新建
从 DFT 先去 DeePKS 再到 DeePMD | DeePKS案例篇 + 增强采样
zhangpengchao@dp.tech
pc-zhang20@foxmail.com
推荐镜像 :Basic Image:bohrium-notebook:2023-04-07
推荐机型 :c2_m4_cpu
赞 9
7
29
目录
数据集
deepks_fundamental_tutorial_gly(v1)
从 DFT 先去 DeePKS 再到 DeePMD | DeePKS案例篇 + 增强采样
代码
文本
©️ Copyright 2023 @ Authors
作者：章鹏超 📨
日期：2023-11
共享协议：本作品采用知识共享署名-非商业性使用-相同方式共享 4.0 国际许可协议进行许可。
代码
文本
🎉 欢迎来到这个关于DeePKS的Notebook！
这是从第一性原理计算到分子动力学模拟的超详细「DeePKS+DeePMD」计算上手指南，主要面向零基础用户。
阅读前，请先了解 DeePKS的基本思想，戳 👉 DeePKS基础篇
代码
文本
在 Bohrium Notebook 界面，你可以点击界面上方蓝色按钮 开始连接，选择 xxx 镜像及 c32_m128_CPU 节点配置，稍等片刻即可运行。
注：镜像可能要制作一个，本工作采用了plumed_v2.8.1_patch (其中额外加入了Voronoi CVs代码,并需要安装OPES模块)，deepmd-kit_v2.1.5 (with lammps and plumed)，deepks-kit_v0.1，abacus_v3.0.5，cp2k_v9.1 (with plumed)。所以目前的notebook还只是输入输出文件介绍，等镜像制作好了再详细更新该notebook...
代码
文本
背景
在本教程中，我们将以甘氨酸分子在纯水溶液中的质子转移现象为例，详细介绍DeePKS模型和DP模型的训练和应用。DeePKS模型是一种基于深度学习校正的泛函，用于精确高效获得训练DP模型的数据集。DP模型是一种基于深度学习形式的力场，用于精确描述原子间相互作用的势能面。一般用DP模型跑MD包含上图右侧的4个步骤，本案例加入了DeePKS，因此增加了2个步骤，对应上图右侧所示，即具体对应以下6节：
代码
文本
代码
文本
代码
文本
1.1 下载教程资源
我们已经在 deepks_fundamental_tutorial_gly 中准备了简化的文件，你可以点击左侧数据集查看（最终用于发表的数据集和模型已上传至AIS Square和Google Drive和Github）：
代码
文本
[48]
# 出于安全考虑，我们没有数据集所在文件夹的写入权限，因此我们将其复制到 `/data/` 目录下:
! cp -r /bohr/1019-sjiy/v1/deepks_fundamental_tutorial_gly /data/
# 我们在这里定义一些路径，并切换到工作路径，方便后续调用：
import os
bohr_dataset_url = "/deepks_fundamental_tutorial_gly" # 数据集的 url 可从左侧数据集复制
work_path = os.path.join("/data", bohr_dataset_url[1:]) # 进行一个切片以删除上述路径中最开始的“/”
os.chdir(work_path)
print(f"当前路径为：{os.getcwd()}")
当前路径为：/data/deepks_fundamental_tutorial_gly
代码
文本
[49]
! ls
01-Sampling_configurations 03-DeePKS_Interation 05-DP_Training 02-DFT_Labeling 04-DeePKS_Labeling 06-DPMD+Enhanced_sampling
代码
文本
[50]
! tree 01-Sampling_configurations
01-Sampling_configurations ├── COLVAR ├── cp2k.inp ├── init.xyz ├── plumed.dat ├── sub.sh └── xyz.dat 0 directories, 6 files
代码
文本
在 01-Sampling_configurations 文件夹包含
输入文件：
cp2k.inp: CP2K 输入文件，用于控制 MD 细节；xyz.dat: 用于存放 MD 模拟的初始构型；plumed.dat: PLUMED 增强采样的输入文件；sub.sh: 提交作业的脚本。
代码
文本
1.2 CP2K跑MD输入文件
输入文件的作用是告诉 CP2K 软件其应该如何计算。关键词以&XXX开始，&END XXX结束，#后面是注释。
下面是一个示例（CP2K手写输入文件入门难度不低，也可选用Multiwfn生成输入文件然后再改）：
代码
文本
[51]
! cat 01-Sampling_configurations/cp2k.inp
&GLOBAL
PROJECT Gly54H2O # Project name
PRINT_LEVEL LOW # Print level
RUN_TYPE MD # Run type
&END GLOBAL
&FORCE_EVAL
METHOD Quickstep # Method
&SUBSYS
&CELL
A 12.02800000 0.00000000 0.00000000 # Lattice
B 0.00000000 12.02800000 0.00000000
C 0.00000000 0.00000000 12.02800000
PERIODIC XYZ # Periodic boundary conditions
&END CELL
&TOPOLOGY
COORD_FILE_FORMAT XYZ # Coordinate file format
COORD_FILE_NAME init.xyz # Initial coordinate file name
&END TOPOLOGY
&END SUBSYS
&DFT
