v3.6.3

What's Changed

Feature

• Feature : descriptors for equivariant DeePKS by @wenfei-li in #3894
• Feature : reading and writing of h(r) and dm(r) in npz format by @wenfei-li in #4066
• Feature: write local-XC and EXX part of band(orbital) energy along with the XC matrix by @maki49 in #4013
• Feature: add device error check for functions within device.cpp by @denghuilu in #4106
• Improve output of H and S matrices: output matrices when the self-consistent iteration steps reach scf_nmax, even if the charge density doesn't converge. by @LiuXiaohui123321 in #4147

Docs

• Docs: add docs about parameter symmetry by @WHUweiqingzhou in #4073
• Docs: remove extra ``` by @caic99 in #4133
• Docs: delay release of deltaspin due to bugs by @hongriTianqi in #4146

Fix

• Fix: use std::abs() instead of abs() by @WHUweiqingzhou in #4076
• Add a warning quit when parameters do not match in read_wfc_nao.cpp by @AsTonyshment in #4087
• Fix: remove non-used pointer in Gint by @maki49 in #4009
• Fix: dcu cmake compilation error by @denghuilu in #4097
• Fix: add some hint about error by @WHUweiqingzhou in #4100
• Fix: solve the convergence problem of E value in dav_subspace method by @haozhihan in #4052
• Fix: bug of plot-tools for nspin=2 and shift=True case by @1041176461 in #4116
• Fix: No warning is given if KPT is ill set by @Luc1anoo in #4134

Refactor

• Refactor istate_charge.cpp to reduce dependence on other modules by @AsTonyshment in #4086
• Refactor: optimized cubic spline by @jinzx10 in #4081
• Replace “lm->paraV” with “paraV” to simpifier parameters by @A-006 in #4110

Build & Test

• Test: increase the time limit of CI test by @WHUweiqingzhou in #4098
• fix: pin the version of intel mpi to 2021.11 by @pxlxingliang in #4105
• CI: modify CUDA test by @dzzz2001 in #4083
• Build: pin Intel MPI version to avoid performance degradation by @caic99 in #4065
• CI: update runner config by @caic99 in #4088
• CI: build images with tags on self-hosted runner by @caic99 in #4089
• CI: fix the incorrect argument order in Autotest.sh by @dzzz2001 in #4118

New Contributors

• @Luc1anoo made their first contribution in #4134
• @LiuXiaohui123321 made their first contribution in #4147

Full Changelog: v3.6.2...v3.6.3

v3.6.2

Feature

• output times for lambda inner loop by @hongriTianqi in #3982
• output mag force in spin constraint calculation by @hongriTianqi in #4037
• add pw_diag_ndim parameter to adjust efficiency for users by @haozhihan in #3998

Fix

• fix issue 3985 about dcu compiling by @haozhihan in #3986
• bug in plot-tools by @1041176461 in #3990
• remove the dependence of EXX on RPA by @maki49 in #3969
• remove parameter_pool.cpp in CMakeLists.txt by @WHUweiqingzhou in #3983
• bug of out_band when kpar > 1 by @Qianruipku in #3963
• multiple memory leak and array access out of bound by @caic99 in #3997
• memory leak in tddft by @caic99 in #4012
• tddft memory leak by @caic99 in #4020
• some variables from pointers to vector by @AmengFromCDH in #3977
• pointers to vectors by @DylanWRh in #3980
• new-delete with stl::vector in H_Ewald_pw.cpp by @QztOc07 in #3996
• pointers to vectors in bessel_basis.cpp by @Fisheory in #3993
• output final mag instead of initial mag in STRU_ION_D by @WHUweiqingzhou in #4044
• add a check_rho() to avoid negative charge density by @WHUweiqingzhou in #4053

Test and Docs

• case test for noncolin force by @hongriTianqi in #4048
• update cuda tests by @caic99 in #4032
• add some docs by @WHUweiqingzhou in #4039
• update clang-tidy config by @caic99 in #4010
• format CMakeLists.txt using cmake_format by @caic99 in #4014
• update version by @pxlxingliang in #4061
• add docs about using MPI in docker. by @grysgreat in #4060

