/* This file is part of the OpenLB library * * Copyright (C) 2006, 2007 Jonas Latt * E-mail contact: info@openlb.net * The most recent release of OpenLB can be downloaded at * * * 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 2 * 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, write to the Free * Software Foundation, Inc., 51 Franklin Street, Fifth Floor, * Boston, MA 02110-1301, USA. */ /** \file * Template specializations for some computationally intensive LB * functions of the header file lbHelpers.h, for the D2Q9 grid. */ #ifndef LB_HELPERS_D2Q9_H #define LB_HELPERS_D2Q9_H namespace olb { // Efficient specialization for D2Q9 base lattice template struct lbDynamicsHelpers > { using SpecializedCellBase = CellBase>; using SpecializedDescriptor = descriptors::D2Q9; static T equilibrium(int iPop, T rho, const T u[2], T uSqr) { typedef descriptors::D2Q9<> L; T c_u = descriptors::c(iPop,0)*u[0] + descriptors::c(iPop,1)*u[1]; return rho * descriptors::t(iPop) * ( 1. + 3.*c_u + 4.5*c_u*c_u - 1.5*uSqr ) - descriptors::t(iPop); } static T equilibriumFirstOrder(int iPop, T rho, const T u[2]) { typedef descriptors::D2Q9<> L; T c_u = descriptors::c(iPop,0) * u[0] + descriptors::c(iPop,1) * u[1]; return rho * descriptors::t(iPop) * ( ( T )1 + c_u * descriptors::invCs2() ) - descriptors::t(iPop); } static T incEquilibrium(int iPop, const T j[2], const T jSqr, const T pressure) { typedef descriptors::D2Q9<> L; T c_j = descriptors::c(iPop,0)*j[0] + descriptors::c(iPop,1)*j[1]; return descriptors::t(iPop) * ( 3.*pressure + 3.*c_j + 4.5*c_j*c_j - 1.5*jSqr ) - descriptors::t(iPop); } static void computeFneq(SpecializedCellBase const& cell, T fNeq[9], T rho, const T u[2]) { const T uSqr = u[0]*u[0] + u[1]*u[1]; for (int iPop=0; iPop < 9; ++iPop) { fNeq[iPop] = cell[iPop] - equilibrium(iPop, rho, u, uSqr); } } static T bgkCollision(SpecializedCellBase& cell, T const& rho, const T u[2], T const& omega) { T uxSqr = u[0]*u[0]; T uySqr = u[1]*u[1]; T ux_ = (T)3 * u[0]; T uy_ = (T)3 * u[1]; T uxSqr_ = (T)3 * uxSqr; T uySqr_ = (T)3 * uySqr; T uxSqr__ = (T)3/(T)2 * uxSqr; T uySqr__ = (T)3/(T)2 * uySqr; T uSqr_ = uxSqr__ + uySqr__; T uxPySqr_ = (T)9/(T)2 * (u[0]+u[1])*(u[0]+u[1]); T uxMySqr_ = (T)9/(T)2 * (u[0]-u[1])*(u[0]-u[1]); T rho_ = (T)4/(T)9 * rho; T cf_ = (T)4/(T)9 * (rho-(T)1); cell[0] *= (T)1-omega; cell[0] += omega*(cf_ + rho_*(- uxSqr__ - uySqr__)); rho_ = (T)1/(T)9 * rho; cf_ = (T)1/(T)9 * (rho-(T)1); cell[6] *= (T)1-omega; cell[6] += omega*(cf_ + rho_*(ux_ + uxSqr_ - uySqr__)); cell[8] *= (T)1-omega; cell[8] += omega*(cf_ + rho_*(uy_ + uySqr_ - uxSqr__)); cell[2] *= (T)1-omega; cell[2] += omega*(cf_ + rho_*(-ux_ + uxSqr_ - uySqr__)); cell[4] *= (T)1-omega; cell[4] += omega*(cf_ + rho_*(-uy_ + uySqr_ - uxSqr__)); rho_ = (T)1/(T)36 * rho; cf_ = (T)1/(T)36 * (rho-(T)1); cell[7] *= (T)1-omega; cell[7] += omega*(cf_ + rho_*(ux_ + uy_ + uxPySqr_ - uSqr_)); cell[1] *= (T)1-omega; cell[1] += omega*(cf_ + rho_*(-ux_ + uy_ + uxMySqr_ - uSqr_)); cell[3] *= (T)1-omega; cell[3] += omega*(cf_ + rho_*(-ux_ - uy_ + uxPySqr_ - uSqr_)); cell[5] *= (T)1-omega; cell[5] += omega*(cf_ + rho_*(ux_ - uy_ + uxMySqr_ - uSqr_)); return uxSqr + uySqr; } static T incBgkCollision(SpecializedCellBase& cell, T pressure, const T j[2], T omega) { const T jSqr = util::normSqr::d>(j); for (int iPop=0; iPop < descriptors::D2Q9<>::q; ++iPop) { cell[iPop] *= (T)1-omega; cell[iPop] += omega * lbHelpers::incEquilibrium ( iPop, j, jSqr, pressure ); } return jSqr; } static T constRhoBgkCollision(SpecializedCellBase& cell, T rho, const T u[2], T ratioRho, T omega) { const T uSqr = util::normSqr::d>(u); for (int iPop=0; iPop < descriptors::D2Q9<>::q; ++iPop) { T feq = lbHelpers::equilibrium(iPop, rho, u, uSqr ); cell[iPop] = ratioRho*(feq+descriptors::t(iPop)) -descriptors::t(iPop) + ((T)1-omega)*(cell[iPop]-feq); } return uSqr; } static void partial_rho(SpecializedCellBase const& cell, T& lineX_P1, T& lineX_0, T& lineX_M1, T& lineY_P1, T& lineY_M1) { lineX_P1 = cell[5] + cell[6] + cell[7]; lineX_0 = cell[0] + cell[4] + cell[8]; lineX_M1 = cell[1] + cell[2] + cell[3]; lineY_P1 = cell[7] + cell[8] + cell[1]; lineY_M1 = cell[3] + cell[4] + cell[5]; } static T computeRho(SpecializedCellBase const& cell) { T rho = cell[0] + cell[1] + cell[2] + cell[3] + cell[4] + cell[5] + cell[6] + cell[7] + cell[8] + (T)1; return rho; } static void computeRhoU(SpecializedCellBase const& cell, T& rho, T u[2]) { T lineX_P1, lineX_0, lineX_M1, lineY_P1, lineY_M1; partial_rho(cell, lineX_P1, lineX_0, lineX_M1, lineY_P1, lineY_M1); rho = lineX_P1 + lineX_0 + lineX_M1 + (T)1; T invRho= 1./rho; u[0] = (lineX_P1 - lineX_M1)*invRho; u[1] = (lineY_P1 - lineY_M1)*invRho; } static void computeRhoJ(SpecializedCellBase const& cell, T& rho, T j[2]) { T lineX_P1, lineX_0, lineX_M1, lineY_P1, lineY_M1; partial_rho(cell, lineX_P1, lineX_0, lineX_M1, lineY_P1, lineY_M1); rho = lineX_P1 + lineX_0 + lineX_M1 + (T)1; j[0] = (lineX_P1 - lineX_M1); j[1] = (lineY_P1 - lineY_M1); } static void computeJ(SpecializedCellBase const& cell, T j[2] ) { T lineX_P1, lineX_M1, lineY_P1, lineY_M1; lineX_P1 = cell[5] + cell[6] + cell[7]; lineX_M1 = cell[1] + cell[2] + cell[3]; lineY_P1 = cell[7] + cell[8] + cell[1]; lineY_M1 = cell[3] + cell[4] + cell[5]; j[0] = (lineX_P1 - lineX_M1); j[1] = (lineY_P1 - lineY_M1); } static void computeStress(SpecializedCellBase const& cell, T rho, const T u[2], T pi[3]) { typedef descriptors::D2Q9<> L; // Workaround for Intel(r) compiler 9.1; // "using namespace util::tensorIndices2D" is not sufficient using util::tensorIndices2D::xx; using util::tensorIndices2D::yy; using util::tensorIndices2D::xy; T lineX_P1, lineX_0, lineX_M1, lineY_P1, lineY_M1; partial_rho(cell, lineX_P1, lineX_0, lineX_M1, lineY_P1, lineY_M1); pi[xx] = lineX_P1+lineX_M1 - 1./descriptors::invCs2()*(rho-(T)1) - rho*u[0]*u[0]; pi[yy] = lineY_P1+lineY_M1 - 1./descriptors::invCs2()*(rho-(T)1) - rho*u[1]*u[1]; pi[xy] = -cell[1] + cell[3] - cell[5] + cell[7] - rho*u[0]*u[1]; } static void computeAllMomenta(SpecializedCellBase const& cell, T& rho, T u[2], T pi[3] ) { typedef descriptors::D2Q9<> L; // Workaround for Intel(r) compiler 9.1; // "using namespace util::tensorIndices2D" is not sufficient using util::tensorIndices2D::xx; using util::tensorIndices2D::yy; using util::tensorIndices2D::xy; T