Refactor: the force and stress calculations related to two-center integrals

WHUweiqingzhou commented 5 months ago

Background

In the current code, the part related to the density matrix has already been substituted with the reconstructed HContainer module from version 3.4.0. However, other physical quantities are still utilizing the old data structures. This leads to:

The new and old data structures coexist within the Force/Stress module.
The old data structures related to force computation in LCAO_Matrix cannot be eliminated, taking up a significant amount of memory.

Intended file deletion:

/module_hamilt_lcao/hamilt_lcaodft/FORCE_gamma_edm.cpp  
/module_hamilt_lcao/hamilt_lcaodft/FORCE_gamma_tvnl.cpp   
/module_hamilt_lcao/hamilt_lcaodft/FORCE_gamma_vl.cpp      
/module_hamilt_lcao/hamilt_lcaodft/foverlap_k.cpp     
/module_hamilt_lcao/hamilt_lcaodft/ftvnl_dphi_k.cpp   
/module_hamilt_lcao/hamilt_lcaodft/fvl_dphi_k.cpp      
/module_hamilt_lcao/hamilt_lcaodft/fvnl_dbeta_k.cpp

Intended data deletion:

//========================================
    // FORCE
    //========================================

    //-----------------------------------------
    // force in LCAO
    // used in gamma only algorithm.
    //-----------------------------------------
    double* DSloc_x;
    double* DSloc_y;
    double* DSloc_z;

    //-----------------------------------------
    // force in LCAO
    // used in k-points algorithm.
    //-----------------------------------------
    double* DSloc_Rx;
    double* DSloc_Ry;
    double* DSloc_Rz;

    //-----------------------------------------
    // dT + part of dVNL
    // used in gamma only algorithm.
    //-----------------------------------------
    double* DHloc_fixed_x;
    double* DHloc_fixed_y;
    double* DHloc_fixed_z;

    //-----------------------------------------
    // dT + part of dVNL
    // used in kpoint algorithm.
    //-----------------------------------------
    double* DHloc_fixedR_x;
    double* DHloc_fixedR_y;
    double* DHloc_fixedR_z;

    //----------------------------------------
    // r_mu - r_nu
    //----------------------------------------

    double* DH_r;//zhengdy added 2017-07

    double* stvnl11;
    double* stvnl12;
    double* stvnl13;
    double* stvnl22;
    double* stvnl23;
    double* stvnl33;

    double* DSloc_11;
    double* DSloc_12;
    double* DSloc_13;
    double* DSloc_22;
    double* DSloc_23;
    double* DSloc_33;

    double* DHloc_fixed_11;
    double* DHloc_fixed_12;
    double* DHloc_fixed_13;
    double* DHloc_fixed_22;
    double* DHloc_fixed_23;
    double* DHloc_fixed_33;

Describe the solution you'd like

Add a cal_force_stress function in each operator, for example: hamilt::EkineticNew<hamilt::OperatorLCAO<TK, TR>>::cal_force_stress For specifics, refer to DFTU<OperatorLCAO<TK, TR>>::cal_force_stress
Remove cal_foverlap, cal_ftvnl_dphi, cal_fvnl_dbeta from ftable_gamma and ftable_k, and directly call cal_force_stress in Force_Stress_LCAO<T>::getForceStress() to calculate the forces and stresses corresponding to each operator
After refactoring, ftable_gamma and ftable_k should only contain the contributions from the local potential grid integration.

Task list only for developers

[ ] Notice possible changes of behavior
[ ] Explain the changes of codes in core modules of ESolver, HSolver, ElecState, Hamilt, Operator or Psi

Notice Possible Changes of Behavior (Reminder only for developers)

No response

Notice any changes of core modules (Reminder only for developers)

No response

Notice Possible Changes of Core Modules (Reminder only for developers)

No response

Additional Context

No response

Task list for Issue attackers (only for developers)

[ ] Review and understand the proposed feature and its importance.
[ ] Research on the existing solutions and relevant research articles/resources.
[ ] Discuss with the team to evaluate the feasibility of implementing the feature.
[ ] Create a design document outlining the proposed solution and implementation details.
[ ] Get feedback from the team on the design document.
[ ] Develop the feature following the agreed design.
[ ] Write unit tests and integration tests for the feature.
[ ] Update the documentation to include the new feature.
[ ] Perform code review and address any issues.
[ ] Merge the feature into the main branch.
[ ] Monitor for any issues or bugs reported by users after the feature is released.
[ ] Address any issues or bugs reported by users and continuously improve the feature.

maki49 commented 5 months ago

Good idea, maybe the local potential grid integration can also be moved into operator Veff and finally ftable can be deleted?

Another suggestion is that, the two common interfaces can be designed like cal_force_stress(const DensityMatrix<T>& dm) and cal_force_stress(const double** rho) declared in the base class OperatorLCAO to be overridden , while other dependences are passed into the constructor of each operator. The inspiration comes from my need to call each term twice with everything the same except for the density (matrix) in #2460.

mohanchen commented 5 months ago

How to leave extra freedom for GPU implementation of force should be considerd.

deepmodeling / abacus-develop