Merge pull request #184 from deepmodeling/develop

Merge: v2.3.5

CMakeLists.txt

@@ -336,6 +336,7 @@ target_link_libraries(${ABACUS_BIN_NAME}
     driver
     xc_
     hsolver
+    genelpa
     elecstate
     hamilt
     psi

README.md

@@ -56,6 +56,7 @@ ABACUS provides the following features and functionalities:
 20. (subsidiary tool)Generator for second generation numerical orbital basis.
 21. Interface with DPGEN
 22. Interface with phonopy
+23. Implicit solvation model
 
 [back to top](#readme-top)
 
@@ -161,6 +162,7 @@ The following provides basic sample jobs in ABACUS. More can be found in the dir
 - [BSSE for molecular formation energy](docs/examples/BSSE.md)
 - [ABACUS-DPGEN interface](docs/examples/dpgen.md)
 - [ABACUS-phonopy interface](docs/examples/phonopy.md)
+- [Implicit solvation model](docs/examples/implicit-sol.md)
 
 [back to top](#readme-top)
 

docs/examples/implicit-sol.md

@@ -0,0 +1,63 @@
+# Implicit solvation model
+
+[back to main page](../../README.md)
+
+Solid-liquid interfaces are ubiquitous in nature and frequently encountered and employed in materials simulation. The solvation effect should be taken into account in accurate first-principles calculations of such systems.  
+Implicit solvation model is a well-developed method to deal with solvation effects, which has been widely used in finite and periodic systems. This approach treats the solvent as a continuous medium instead of individual “explicit” solvent molecules, which means that the solute embedded in an implicit solvent and the average over the solvent degrees of freedom becomes implicit in the properties of the solvent bath.
+
+## Input
+```
+INPUT_PARAMETERS
+imp_sol                 1
+eb_k                    80
+tau                     0.000010798
+sigma_k                 0.6
+nc_k                    0.00037
+```
+- imp_sol  
+
+    If set to 1, an implicit solvation correction is considered. 0：vacuum calculation(default).
+- eb_k  
+
+    The relative permittivity of the bulk solvent, 80 for water. Used only if `imp_sol` == true.
+- tau 
+
+    The effective surface tension parameter, which describes the cavitation, the dispersion, and the repulsion interaction between the solute and the solvent that are not captured by the electrostatic terms.
+    We use the values of `tau`, `sigma_k`, `nc_k` that were obtained by a fit of the model to experimental solvation energies for molecules in water. tau = 0.525 $meV/Å^{2}$ = 1.0798e-05 $Ry/Bohr^{2}$.
+- sigma_k 
+
+    We assume a diffuse cavity that is implicitly determined by the electronic structure of the solute. 
+    `sigma_k` is the parameter that describes the width of the diffuse cavity. The specific value is sigma_k = 0.6.
+- nc_k
+
+    `nc_k` determines at what value of the electron density the dielectric cavity forms. 
+    The specific value is nc_k = 0.0025 $Å^{-3}$ = 0.00037 $Bohr^{-3}$.
+
+## Output
+In this example, we calculate the implicit solvation correction for H2O.
+The results of the energy calculation are written in the “running_nscf.log” in the OUT folder.
+```
+       Energy                       Rydberg                            eV
+   E_KohnSham                -34.3200995971                -466.948910448
+     E_Harris                -34.2973698556                -466.639656449
+       E_band                -7.66026117767                -104.223200184
+   E_one_elec                -56.9853883251                -775.325983964
+    E_Hartree                +30.0541108968                +408.907156521
+         E_xc                -8.32727420734                -113.298378028
+      E_Ewald               +0.961180728747                +13.0775347188
+      E_demet                            +0                            +0
+      E_descf                            +0                            +0
+     E_efield                            +0                            +0
+        E_exx                            +0                            +0
+     E_sol_el              -0.0250553663339               -0.340895747619
+    E_sol_cav             +0.00232667606131               +0.031656051834
+      E_Fermi               -0.499934383866                 -6.8019562467
+
+```
+- E_sol_el: Electrostatic contribution to the solvation energy.
+- E_sol_cav: Cavitation and dispersion contributions to the solvation energy.
+Both `E_sol_el` and `E_sol_cav` corrections are included in `E_KohnSham`. 
+
+
+
+[back to top](#implicit-solvation-model)

docs/examples/phonopy.md

@@ -3,7 +3,7 @@
 [back to main page](../../README.md)
 