New Contributors

• @QztOc07 made their first contribution in #3996

Full Changelog: v3.6.1...v3.6.2

v3.6.1

What's Changed

Feature

• Add INPUT parameter bands_to_print for LCAO basis set by @AsTonyshment in #3934
• Add a new Davidson iteration method called subspace davidson for pw basis by @haozhihan in #3903
• Enable 'calculation=test_memory" function again by @mohanchen in #3919
• add force and stress output for noncollinear calculation with soc by @hongriTianqi in #3933
• add so-called U-Ramping method to improve DFTU convergence by @WHUweiqingzhou in #3890
• read and write Hexx(R) in CSR format by @maki49 in #3727

Fix

• Bug when kpar > 1, nspin=2, and cell-relax by @Qianruipku in #3957
• Fix mistakes in TDDFT docs by @ESROAMER in #3841
• Fix the hexadecimal output in MD with OFDFT by @sunliang98 in #3896
• Initialize an int variable "nrotkm" in klist.cpp by @randomname23333 in #3908
• Initialize variable "no_use_c4" in bessel_basis.cpp by @Fisheory in #3910
• Initialized some variables in projgen.cpp by @Haerxile in #3905
• Update bessel_basis.cpp by @bot10032 in #3911
• array overreach in psi2rho of uspp by @YuLiu98 in #3961
• change the unit of decay gradient by @hongriTianqi in #3826
• compiling warning in veff_lcao.cpp by @dyzheng in #3901
• delete an extra bracket by @WHUweiqingzhou in #3863
• dos plot for nspin=2 by @1041176461 in #3928
• fix a bug in get_dmk_vector() and make some refactor of get-function by @WHUweiqingzhou in #3893
• initial the file : source/module_ri/RI_2D_Comm.cpp and ABF_Construct by @RobertRainbow in #3898
• initialize some variables by @DylanWRh in #3846
• initialize some variables by @Esperanzalukas in #3904
• initialize some variables by @yang2022i in #3917
• initialize variables by @Zjhjunhao in #3887
• initialized winput::imp_pao = 0 in winput.cpp by @charches in #3897
• llk_initialization of read_atoms.cpp by @Luckken21 in #3900
• memory record for wrapper data in HContainer by @dyzheng in #3962
• potentially uninitialized variables in calculation by @OldDriver233 in #3834
• some variables initialized by @AmengFromCDH in #3858

CI

• Mirror the repository to Gitee by @njzjz-bot in #3952

Docs

• A clearer notification for the easy installation part by @flinky-ai in #3871
• Update description for ESolver by @mohanchen in #3828
• update the default recommended requirements for conda by @caic99 in #3924

Perf

• Improve the performence of dav-subspace method in pw basis by @haozhihan in #3953
• use converged occupation matrix for initial DFTU calculation in MD calculation by @dyzheng in #3837

Refactor

• Improve the C++ code quality by @mohanchen in #3945
• Keep updating esolver and related source files by @mohanchen in #3853
• LCAO refactor step 4 by @mohanchen in #3921
• Refactor LCAO code Step 3 by @mohanchen in #3914
• Refactor LCAO step 2 by @mohanchen in #3874
• Refactoring LCAO codes. by @mohanchen in #3860
• Remove direct references to GlobalV related to function read_cube. by @Cstandardlib in #3824
• Update ORB_atomic.cpp by @aFlyingOctopus in #3835
• Update berryphase.cpp by @Pureorangeade in #3876
• Vector instead of New, changed by Linbo Cao by @RobertRainbow in #3975
• change namespace A::B{} to namespace A{namespace B{}} by @PeizeLin in #3960
• exchange some pointers into vectors by @Zjhjunhao in #3965
• replace with vector by @jwsyzy in #3972

Test

• Save simpson test results to file by @jieli-matrix in #3842
• add unit test of DFTUNew and refactor some code by @dyzheng in #3814
• modify the precision criteria of Simpson algorithm by @jieli-matrix in #3849
• save precision results to file when building unit tests by @jieli-matrix in #3829

