Merge branch 'develop' into rdmft_PR

Author: mohanchen

docs/advanced/input_files/input-main.md

@@ -3960,6 +3960,13 @@ Currently supported: `RPA`, `LDA`, `PBE`, `HSE`, `HF`.
 - **Description**:  The number of 2-particle states to be solved
 - **Default**: 0
 
+### lr_unrestricted
+- **Type**: Boolean
+- **Description**: Whether to use unrestricted construction for LR-TDDFT (the matrix size will be doubled).
+  - True:  Always use unrestricted LR-TDDFT. 
+  - False: Use unrestricted LR-TDDFT only when the system is open-shell.
+- **Default**: False
+
 ### abs_wavelen_range
 
 - **Type**: Real Real

source/Makefile.Objects

@@ -291,6 +291,7 @@ OBJS_GINT=gint.o\
       init_orb.o\
 
 OBJS_HAMILT=hamilt_pw.o\
+    hamilt_sdft_pw.o\
     operator.o\
     operator_pw.o\
     ekinetic_pw.o\
@@ -648,7 +649,6 @@ OBJS_SRCPW=H_Ewald_pw.o\
     structure_factor_k.o\
     soc.o\
     sto_iter.o\
-    sto_hchi.o\
     sto_che.o\
     sto_wf.o\
     sto_func.o\
@@ -728,7 +728,6 @@ OBJS_TENSOR=tensor.o\
     operator_lr_exx.o\
     kernel_xc.o\
     pot_hxc_lrtd.o\
-    hsolver_lrtd.o\
     lr_spectrum.o\
     hamilt_casida.o\
     esolver_lrtd_lcao.o\

source/main.cpp

@@ -27,13 +27,14 @@ int main(int argc, char** argv)
     */
     int nproc = 1;
     int my_rank = 0;
-    Parallel_Global::read_mpi_parameters(argc, argv, nproc, my_rank);
+    int nthread_per_proc = 1;
+    Parallel_Global::read_pal_param(argc, argv, nproc, nthread_per_proc, my_rank);
 #ifdef _OPENMP
     // ref: https://www.fftw.org/fftw3_doc/Usage-of-Multi_002dthreaded-FFTW.html
     fftw_init_threads();
     fftw_plan_with_nthreads(omp_get_max_threads());
 #endif
-    PARAM.set_rank_nproc(my_rank, nproc);
+    PARAM.set_pal_param(my_rank, nproc, nthread_per_proc);
 
     /*
     main program for doing electronic structure calculations.

source/module_base/math_chebyshev.cpp

@@ -575,6 +575,7 @@ bool Chebyshev<REAL>::checkconverge(
     std::function<void(std::complex<REAL>* in, std::complex<REAL>* out, const int)> funA,
     std::complex<REAL>* wavein,
     const int N,
+    const int LDA,
     REAL& tmax,
     REAL& tmin,
     REAL stept)
@@ -584,9 +585,9 @@ bool Chebyshev<REAL>::checkconverge(
     std::complex<REAL>* arrayn;
     std::complex<REAL>* arrayn_1;
 
-    arraynp1 = new std::complex<REAL>[N];
-    arrayn = new std::complex<REAL>[N];
-    arrayn_1 = new std::complex<REAL>[N];
+    arraynp1 = new std::complex<REAL>[LDA];
+    arrayn = new std::complex<REAL>[LDA];
+    arrayn_1 = new std::complex<REAL>[LDA];
 
     ModuleBase::GlobalFunc::DCOPY(wavein, arrayn_1, N);
 

source/module_base/math_chebyshev.h

@@ -191,6 +191,7 @@ class Chebyshev
     bool checkconverge(std::function<void(std::complex<REAL>* in, std::complex<REAL>* out, const int)> funA,
                        std::complex<REAL>* wavein,
                        const int N,
+                       const int LDA,
                        REAL& tmax,  // trial number for upper bound
                        REAL& tmin,  // trial number for lower bound
                        REAL stept); // tmax = max() + stept, tmin = min() - stept

source/module_base/parallel_global.cpp

@@ -14,6 +14,8 @@
 #include "module_base/global_function.h"
 #include "module_base/parallel_common.h"
 #include "module_base/parallel_reduce.h"
+#include "module_parameter/parameter.h"
+// #include "module_base/tool_quit.h"
 #include "version.h"
 
 #include <iostream>
@@ -85,7 +87,12 @@ void Parallel_Global::split_grid_world(const int diag_np, const int& nproc, cons
     return;
 }
 
-void Parallel_Global::read_mpi_parameters(int argc, char** argv, int& NPROC, int& MY_RANK)
+// changed from read_mpi_parameters in 2024-1018
+void Parallel_Global::read_pal_param(int argc, 
+                                          char** argv, 
+                                          int& NPROC, 
+                                          int& NTHREAD_PER_PROC, 
+                                          int& MY_RANK)
 {
 #ifdef __MPI
 #ifdef _OPENMP
@@ -150,6 +157,10 @@ void Parallel_Global::read_mpi_parameters(int argc, char** argv, int& NPROC, int
         // the user may take their own risk by set the OMP_NUM_THREADS env var.
         if (std::getenv("OMP_NUM_THREADS") == nullptr)
         {
+            // usage of WARNING_QUIT need module_base/tool_quit.cpp
+            // lead to undefined error in unit_test building
+            // ModuleBase::WARNING_QUIT( "Parallel_Global::read_pal_param","OMP_NUM_THREADS setting is invalid. Please set it to a proper value.");
+            std::cerr << "ERROR: OMP_NUM_THREADS setting is invalid. Please set it to a proper value." << std::endl;
             exit(1);
         }
     }
@@ -162,6 +173,8 @@ void Parallel_Global::read_mpi_parameters(int argc, char** argv, int& NPROC, int
                   << "Local thread limit: " << max_thread_num << std::endl;
     }
 
