Merge pull request #19 from deepmodeling/develop

Merge With Latest

Author: Flying-dragon-boxing

docs/advanced/acceleration/cuda.md

@@ -46,6 +46,7 @@ We provides [examples](https://github.com/deepmodeling/abacus-develop/tree/devel
 PW basis:
 - Only k point parallelization is supported, so the input keyword `kpar` will be set to match the number of MPI tasks automatically.
 - By default, CUDA architectures 60, 70, 75, 80, 86, and 89 are compiled (if supported). It can be overriden using the CMake variable [`CMAKE_CUDA_ARCHITECTURES`](https://cmake.org/cmake/help/latest/variable/CMAKE_CUDA_ARCHITECTURES.html) or the environmental variable [`CUDAARCHS`](https://cmake.org/cmake/help/latest/envvar/CUDAARCHS.html).
+
 LCAO basis:
 - Unless there is a specific reason, avoid using multiple GPUs, as it can be slower than using a single GPU. This is because the generalized eigenvalue solution of the LCAO basis set will incur additional communication overhead when calculated on multiple cards. When the memory limit of a GPU card makes it insufficient to complete the task, it is recommended to use multiple cards for calculation.
 - When using elpa on GPUs, some ELPA internal logs will be output.

docs/advanced/input_files/input-main.md

@@ -140,6 +140,7 @@
     - [out\_wfc\_r](#out_wfc_r)
     - [out\_wfc\_lcao](#out_wfc_lcao)
     - [out\_dos](#out_dos)
+    - [out\_ldos](#out_ldos)
     - [out\_band](#out_band)
     - [out\_proj\_band](#out_proj_band)
     - [out\_stru](#out_stru)
@@ -153,6 +154,8 @@
     - [out\_mat\_t](#out_mat_t)
     - [out\_mat\_dh](#out_mat_dh)
     - [out\_mat\_xc](#out_mat_xc)
+    - [out\_mat\_xc2](#out_mat_xc2)
+    - [out\_mat\_l](#out_mat_l)
     - [out\_eband\_terms](#out_eband_terms)
     - [out\_hr\_npz/out\_dm\_npz](#out_hr_npzout_dm_npz)
     - [dm\_to\_rho](#dm_to_rho)
@@ -173,6 +176,7 @@
     - [dos\_emin\_ev](#dos_emin_ev)
     - [dos\_emax\_ev](#dos_emax_ev)
     - [dos\_nche](#dos_nche)
+    - [stm\_bias](#stm_bias)
   - [NAOs](#naos)
     - [bessel\_nao\_ecut](#bessel_nao_ecut)
     - [bessel\_nao\_tolerence](#bessel_nao_tolerence)
@@ -1705,6 +1709,12 @@ These variables are used to control the output of properties.
     - lcao-only: output the density of states (DOS) and the projected density of states (PDOS)
 - **Default**: 0
 
+### out_ldos
+
+- **Type**: Boolean
+- **Description**: Whether to output the local density of states for given bias in cube file format, which is controlled by [stm_bias](#stm_bias). 
+- **Default**: False
+
 ### out_band
 
 - **Type**: Boolean \[Integer\](optional)
@@ -1799,6 +1809,20 @@ These variables are used to control the output of properties.
 The band (KS orbital) energy for each (k-point, spin, band) will be printed in the file `OUT.${suffix}/vxc_out.dat`. If EXX is calculated, the local and EXX part of band energy will also be printed in `OUT.${suffix}/vxc_local_out.dat`and `OUT.${suffix}/vxc_exx_out.dat`, respectively. All the `vxc*_out.dat` files contains 3 integers (nk, nspin, nband) followed by nk\*nspin\*nband lines of energy Hartree and eV.
 - **Default**: False
 
