Merge branch 'develop' into Charge_template

linpeize · web-flow · commit 430d19f7f7bf · 2026-01-22T00:26:17.000+08:00
diff --git a/source/source_base/module_container/ATen/kernels/cuda/lapack.cu b/source/source_base/module_container/ATen/kernels/cuda/lapack.cu
@@ -289,7 +289,7 @@ struct lapack_hegvd<T, DEVICE_GPU> {
     {
         const int itype = 1;
         const char jobz = 'V';
-        const char uplo = 'L';
+        const char uplo = 'U';
         cudaErrcheck(cudaMemcpy(eigen_vec, Mat_A, sizeof(T) * dim * lda, cudaMemcpyDeviceToDevice));
 
         // prevent B from being overwritten by Cholesky
diff --git a/source/source_hamilt/module_vdw/vdwd2.h b/source/source_hamilt/module_vdw/vdwd2.h
@@ -36,9 +36,9 @@ class Vdwd2 : public Vdw
         int yidx = para_.period().y / 2;
         int zidx = para_.period().z / 2;
 
-        for (int it1 = 0; it1 != ucell_.ntype; ++it1)
+        for (int it1 = 0; it1 < ucell_.ntype; ++it1)
         {
-            for (int it2 = 0; it2 != ucell_.ntype; ++it2)
+            for (int it2 = 0; it2 < ucell_.ntype; ++it2)
             {
                 const double C6_product
                     = sqrt(para_.C6().at(ucell_.atoms[it1].ncpp.psd) * para_.C6().at(ucell_.atoms[it2].ncpp.psd))
diff --git a/source/source_hsolver/diago_lapack.cpp b/source/source_hsolver/diago_lapack.cpp
@@ -58,7 +58,7 @@ void DiagoLapack<std::complex<double>>::diag(hamilt::Hamilt<std::complex<double>
     Real* eigenvalue_in,
     MPI_Comm& comm)
 {
-    ModuleBase::TITLE("DiagoScalapack", "diag_pool");
+    ModuleBase::TITLE("DiagoLapack", "diag_pool");
     assert(h_mat.col == s_mat.col && h_mat.row == s_mat.row && h_mat.desc == s_mat.desc);
     std::vector<double> eigen(PARAM.globalv.nlocal, 0.0);
     this->dsygvx_diag(h_mat.col, h_mat.row, h_mat.p, s_mat.p, eigen.data(), psi);
@@ -72,7 +72,7 @@ void DiagoLapack<std::complex<double>>::diag(hamilt::Hamilt<std::complex<double>
     Real* eigenvalue_in,
     MPI_Comm& comm)
 {
-    ModuleBase::TITLE("DiagoScalapack", "diag_pool");
+    ModuleBase::TITLE("DiagoLapack", "diag_pool");
     assert(h_mat.col == s_mat.col && h_mat.row == s_mat.row && h_mat.desc == s_mat.desc);
     std::vector<double> eigen(PARAM.globalv.nlocal, 0.0);
     this->zhegvx_diag(h_mat.col, h_mat.row, h_mat.p, s_mat.p, eigen.data(), psi);
@@ -207,9 +207,7 @@ std::pair<int, std::vector<int>> DiagoLapack<T>::zhegvx_once(const int ncol,
     const int itype = 1, il = 1, iu = PARAM.inp.nbands, one = 1;
     int M = 0, NZ = 0, lwork = -1, lrwork = -1, liwork = -1, info = 0;
     const double abstol = 0, orfac = -1;
-    //Note: pzhegvx_ has a bug
-    //      We must give vl,vu a value, although we do not use range 'V'