 
-[Phonopy](https://github.com/phonopy/phonopy) is a powerful package to calculate phonon and related properties. It has provided interface with ABACUS. In the following, we take the FCC aluminum as an example:
+[Phonopy](https://github.com/phonopy/phonopy) (Note: please use the `develop` branch, rather than the `master` branch until the abacus interface has been merged into phonopy's `master` branch.) is a powerful package to calculate phonon and related properties. It has provided interface with ABACUS. In the following, we take the FCC aluminum as an example:
 
 
 1. Prepare a 'setting.conf' with following tags:
@@ -39,4 +39,4 @@ PRIMITIVE_AXES = 0 1/2 1/2  1/2 0 1/2  1/2 1/2 0
 BAND= 1 1 1  1/2 1/2 1  3/8 3/8 3/4  0 0 0   1/2 1/2 1/2
 BAND_POINTS = 21
 BAND_CONNECTION = .TRUE.
-```
+```

docs/features.md

@@ -50,9 +50,14 @@ ATOMIC_SPECIES
 Si 28.00 Si_ONCV_PBE-1.0.upf
 ```
 
-The user can download the pseudopotential files from our [website](http://abacus.ustc.edu.cn/pseudo.html).
+You can download the pseudopotential files from our [website](http://abacus.ustc.edu.cn/pseudo/list.htm).
 
-For more information of different types of pseudopotentials, please visit the Quantum espresso [website](http://www.quantum-espresso.org/pseudopotentials/).
+There are pseudopotential files in these websites which are also supported by ABACUS:
+1. [Quantum ESPRESSO](http://www.quantum-espresso.org/pseudopotentials/).
+2. [SG15-ONCV](http://quantum-simulation.org/potentials/sg15_oncv/upf/).
+3. [DOJO](http://www.pseudo-dojo.org/).
+
+If LCAO base is used, the numerical orbital files should match the pseudopotential files. The [official orbitals package](http://abacus.ustc.edu.cn/pseudo/list.htm) only matches SG15-ONCV pseudopotentials.
 
 [back to top](#features)
 

docs/input-main.md

@@ -82,6 +82,10 @@
 
     [cal_cond](#cal_cond) | [cond_nche](#cond_nche) | [cond_dw](#cond_dw) | [cond_wcut](#cond_wcut) | [cond_wenlarge](#cond_wenlarge) | [cond_fwhm ](#cond_fwhm )
 
+- [Implicit solvation model](#implicit-solvation-model)
+
+    [imp_sol](#imp_sol) | [eb_k](#eb_k) | [tau](#tau) | [sigma_k](#sigma_k) | [nc_k](#nc_k) 
+
 [back to main page](../README.md)
 
 ## Structure of the file
@@ -1662,3 +1666,39 @@ Thermal conductivities: $\kappa = \lim_{\omega\to 0}\kappa(\omega)$
 - **Type**: Integer
 - **Description**: We use gaussian functions to approxiamte $\delta(E)\approx \frac{1}{\sqrt{2\pi}\Delta E}e^{-\frac{E^2}{2{\Delta E}^2}}$. FWHM for conductivities, $FWHM=2*\sqrt{2\ln2}\cdot \Delta E$. The unit is eV.
 - **Default**: 0.3
+
+### Implicit solvation model
+
+This part of variables are used to control the usage of implicit solvation model. This approach treats the solvent as a continuous medium instead of individual “explicit” solvent molecules, which means that the solute embedded in an implicit solvent and the average over the solvent degrees of freedom becomes implicit in the properties of the solvent bath.
+
+#### imp_sol
+
+- **Type**: Boolean
+- **Description**: If set to 1, an implicit solvation correction is considered.
+- **Default**: 0
+
+#### eb_k
+
+- **Type**: Real
+- **Description**: The relative permittivity of the bulk solvent, 80 for water. Used only if `imp_sol` == true.
+- **Default**: 80
+
+#### tau
+
+- **Type**: Real
+- **Description**: The effective surface tension parameter, which describes the cavitation, the dispersion, and the repulsion interaction between the solute and the solvent that are not captured by the electrostatic terms. The unit is $Ry/Bohr^{2}$.