New Contributors

• @Cstandardlib made their first contribution in #3824
• @AmengFromCDH made their first contribution in #3858
• @OldDriver233 made their first contribution in #3834
• @Pureorangeade made their first contribution in #3876
• @flinky-ai made their first contribution in #3871
• @aFlyingOctopus made their first contribution in #3835
• @DylanWRh made their first contribution in #3846
• @randomname23333 made their first contribution in #3908
• @Haerxile made their first contribution in #3905
• @Luckken21 made their first contribution in #3900
• @Zjhjunhao made their first contribution in #3887
• @charches made their first contribution in #3897
• @Esperanzalukas made their first contribution in #3904
• @Fisheory made their first contribution in #3910
• @bot10032 made their first contribution in #3911
• @yang2022i made their first contribution in #3917
• @RobertRainbow made their first contribution in #3898
• @njzjz-bot made their first contribution in #3952
• @jwsyzy made their first contribution in #3972

Full Changelog: v3.6.0...v3.6.1

v3.6.0

What's Changed

Feature

• Feature: release numerically stable QO with hands-on manual in tools/qo/ by @kirk0830 in #3669
• Feature: make mixing restarts if mixing_restart is smaller than drho by @WHUweiqingzhou in #3681
• Feature: Pythonization of ModuleNAO by @jieli-matrix in #3642
• Feature: Add readin-json support by @grysgreat in #3648
• Feature: Pythonization of NumericalRadial by @jieli-matrix in #3712
• Feature: output dkin for metaGGA calculations by @WHUweiqingzhou in #3700
• Feature: Visualize numerical orbital in pyabacus by @jieli-matrix in #3726
• Feature: enable the construction of RadialSet from individual spherical Bessel function by @jinzx10 in #3721
• Feature: QO MPI parallelization on both k and cell vector R when calculating overlap matrix by @kirk0830 in #3707
• Feature & Refactor: restart from Hexx by @maki49 in #3451
• Feature: new method of DFT+U calculation, add parametor "onsite_radius" for projections modulation. by @dyzheng in #3682

Tests

• Test: upload file for unittest of toQO::read_ovlp by @kirk0830 in #3724
• Test: add tests for jY-jY two-center integrals by @jinzx10 in #3734
• Test: add pyabacus CI test by @jieli-matrix in #3750
• test: in integrate test, skip running test if test folder do not exist. by @pxlxingliang in #3793
• test: set ENABLE_RAPIDJSON=ON in dockerfile to output abacus.json by @pxlxingliang in #3788

Refactor

• Refactor: move Module_IO::output_radial into RadialSet class to make module_io and module_nao decoupled by @kirk0830 in #3680
• Reflactor: new json array support and add json tree of abacus-output by @grysgreat in #3687
• Refactor: Removed GlobalVs related to function ModuleIO::write_dm by @AsTonyshment in #3797
• Refactor: Removed GlobalVs related to module_io/istate_envelope.cpp by @AsTonyshment in #3801
• Refactor: remove some global variables in read_dm and read_wfc_nao by @dzzz2001 in #3794

Bug Fixes

• Fix: standalone FindBLAS.cmake, mislink to global-visible lib by @yizeyi18 in #3623
• Fix outdated comment by @jinzx10 in #3694
• Fix: Support pseudopotential with dense grid by @sunliang98 in #3678
• Fix: fix bugs in TB2J docs by @WHUweiqingzhou in #3709
• Fix: delete useless version info in CMakeList.txt by @WHUweiqingzhou in #3731
• Fix: no cal_nlm_tot in cal_force_stress and onsite_radius > rcut_NAOs by @dyzheng in #3732
• Bug: fix undefined symbol in pyabacus by @jieli-matrix in #3740
• Bug: Fix intel undefined error in pyabacus by @jieli-matrix in #3742
• Fix: error with skipping no_U elements by @dyzheng in #3748
• Fix: automatically turn off symmetry when not all atoms are moveable by @maki49 in #3738
• Bug: Fix Intel compile warnings by @jieli-matrix in #3756
• Fix: fix a json in init-element. by @grysgreat in #3818
• fix: the natom in abacus.json by @pxlxingliang in #3822
• Fix: Support LPS in psp6 format by @sunliang98 in #3741
• Fix FindLibxc.cmake to use /usr/include as include directory by @yizeyi18 in #3771
• Fix: update the default two-center integration algorithm for Makefile build by @jinzx10 in #3773
• Fix: warn not to set nupdown in deltaspin calcualtion by @hongriTianqi in #3765
• Fix: ELPA test compile by @yizeyi18 in #3774
• fix the output value of ediff in abaucs.json by @pxlxingliang in #3787
• Fix: fix cases by @WHUweiqingzhou in #3804
• Fix: warning message of nspin=4 by @dyzheng in #3791
• Fix: timer error of Veff::contributeHR in gamma_only calculation by @dyzheng in #3799
• Fix: fix scf_nmax restart by @WHUweiqingzhou in #3808
• Refactor&Fix: remove normalization of numerical atomic orbitals in psi_initializer by @kirk0830 in #3716