+### out_mat_xc2
+
+- **Type**: Boolean
+- **Availability**: Numerical atomic orbital (NAO) basis
+- **Description**: Whether to print the exchange-correlation matrices in **numerical orbital representation** (unit: Ry): $\braket{\phi_i|V_\text{xc}^\text{(semi-)local}+V_\text{exx}+V_\text{DFTU}|\phi_j}(\mathbf{R})$ in CSR format (the same format as [out_mat_hs2](../elec_properties/hs_matrix.md#out_mat_hs2)) in the directory `OUT.${suffix}`. (Note that currently DeePKS term is not included. ) The files are named `Vxc_R_spin$s`.
+- **Default**: False
+
+### out_mat_l
+
+- **Type**: Boolean [Integer\](optional)
+- **Availability**: Numerical atomic orbital (NAO) basis
+- **Description**: Whether to print the expectation value of the angular momentum operator $\hat{L}_x$, $\hat{L}_y$, and $\hat{L}_z$ in the basis of the localized atomic orbitals. The files are named `OUT.${suffix}/${suffix}_Lx.dat`, `OUT.${suffix}/${suffix}_Ly.dat`, and `OUT.${suffix}/${suffix}_Lz.dat`. The second integer controls the precision of the output.
+- **Default**: False 8
+
 ### out_eband_terms
 
 - **Type**: Boolean
@@ -1947,9 +1971,16 @@ These variables are used to control the calculation of DOS. [Detailed introducti
 ### dos_nche
 
 - **Type**: Integer
-The order of Chebyshev expansions when using Stochastic Density Functional Theory (SDFT) to calculate DOS.
+- **Description**: The order of Chebyshev expansions when using Stochastic Density Functional Theory (SDFT) to calculate DOS.
 - **Default**: 100
 
+### stm_bias
+
+- **Type**: Real
+- **Description**: The bias voltage used to calculate local density of states to simulate scanning tunneling microscope, see details in [out_ldos](#out_ldos).
+- **Default**: 1.0
+- **Unit**: V
+
 [back to top](#full-list-of-input-keywords)
 
 ## NAOs

docs/quick_start/output.md

@@ -6,15 +6,21 @@ The following files are the central output files for ABACUS. After executing the
 
 Different from `INPUT` given by the users, `OUT.suffix/INPUT` contains all parameters in ABACUS.
 
-> **Note:** `OUT.suffix/INPUT` contain the initial default of ABACUS instead of the real parameters used in calculations. This file is stored for reproduction in case the default value is changed during development. If you want to figure out the real parameters used in calculations, you can open `OUT.suffix/runing_scf.log` and research corresponding parameter you are interested.
+> **Note:** `OUT.suffix/INPUT` contains the **actual parameters used in the calculation**, including:
+> 1. **User-specified parameters** (explicitly defined in your input file or command-line arguments, overriding default parameters).
+> 2. **System default parameters** (automatically applied when not explicitly provided by the user).
+
+
+This file ensures calculations can be fully reproduced, even if default values change in future ABACUS versions.
+Also notice that in rare cases, a small number of parameters may be dynamically reset to appropriate values during runtime.
 
 For a complete list of input parameters, please consult this [instruction](../advanced/input_files/input-main.md).
 
 ## *running_scf.log*
 
 `running_scf.log` contains information on nearly all function calls made during the execution of ABACUS.
 
-## *KPT*
+## *KPT.info*
 
 This file contains the information of all generated k-points, as well as the list of k-points actually used for calculations after considering symmetry.
 
@@ -39,4 +45,4 @@ ABACUS generates a `.cif` format structure file based on the input file `STRU`,
 
 ## *warning.log*
 
-The file contains all the warning messages generated during the ABACUS run.
+The file contains all the warning messages generated during the ABACUS run.

source/CMakeLists.txt

@@ -104,7 +104,7 @@ if(USE_ROCM)
     module_hamilt_pw/hamilt_pwdft/kernels/rocm/wf_op.hip.cu
     module_hamilt_pw/hamilt_pwdft/kernels/rocm/vnl_op.hip.cu
     module_base/kernels/rocm/math_kernel_op.hip.cu
-    module_base/kernels/rocm/math_kernel_op.hip_vec.cu
+    module_base/kernels/rocm/math_kernel_op_vec.hip.cu
     module_base/kernels/rocm/math_ylm_op.hip.cu
     module_hamilt_general/module_xc/kernels/rocm/xc_functional_op.hip.cu
   )

