Releases: deepmodeling/abacus-develop
Releases · deepmodeling/abacus-develop
ABACUS-v2.2.1 : Hotfix for v2.2.0
Feature changed:
- deepks : allow reading embedding model by @wenfei-li in #822
- Feature: Add module surchem. by @Asuna981002 in #846
Bug fixed:
- Fix uninited rcutmax by @maki49 in #821
- Fix bug: the opended file 'data-H' is not closed by @pxlxingliang in #820
- Fix a delete in parallel_orbitals by @maki49 in #827
- Fix : attempt to fix search_radius error in 'unit' test by @wenfei-li in #825
- Fix: rewrite MD_dump when step=0 by @YuLiu98 in #824
- Fix: add
.gitignoreitems for generated files intests/deepks/*. by @caic99 in #829 - Fix: gint : removed obsolete codes by @wenfei-li in #831
- Build: fix sanitizer test logic. by @caic99 in #832
- Fix: deleted useless file, output right occupied band for soc case by @dyzheng in #826
- Build: Add
libnpy,googletestin docker env. by @caic99 in #840 - Fix: potential array access overflow in gamma vl by @caic99 in #845
- Build: fix static analysis workflow logic. by @caic99 in #847
- Fix: misuse of '=' in assertion by @caic99 in #850
- Fix: variable not initialized in for-loop by @caic99 in #851
- Doc: Format markdown files. by @caic99 in #855
- docs: move
developers.mdtoCONTRIBUTING.md. by @caic99 in #856 - Fix: update the default value of md_tfreq by @YuLiu98 in #857
- docs: fix path link to
docs/by @caic99 in #858 - Fix: modify documentation, comment, and HSE input by @wenfei-li in #859
- Fix: remove invalid warning in
Pseudopot_upf::read_pseudo_upf201. by @sunliang98 in #862 - Fix: 1. only init wfc once for MD and relax; 2. reset PW_DIAG_THR every Ionic step. 3. 17 test cases reference updated by @dyzheng in #864
- Fix: 1, 2D-block allocation error. 2, not print charge density cube file if out_chg is 0 by @dyzheng in #868
- Fix: 1. update Makefile.system and Makefile.vars by @PeizeLin in abacusmodeling#122
- Fix: Adjust bands output for DFT+DMFT by @Qx80610702 in abacusmodeling#123
- Fix: 1. fix bug for 'dft_functional = hse' in LOOP_elec::solver() by @PeizeLin in abacusmodeling#124
- Fix：Added support for makefile to module_surchem. by @Asuna981002 in #871
Test updates:
- Test: update UT branches by @caic99 in #835
- Test: change testing target to merged reference by @caic99 in #838
- Test: add the unit test of parallel diagonalization by ELPA by @pxlxingliang in #839
- Test(pw): cg of HSolver with larger H matrices by @hongriTianqi in #863
- Test: rename CI step by @caic99 in #853
- CI: Enable static code analysis. by @caic99 in #854
- Build: update testing strategy by @caic99 in #852
- Test(pw): add UT of
Charge::sum_band()by @hongriTianqi in #860 - Test(LCAO): add the UT of ELPA/SCALAPACK diagonalization by @pxlxingliang in #861
Refactors:
- Update esolver.h by @Qianruipku in #843
- Refactor: move Global.h dependency from header to cpp files, by @dyzheng in #842
- Merge: HSolver to develop, diagonalization for PW and LCAO has finished. by @dyzheng in #848
Full Changelog: v2.2.0...v2.2.1
Contributors
hongriTianqi, maki49, and 10 other contributors
Assets 2
v2.2.0
New features for users:
- Feature: Optimized the calculation of structure factor with B-spline interpolation method. (Qianruipku #370 #371 #375 )
- Feature: The function of using the same random number seed to produce the initial trial wave function is added to the plane wave code. (Qianruipku #285 #284 #278 dyzheng #308 )
- Feature: SOC modulation function is added, used by set parameter "soc_lambda" (dyzheng #232 #258 #333 #336 #340 )
- Feature: Added parameter "symmetry_prec" to control symmetry precision. (LiuXiaohui123321 #333 )
- Feature: Added independent code for drawing electric band. (jiyuyang #333 )