CHARGE 0 # Charge of the system
MULTIPLICITY 1
&QS
EPS_DEFAULT 1E-10 # Leading to an energy correct up to EPS_DEFAULT
EXTRAPOLATION ASPC # Always stable predictor corrector
EXTRAPOLATION_ORDER 3 # Order for the PS or ASPC extrapolation (typically 2-4)
METHOD xTB
&xTB
DO_EWALD T # Use Ewald summation for long-range interactions
CHECK_ATOMIC_CHARGES F # Don't stop calculation if atomic charges are outside chemical range
&PARAMETER # parameter file
DISPERSION_PARAMETER_FILE dftd3.dat
PARAM_FILE_NAME xTB_parameters
&END PARAMETER
&END xTB
&END QS
&POISSON # Periodic boundary conditions for Poisson solver
PERIODIC XYZ
PSOLVER PERIODIC
&END POISSON
&SCF
MAX_SCF 200 # Maximum number of SCF iterations
EPS_SCF 1.0E-05 # SCF convergence criterion
&OT # Sets the various options for the orbital transformation (OT) method.
PRECONDITIONER FULL_SINGLE_INVERSE # Preconditioning method. Based on H-eS cholesky inversion.
MINIMIZER DIIS # Minimization method. Direct inversion in the iterative subspace
LINESEARCH 2PNT # Line search method. extrapolate based on 2 points
&END OT
&PRINT
&RESTART # Controls the dumping of the MO restart file during SCF
BACKUP_COPIES 0 # Specifies the maximum number of backup copies
&END RESTART
&END PRINT
&END SCF
&END DFT
&END FORCE_EVAL
&MOTION
&MD # molecular dynamics
ENSEMBLE NVT # Ensemble type
STEPS 200000 # Number of MD steps
TIMESTEP 1.0 # MD time step 1 fs
TEMPERATURE 300 # Temperature 300 K
&THERMOSTAT
TYPE CSVR # Thermostat type
&CSVR
TIMECON 200 # Thermostat time constant 200 fs
&END CSVR
&END THERMOSTAT
&END MD
&FREE_ENERGY
&METADYN
USE_PLUMED .TRUE. # Use PLUMED for enhanced sampling
PLUMED_INPUT_FILE ./plumed.dat # PLUMED input file location
&END METADYN
&END FREE_ENERGY
&PRINT
&TRAJECTORY
&EACH
MD 10 # Trajectory output frequency
&END EACH
FORMAT xyz # Trajectory output format
&END TRAJECTORY
&VELOCITIES
&EACH
MD 10 # Velocities output frequency
&END EACH
&END VELOCITIES
&RESTART
BACKUP_COPIES 0
&EACH
MD 50000 # Restart output frequency
&END EACH
&END RESTART
&END PRINT
&END MOTION
#&EXT_RESTART
# RESTART_FILE_NAME xxx.restart
#&END EXT_RESTART
代码
文本
为了缩短运行所需的时间，我们适当降低了运行步数。
在采样构象阶段，可以同时跑多个独立的MD模拟加速构象收集过程。本工作跑了40个左右的独立MD，即从不同初始结构和不同随机数出发，根据需要还可以改变温度和压力等参数。
代码
文本
1.3 PLUMED 增强采样收集构象的输入文件
下面是一个示例：
代码
文本
[52]
! cat 01-Sampling_configurations/plumed.dat