-    //      We must give rwork at least a memory of sizeof(double) * 3
+    
     const double vl = 0, vu = 0;
     std::vector<std::complex<double>> work(1, 0);
     std::vector<double> rwork(3, 0);
diff --git a/source/source_hsolver/hsolver_lcao.cpp b/source/source_hsolver/hsolver_lcao.cpp
@@ -61,7 +61,7 @@ void HSolverLCAO<TK, Device>::solve(hamilt::Hamilt<TK>* pHamilt,
         }else 
     #endif
         if (PARAM.globalv.kpar_lcao > 1
-            && (this->method == "genelpa" || this->method == "elpa" || this->method == "scalapack_gvx"))
+            && (this->method == "genelpa" || this->method == "elpa" || this->method == "scalapack_gvx" || this->method == "lapack"))
         {
             this->parakSolve(pHamilt, psi, pes, PARAM.globalv.kpar_lcao);
         } else
@@ -254,6 +254,11 @@ void HSolverLCAO<T, Device>::parakSolve(hamilt::Hamilt<T>* pHamilt,
                 DiagoScalapack<T> sa;
                 sa.diag_pool(hk_pool, sk_pool, psi_pool, &(pes->ekb(ik_global, 0)), k2d.POOL_WORLD_K2D);
             }
+            else if (this->method == "lapack")
+            {
+                DiagoLapack<T> la;
+                la.diag_pool(hk_pool, sk_pool, psi_pool, &(pes->ekb(ik_global, 0)), k2d.POOL_WORLD_K2D);
+            }
 #ifdef __ELPA
             else if (this->method == "genelpa")
             {
diff --git a/source/source_main/version.h b/source/source_main/version.h
@@ -1,3 +1,3 @@
 #ifndef VERSION
-#define VERSION "v3.9.0.22"
+#define VERSION "v3.9.0.23"
 #endif
diff --git a/source/source_psi/psi_init.cpp b/source/source_psi/psi_init.cpp
@@ -44,17 +44,28 @@ void PSIInit<T, Device>::prepare_init(const int& random_seed)
     if (this->init_wfc == "random")
     {
         this->psi_initer = std::unique_ptr<psi_initializer<T>>(new psi_initializer_random<T>());
+        GlobalV::ofs_running << "\n Using RANDOM starting wave functions for all " << PARAM.inp.nbands << " bands\n";
     }
     else if (this->init_wfc == "file")
     {
         this->psi_initer = std::unique_ptr<psi_initializer<T>>(new psi_initializer_file<T>());
+        GlobalV::ofs_running << "\n Using FILE starting wave functions\n";
     }
     else if ((this->init_wfc.substr(0, 6) == "atomic") && (this->ucell.natomwfc == 0))
     {
-        ModuleBase::WARNING_QUIT("PSIInit::prepare_init",
-            "init_wfc = atomic requires pseudo wavefunctions, but number of atom wfc = 0. \n"
-            " Please use init_wfc = random or try a pseudopotential with pseudo wavefunctions.");
+        std::cout << " WARNING: init_wfc = " + this->init_wfc +
+            " requires atomic pseudo wavefunctions(PP_PSWFC),\n but none available."