- Feature: The SCAN exchange-correlation functional was added in the plane wave basis set. (wenfei-li #231 #236 #245 #253 #257 #262 #264 #292 #294 #296 #309 dyzheng #238 caic99 #268 )
- Feature: Added sparse matrix format for Hamiltonian matrix.(jingan #159 jingan-181 #75 )
- Feature: Molecular dynamics calculation in plane wave code is added. (dyzheng #99 )
- Feature: The DeePKS descriptor function was added. (maki49 #6 #15 #21 #24 #44 #67 #197 )
- Feature: Classical force field is added for molecular dynamics, including LJ potential and DP potential. (YuLiu98 #169 #306 #311 #368 #327 )
- Feature: Modified STRU file format, now you can import initial velocities and magnetic moments. (lyb9812 #234 #287 #295 #307 #355 ZhouGD2000 #312 #342 )
- Feature: BLPS pseudopotential file is supported. (sunliang98 #316 #318 #323 )
- Feature: BSSE function is added. (PeizeLin #4 )
- Feature: Stochastic DFT method is added. (Qianruipku #38 )
- Feature: The SCAN exchange-correlation functional was added in the plane wave basis set. (wenfei-li #309 #296 #294 #292 #264 #262 #257 #253 #245 #236 #231 dyzheng #238 caic99 #268 )
- Feature: DeePKS Function with LCAO code is added, multi core test is still in developing. (maki49 #52 #60 #65 #67 #177 #178 #197 #218 #346 )
- Feature: Time dependent DFT(TDDFT) is added, still in test.(HeFuxiang94 #30 #33 #34 #45 #51 #54 #280 #335 #361 #363 )
- Feature: A linear algorithm for Structure Factor by @Qianruipku in #370
- Feature: add stress calculation for LJ potential by @YuLiu98 in #368
- Feature: Newelpa and reconstruction merge to develop branch, compatible for current deepks branch. by @dyzheng in #395
- Feature: Deepks parallelization : gamma point by @wenfei-li in #404
- Feature: update scalapack_gvx from Peizelin and update openmp for gint_k_vl.cpp from Liu Xiaohui by @dyzheng in #434
- Feature: merge openmp update and PDOS+soc by @dyzheng in #410
- Feature: Update distribution method 1. by @sunliang98 in #431
- Feature: Add cudafft by @Qianruipku in #435
- Feature: merge scalapack update to develop branch by @dyzheng in #436
- Feature: Add variables(nsz2ip, ...) to header file, and update distribution_method1. by @sunliang98 in #433
- Feature: multi-k descriptor; fix small bug by @wenfei-li in #413
- Feature: Deepks multi-k : energy by @wenfei-li in #419
- Feature: Deepks : multi-k, force by @wenfei-li in #426
- Deepks : add comment; rearrange code; test case by @wenfei-li in #455
- Deepks : remove lcao_descriptor_new.cpp by @wenfei-li in #459
- Deepks : move libnpy dependence by @wenfei-li in #461
- Deepks : fix compile with cmake; set enable deepks as default; modify autotest by @wenfei-li in #466
- Deepks : multi-k and paralleliation for grad_vx by @wenfei-li in #472
- Feature: Add distributeg_method2 and corresponding unit tests. by @sunliang98 in #510
- Feature: by @Qianruipku in #554
- Print: add absolute k spacing in BANDS_1.dat by @hongriTianqi in #571
- Feature: Langevin method in MD by @YuLiu98 in #572
- Feature: Merge from cuda11 to develop by @Asuna981002 in #622
- Feature: enabling the orbital (band gap) training in DeePKS by @ouqi0711 in #617
- Feature: INPUT parameters refactor by @YuLiu98 in #810 #813
- Feature: Sto dft by @Qianruipku in abacusmodeling#4
- Feature: output jle(r) by @maki49 in abacusmodeling#8
- Feature: read descriptor basis by @maki49 in abacusmodeling#10
- Feature: output 5D overlap table between lcao and descriptor orbitals by @maki49 in abacusmodeling#11
- Feature: out lcao-line descriptor by @maki49 in abacusmodeling#15
- Feature: add the input parameter for accuracy of symmetry: symmetry_prec by @LiuXiaohui123321 in abacusmodeling#47
- Feature: add band structure and DOS plotting tools by @1041176461 in abacusmodeling#51
- Feature: Newelpa by @pplab in abacusmodeling#60