source/Makefile.Objects

@@ -35,7 +35,7 @@ VPATH=./src_global:\
 ./module_hsolver/module_pexsi:\
 ./module_elecstate:\
 ./module_elecstate/kernels:\
-./module_elecstate/potentials:\
+./module_elecstate/module_pot:\
 ./module_elecstate/module_charge:\
 ./module_elecstate/module_dm:\
 ./module_psi:\
@@ -208,7 +208,7 @@ OBJS_DEEPKS=LCAO_deepks.o\
         deepks_orbpre.o\
         deepks_vdelta.o\
         deepks_vdpre.o\
-        deepks_hmat.o\
+        deepks_vdrpre.o\
         deepks_pdm.o\
         deepks_phialpha.o\
         LCAO_deepks_io.o\
@@ -266,6 +266,7 @@ OBJS_ESOLVER_LCAO=esolver_ks_lcao.o\
       lcao_after_scf.o\
       esolver_gets.o\
       lcao_others.o\
+      esolver_dm2rho.o\
 
 OBJS_GINT=gint.o\
       gint_gamma_env.o\
@@ -461,6 +462,14 @@ OBJS_SURCHEM=surchem.o\
 
 OBJS_SYMMETRY=symm_other.o\
     symmetry_basic.o\
+    symm_analysis.o\
+    symm_check.o\
+    symm_getgroup.o\
+    symm_hermite.o\
+    symm_lattice.o\
+    symm_magnetic.o\
+    symm_pricell.o\
+    symm_rho.o\
     symmetry.o\
 
 OBJS_XC=xc_functional.o\
@@ -488,6 +497,7 @@ OBJS_IO=input_conv.o\
     write_dos_pw.o\
     nscf_band.o\
     cal_dos.o\
+    cal_ldos.o\
     cif_io.o\
     dos_nao.o\
     numerical_descriptor.o\
@@ -545,6 +555,7 @@ OBJS_IO=input_conv.o\
     read_input_item_output.o\
     read_set_globalv.o\
     orb_io.o\
+    cal_pLpR.o\
 
 OBJS_IO_LCAO=cal_r_overlap_R.o\
       write_orb_info.o\

source/driver.cpp

@@ -23,24 +23,27 @@ Driver::~Driver()
 
 void Driver::init()
 {
-    ModuleBase::TITLE("Driver", "init");
+    // 1) Let's start by printing a title.
+    ModuleBase::TITLE("Driver", "ABACUS_begins");
 
+    // 2) Print the current time, since it may run a long time.
     time_t time_start = std::time(nullptr);
     ModuleBase::timer::start();
 
-    // (1) read the input parameters.
-    // INPUT should be initalized here and then pass to atomic world, mohan
-    // 2024-05-12 INPUT should not be GlobalC, mohan 2024-05-12
-    Driver::reading();
-
-    // (2) welcome to the atomic world!
+    // 3) Welcome to the atomic world! Let's do some fancy stuff here.
     this->atomic_world();
 
-    // (3) output information
+    // 4) All timers recorders are printed.
+    ModuleBase::timer::finish(GlobalV::ofs_running);
+
+    // 5) All memory recorders are printed.
+    ModuleBase::Memory::print_all(GlobalV::ofs_running);
+
+    // 6) Print the final time, hopefully it will not cost too long. 
     time_t time_finish = std::time(nullptr);
     ModuleIO::print_time(time_start, time_finish);
 
-    // (4) close all of the running logs
+    // 7) Clean up: close all of the running logs
     ModuleBase::Global_File::close_all_log(GlobalV::MY_RANK, PARAM.inp.out_alllog,PARAM.inp.calculation);
 
 }
@@ -169,16 +172,13 @@ void Driver::reading()
 void Driver::atomic_world()
 {
     ModuleBase::TITLE("Driver", "atomic_world");
-    //--------------------------------------------------
-    // choose basis sets:
-    // pw: plane wave basis set
-    // lcao_in_pw: LCAO expaned by plane wave basis set
-    // lcao: linear combination of atomic orbitals
-    //--------------------------------------------------
+    ModuleBase::timer::tick("Driver", "atomic_world");
+
+    // reading information 
+    this->reading();
 