+            " Automatically switch to random initialization." << std::endl;
+        GlobalV::ofs_running << "\n Using RANDOM starting wave functions for all " << PARAM.inp.nbands << " bands\n";
+        GlobalV::ofs_running << "\n WARNING:\n init_wfc = " + this->init_wfc + " requires atomic pseudo wavefunctions(PP_PSWFC), but none available. \n"
+            " Automatically switch to random initialization.\n"
+            " Note: Random starting wavefunctions may slow down convergence.\n"
+            "      For faster convergence, consider using:\n"
+            "      1) A pseudopotential file that includes atomic wavefunctions (with PP_PSWFC), or\n"
+            "      2) Numerical atomic orbitals with 'init_wfc = nao' or 'nao+random' if available.\n"
+            << std::endl;
         this->psi_initer = std::unique_ptr<psi_initializer<T>>(new psi_initializer_random<T>());
+
     }
     else if (this->init_wfc == "atomic"
              || (this->init_wfc == "atomic+random" && this->ucell.natomwfc < PARAM.inp.nbands))
@@ -64,26 +75,30 @@ void PSIInit<T, Device>::prepare_init(const int& random_seed)
             int nrandom = PARAM.inp.nbands - this->ucell.natomwfc;
             GlobalV::ofs_running << "\n Using ATOMIC starting wave functions with " << this->ucell.natomwfc << " atomic orbitals"
             << " + " << nrandom << " random orbitals"
-            << " (total " << PARAM.inp.nbands << " bands)";
+            << " (total " << PARAM.inp.nbands << " bands)\n";
         }
         else
         {
             GlobalV::ofs_running << "\n Using ATOMIC starting wave functions for all " << this->ucell.natomwfc << " atomic orbitals"
-                << " (covers " << PARAM.inp.nbands << " bands)";
+                << " (covers " << PARAM.inp.nbands << " bands)\n";
         }
         this->psi_initer = std::unique_ptr<psi_initializer<T>>(new psi_initializer_atomic<T>());
     }
     else if (this->init_wfc == "atomic+random")
     {
         this->psi_initer = std::unique_ptr<psi_initializer<T>>(new psi_initializer_atomic_random<T>());
+        GlobalV::ofs_running << "\n Using ATOMIC+RANDOM starting wave functions with "
+                             << this->ucell.natomwfc << " atomic orbitals\n";
     }
     else if (this->init_wfc == "nao")
     {
         this->psi_initer = std::unique_ptr<psi_initializer<T>>(new psi_initializer_nao<T>());
+        GlobalV::ofs_running << "\n Using NAO starting wave functions\n";
     }
     else if (this->init_wfc == "nao+random")
     {
         this->psi_initer = std::unique_ptr<psi_initializer<T>>(new psi_initializer_nao_random<T>());
+        GlobalV::ofs_running << "\n Using NAO+RANDOM starting wave functions\n";
     }
     else
     {
diff --git a/tests/04_FF/101_NEP_HfO2/INPUT b/tests/04_FF/101_NEP_HfO2/INPUT
@@ -10,12 +10,12 @@ pot_file            ../../PP_ORB/nep_hfo2.txt
 cal_force           1
 cal_stress          1
 
-md_nstep            3
+md_nstep            4
 md_type             npt
 md_dt               1
 md_tfirst           300
 md_thermostat       nhc  
 md_dumpfreq         1
 md_seed             1
 
-init_vel            1