Build

• add codecov token by @njzjz in #3683
• install: write the commit to commit.h instead of setting the macro COMMIT by @pxlxingliang in #3692
• Toolchain modification by @QuantumMisaka in #3733
• Build: fix clang-format file format by @caic99 in #3736
• add lapack/blas wrapper as suggested in #3706 by @yizeyi18 in #3757

Documentation

• Docs: Add docs for pyabacus by @jieli-matrix in #3697
• add TB2J interface docs (#3702) by @Shen-Zhen-Xiong in #3703
• docs: add the link for conda FAQ by @njzjz in #3728
• Docs: add a faq by @WHUweiqingzhou in #3723
• Doc: update doc about EXX restart and symmetry by @maki49 in #3764
• Doc: add a warning in pull request template by @hongriTianqi in #3768
• Docs: Add json docs. by @grysgreat in #3754
• docs: fix the comment in abacus.json by @pxlxingliang in #3784

New Contributors

• @dzzz2001 made their first contribution in #3794

Full Changelog: v3.5.4...v3.6.0

v3.5.4

Feature

• Feature: Added json array support. Implement init json by @grysgreat in #3612
• Feature: extended support on arbitrary atom symbols defined in STRU to calculate QO_OVLP data by @kirk0830 in #3651
• Feature: add a new parameter nelec_delta by @WHUweiqingzhou in #3641
• Feature: including the Slater screening to more accurately estimate the hydrogen radial function by @kirk0830 in #3633

Refactor

• Test: add a new UnitTest to protect save_DMR() by @WHUweiqingzhou in #3659

Bug Fixes

• Fix: fix accessing out-of-bounds in save_DMR() with changed nnr by @WHUweiqingzhou in #3620
• Fix: correct the way to use two-center-intergral and support more hydrogen-like orbitals strategy by @kirk0830 in #3606
• Fix: mulliken_charge error for magnetism_atom_y by @dyzheng in #3635
• Fix: fix nbands<ncpus in atom example by @WHUweiqingzhou in #3636
• Fix: error in reading sc.json file by @hongriTianqi in #3638
• Fix: reuse elpa_handle to avoid memory_leak by @dyzheng in #3637
• Fix: fix the bug when reuse the numerical_basis class instance by @kirk0830 in #3615
• Fix: the bug of possible large movement of atom position caused by periodicity in BFGS method by @pxlxingliang in #3658
• Fix: analyze the spin-up part's symmetry for AFM cases by @maki49 in #3666
• Fix: change error to warning in Symmetry for user friendliness by @maki49 in #3670

Build

• Toolchain 2024.1 by @QuantumMisaka in #3619

Full Changelog: v3.5.3...v3.5.4

v3.5.3

Feature

• Refactor&Feature: Rebuild rapidjson module. by @grysgreat in #3545
• Feature: Add json array support and unit test. by @grysgreat in #3584
• Feature: support double_grid in mix_rho_recip_new() by @WHUweiqingzhou in #3577
• Feature: TDDFT velocity gague by @ESROAMER in #3576
• Feature: Pythonization of math_integral in Module Base by @jieli-matrix in #3574

Refactor

• Refactor: namespace ABFs_Construct::PCA by @PeizeLin in #3546
• Refactor: charge_mixing::set_mixing() by @WHUweiqingzhou in #3596
• Refactor: split Charge_Mixing::set_mixing into set_mixing and init_mixing by @WHUweiqingzhou in #3601
• Refactor: let Charge_Mixing::inner_product_recip_hartree supports all cases by @WHUweiqingzhou in #3604
• Refactor: rename Kerker_screen_recip() and add more UnitTests to protect Charge_Mixing by @WHUweiqingzhou in #3609