     // where the actual stuff is done
     this->driver_run();
 
-    ModuleBase::timer::finish(GlobalV::ofs_running);
-    ModuleBase::Memory::print_all(GlobalV::ofs_running);
+    ModuleBase::timer::tick("Driver", "atomic_world");
 }

source/driver_run.cpp

@@ -25,13 +25,13 @@
  */
 void Driver::driver_run()
 {
-    ModuleBase::TITLE("Driver", "driver_line");
-    ModuleBase::timer::tick("Driver", "driver_line");
+    ModuleBase::TITLE("Driver", "driver_run");
 
     //! 1: setup cell and atom information
     // this warning should not be here, mohan 2024-05-22
 #ifndef __LCAO
     if (PARAM.inp.basis_type == "lcao_in_pw" || PARAM.inp.basis_type == "lcao") {
+        ModuleBase::timer::tick("Driver","driver_run");
         ModuleBase::WARNING_QUIT("driver",
                                  "to use LCAO basis, compile with __LCAO");
     }
@@ -93,6 +93,6 @@ void Driver::driver_run()
 
     //! 6: output the json file
     Json::create_Json(&ucell, PARAM);
-    ModuleBase::timer::tick("Driver", "driver_line");
+
     return;
 }

source/module_base/blas_connector.cpp

@@ -820,7 +820,7 @@ void vector_add_vector(const int& dim, float *result, const float *vector1, cons
 	}
 	else if (device_type == base_device::GpuDevice){
 #ifdef __CUDA
-		ModuleBase::constantvector_addORsub_constantVector_op<float, base_device::DEVICE_GPU>()(dim, result, vector1, constant1, vector2, constant2);
+		ModuleBase::vector_add_vector_op<float, base_device::DEVICE_GPU>()(dim, result, vector1, constant1, vector2, constant2);
 #endif
 	}
 }
@@ -838,7 +838,7 @@ void vector_add_vector(const int& dim, double *result, const double *vector1, co
 	}
 	else if (device_type == base_device::GpuDevice){
 #ifdef __CUDA
-		ModuleBase::constantvector_addORsub_constantVector_op<double, base_device::DEVICE_GPU>()(dim, result, vector1, constant1, vector2, constant2);
+		ModuleBase::vector_add_vector_op<double, base_device::DEVICE_GPU>()(dim, result, vector1, constant1, vector2, constant2);
 #endif
 	}
 }
@@ -856,7 +856,7 @@ void vector_add_vector(const int& dim, std::complex<float> *result, const std::c
 	}
 	else if (device_type == base_device::GpuDevice){
 #ifdef __CUDA
-		ModuleBase::constantvector_addORsub_constantVector_op<std::complex<float>, base_device::DEVICE_GPU>()(dim, result, vector1, constant1, vector2, constant2);
+		ModuleBase::vector_add_vector_op<std::complex<float>, base_device::DEVICE_GPU>()(dim, result, vector1, constant1, vector2, constant2);
 #endif
 	}
 }
@@ -874,7 +874,7 @@ void vector_add_vector(const int& dim, std::complex<double> *result, const std::
 	}
 	else if (device_type == base_device::GpuDevice){
 #ifdef __CUDA
-		ModuleBase::constantvector_addORsub_constantVector_op<std::complex<double>, base_device::DEVICE_GPU>()(dim, result, vector1, constant1, vector2, constant2);
+		ModuleBase::vector_add_vector_op<std::complex<double>, base_device::DEVICE_GPU>()(dim, result, vector1, constant1, vector2, constant2);
 #endif
 	}
 }