+init_vel            1
diff --git a/tests/04_FF/50_DP_Al/INPUT b/tests/04_FF/50_DP_Al/INPUT
@@ -10,7 +10,7 @@ pot_file            ../../PP_ORB/Al-SCAN.pb
 cal_force           1
 cal_stress          1
 
-md_nstep            3
+md_nstep            4
 md_type             msst
 md_dt               1
 md_tfirst           200
diff --git a/tests/11_PW_GPU/002_PW_CG_GPU/README b/tests/11_PW_GPU/002_PW_CG_GPU/README
@@ -4,6 +4,6 @@ This test for:
 *kpoints 2*2*2
 *sg15 pseudopotential
 *smearing_method gauss
-*ks_solver bpcg
+*ks_solver cg
 *mixing_type broyden-kerker
 *mixing_beta 0.4
diff --git a/tests/11_PW_GPU/003_PW_DA_GPU/README b/tests/11_PW_GPU/003_PW_DA_GPU/README
@@ -4,6 +4,6 @@ This test for:
 *kpoints 2*2*2
 *sg15 pseudopotential
 *smearing_method gauss
-*ks_solver bpcg
+*ks_solver dav
 *mixing_type broyden-kerker
 *mixing_beta 0.4
diff --git a/tests/11_PW_GPU/004_PW_DS_GPU/INPUT b/tests/11_PW_GPU/004_PW_DS_GPU/INPUT
@@ -0,0 +1,33 @@
+INPUT_PARAMETERS
+#Parameters	(General)
+suffix              autotest
+pseudo_dir          ../../PP_ORB
+
+gamma_only          0
+calculation         scf
+symmetry            1
+relax_nmax          1
+out_level           ie
+smearing_method     gaussian
+smearing_sigma      0.02
+
+#Parameters (3.PW)
+ecutwfc             40
+scf_thr             1e-7
+scf_nmax            100
+
+#Parameters (LCAO)
+basis_type          pw
+ks_solver           dav_subspace
+device              gpu
+chg_extrap          second-order
+pw_diag_thr         0.00001
+
+cal_force           1
+#test_force         1
+cal_stress          1
+#test_stress        1
+
+mixing_type         broyden
+mixing_beta         0.4
+mixing_gg0          1.5
diff --git a/tests/11_PW_GPU/004_PW_DS_GPU/KPT b/tests/11_PW_GPU/004_PW_DS_GPU/KPT
@@ -0,0 +1,4 @@
+K_POINTS
+0
+Gamma
+2 2 2  0 0 0
diff --git a/tests/11_PW_GPU/004_PW_DS_GPU/README b/tests/11_PW_GPU/004_PW_DS_GPU/README
@@ -0,0 +1,9 @@
+This test for:
+*GaAs-deformation
+*PW
+*kpoints 2*2*2
+*sg15 pseudopotential
+*smearing_method gauss
+*ks_solver dav_subspace
+*mixing_type broyden-kerker
+*mixing_beta 0.4
diff --git a/tests/11_PW_GPU/004_PW_DS_GPU/STRU b/tests/11_PW_GPU/004_PW_DS_GPU/STRU
@@ -0,0 +1,23 @@
+ATOMIC_SPECIES
+As 1   As_dojo.upf upf201
+Ga 1   Ga_dojo.upf upf201
+
+LATTICE_CONSTANT
+1  // add lattice constant, 10.58 ang
+
+LATTICE_VECTORS
+5.33 5.33  0.0
+0.0  5.33 5.33
+5.33  0.0  5.33
+ATOMIC_POSITIONS
+Direct //Cartesian or Direct coordinate.
+
+As
+0
+1
+0.300000          0.3300000          0.27000000     0 0 0
+
+Ga              //Element Label
+0
+1              //number of atom
+0.00000          0.00000          0.000000     0 0 0
diff --git a/tests/11_PW_GPU/004_PW_DS_GPU/result.ref b/tests/11_PW_GPU/004_PW_DS_GPU/result.ref
@@ -0,0 +1,8 @@
+etotref -4869.7470518365098542
+etotperatomref -2434.8735259183
+totalforceref 5.200640
+totalstressref 37241.467259
+pointgroupref C_1
+spacegroupref C_1
+nksibzref 8
+totaltimeref 4.79
diff --git a/tests/11_PW_GPU/004_PW_DS_GPU/threshold b/tests/11_PW_GPU/004_PW_DS_GPU/threshold
@@ -0,0 +1,4 @@
+threshold 1
+force_threshold 1 
+stress_threshold 1
+fatal_threshold 1
diff --git a/tests/11_PW_GPU/CASES_GPU.txt b/tests/11_PW_GPU/CASES_GPU.txt
