PositionCellProcessorRMM.h

/*
 * PositionCellProcessorRMM.h
 *
 *  Created on: 5 Oct 2017
 *      Author: tchipevn
 */
 
#ifndef SRC_INTEGRATORS_POSITIONCELLPROCESSORRMM_H_
#define SRC_INTEGRATORS_POSITIONCELLPROCESSORRMM_H_
 
#include "particleContainer/ParticleCell.h"
#include "particleContainer/adapter/CellProcessor.h"
 
// explicit Leapfrog position update rule in a vectorized fashion for the RMM
 
class PositionCellProcessorRMM : public CellProcessor {
public:
    PositionCellProcessorRMM(double timeStep) : CellProcessor(0.0, 0.0), _timeStep(static_cast<vcp_real_calc>(timeStep)) {}
    void initTraversal() {}
 
    void preprocessCell(ParticleCell& cell) {}
 
    void processCellPair(ParticleCell& cell1, ParticleCell& cell2, bool sumAll = false) {} // does this need bool?
 
    double processSingleMolecule(Molecule* m1, ParticleCell& cell2) { return 0.0; }
 
    void postprocessCell(ParticleCell& cell) {}
 
    void endTraversal() {}
 
    void processCell(ParticleCell& cell) {
        ParticleCellRMM & c = downcastCellReferenceRMM(cell);
        CellDataSoARMM & soa = c.getCellDataSoA();
        const size_t end_i = vcp_floor_to_vec_size(soa.getMolNum());
 
               vcp_real_calc * const soa_r_x = soa.r_xBegin();
               vcp_real_calc * const soa_r_y = soa.r_yBegin();
               vcp_real_calc * const soa_r_z = soa.r_zBegin();
        const vcp_real_accum * const soa_v_x = soa.v_xBegin();
        const vcp_real_accum * const soa_v_y = soa.v_yBegin();
        const vcp_real_accum * const soa_v_z = soa.v_zBegin();
 
        const RealCalcVec dt = RealCalcVec::set1(_timeStep);
 
        size_t i = 0;
        for (; i < end_i; i += VCP_VEC_SIZE) {
            RealCalcVec r_x = RealCalcVec::aligned_load(soa_r_x + i);
            RealCalcVec r_y = RealCalcVec::aligned_load(soa_r_y + i);
            RealCalcVec r_z = RealCalcVec::aligned_load(soa_r_z + i);
 
            const RealAccumVec v_x = RealAccumVec::aligned_load(soa_v_x + i);
            const RealAccumVec v_y = RealAccumVec::aligned_load(soa_v_y + i);
            const RealAccumVec v_z = RealAccumVec::aligned_load(soa_v_z + i);
 
            const RealCalcVec c_v_x = RealAccumVec::convertAccumToCalc(v_x);
            const RealCalcVec c_v_y = RealAccumVec::convertAccumToCalc(v_y);
            const RealCalcVec c_v_z = RealAccumVec::convertAccumToCalc(v_z);
 
            r_x = RealCalcVec::fmadd(dt, c_v_x, r_x);
            r_y = RealCalcVec::fmadd(dt, c_v_y, r_y);
            r_z = RealCalcVec::fmadd(dt, c_v_z, r_z);
 
            r_x.aligned_store(soa_r_x + i);
            r_y.aligned_store(soa_r_y + i);
            r_z.aligned_store(soa_r_z + i);
 
        }
        const MaskCalcVec remainderMask = vcp_simd_getRemainderMask(soa.getMolNum());
        if (remainderMask.movemask()) {
            RealCalcVec r_x = RealCalcVec::aligned_load_mask(soa_r_x + i, remainderMask);
            RealCalcVec r_y = RealCalcVec::aligned_load_mask(soa_r_y + i, remainderMask);
            RealCalcVec r_z = RealCalcVec::aligned_load_mask(soa_r_z + i, remainderMask);
 
            const RealAccumVec v_x = RealAccumVec::aligned_load_mask(soa_v_x + i, remainderMask);
            const RealAccumVec v_y = RealAccumVec::aligned_load_mask(soa_v_y + i, remainderMask);
            const RealAccumVec v_z = RealAccumVec::aligned_load_mask(soa_v_z + i, remainderMask);
 
            const RealCalcVec c_v_x = RealAccumVec::convertAccumToCalc(v_x);
            const RealCalcVec c_v_y = RealAccumVec::convertAccumToCalc(v_y);
            const RealCalcVec c_v_z = RealAccumVec::convertAccumToCalc(v_z);
 
            r_x = RealCalcVec::fmadd(dt, c_v_x, r_x);
            r_y = RealCalcVec::fmadd(dt, c_v_y, r_y);
            r_z = RealCalcVec::fmadd(dt, c_v_z, r_z);
 
            // TODO: handle aligned masked store properly in intrinsics.
            // For now just store, not caring about mask. Due to internal padding, this will work without problems and will not overwrite stuff.
            r_x.aligned_store(soa_r_x + i);
            r_y.aligned_store(soa_r_y + i);
            r_z.aligned_store(soa_r_z + i);
        }
    }
 
private:
    vcp_real_calc _timeStep;
};
 
#endif /* SRC_INTEGRATORS_POSITIONCELLPROCESSORRMM_H_ */