+    NTHREAD_PER_PROC = current_thread_num;
+
     if (MY_RANK == 0)
     {
 #ifdef VERSION
@@ -192,7 +205,7 @@ void Parallel_Global::read_mpi_parameters(int argc, char** argv, int& NPROC, int
                   << std::endl
                   << "                      Commit: " << commit << std::endl
                   << std::endl;
-        time_t time_now = time(NULL);
+        time_t time_now = time(nullptr);
         std::cout << " " << ctime(&time_now);
     }
 

source/module_base/parallel_global.h

@@ -18,7 +18,9 @@ extern int omp_number;
 //---------------------------
 // call at the very first.
 //---------------------------
-void read_mpi_parameters(int argc, char** argv, int& NPROC, int& MY_RANK);
+
+// changed from read_mpi_parameters in 2024-1018
+void read_pal_param(int argc, char** argv, int& NPROC, int& NTHREAD_PER_PROC, int& MY_RANK);
 #ifdef __MPI
 void myProd(std::complex<double>* in, std::complex<double>* inout, int* len, MPI_Datatype* dptr);
 #endif

source/module_base/test/math_chebyshev_test.cpp

@@ -346,15 +346,15 @@ TEST_F(MathChebyshevTest, checkconverge)
     double tmin = -1.1;
     double tmax = 1.1;
     bool converge;
-    converge = p_chetest->checkconverge(fun_sigma_y, v, 2, tmax, tmin, 0.2);
+    converge = p_chetest->checkconverge(fun_sigma_y, v, 2, 2, tmax, tmin, 0.2);
     EXPECT_TRUE(converge);
-    converge = p_chetest->checkconverge(fun_sigma_y, v + 2, 2, tmax, tmin, 0.2);
+    converge = p_chetest->checkconverge(fun_sigma_y, v + 2, 2, 2, tmax, tmin, 0.2);
     EXPECT_TRUE(converge);
     EXPECT_NEAR(tmin, -1.1, 1e-8);
     EXPECT_NEAR(tmax, 1.1, 1e-8);
 
     tmax = -1.1;
-    converge = p_chetest->checkconverge(fun_sigma_y, v, 2, tmax, tmin, 2.2);
+    converge = p_chetest->checkconverge(fun_sigma_y, v, 2, 2, tmax, tmin, 2.2);
     EXPECT_TRUE(converge);
     EXPECT_NEAR(tmin, -1.1, 1e-8);
     EXPECT_NEAR(tmax, 1.1, 1e-8);
@@ -363,12 +363,12 @@ TEST_F(MathChebyshevTest, checkconverge)
     v[0] = std::complex<double>(0, 1), v[1] = 1;
     fun.factor = 1.5;
     tmin = -1.1, tmax = 1.1;
-    converge = p_chetest->checkconverge(fun_sigma_y, v, 2, tmax, tmin, 0.2);
+    converge = p_chetest->checkconverge(fun_sigma_y, v, 2, 2, tmax, tmin, 0.2);
     EXPECT_FALSE(converge);
 
     fun.factor = -1.5;
     tmin = -1.1, tmax = 1.1;
-    converge = p_chetest->checkconverge(fun_sigma_y, v, 2, tmax, tmin, 0.2);
+    converge = p_chetest->checkconverge(fun_sigma_y, v, 2, 2, tmax, tmin, 0.2);
     EXPECT_FALSE(converge);
     fun.factor = 1;
 
@@ -632,9 +632,9 @@ TEST_F(MathChebyshevTest, checkconverge_float)
 
     auto fun_sigma_yf
         = [&](std::complex<float>* in, std::complex<float>* out, const int m = 1) { fun.sigma_y(in, out, m); };
-    converge = p_fchetest->checkconverge(fun_sigma_yf, v, 2, tmax, tmin, 0.2);
+    converge = p_fchetest->checkconverge(fun_sigma_yf, v, 2, 2, tmax, tmin, 0.2);
     EXPECT_TRUE(converge);
-    converge = p_fchetest->checkconverge(fun_sigma_yf, v + 2, 2, tmax, tmin, 0.2);
+    converge = p_fchetest->checkconverge(fun_sigma_yf, v + 2, 2, 2, tmax, tmin, 0.2);
     EXPECT_TRUE(converge);
     EXPECT_NEAR(tmin, -1.1, 1e-6);
     EXPECT_NEAR(tmax, 1.1, 1e-6);

source/module_basis/module_ao/parallel_2d.h

@@ -15,6 +15,7 @@ class Parallel_2D
     ~Parallel_2D() = default;
 
     Parallel_2D& operator=(Parallel_2D&& rhs) = default;
+    Parallel_2D(Parallel_2D&& rhs) = default;
 
     /// number of local rows
     int get_row_size() const

source/module_cell/module_symmetry/run_symmetry.cpp

@@ -14,7 +14,7 @@ int main(int argc, char **argv)
 	std::cout << "Hello, this is the 'symmetry' module of ABACUS." << std::endl;
 
 	std::cout << "The module does symmetry analysis for an input geometry." << std::endl;
-	Parallel_Global::read_mpi_parameters(argc,argv);
+	Parallel_Global::read_pal_param(argc,argv);
 	//std::cout << "Right now, the module is still empty, soon we will have more tests." << std::endl;
 
     calculate();