Bug Fixes

• Fix: add docs and example to show how to perform magnetic calculations of atomic systems by @WHUweiqingzhou in #3571
• Fix the bug that the Wannier90 interface does not report error informations when reading files under MPI. by @jingan-181 in #3563
• Fix: Address compilation warnings in clang compiler by @denghuilu in #3593
• Fix: memory leak in module_nao by @dyzheng in #3602

Build

• CUDA: Add device error check for CUDA API and device function by @denghuilu in #3575
• GPU: Add rocm error check by @denghuilu in #3585

Documentation

• Docs: add introductions to ci/cd and new modules by @caic99 in #3610

Full Changelog: v3.5.2...v3.5.3

v3.5.2

What's Changed

Feature

• Accelerate: reduce unnecessary computation due to the calculation of the overlap matrix in the cg method by @haozhihan in #3556
• Feature: enable restart charge density mixing during SCF by @WHUweiqingzhou in #3542
• Feature: Add ENABLE_RAPIDJSON option to control the output of abacus.json by @pxlxingliang in #3519
• Feature : printing band density by @wenfei-li in #3501
• Feature: support segment split in kline mode in KPT file and out_band band output precision control, 8 as default by @kirk0830 in #3493
• Feature: add python wrapper for math sphbes by @jieli-matrix in #3475
• Build: add ccache to accelerate the testing process by @caic99 in #3509
• Build: Improving CMake performance for finding LibXC and ELPA by @yizeyi18 in #3478
• Feature: enable the computation of zeros of multiple spherical Bessel functions in one call by @jinzx10 in #3449

Bug Fixes

• fix: bug in Autotest.sh when result.ref has no totaltimeref by @pxlxingliang in #3523
• Fix : unit test of module_xc by @wenfei-li in #3524
• Fix: omit small magnetic moments to avoid numerical instability by @hongriTianqi in #3530
• Fix: fix multiple compiler warnings by @caic99 in #3515
• Fix: print the default precision flag by @denghuilu in #3496
• Fix: add blas support for FindLAPACK.cmake by @haozhihan in #3497
• Fix: fix bug in mulliken charge calculation by @hongriTianqi in #3503
• Fix: avoid infinite loop when meeting badbit or failbit in reading STRU by @WHUweiqingzhou in #3506
• Fix: cuda build target by @caic99 in #3276
• Build: fix compatibility issue against toolchain install by @yizeyi18 in #3540

Refactor

• Refactor: namespace Conv_Coulomb_Pot_K by @PeizeLin in #3446

Tests

• Test: run INPUT.Default() in every process in InputParaTest by @WHUweiqingzhou in #3490
• Test: more unittests for QO - arbitrary orbital selection by @kirk0830 in #3499
• Test: Configure performance tests for math libraries by @jieli-matrix in #3511

Documentation

• Docs: to avoid the misunderstanding in docs by @WHUweiqingzhou in #3518
• Docs: fix a missing depencency in conda build env by @caic99 in #3508
• Docs: correct some docs about mp2 smearing method by @WHUweiqingzhou in #3533
• Docs: add some docs for PR#3501 by @WHUweiqingzhou in #3537
• Update input-main.md by @jingan-181 in #3551

New Contributors

• @yizeyi18 made their first contribution in #3478

Full Changelog: v3.5.1...v3.5.2

v3.5.1

Feature

• Feature: Quasiatomic Orbital (QO) implement - immediate by @kirk0830 in #3236
• Feature: autoset efield related parameters by @YuLiu98 in #3432
• add note if no INPUT_PARAMETERS in INPUT card by @Satinelamp in #3407
• Refactor&Feature: add default value controlling decimal length of output for out_mat_hs 1 case and expose user control keyword out_ndigits by @kirk0830 in #3377
• Feature: Pythonization of Two-Center Integral Module by @jieli-matrix in #3349
• Feature: enable using spherical Bessel coefficients as an alternative to numerical atomic orbital file by @jinzx10 in #3404
• Feature: support list input parse and multiple values for bessel_nao_rcut by @kirk0830 in #3454
• Feature: support QO can be installed like python setup.py install and use as import abacus2qo by @kirk0830 in #3463
• Feature: support precision control for out_mat_hs keyword and default value set to 8 by @kirk0830 in #3468
• Feature & Refactor: write XC matrix in KS-wfc representation and unify matrix-writing functions by @maki49 in #3441
• Feature: Add CUDA support for stress_mgga function by @denghuilu in #3474
• Feature: Autoset a finite magmon if STRU does not specify a finite one by @WHUweiqingzhou in #3453
• Feature: Add rocm implementation for mgga by @denghuilu in #3479
• Feature: add cg method for lcao basis by @haozhihan in #3473
• Feature&Refactor: multiple values input now is supported by QO by @kirk0830 in #3483

