i0data_sorted(new double[in(0)->get_size(x0, a1, x1, x2)]);
    sort_indices<1,0,2,3,0,1,1,1>(i0data, i0data_sorted, x0.size(), a1.size(), x1.size(), x2.size());
    // tensor label: l2
    std::unique_ptr<double[]> i1data = in(1)->get_block(x5, a1, x4, x3);
    std::unique_ptr<double[]> i1data_sorted(new double[in(1)->get_size(x5, a1, x4, x3)]);
    sort_indices<1,0,2,3,0,1,1,2>(i1data, i1data_sorted, x5.size(), a1.size(), x4.size(), x3.size());
    dgemm_("T", "N", x0.size()*x1.size()*x2.size(), x3.size()*x4.size()*x5.size(), a1.size(),
           1.0, i0data_sorted, a1.size(), i1data_sorted, a1.size(),
           1.0, odata_sorted, x0.size()*x1.size()*x2.size());
  }
  sort_indices<5,4,3,0,1,2,1,1,1,1>(odata_sorted, odata, x0.size(), x1.size(), x2.size(), x5.size(), x4.size(), x3.size());
  out()->add_block(odata, x3, x4, x5, x0, x1, x2);
}

void Task562::Task_local::compute() {
  const Index x5 = b(0);
  const Index x2 = b(1);
  const Index x4 = b(2);
  const Index x3 = b(3);
  const Index x1 = b(4);
  const Index x0 = b(5);
  // tensor label: I859
  std::unique_ptr<double[]> i0data = in(0)->get_block(x3, x4, x5, x0, x1, x2);
  std::unique_ptr<double[]> i0data_sorted(new double[in(0)->get_size(x3, x4, x5, x0, x1, x2)]);
  sort_indices<2,5,1,0,4,3,0,1,1,1>(i0data, i0data_sorted, x3.size(), x4.size(), x5.size(), x0.size(), x1.size(), x2.size());
  // tensor label (calculated on-the-fly): Gamma161
  {
    if (x1 == x2) {
      std::unique_ptr<double[]> o2data(new double[out(2)->get_size(x5, x0, x4, x3)]);
      std::fill_n(o2data.get(), out(2)->get_size(x5, x0, x4, x3), 0.0);
      for (int ix0 = 0; ix0 != x0.size(); ++ix0) {
        for (int ix3 = 0; ix3 != x3.size(); ++ix3) {
          for (int ix4 = 0; ix4 != x4.size(); ++ix4) {
            for (int ix2 = 0; ix2 != x2.size(); ++ix2) {
              for (int ix5 = 0; ix5 != x5.size(); ++ix5) {
                o2data[ix5+x5.size()*(ix0+x0.size()*(ix4+x4.size()*(ix3)))] +=
                  1.0 * i0data_sorted[ix5+x5.size()*(ix2+x2.size()*(ix4+x4.size()*(ix3+x3.size()*(ix2+x1.size()*(ix0)))))];
              }
            }
          }
        }
      }
      out(2)->add_block(o2data, x5, x0, x4, x3);
    }
  }
  {
    if (x1 == x3) {
      std::unique_ptr<double[]> o2data(new double[out(2)->get_size(x5, x2, x4, x0)]);
      std::fill_n(o2data.get(), out(2)->get_size(x5, x2, x4, x0), 0.0);
      for (int ix0 = 0; ix0 != x0.size(); ++ix0) {
        for (int ix3 = 0; ix3 != x3.size(); ++ix3) {
          for (int ix4 = 0; ix4 != x4.size(); ++ix4) {
            for (int ix2 = 0; ix2 != x2.size(); ++ix2) {
              for (int ix5 = 0; ix5 != x5.size(); ++ix5) {
                o2data[ix5+x5.size()*(ix2+x2.size()*(ix4+x4.size()*(ix0)))] +=
                  1.0 * i0data_sorted[ix5+x5.size()*(ix2+x2.size()*(ix4+x4.size()*(ix3+x3.size()*(ix3+x1.size()*(ix0)))))];
              }
            }
          }
        }
      }
      out(2)->add_block(o2data, x5, x2, x4, x0);
    }
  }
  {
    std::unique_ptr<double[]> o3data(new double[out(3)->get_size(x5, x2, x4, x3, x1, x0)]);
    std::fill_n(o3data.get(), out(3)->get_size(x5, x2, x4, x3, x1, x0), 0.0);
    sort_indices<0,1,2,3,4,5,1,1,1,1>(i0data_sorted, o3data, x5.size(), x2.size(), x4.size(), x3.size(), x1.size(), x0.size());
    out(3)->add_block(o3data, x5, x2, x4, x3, x1, x0);
  }
}