# this is using Angstrom UNITS LENGTH=A # set atom groups GlyN: GROUP ATOMS=163 GlyOhasH: GROUP ATOMS=171 GlyOnotH: GROUP ATOMS=169 GlyH: GROUP ATOMS=164,170,172,2 WaterO: GROUP ATOMS=1-160:3 WaterH: GROUP ATOMS=5-161:3,3-162:3 # define voronoi CVs s05: VORONOIS1 GROUPA=WaterO,GlyN,GlyOhasH,GlyOnotH GROUPB=WaterH,GlyH NRX=3 LAMBDA=-5 D_0=108 D_1=3 NLIST NL_CUTOFF=2.4 NL_STRIDE=1 d05: VORONOID2 GROUPA=WaterO,GlyN,GlyOhasH,GlyOnotH GROUPB=WaterH,GlyH NRX=3 LAMBDA=-5 D_0=2 D_1=2 D_2=0.5 D_3=0.5 NLIST NL_CUTOFF=2.4 NL_STRIDE=1 # define opes-explore OPES_METAD_EXPLORE ... RESTART=NO LABEL=opes ARG=s05,d05 FILE=HILLS # a file in which the list of all deposited kernels is stored TEMP=300 # temperature(K). If not set, it is taken from MD engine PACE=500 # the frequency for kernel deposition BARRIER=80 # the free energy barrier to be overcome STATE_WFILE=STATE # write to this file the compressed kernels and all the info needed to RESTART the simulation STATE_WSTRIDE=5000 # number of MD steps between writing the STATE_WFILE STORE_STATES # append to STATE_WFILE instead of ovewriting it each time WALKERS_MPI # switch on MPI version of multiple walkers ... OPES_METAD_EXPLORE # print CVs and bias FLUSH STRIDE=100 PRINT ... ARG=s05,d05,*.bias # the outputs you want to print STRIDE=1 # number of MD steps between writing ARGs FILE=COLVAR # a file in which the ARGs is stored RESTART=NO ... PRINT ENDPLUMED
代码
文本
这里所用的集合变量（collective variables, CVs），简单来说是为了将我们所研究的质子转移坐标从高维空间降维，以便可视化质子转移反应的自由能面。基于CV的增强采样，即增强了对所设反应坐标的采样，这样可以大幅缩短模拟时间。此处涉及的篇幅可能过长，我们会另写notebook具体介绍on-the-fly probability enhanced sampling (OPES)与Voronoi CVs
代码
文本
1.4 初始结构的输入文件
初始结构文件init.xyz需要在cp2k.inp中指定，CV所涉及的原子索引号则需要plumed.dat中指定。当然也可以用.pdb格式，对应修改cp2k.inp中的&TOPOLOGY
下面是一个172原子体系的示例（包含，只展示了前10个原子）：
代码
文本
[53]
! head -n 12 01-Sampling_configurations/init.xyz
172 init xyz O 4.299293 4.776775 9.737641 H 6.775417 5.071937 0.136651 H 4.246821 3.309147 1.263330 O 11.063144 10.339872 1.599614 H 9.168386 1.532002 2.761902 H 0.739205 4.121296 7.187894 O 8.877948 5.856919 7.472394 H 8.492142 11.270641 3.570705 H 5.342224 1.546288 3.804414 O 10.399652 8.269867 5.802673
代码
文本
1.5 运行 GFN1-xTB 级别下的分子动力学模拟
笔者所用的是PBS作业调度系统，执行qsub sub.sh运行作业。
主要的输出文件（文件较大只截取了部分作为示例）有：
xyz.dat: 结构轨迹文件；COLVAR: 记录模拟时间、CV值s05和d05、以及其他增强采样输出参数（此处为偏置势opes.bias）的文件。
代码
文本
[54]
! head -n 12 01-Sampling_configurations/xyz.dat
172 frame 1 O 4.610098 4.684042 9.775987 H 6.329322 4.861331 11.283769 H 4.148137 3.248957 0.906027 O 11.552249 10.303507 1.766647 H 9.560211 1.320587 2.804950 H 1.338245 3.731550 7.389257 O 8.983176 5.859800 7.711576 H 8.339740 11.099571 3.989882 H 5.054429 2.055170 3.794269 O 10.508759 8.508986 5.898236
代码
文本
[55]
! cat 01-Sampling_configurations/COLVAR