@@ -1,6 +1,7 @@
 001_PW_BPCG_GPU
 002_PW_CG_GPU
 003_PW_DA_GPU
+004_PW_DS_GPU
 005_PW_CG_GPU_float
 006_PW_get_wf_GPU
 007_PW_OW_GPU
diff --git a/toolchain/scripts/stage1/install_openmpi.sh b/toolchain/scripts/stage1/install_openmpi.sh
@@ -91,7 +91,12 @@ case "${with_openmpi}" in
             #   else
             #     EXTRA_CONFIGURE_FLAGS=""
             #   fi
-            ./configure CFLAGS="${CFLAGS}" \
+            ./configure \
+                CC=gcc \
+                CXX=g++ \
+                FC=gfortran \
+                F77=gfortran \
+                CFLAGS="${CFLAGS}" \
                 --prefix=${pkg_install_dir} \
                 --libdir="${pkg_install_dir}/lib" \
                 --with-libevent=internal \
diff --git a/toolchain/scripts/stage4/install_nep.sh b/toolchain/scripts/stage4/install_nep.sh
@@ -63,7 +63,7 @@ CXXFLAGS = -O2 -fPIC -std=c++11
 INCLUDES = -I./src
 
 # Source files
-SRCS = ./src/nep.cpp ./src/ewald.cpp ./src/neighbor.cpp
+SRCS = ./src/nep.cpp ./src/ewald_nep.cpp ./src/neighbor_nep.cpp
 
 # Object files
 OBJS = \$(SRCS:.cpp=.o)
@@ -154,4 +154,4 @@ load "${BUILDDIR}/setup_nep"
 write_toolchain_env "${INSTALLDIR}"
 
 cd "${ROOTDIR}"
-report_timing "nep"
+report_timing "nep"

Original file line number	Diff line number	Diff line change
`@@ -289,7 +289,7 @@ struct lapack_hegvd<T, DEVICE_GPU> {`
`289`	`289`	`{`
`290`	`290`	`const int itype = 1;`
`291`	`291`	`const char jobz = 'V';`
`292`		`- const char uplo = 'L';`
	`292`	`+ const char uplo = 'U';`
`293`	`293`	`cudaErrcheck(cudaMemcpy(eigen_vec, Mat_A, sizeof(T) * dim * lda, cudaMemcpyDeviceToDevice));`
`294`	`294`
`295`	`295`	`// prevent B from being overwritten by Cholesky`
Original file line number	Diff line number	Diff line change
`@@ -36,9 +36,9 @@ class Vdwd2 : public Vdw`
`36`	`36`	`int yidx = para_.period().y / 2;`
`37`	`37`	`int zidx = para_.period().z / 2;`
`38`	`38`
`39`		`- for (int it1 = 0; it1 != ucell_.ntype; ++it1)`
	`39`	`+ for (int it1 = 0; it1 < ucell_.ntype; ++it1)`
`40`	`40`	`{`
`41`		`- for (int it2 = 0; it2 != ucell_.ntype; ++it2)`
	`41`	`+ for (int it2 = 0; it2 < ucell_.ntype; ++it2)`
`42`	`42`	`{`
`43`	`43`	`const double C6_product`
`44`	`44`	`= sqrt(para_.C6().at(ucell_.atoms[it1].ncpp.psd) * para_.C6().at(ucell_.atoms[it2].ncpp.psd))`
Original file line number	Diff line number	Diff line change
`@@ -44,17 +44,28 @@ void PSIInit<T, Device>::prepare_init(const int& random_seed)`
`44`	`44`	`if (this->init_wfc == "random")`
`45`	`45`	`{`
`46`	`46`	`this->psi_initer = std::unique_ptr<psi_initializer<T>>(new psi_initializer_random<T>());`
	`47`	`+ GlobalV::ofs_running << "\n Using RANDOM starting wave functions for all " << PARAM.inp.nbands << " bands\n";`
`47`	`48`	`}`
`48`	`49`	`else if (this->init_wfc == "file")`
`49`	`50`	`{`
`50`	`51`	`this->psi_initer = std::unique_ptr<psi_initializer<T>>(new psi_initializer_file<T>());`
	`52`	`+ GlobalV::ofs_running << "\n Using FILE starting wave functions\n";`
`51`	`53`	`}`
`52`	`54`	`else if ((this->init_wfc.substr(0, 6) == "atomic") && (this->ucell.natomwfc == 0))`
`53`	`55`	`{`
`54`		`- ModuleBase::WARNING_QUIT("PSIInit::prepare_init",`