void Task563::Task_local::compute() {
  const Index x3 = b(0);
  const Index x4 = b(1);
  const Index x5 = b(2);
  const Index x0 = b(3);
  const Index x1 = b(4);
  const Index x2 = b(5);
  // tensor label: I859
  std::unique_ptr<double[]> odata(new double[out()->get_size(x3, x4, x5, x0, x1, x2)]);
  std::fill_n(odata.get(), out()->get_size(x3, x4, x5, x0, x1, x2), 0.0);
  std::unique_ptr<double[]> odata_sorted(new double[out()->get_size(x3, x4, x5, x0, x1, x2)]);
  std::fill_n(odata_sorted.get(), out()->get_size(x3, x4, x5, x0, x1, x2), 0.0);
  for (auto& a1 : *range_[2]) {
    // tensor label: v2
    std::unique_ptr<double[]> i0data = in(0)->get_block(x0, x1, x2, a1);
    std::unique_ptr<double[]> i0data_sorted(new double[in(0)->get_size(x0, x1, x2, a1)]);
    sort_indices<3,0,1,2,0,1,1,1>(i0data, i0data_sorted, x0.size(), x1.size(), x2.size(), a1.size());
    // tensor label: l2
    std::unique_ptr<double[]> i1data = in(1)->get_block(x5, a1, x4, x3);
    std::unique_ptr<double[]> i1data_sorted(new double[in(1)->get_size(x5, a1, x4, x3)]);
    sort_indices<1,0,2,3,0,1,1,2>(i1data, i1data_sorted, x5.size(), a1.size(), x4.size(), x3.size());
    dgemm_("T", "N", x0.size()*x1.size()*x2.size(), x3.size()*x4.size()*x5.size(), a1.size(),
           1.0, i0data_sorted, a1.size(), i1data_sorted, a1.size(),
           1.0, odata_sorted, x0.size()*x1.size()*x2.size());
  }
  sort_indices<5,4,3,0,1,2,1,1,1,1>(odata_sorted, odata, x0.size(), x1.size(), x2.size(), x5.size(), x4.size(), x3.size());
  out()->add_block(odata, x3, x4, x5, x0, x1, x2);
}

void Task564::Task_local::compute() {
  // tensor label: I816
  std::unique_ptr<double[]> odata(new double[out()->get_size()]);
  std::fill_n(odata.get(), out()->get_size(), 0.0);
  // tensor label: I862
  std::unique_ptr<double[]> i0data = in(0)->get_block();
  odata[0] += -2.0 * i0data[0];
  out()->add_block(odata);
}