#! FIELDS time s05 d05 opes.bias 0.001000 -0.001057 3.023400 -80.000000 0.002000 -0.000282 3.020795 -80.000000 0.003000 0.000240 3.015184 -80.000000 0.004000 0.000431 3.010100 -80.000000 0.005000 0.000129 3.007531 -80.000000 0.006000 -0.000583 3.006861 -80.000000 0.007000 -0.001239 3.008498 -80.000000 0.008000 -0.001364 3.014579 -80.000000 0.009000 -0.000905 3.024676 -80.000000 0.010000 -0.000156 3.034457 -80.000000 0.011000 0.000601 3.039903 -80.000000 199.989000 -0.001910 1.734044 92.918086 199.990000 -0.001846 1.834772 95.331408 199.991000 -0.001584 1.881821 96.663240 199.992000 -0.001157 1.867308 96.241737 199.993000 -0.000667 1.792590 94.243802 199.994000 -0.000261 1.673940 91.796998 199.995000 -0.000062 1.558311 90.434774 199.996000 -0.000083 1.510852 90.201213 199.997000 -0.000179 1.555498 90.417191 199.998000 -0.000209 1.646575 91.377599 199.999000 -0.000180 1.710271 92.440309 200.000000 -0.000139 1.696906 92.191787
代码
文本
在得到以上输出文件之后，我们可以根据CV值来选取组成数据集的构象，这些构象需要尽可能涵盖我们研究问题对应的构象空间。如下图所示，数据集中的每个构象对应图中的一个点，基本覆盖了最后的自由能面。然而自由能面我们在初始时并不知道，所以这个过程可能需要迭代多轮才能收集充分的构象数量。
代码
文本
代码
文本
2.1 下载教程资源
在本章节中，我们已经在 02-DFT_Labeling 中准备了需要的文件。
输入文件有
m062x.inp：设定单点计算；geo.xyz：所选结构；sub.sh：提交脚本；
代码
文本
[56]
! tree 02-DFT_Labeling
02-DFT_Labeling ├── geo.xyz ├── m062x.inp ├── m062x.out └── sub.sh 0 directories, 4 files
代码
文本
与章节1不同的是我们更换了计算任务类型为单点计算，即通过自洽场迭代（SCF）获取能量和受力。
以及所选计算级别换成了高精度方法，当然更耗时，这也是我们后面采用DeePKS的原因。
代码
文本
[57]
! cat 02-DFT_Labeling/m062x.inp
&GLOBAL
PROJECT M062X
PRINT_LEVEL LOW
RUN_TYPE ENERGY_FORCE # print energy and force
&END GLOBAL
&FORCE_EVAL
METHOD Quickstep
&PRINT
&FORCES ON
&END FORCES
&END PRINT
&SUBSYS
&CELL
A 12.02800000 0.00000000 0.00000000
B 0.00000000 12.02800000 0.00000000
C 0.00000000 0.00000000 12.02800000
PERIODIC XYZ
&END CELL
&TOPOLOGY
COORD_FILE_FORMAT XYZ
COORD_FILE_NAME geo.xyz
&END TOPOLOGY
&KIND O # basis sets
ELEMENT O
BASIS_SET QZV3P-GTH-q6
BASIS_SET AUX_FIT admm-dzp-q6
POTENTIAL GTH-PBE
&END KIND
&KIND H
ELEMENT H
BASIS_SET QZV3P-GTH-q1
BASIS_SET AUX_FIT admm-dzp-q1
POTENTIAL GTH-PBE
&END KIND
&KIND N
ELEMENT N
BASIS_SET QZV3P-GTH-q5
BASIS_SET AUX_FIT admm-dzp-q5
POTENTIAL GTH-PBE
&END KIND
&KIND C
ELEMENT C
BASIS_SET QZV3P-GTH-q4
BASIS_SET AUX_FIT admm-dzp-q4
POTENTIAL GTH-PBE
&END KIND
&END SUBSYS
&DFT
BASIS_SET_FILE_NAME GTH_BASIS_SETS # Specifies the basis set
BASIS_SET_FILE_NAME BASIS_ADMM_UZH # ADMM calculations
POTENTIAL_FILE_NAME POTENTIAL # Specifies the file containing the potential
#WFN_RESTART_FILE_NAME ../pbe/PBE-RESTART.wfn # Specifies the path to the wavefunction restart file
CHARGE 0
MULTIPLICITY 1
&QS
EPS_DEFAULT 1.0E-11 # leading to an energy correct up
EPS_PGF_ORB 1E-12 # Sets precision of the overlap matrix elements
&END QS
&POISSON
PERIODIC XYZ
PSOLVER PERIODIC
&END POISSON
&AUXILIARY_DENSITY_MATRIX_METHOD
METHOD BASIS_PROJECTION # Construct auxiliary density matrix from auxiliary basis
ADMM_PURIFICATION_METHOD MO_DIAG # Calculate MO derivatives via Cauchy representation by diagonalization