Refactor

• Refactor: delete tmp files in md/relax by @YuLiu98 in #3427
• Refactor: Exx_LRI::cal_exx_elec(Ds) by @PeizeLin in #3429
• Refactor: move rho_mag from Charge into Charge_Mixing, making them independent of each other by @WHUweiqingzhou in #3435
• Refactor：Support custom HexxR to OperatorEXX by @maki49 in #3470

Bug fix

• Fix: bug of training force and stress with DeePKS by @dyzheng in #3444
• Fix: the bug of garbled output by @Zhuxuegang2022 in #3458
• Fix: compare k-lattice in ibz_kpoint only when k-point is generated by Monkhorst-Pack by @maki49 in #3442
• Fix: elpa memory leak by @caic99 in #3472
• Fix: Remove backward-cpp tools by @denghuilu in #3480
• Fix: Fix single precision md bug by @denghuilu in #3481
• Fix : some minor modifications of PAW by @wenfei-li in #3476

Test

• Test: add a UnitTest for AutosetMag in read_atom_position() by @WHUweiqingzhou in #3471
• Test: add precision tests for zeros and derivatives in math sphbes by @jieli-matrix in #3445

Documentation

• Docs: Add instructions on using conda as build env by @caic99 in #3466
• update version to 3.5.1 by @hongriTianqi in #3489

Full Changelog: v3.5.0...v3.5.1

v3.5.0

What's Changed

Features:

• Feature: read in wfc from binary file for PW basis by @YuLiu98 in #3338
• Feature: read rho in G space from binary file in qe format by @YuLiu98 in #3303
• Feature: add a new parameter 'mixing_gg0_min' by @WHUweiqingzhou in #3305
• Feature: Building numerical atomic orbitals from spherical Bessel coefficients by @jinzx10 in #3296
• Feature: mixing density and magnetism {mx,my,mz} separately for non-colinear calculation by @WHUweiqingzhou in #3330
• Feature: change default value of smearing from fixed to gauss by @WHUweiqingzhou in #3337
• Feature : implementing PAW force by @wenfei-li in #3306
• Feature: a promising angle mixing method for non-colinear calculations by @WHUweiqingzhou in #3356
• Feature: enable lcao_in_pw function on present version by @kirk0830 in #3057

Refactors:

• Perf: save HR for spin-up and spin-down both in hamilt_lcao.cpp for delta-spin by @dyzheng in #3299
• Refactor: refactor cg interface by @denghuilu in #3293
• Refactor: formatting kpoint information in running_${calculation}.log file and add kpoint reduction information output by @kirk0830 in #3325
• Refactor: add values to hK rather than LM.Hloc in OperatorEXX::contribute_Hk by @maki49 in #3322
• Refactor: Modify the wannier90 interface to support PW and LCAO basis sets. (support nonlinear calculation) by @jingan-181 in #3097
• Refactor: interface between new & old NAO (1/3): NumericalRadial & Numerical_Orbital_Lm by @jinzx10 in #3353
• Refactor: output total pressure by @YuLiu98 in #3360
• Refactor: update toWannier90 interface with psi_initializer_nao by @kirk0830 in #3371

Tests:

• Test: cancel previous runs on new commit pushed by @caic99 in #3288
• Test: modify the reference of 184_PW_BNDKPAR_SDFT_MALL/ALL by @pxlxingliang in #3369

Fixs:

• Fix: bug of PR #3299 by @dyzheng in #3304
• Fix: get_wf with nspin=4 by @dyzheng in #3302
• Fix: fix print Error of nspin with noncolin=1 by @WHUweiqingzhou in #3307
• Fix: cal_ux before init_scf to make first step mag rho consistent by @hongriTianqi in #3308
• Fix: output lambda after lambda_loop by @hongriTianqi in #3317
• Fix: Fix convergence problem of spin-polarized metaGGA calculation for nspin=2 by @WHUweiqingzhou in #3328
• Fix: fix default setting of constrain by @hongriTianqi in #3326
• Fix: delete threshold for weight of bands by @dyzheng in #3345
• Fix: uspp bug caused by kpar>1 by @YuLiu98 in #3342
• Fix: setting the highest required cpp standard version by @caic99 in #3277
• Fix: several bugs exceeded the upper limit of int by @PeizeLin in #3341
• Fix test-warnings related to symmetry by @maki49 in #3355
• Fix: move template LCAO_Matrix::set_mat2d() from LCAO_matrix.cpp to LCAO_matrix.hpp by @PeizeLin in #3354
• Fix: Turn off two confusing warnings by @sunliang98 in #3357
• Fix: add a Waring_Quit if LATTICE_VECTORS is missed in STRU for non-latName case by @WHUweiqingzhou in #3359
• Fix: wrong override in veff_lcao, and some error annotations by @dyzheng in #3367
• Fix: gamma_only label calculating error for DeePKS by @dyzheng in #3375
• Fix: autoset cond_nche in sKG by @Qianruipku in #3372
• Fix: non-hermitian problem of h lambda by @hongriTianqi in #3381
• Fix: mismatch of DM with DeePKS in gamma_only case by @dyzheng in #3387
• Fix: memory leak in Gint by @dyzheng in #3399
• Fix: memory leak in tensor_map by @dyzheng in #3405
• Fxi: fix bug in input_test_para.cpp by @WHUweiqingzhou in #3403
• Fix: set default value of out_wannier_unk to false by @dyzheng in #3411

Docs:

• Doc: correct smearing_sigma in doc by @WHUweiqingzhou in #3312
• Docs: correct some typos in docs. by @WHUweiqingzhou in #3348
• Docs: update docs about charge mixing by @WHUweiqingzhou in #3368
• Docs: add Brief Introduction of the Output Files by @WHUweiqingzhou in #3370
• Docs: correct a typo in contributing.md by @WHUweiqingzhou in #3393

Full Changelog: v3.4.4...v3.5.0

v3.4.4

What's Changed

Feature

• a new mixing method for magnetic calculations by @WHUweiqingzhou in #3226
• implement deltaspin for collinear case by @hongriTianqi in #3220
• add interface to paw_force by @wenfei-li in #3238
• implement stress for mgga nspin = 2 by @wenfei-li in #3263
• cond_smear for sKG & Fix: catch_properties.sh by @Qianruipku in #3282

Refactor

• Use hartree-like sum to unify the definition of drho by @WHUweiqingzhou in #3258
• Simplify esolver_of.cpp by @sunliang98 in #3248
• modify all HContainer objects to wrapper case by @dyzheng in #3213
• charge_mixing by @Qianruipku in #3250
• Better zero-finding algorithms for spherical Bessel functions by @jinzx10 in #3269
• add sc_scf_nmin to control min scf steps before spin constraining by @hongriTianqi in #3270
• Newcusolver by @A-006 in #3261
• opitimize performance of Deepks by @dyzheng in #3256
• Structure symmetrization during relax by @maki49 in #3247

Fix

• add float_fftw error message by @denghuilu in #3287
• two bugs in PAW by @wenfei-li in #3217
• optimize the IO of lcao-wfc and charge by @maki49 in #3264
• cuda and rocm bugs by @denghuilu in #3286
• right criterion in Nonlocal Operator for LCAO code by @dyzheng in #3189

Build & Test

• update MKL version for ELPA by @caic99 in #3227
• Toolchain-2023.5-hotfix：LibRI, Libnpy and others link update by @QuantumMisaka in #3224
• change workflows related to intel image to fix the CI errors. by @dyzheng in #3231

New Contributors

• @A-006 made their first contribution in #3261

Full Changelog: v3.4.3...v3.4.4