void Task565::Task_local::compute() {
  // tensor label: I862
  std::unique_ptr<double[]> odata(new double[out()->get_size()]);
  std::fill_n(odata.get(), out()->get_size(), 0.0);
  std::unique_ptr<double[]> odata_sorted(new double[out()->get_size()]);
  std::fill_n(odata_sorted.get(), out()->get_size(), 0.0);
  const Index c1 = b(0);
  const Index a4 = b(1);
  const Index c3 = b(2);
  const Index a2 = b(3);
  // tensor label: v2
  std::unique_ptr<double[]> i0data = in(0)->get_block(c1, a4, c3, a2);
  std::unique_ptr<double[]> i0data_sorted(new double[in(0)->get_size(c1, a4, c3, a2)]);
  sort_indices<0,1,2,3,0,1,1,1>(i0data, i0data_sorted, c1.size(), a4.size(), c3.size(), a2.size());
  // tensor label: l2
  std::unique_ptr<double[]> i1data = in(1)->get_block(c1, a2, c3, a4);
  std::unique_ptr<double[]> i1data_sorted(new double[in(1)->get_size(c1, a2, c3, a4)]);
  sort_indices<0,3,2,1,0,1,-2,1>(i1data, i1data_sorted, c1.size(), a2.size(), c3.size(), a4.size());
  odata_sorted[0] += ddot_(a4.size()*c3.size()*a2.size()*c1.size(), i0data_sorted, 1, i1data_sorted, 1);
  sort_indices<1,1,1,1>(odata_sorted, odata);
  out()->add_block(odata);
}

void Task566::Task_local::compute() {
  // tensor label: I816
  std::unique_ptr<double[]> odata(new double[out()->get_size()]);
  std::fill_n(odata.get(), out()->get_size(), 0.0);
  // tensor label: I865
  std::unique_ptr<double[]> i0data = in(0)->get_block();
  odata[0] += 4.0 * i0data[0];
  out()->add_block(odata);
}

void Task567::Task_local::compute() {
  // tensor label: I865
  std::unique_ptr<double[]> odata(new double[out()->get_size()]);
  std::fill_n(odata.get(), out()->get_size(), 0.0);
  std::unique_ptr<double[]> odata_sorted(new double[out()->get_size()]);
  std::fill_n(odata_sorted.get(), out()->get_size(), 0.0);
  const Index c1 = b(0);
  const Index a2 = b(1);
  const Index c3 = b(2);
  const Index a4 = b(3);
  // tensor label: v2
  std::unique_ptr<double[]> i0data = in(0)->get_block(c1, a2, c3, a4);
  std::unique_ptr<double[]> i0data_sorted(new double[in(0)->get_size(c1, a2, c3, a4)]);
  sort_indices<0,1,2,3,0,1,1,1>(i0data, i0data_sorted, c1.size(), a2.size(), c3.size(), a4.size());
  // tensor label: l2
  std::unique_ptr<double[]> i1data = in(1)->get_block(c1, a2, c3, a4);
  std::unique_ptr<double[]> i1data_sorted(new double[in(1)->get_size(c1, a2, c3, a4)]);
  sort_indices<0,1,2,3,0,1,4,1>(i1data, i1data_sorted, c1.size(), a2.size(), c3.size(), a4.size());
  odata_sorted[0] += ddot_(a4.size()*c3.size()*a2.size()*c1.size(), i0data_sorted, 1, i1data_sorted, 1);
  sort_indices<1,1,1,1>(odata_sorted, odata);
  out()->add_block(odata);
}

void Task568::Task_local::compute() {
  const Index x1 = b(0);
  const Index x0 = b(1);
  // tensor label: I868
  std::unique_ptr<double[]> i0data = in(0)->get_block(x1, x0);
  std::unique_ptr<double[]> i0data_sorted(new double[in(0)->get_size(x1, x0)]);
  sort_indices<0,1,0,1,1,1>(i0data, i0data_sorted, x1.size(), x0.size());
  // tensor label (calculated on-the-fly): Gamma148
  {
    std::unique_ptr<double[]> o1data(new double[out(1)->get_size(x1, x0)]);
    std::fill_n(o1data.get(), out(1)->get_size(x1, x0), 0.0);
    sort_indices<0,1,1,1,1,1>(i0data_sorted, o1data, x1.size(), x0.size());
    out(1)->add_block(o1data, x1, x0);
  }
}