+- **Default**: 1.0798e-05
+
+#### sigma_k
+
+- **Type**: Real
+- **Description**: We assume a diffuse cavity that is implicitly determined by the electronic structure of the solute.
+`sigma_k` is the parameter that describes the width of the diffuse cavity.
+- **Default**: 0.6
+
+#### nc_k
+
+- **Type**: Real
+- **Description**: It determines at what value of the electron density the dielectric cavity forms. 
+The unit is $Bohr^{-3}$.
+- **Default**: 0.00037

modules/FindELPA.cmake

@@ -38,17 +38,4 @@ if(ELPA_FOUND)
 endif()
 
 set(CMAKE_REQUIRED_INCLUDES ${CMAKE_REQUIRED_INCLUDES} ${ELPA_INCLUDE_DIR})
-include(CheckCXXSourceCompiles)
-check_cxx_source_compiles("
-#include <elpa/elpa_version.h>
-#if ELPA_API_VERSION < 20210430
-#error ELPA version is too old.
-#endif
-int main(){}
-"
-ELPA_VERSION_SATISFIES
-)
-if(NOT ELPA_VERSION_SATISFIES)
-    message(FATAL_ERROR "ELPA version is too old. We support version 2017 or higher.")
-endif()
 mark_as_advanced(ELPA_INCLUDE_DIR ELPA_LIBRARY)

source/Makefile

@@ -20,6 +20,7 @@ VPATH=./src_global\
 :./module_xc\
 :./module_esolver\
 :./module_hsolver\
+:./module_hsolver/genelpa\
 :./module_elecstate\
 :./module_psi\
 :./module_hamilt\

source/Makefile.Objects

@@ -277,6 +277,11 @@ hsolver_lcao.o\
 hsolver_pw.o\
 hsolver_pw_sdft.o
 
+OBJ_GENELPA=elpa_new_complex.o\
+elpa_new_real.o\
+elpa_new.o\
+utils.o
+
 OBJ_ELECSTATES=elecstate.o\
 dm2d_to_grid.o\
 elecstate_lcao.o\
@@ -304,6 +309,7 @@ $(OBJ_HSOLVER)\
 $(OBJ_ELECSTATES)\
 $(OBJ_PSI)\
 ${OBJ_OPERATOR}\
+${OBJ_GENELPA}\
 charge.o \
 charge_mixing.o \
 charge_pulay.o \

source/input_conv.cpp

@@ -373,6 +373,7 @@ void Input_Conv::Convert(void)
 
     if (GlobalC::exx_global.info.hybrid_type != Exx_Global::Hybrid_Type::No)
     {
+        //EXX case, convert all EXX related variables 
         GlobalC::exx_global.info.hybrid_alpha = INPUT.exx_hybrid_alpha;
         XC_Functional::get_hybrid_alpha(INPUT.exx_hybrid_alpha);
         GlobalC::exx_global.info.hse_omega = INPUT.exx_hse_omega;
@@ -406,6 +407,9 @@ void Input_Conv::Convert(void)
         Exx_Abfs::Jle::Lmax = INPUT.exx_opt_orb_lmax;
         Exx_Abfs::Jle::Ecut_exx = INPUT.exx_opt_orb_ecut;
         Exx_Abfs::Jle::tolerence = INPUT.exx_opt_orb_tolerence;
+
+        //EXX does not support any symmetry analyse, force symmetry setting to -1
+        ModuleSymmetry::Symmetry::symm_flag = -1;
     }
 #endif
 #endif

source/module_deepks/test/CMakeLists.txt

@@ -8,7 +8,7 @@ target_link_libraries(
     test_deepks
     base cell symmetry md surchem xc_
     neighbor orb io relax gint lcao parallel mrrr pdiag pw ri driver esolver hsolver psi elecstate hamilt planewave
-    pthread
+    pthread genelpa
     deepks
     ${ABACUS_LINK_LIBRARIES}
 )

source/module_elecstate/elecstate_lcao.cpp

@@ -137,7 +137,9 @@ void ElecStateLCAO::print_psi(const psi::Psi<double>& psi_in)
 
     // output but not do "2d-to-grid" conversion
     double** wfc_grid = nullptr;
+#ifdef __MPI
     this->lowf->wfc_2d_to_grid(ElecStateLCAO::out_wfc_lcao, psi_in.get_pointer(), wfc_grid, this->ekb, this->wg);
+#endif
     return;
 }
 void ElecStateLCAO::print_psi(const psi::Psi<std::complex<double>>& psi_in)
@@ -159,7 +161,7 @@ void ElecStateLCAO::print_psi(const psi::Psi<std::complex<double>>& psi_in)
     {
         for (int iw = 0; iw < GlobalV::NLOCAL; iw++)
         {
-            this->lowf->wfc_k_grid[ik][ib][iw] = psi(ib,iw);
+            this->lowf->wfc_k_grid[ik][ib][iw] = psi_in(ib, iw);
         }
     }
 #endif