source/module_base/formatter.h

@@ -191,17 +191,22 @@ class FmtTable
      * 
      * @param titles titles, its size should be the same as the number of columns
      * @param nrows number of rows
-     * @param aligns Alignments instance, can be constructed with initializer_list<char> like {'r', 'c'}, for right and center alignment for values and titles
+     * @param fmts format strings for each column, its size should be the same as the number of columns
+     * @param indent indent for each column, default is 0
+     * @param aligns Alignments instance, for alignment of values and titles, e.g. {Align::LEFT, Align::RIGHT} for left alignment of values and right alignment of titles
      * @param frames Frames instance, can be constructed with initializer_list<char> like {'-', '-', '-', ' ', ' '}, for up, middle, down, left and right frames
      * @param delimiters Delimiters instance, can be constructed with initializer_list<char> like {'-', ' '}, for horizontal and vertical delimiters
      */
     FmtTable(const std::vector<std::string>& titles, 
-             const size_t& nrows, 
+             const size_t nrows, 
              const std::vector<std::string>& fmts,
+             const size_t indent = 0,
              const Alignments& aligns = {},
              const Frames& frames = {},
-             const Delimiters& delimiters = {}): titles_(titles), data_(nrows, titles.size()), fmts_(fmts), aligns_(aligns), frames_(frames), delimiters_(delimiters)
-    { assert(titles.size() == fmts.size()); };
+             const Delimiters& delimiters = {}): 
+             titles_(titles), data_(nrows, titles.size()), // data
+             fmts_(fmts), indent_(indent), aligns_(aligns), frames_(frames), delimiters_(delimiters) // styles
+    { assert(titles.size() == fmts.size()||titles.size() == 0); };
     ~FmtTable() {};
     /**
      * @brief import data from std::vector
@@ -269,22 +274,24 @@ class FmtTable
      */
     std::string concat_title(const std::vector<std::string>& titles) const
     {
-        std::string dst;
+        std::string dst = "";
         // first sum width of all titles
         size_t width = std::accumulate(titles.begin(), titles.end(), 0, [](const size_t& acc, const std::string& s) { return acc + s.size(); });
         // add width of delimiters
         width += titles.size() - 1;
         // add width of left and right frames
         width += 2;
-        dst += std::string(width, frames_.up_) + "\n" + std::string(1, frames_.l_);
+        dst += std::string(indent_, ' ') + std::string(width, frames_.up_) + "\n"; // first line: the upper frame
+        dst += std::string(indent_, ' ') + std::string(1, frames_.l_); // second line: the left frame + titles + right frame
         for(size_t i = 0; i < titles.size(); i++)
         {
             dst += titles[i];
             if (i != titles.size() - 1) {
                 dst += delimiters_.v_;
             }
         }
-        dst += std::string(1, frames_.r_) + "\n" + std::string(width, frames_.mid_) + "\n";
+        dst += std::string(1, frames_.r_) + "\n";
+        dst += std::string(indent_, ' ') + std::string(width, frames_.mid_) + "\n"; // third line: the middle frame
         return dst;
     }
     /**
@@ -303,10 +310,10 @@ class FmtTable
         width += row.size() - 1;
         // for the left and right frame
         width += 2;
-        if (pos == 't') {
-            dst += std::string(width, frames_.up_) + "\n";
+        if (pos == 't') { // 't' for top
+            dst += std::string(indent_, ' ') + std::string(width, frames_.up_) + "\n";
         }
-        dst += std::string(1, frames_.l_);
+        dst += std::string(indent_, ' ') + std::string(1, frames_.l_);
         for(size_t i = 0; i < row.size(); i++)
         {
             dst += row[i];
@@ -315,8 +322,8 @@ class FmtTable
             }
         }
         dst += std::string(1, frames_.r_) + "\n";
-        if (pos == 'b') {
-            dst += std::string(width, frames_.dw_) + "\n";
+        if (pos == 'b') { // 'b' for bottom
+            dst += std::string(indent_, ' ') + std::string(width, frames_.dw_) + "\n"; // the last line
         }
         return dst;
     }
@@ -397,6 +404,7 @@ class FmtTable
     std::vector<std::string> titles_;
     NDArray<std::string> data_; // data
     std::vector<std::string> fmts_; // format strings for each column
+    size_t indent_ = 0; // indent for each column
 };
 
 #endif