void Task569::Task_local::compute() {
  const Index x1 = b(0);
  const Index x0 = b(1);
  // tensor label: I868
  std::unique_ptr<double[]> odata(new double[out()->get_size(x1, x0)]);
  std::fill_n(odata.get(), out()->get_size(x1, x0), 0.0);
  std::unique_ptr<double[]> odata_sorted(new double[out()->get_size(x1, x0)]);
  std::fill_n(odata_sorted.get(), out()->get_size(x1, x0), 0.0);
  for (auto& a3 : *range_[2]) {
    for (auto& c2 : *range_[0]) {
      for (auto& a1 : *range_[2]) {
        // tensor label: v2
        std::unique_ptr<double[]> i0data = in(0)->get_block(x0, a3, c2, a1);
        std::unique_ptr<double[]> i0data_sorted(new double[in(0)->get_size(x0, a3, c2, a1)]);
        sort_indices<1,2,3,0,0,1,1,1>(i0data, i0data_sorted, x0.size(), a3.size(), c2.size(), a1.size());
        // tensor label: l2
        std::unique_ptr<double[]> i1data = in(1)->get_block(x1, a1, c2, a3);
        std::unique_ptr<double[]> i1data_sorted(new double[in(1)->get_size(x1, a1, c2, a3)]);
        sort_indices<3,2,1,0,0,1,-1,1>(i1data, i1data_sorted, x1.size(), a1.size(), c2.size(), a3.size());
        dgemm_("T", "N", x0.size(), x1.size(), a3.size()*c2.size()*a1.size(),
               1.0, i0data_sorted, a3.size()*c2.size()*a1.size(), i1data_sorted, a3.size()*c2.size()*a1.size(),
               1.0, odata_sorted, x0.size());
      }
    }
  }
  sort_indices<1,0,1,1,1,1>(odata_sorted, odata, x0.size(), x1.size());
  out()->add_block(odata, x1, x0);
}

void Task570::Task_local::compute() {
  const Index x1 = b(0);
  const Index x0 = b(1);
  // tensor label: I868
  std::unique_ptr<double[]> odata(new double[out()->get_size(x1, x0)]);
  std::fill_n(odata.get(), out()->get_size(x1, x0), 0.0);
  std::unique_ptr<double[]> odata_sorted(new double[out()->get_size(x1, x0)]);
  std::fill_n(odata_sorted.get(), out()->get_size(x1, x0), 0.0);
  for (auto& a1 : *range_[2]) {
    for (auto& c2 : *range_[0]) {
      for (auto& a3 : *range_[2]) {
        // tensor label: v2
        std::unique_ptr<double[]> i0data = in(0)->get_block(x0, a1, c2, a3);
        std::unique_ptr<double[]> i0data_sorted(new double[in(0)->get_size(x0, a1, c2, a3)]);
        sort_indices<1,2,3,0,0,1,1,1>(i0data, i0data_sorted, x0.size(), a1.size(), c2.size(), a3.size());
        // tensor label: l2
        std::unique_ptr<double[]> i1data = in(1)->get_block(x1, a1, c2, a3);
        std::unique_ptr<double[]> i1data_sorted(new double[in(1)->get_size(x1, a1, c2, a3)]);
        sort_indices<1,2,3,0,0,1,2,1>(i1data, i1data_sorted, x1.size(), a1.size(), c2.size(), a3.size());
        dgemm_("T", "N", x0.size(), x1.size(), a3.size()*c2.size()*a1.size(),
               1.0, i0data_sorted, a3.size()*c2.size()*a1.size(), i1data_sorted, a3.size()*c2.size()*a1.size(),
               1.0, odata_sorted, x0.size());
      }
    }
  }
  sort_indices<1,0,1,1,1,1>(odata_sorted, odata, x0.size(), x1.size());
  out()->add_block(odata, x1, x0);
}