- Feature: add PDOS with soc by @1041176461 in abacusmodeling#70
- Feature: add class Diag_Scalapack_gvx by @PeizeLin in abacusmodeling#74
- Feature: add opt_orb_pytorch_dpsi by @PeizeLin in abacusmodeling#80
- Feature: Integrate pytorch+dpsi method into original orbital-generation-script by @wszhang in abacusmodeling#86
- Feature: Output SR sparse matrix separately by @jingan-181 in abacusmodeling#95
- Feature: DFT+DMFT and DFT+U by @Qx80610702 in abacusmodeling#98
Performance updates for users:
- Perf: The efficiency of plane-wave code was optimized by updating the CG method in self-consistent calculations. (by Qianruipku before #1 )
- Perf: Improved cache memory for wave function for gamma_only method in LCAO code. (Liuxiaohui123321 dyzheng #367)
- Perf: Improved DFT+U memory cost and efficiency. (Quxin #260 #222 Quxin80610702 #222 ZhouGD2000 #357 #249 )
- Perf: Improved the efficiency of non-local pseudopotential part in Hamiltonian matrix, still in test now. (wenfei-li #352 #354 #365 #372 )
- Perf: LCAO nonlocal force & stress: switch to new method by @wenfei-li in #365
- Perf: LCAO : new method for nonlocalPP, k points by @wenfei-li in #372
- Perf: vnl : make it faster for large system by @wenfei-li in #427
- Perf: deepks : acceleration for multi-k by @wenfei-li in #524
- Perf: by @Qianruipku in #535
- Perf: by @Qianruipku in #540
- Perf: optimize the program by @Qianruipku in #542
- Perf: optimize efficiency by @Qianruipku in #544
- Perf: by @Qianruipku in #548
- Perf: by @Qianruipku in #549
- Perf optimized non-local part for Stress and Forces calculation in PW base by @dyzheng in #559
- Perf: sum_band() in pw base; fix: memory leak by @dyzheng in #586
- Perf: Modified the recommended values for exx-related input parameters in input-main.md by @ouqi0711 in #723
- Perf: optimize hpsi and cgdiag by @Qianruipku in abacusmodeling#9
- Perf: improve cinitcgg function by @Qianruipku in abacusmodeling#6
- Perf: optimize src_lcao/LCAO_gen_fixedH.cpp by @YuLiu98 in abacusmodeling#16
- Perf: 1. optimize Numerical_Basis::cal_overlap_Sq() to gemm by @PeizeLin in abacusmodeling#65
- Perf: OpenMP optimization for local potential and charge density under k-points line by @LiuXiaohui123321 in abacusmodeling#76
- Perf: 1. optimize Write_Wfc_Realspace::write_charge_realspace_1() by @PeizeLin in abacusmodeling#103
- Perf: OpenMP optimization for nonlocal pseudo-potential under gamma-only line by @LiuXiaohui123321 in abacusmodeling#108
Documentation updates for users:
Contributors
amcadmus, pplab, and 25 other contributors
Assets 2
Checkpoint tag for DCU and GPU
What's Changed
- Merge recent abacus changes to deepmodeling by @amcadmus in #1
- Merge recent changes on master by @amcadmus in #2
- update soc class for non-local pseudopotential initial by @dyzheng in #3
-
- add bcast flag_empty_element in Atom::bcast_atom() by @mohanchen in #4
- update ORB module by @mohanchen in #5
- Peize fix a bug in Matrix3, found by Jiyuyang by @mohanchen in #7
- extract common parts in 'Read_PAO' and 'Read_Descriptor' to a new function by @maki49 in #6
- merge last week's developments by @mohanchen in #10
- update files in terms of pseudopotentials by @mohanchen in #11
- divide read_pseudopotentials to three formats: upf100 stands for vers… by @mohanchen in #12
- add vdwd3 by @1041176461 in #13
- remove ucell.label in Read_PAO by @maki49 in #15
- fixed bug of ewald term in STRESS by @dyzheng in #16
- update some modifications for the ORB module by @mohanchen in #17
- modify src_pw/wf_atomic.cpp for pseudopotentials containing f electron by @YuLiu98 in #20
- the ORB module can be compiled successfully by @mohanchen in #22
- deepks: divide S_mu_alpha and PDM into small matrices with same I,n,l by @maki49 in #21