`55`		`- "init_wfc = atomic requires pseudo wavefunctions, but number of atom wfc = 0. \n"`
`56`		`- " Please use init_wfc = random or try a pseudopotential with pseudo wavefunctions.");`
	`56`	`+ std::cout << " WARNING: init_wfc = " + this->init_wfc +`
	`57`	`+ " requires atomic pseudo wavefunctions(PP_PSWFC),\n but none available."`
	`58`	`+ " Automatically switch to random initialization." << std::endl;`
	`59`	`+ GlobalV::ofs_running << "\n Using RANDOM starting wave functions for all " << PARAM.inp.nbands << " bands\n";`
	`60`	`+ GlobalV::ofs_running << "\n WARNING:\n init_wfc = " + this->init_wfc + " requires atomic pseudo wavefunctions(PP_PSWFC), but none available. \n"`
	`61`	`+ " Automatically switch to random initialization.\n"`
	`62`	`+ " Note: Random starting wavefunctions may slow down convergence.\n"`
	`63`	`+ " For faster convergence, consider using:\n"`
	`64`	`+ " 1) A pseudopotential file that includes atomic wavefunctions (with PP_PSWFC), or\n"`
	`65`	`+ " 2) Numerical atomic orbitals with 'init_wfc = nao' or 'nao+random' if available.\n"`
	`66`	`+ << std::endl;`
`57`	`67`	`this->psi_initer = std::unique_ptr<psi_initializer<T>>(new psi_initializer_random<T>());`
	`68`	`+`
`58`	`69`	`}`
`59`	`70`	`else if (this->init_wfc == "atomic"`
`60`	`71`	`\|\| (this->init_wfc == "atomic+random" && this->ucell.natomwfc < PARAM.inp.nbands))`
`@@ -64,26 +75,30 @@ void PSIInit<T, Device>::prepare_init(const int& random_seed)`
`64`	`75`	`int nrandom = PARAM.inp.nbands - this->ucell.natomwfc;`
`65`	`76`	`GlobalV::ofs_running << "\n Using ATOMIC starting wave functions with " << this->ucell.natomwfc << " atomic orbitals"`
`66`	`77`	`<< " + " << nrandom << " random orbitals"`
`67`		`- << " (total " << PARAM.inp.nbands << " bands)";`
	`78`	`+ << " (total " << PARAM.inp.nbands << " bands)\n";`
`68`	`79`	`}`
`69`	`80`	`else`
`70`	`81`	`{`
`71`	`82`	`GlobalV::ofs_running << "\n Using ATOMIC starting wave functions for all " << this->ucell.natomwfc << " atomic orbitals"`
`72`		`- << " (covers " << PARAM.inp.nbands << " bands)";`
	`83`	`+ << " (covers " << PARAM.inp.nbands << " bands)\n";`
`73`	`84`	`}`
`74`	`85`	`this->psi_initer = std::unique_ptr<psi_initializer<T>>(new psi_initializer_atomic<T>());`
`75`	`86`	`}`
`76`	`87`	`else if (this->init_wfc == "atomic+random")`
`77`	`88`	`{`
`78`	`89`	`this->psi_initer = std::unique_ptr<psi_initializer<T>>(new psi_initializer_atomic_random<T>());`
	`90`	`+ GlobalV::ofs_running << "\n Using ATOMIC+RANDOM starting wave functions with "`
	`91`	`+ << this->ucell.natomwfc << " atomic orbitals\n";`
`79`	`92`	`}`
`80`	`93`	`else if (this->init_wfc == "nao")`
`81`	`94`	`{`
`82`	`95`	`this->psi_initer = std::unique_ptr<psi_initializer<T>>(new psi_initializer_nao<T>());`
	`96`	`+ GlobalV::ofs_running << "\n Using NAO starting wave functions\n";`
`83`	`97`	`}`
`84`	`98`	`else if (this->init_wfc == "nao+random")`
`85`	`99`	`{`
`86`	`100`	`this->psi_initer = std::unique_ptr<psi_initializer<T>>(new psi_initializer_nao_random<T>());`
	`101`	`+ GlobalV::ofs_running << "\n Using NAO+RANDOM starting wave functions\n";`
`87`	`102`	`}`
`88`	`103`	`else`
`89`	`104`	`{`