void Task571::Task_local::compute() {
  const Index x1 = b(0);
  const Index x0 = b(1);
  // tensor label: I868
  std::unique_ptr<double[]> odata(new double[out()->get_size(x1, x0)]);
  std::fill_n(odata.get(), out()->get_size(x1, x0), 0.0);
  std::unique_ptr<double[]> odata_sorted(new double[out()->get_size(x1, x0)]);
  std::fill_n(odata_sorted.get(), out()->get_size(x1, x0), 0.0);
  for (auto& c2 : *range_[0]) {
    for (auto& a1 : *range_[2]) {
      // tensor label: h1
      std::unique_ptr<double[]> i0data = in(0)->get_block(c2, a1);
      std::unique_ptr<double[]> i0data_sorted(new double[in(0)->get_size(c2, a1)]);
      sort_indices<0,1,0,1,1,1>(i0data, i0data_sorted, c2.size(), a1.size());
      // tensor label: l2
      std::unique_ptr<double[]> i1data = in(1)->get_block(x1, a1, c2, x0);
      std::unique_ptr<double[]> i1data_sorted(new double[in(1)->get_size(x1, a1, c2, x0)]);
      sort_indices<2,1,0,3,0,1,-1,1>(i1data, i1data_sorted, x1.size(), a1.size(), c2.size(), x0.size());
      dgemm_("T", "N", 1, x0.size()*x1.size(), c2.size()*a1.size(),
             1.0, i0data_sorted, c2.size()*a1.size(), i1data_sorted, c2.size()*a1.size(),
             1.0, odata_sorted, 1);
    }
  }
  sort_indices<0,1,1,1,1,1>(odata_sorted, odata, x1.size(), x0.size());
  out()->add_block(odata, x1, x0);
}

void Task572::Task_local::compute() {
  const Index x1 = b(0);
  const Index x0 = b(1);
  // tensor label: I868
  std::unique_ptr<double[]> odata(new double[out()->get_size(x1, x0)]);
  std::fill_n(odata.get(), out()->get_size(x1, x0), 0.0);
  std::unique_ptr<double[]> odata_sorted(new double[out()->get_size(x1, x0)]);
  std::fill_n(odata_sorted.get(), out()->get_size(x1, x0), 0.0);
  for (auto& c1 : *range_[0]) {
    for (auto& a2 : *range_[2]) {
      // tensor label: h1
      std::unique_ptr<double[]> i0data = in(0)->get_block(c1, a2);
      std::unique_ptr<double[]> i0data_sorted(new double[in(0)->get_size(c1, a2)]);
      sort_indices<0,1,0,1,1,1>(i0data, i0data_sorted, c1.size(), a2.size());
      // tensor label: l2
      std::unique_ptr<double[]> i1data = in(1)->get_block(c1, a2, x1, x0);
      std::unique_ptr<double[]> i1data_sorted(new double[in(1)->get_size(c1, a2, x1, x0)]);
      sort_indices<0,1,2,3,0,1,2,1>(i1data, i1data_sorted, c1.size(), a2.size(), x1.size(), x0.size());
      dgemm_("T", "N", 1, x0.size()*x1.size(), a2.size()*c1.size(),
             1.0, i0data_sorted, a2.size()*c1.size(), i1data_sorted, a2.size()*c1.size(),
             1.0, odata_sorted, 1);
    }
  }
  sort_indices<0,1,1,1,1,1>(odata_sorted, odata, x1.size(), x0.size());
  out()->add_block(odata, x1, x0);
}

void Task573::Task_local::compute() {
  const Index x3 = b(0);
  const Index x0 = b(1);
  const Index x2 = b(2);
  const Index x1 = b(3);
  // tensor label: I874
  std::unique_ptr<double[]> i0data = in(0)->get_block(x2, x3, x0, x1);
  std::unique_ptr<double[]> i0data_sorted(new double[in(0)->get_size(x2, x3, x0, x1)]);
  sort_indices<1,2,0,3,0,1,1,1>(i0data, i0data_sorted, x2.size(), x3.size(), x0.size(), x1.size());
  // tensor label (calculated on-the-fly): Gamma170
  {
    std::unique_ptr<double[]> o2data(new double[out(2)->get_size(x3, x0, x2, x1)]);
    std::fill_n(o2data.get(), out(2)->get_size(x3, x0, x2, x1), 0.0);
    sort_indices<0,1,2,3,1,1,1,1>(i0data_sorted, o2data, x3.size(), x0.size(), x2.size(), x1.size());
    out(2)->add_block(o2data, x3, x0, x2, x1);
  }
}