- update for the ORB module by @mohanchen in #23
- delete snap_psialpha by @maki49 in #24
- refactor potential and add DELETE_MUL_PTR() by @PeizeLin in #25
- refactor gint_gamma_vl.cpp by @PeizeLin in #26
- fix a pseudopotential bug by qianrui by @mohanchen in #27
- add src_lcao/gint_gamma_common.cpp and refactor src_lcao/gint_gamma_fvl.cpp by @PeizeLin in #29
- Evolve functions added in evolve.cpp for the future change. by @HeFuxiang94 in #30
- Add using_LAPACK_2 function for next use. by @HeFuxiang94 in #33
- Reconstruction MD: by @dyzheng in #35
- Add four input parameters for the next update use. by @HeFuxiang94 in #34
- merge some fixed from qianrui by @mohanchen in #38
- fix a bug in pw distribution by @mohanchen in #39
- fix bugs for pw by @mohanchen in #40
- add the 'tests' directory to include some autotests by @mohanchen in #41
- fix some bugs, add tests to include autotests by @mohanchen in #42
- Minimally working CMake scripts, Dockerfile and CI by @tansongchen in #47
- Fix codes for GNU compiling by @darelbeida in #61
- Fix TO_STRING not defined. by @shishaochen in #69
- [master] Add cmake config for ABACUS MPI. by @shishaochen in #70
- [develop] Find cereal, ScaLAPACK, ELPA, BLAS in CMake. by @shishaochen in #76
- [develop] Support installation in CMake by @shishaochen in #79
- Reconstruction by @dyzheng in #88
- reconstruct the interface to get gcar and kvec_c + gcar by @dyzheng in #89
- update tests 103_PW_15_CS_CF and 104_PW_FM_magnetic by @YuLiu98 in #95
- fixed result.ref in 101_PW_VW_psuedopot test by @Liu-RX in #93
- update tests 103_PW_CF_CS_S1_smallg and 201_NO_KP_DJ_CF_CS_GaAs by @YuLiu98 in #97
- Reconstruction ：added molecular dynamics in PW by @dyzheng in #99
- cmake compile debug by @dyzheng in #102
- Reconstruction by @Liu-RX in #104
- Update ORB_read.cpp by @Liu-RX in #105
- move math data type files to module_base by @Liu-RX in #106
- update module_cell by @YuLiu98 in #109
- Reconstruction by @Qianruipku in #110
- update module_cell by @YuLiu98 in #111
- Reconstruction by @Liu-RX in #112
- refactor Gint_Gamma and fix bug by @PeizeLin in #113
- Reconstruction by @Liu-RX in #117
- Test develop by @tansongchen in #118
- Adding support for build kits: Clang (12.0.5) + OpenMPI by @tansongchen in #121
- documentation by @wenfei-li in #122
- move cal_nelec from module_cell to src_pw/charge.cpp by @YuLiu98 in #123
- src_pw can compile ABACUS-PW.x by itself by @Qianruipku in #127
- fix logo and syntax by @wenfei-li in #129
- Reconstruction by @Liu-RX in #125
- Reconstruction vdwd2 and vdwd3 by @dyzheng in #131
- Reconstruction by @Qianruipku in #130
- Some modifications on src_pw module by @Qianruipku in #133
- Reconstruction by @Liu-RX in #132
- fix compilation error of cmake by @denghuilu in #134
- Reconstruction by @Qianruipku in #135
- Reconstruction of module_cell by @YuLiu98 in #136
- Improve README by @tansongchen in #137
- Reconstruction by @dyzheng in #138
- Revert "Reconstruction" by @dyzheng in #141
- Improve Autotest.sh by @caic99 in #140
- Reconstruction by @Qianruipku in #139
- Reconstruction: fixed some warning and fixed bug of Autotest by @dyzheng in #142
- optimized default set of Autotest.sh and fixed reference of 114_PW_BD_15 by @dyzheng in #143
- Reconstruction of module_cell is finished by @YuLiu98 in #145
- some minor changes on io & documentation by @wenfei-li in #144
- fixed bug of digital error caused ggchg parallel fatal error by @dyzheng in #146
- update some result.ref in tests by @YuLiu98 in #147
- Warning modified in STRESS by @dyzheng in #148
- Reconstruction by @Qianruipku in #150
- fixed valgrind bugs by @dyzheng in #154
- Reconstruction of module_md by @YuLiu98 in #152
- Adding deepks to CM...
Contributors
amcadmus, mohanchen, and 21 other contributors