lineX_P1, lineX_0, lineX_M1, lineY_P1, lineY_M1; partial_rho(cell, lineX_P1, lineX_0, lineX_M1, lineY_P1, lineY_M1); rho = lineX_P1 + lineX_0 + lineX_M1 + (T)1; T rhoU0 = (lineX_P1 - lineX_M1); T rhoU1 = (lineY_P1 - lineY_M1); u[0] = rhoU0/rho; u[1] = rhoU1/rho; pi[xx] = lineX_P1 + lineX_M1 - 1./descriptors::invCs2()*(rho-(T)1) - rhoU0*u[0]; pi[yy] = lineY_P1 + lineY_M1 - 1./descriptors::invCs2()*(rho-(T)1) - rhoU1*u[1]; pi[xy] = -cell[1] + cell[3] - cell[5] + cell[7] - rhoU0*u[1]; } static void modifyVelocity(SpecializedCellBase const& cell, const T newU[2]) { T rho, oldU[2]; computeRhoU(cell, rho, oldU); const T oldUSqr = util::normSqr(oldU); const T newUSqr = util::normSqr(newU); for (int iPop=0; iPop<9; ++iPop) { cell[iPop] = cell[iPop] - equilibrium(iPop, rho, oldU, oldUSqr) + equilibrium(iPop, rho, newU, newUSqr); } } }; //struct lbHelpers // Efficient specialization for D2Q9 lattice with force template struct lbExternalHelpers> { static void addExternalForce( Cell>& cell, const T u[descriptors::D2Q9::d], T omega, T amplitude) { const T* force = cell.template getFieldPointer(); const T mu = amplitude*((T)1-omega/(T)2); cell[0] += mu *(T)4/(T)3 *( force[0] * (- u[0] ) + force[1] * ( - u[1] ) ); cell[1] += mu *(T)1/(T)12 *( force[0] * ( 2*u[0] - 3*u[1] - 1) + force[1] * (-3*u[0] + 2*u[1] + 1) ); cell[2] += mu *(T)1/(T)3 *( force[0] * ( 2*u[0] - 1) + force[1] * ( - u[1] ) ); cell[3] += mu *(T)1/(T)12 *( force[0] * ( 2*u[0] + 3*u[1] - 1) + force[1] * ( 3*u[0] + 2*u[1] - 1) ); cell[4] += mu *(T)1/(T)3 *( force[0] * (- u[0] ) + force[1] * ( + 2*u[1] - 1) ); cell[5] += mu *(T)1/(T)12 *( force[0] * ( 2*u[0] - 3*u[1] + 1) + force[1] * (-3*u[0] + 2*u[1] - 1) ); cell[6] += mu *(T)1/(T)3 *( force[0] * ( 2*u[0] + 1) + force[1] * ( - u[1] ) ); cell[7] += mu *(T)1/(T)12 *( force[0] * ( 2*u[0] + 3*u[1] + 1) + force[1] * ( 3*u[0] + 2*u[1] + 1) ); cell[8] += mu *(T)1/(T)3 *( force[0] * (- u[0] ) + force[1] * ( + 2*u[1] + 1) ); } }; // Efficient specialization for D2Q9 lattice and for forced D2Q9 lattice // (operations applying to the whole lattice) template struct lbLatticeHelpers> { static void swapAndStreamCell ( Cell> **grid, int iX, int iY, int nX, int nY, int iPop, T& fTmp ) { fTmp = grid[iX][iY][iPop]; grid[iX][iY][iPop] = grid[iX][iY][iPop+4]; grid[iX][iY][iPop+4] = grid[nX][nY][iPop]; grid[nX][nY][iPop] = fTmp; } static void swapAndStream2D ( Cell> **grid, int iX, int iY ) { T fTmp; swapAndStreamCell(grid, iX, iY, iX-1, iY+1, 1, fTmp); swapAndStreamCell(grid, iX, iY, iX-1, iY, 2, fTmp); swapAndStreamCell(grid, iX, iY, iX-1, iY-1, 3, fTmp); swapAndStreamCell(grid, iX, iY, iX, iY-1, 4, fTmp); } }; template struct lbLatticeHelpers> { static void swapAndStreamCell ( Cell> **grid, int iX, int iY, int nX, int nY, int iPop, T& fTmp ) { fTmp = grid[iX][iY][iPop]; grid[iX][iY][iPop] = grid[iX][iY][iPop+4]; grid[iX][iY][iPop+4] = grid[nX][nY][iPop]; grid[nX][nY][iPop] = fTmp; } static void swapAndStream2D ( Cell> **grid, int iX, int iY ) { T fTmp; swapAndStreamCell(grid, iX, iY, iX-1, iY+1, 1, fTmp); swapAndStreamCell(grid, iX, iY, iX-1, iY, 2, fTmp); swapAndStreamCell(grid, iX, iY, iX-1, iY-1, 3, fTmp); swapAndStreamCell(grid, iX, iY, iX, iY-1, 4, fTmp); } }; } // namespace olb #endif