void Task574::Task_local::compute() {
  const Index x2 = b(0);
  const Index x3 = b(1);
  const Index x0 = b(2);
  const Index x1 = b(3);
  // tensor label: I874
  std::unique_ptr<double[]> odata(new double[out()->get_size(x2, x3, x0, x1)]);
  std::fill_n(odata.get(), out()->get_size(x2, x3, x0, x1), 0.0);
  std::unique_ptr<double[]> odata_sorted(new double[out()->get_size(x2, x3, x0, x1)]);
  std::fill_n(odata_sorted.get(), out()->get_size(x2, x3, x0, x1), 0.0);
  for (auto& a1 : *range_[2]) {
    for (auto& a2 : *range_[2]) {
      // tensor label: v2
      std::unique_ptr<double[]> i0data = in(0)->get_block(x0, a1, x1, a2);
      std::unique_ptr<double[]> i0data_sorted(new double[in(0)->get_size(x0, a1, x1, a2)]);
      sort_indices<1,3,0,2,0,1,1,1>(i0data, i0data_sorted, x0.size(), a1.size(), x1.size(), a2.size());
      // tensor label: l2
      std::unique_ptr<double[]> i1data = in(1)->get_block(x3, a1, x2, a2);
      std::unique_ptr<double[]> i1data_sorted(new double[in(1)->get_size(x3, a1, x2, a2)]);
      sort_indices<1,3,0,2,0,1,1,1>(i1data, i1data_sorted, x3.size(), a1.size(), x2.size(), a2.size());
      dgemm_("T", "N", x0.size()*x1.size(), x2.size()*x3.size(), a2.size()*a1.size(),
             1.0, i0data_sorted, a2.size()*a1.size(), i1data_sorted, a2.size()*a1.size(),
             1.0, odata_sorted, x0.size()*x1.size());
    }
  }
  sort_indices<3,2,0,1,1,1,1,1>(odata_sorted, odata, x0.size(), x1.size(), x3.size(), x2.size());
  out()->add_block(odata, x2, x3, x0, x1);
}

void Task575::Task_local::compute() {
  const Index x2 = b(0);
  const Index x3 = b(1);
  const Index x0 = b(2);
  const Index x1 = b(3);
  // tensor label: I874
  std::unique_ptr<double[]> odata(new double[out()->get_size(x2, x3, x0, x1)]);
  std::fill_n(odata.get(), out()->get_size(x2, x3, x0, x1), 0.0);
  std::unique_ptr<double[]> odata_sorted(new double[out()->get_size(x2, x3, x0, x1)]);
  std::fill_n(odata_sorted.get(), out()->get_size(x2, x3, x0, x1), 0.0);
  for (auto& a1 : *range_[2]) {
    // tensor label: h1
    std::unique_ptr<double[]> i0data = in(0)->get_block(x0, a1);
    std::unique_ptr<double[]> i0data_sorted(new double[in(0)->get_size(x0, a1)]);
    sort_indices<1,0,0,1,1,1>(i0data, i0data_sorted, x0.size(), a1.size());
    // tensor label: l2
    std::unique_ptr<double[]> i1data = in(1)->get_block(x3, a1, x2, x1);
    std::unique_ptr<double[]> i1data_sorted(new double[in(1)->get_size(x3, a1, x2, x1)]);
    sort_indices<1,0,2,3,0,1,1,1>(i1data, i1data_sorted, x3.size(), a1.size(), x2.size(), x1.size());
    dgemm_("T", "N", x0.size(), x1.size()*x2.size()*x3.size(), a1.size(),
           1.0, i0data_sorted, a1.size(), i1data_sorted, a1.size(),
           1.0, odata_sorted, x0.size());
  }
  sort_indices<2,1,0,3,1,1,1,1>(odata_sorted, odata, x0.size(), x3.size(), x2.size(), x1.size());
  out()->add_block(odata, x2, x3, x0, x1);
}

#endif