&END AUXILIARY_DENSITY_MATRIX_METHOD
&XC
&XC_FUNCTIONAL
&LIBXC
FUNCTIONAL MGGA_C_M06_2X # Specifies the correlation (C) part
&END LIBXC
&LIBXC
FUNCTIONAL HYB_MGGA_X_M06_2X # Specifies the exchange (X) part
&END LIBXC
&END XC_FUNCTIONAL
&HF
FRACTION 0.54 # Specifies the fraction of exact exchange included in the HF exchange
&SCREENING
EPS_SCHWARZ 1E-10 # Screens the near field part of the electronic repulsion integrals using the Schwarz inequality for the given threshold
SCREEN_ON_INITIAL_P T # Screen on an initial density matrix
&END SCREENING
&INTERACTION_POTENTIAL
POTENTIAL_TYPE TRUNCATED # runcated coulomb potential: if (r < R_c) 1/r else 0
CUTOFF_RADIUS 6.0 # Determines cutoff radius (in Angstroms) for the truncated 1/r potential
&END INTERACTION_POTENTIAL
&MEMORY
MAX_MEMORY 3000 # Defines the maximum amount of memory [MB]
EPS_STORAGE_SCALING 0.1 # Scaling factor to scale eps_schwarz.
&END MEMORY
&END HF
&END XC
&MGRID
CUTOFF 1000 # planewave cutoff (default unit is in Ry) for the finest level of the multi-grid
REL_CUTOFF 70 # planewave cutoff of a reference grid covered by a Gaussian with unit standard deviation
&END MGRID
&SCF
MAX_SCF 50 # Maximum number of SCF iteration to be performed for one optimization
EPS_SCF 1.0E-07 # Target accuracy for the SCF convergence
SCF_GUESS RESTART # Use the RESTART file as an initial guess (and ATOMIC if not present)
&OT
PRECONDITIONER FULL_ALL # This preconditioner is recommended for almost all systems
MINIMIZER DIIS # Direct inversion in the iterative subspace: less reliable than CG, but sometimes about 50% faster
LINESEARCH 2PNT # extrapolate based on 2 points
&END OT
&OUTER_SCF
MAX_SCF 10 # The maximum number of outer loops
EPS_SCF 1.0E-07 # The target gradient of the outer SCF variables.
&END OUTER_SCF
&PRINT
&RESTART
BACKUP_COPIES 0
&END RESTART
&END PRINT
&END SCF
&END DFT
&END FORCE_EVAL
代码
文本
值得注意的是我们需要对CUTOFF进行收敛测试，如下表所示，我们最终选择CUTOFF=1000，REL_CUTOFF=70的组合
代码
文本
2.2 运行 M06-2X 级别下的单点计算
笔者所用的是PBS作业调度系统，执行qsub sub.sh运行作业。
主要的输出文件（文件较大只截取了部分作为示例）有：
m062x.out: 记录SCF收敛后的能量和受力文件；
代码
文本
[58]
#能量
! grep 'ENERGY| Total FORCE_EVAL' 02-DFT_Labeling/m062x.out
ENERGY| Total FORCE_EVAL ( QS ) energy [a.u.]: -987.850379144107478
代码
文本
[59]
#打印前3个原子受力
!grep -A 5 'ATOMIC FORCES in' 02-DFT_Labeling/m062x.out
ATOMIC FORCES in [a.u.]
# Atom Kind Element X Y Z
1 1 O 0.01317008 -0.02281718 -0.01692091
2 2 H 0.01592927 -0.01750254 -0.02033401
3 2 H -0.02224085 -0.00953753 0.01490330
代码
文本
3.1 下载教程资源
在本章节中，我们已经在 03-DeePKS_Interation 中准备了需要的文件。
输入文件有
machines.yaml：机器配置；params.yaml：训练DeePKS模型的参数；scf_abacus.yaml：SCF迭代参数，其输出的密度矩阵、轨道等信息会额外作为DeePKS模型的输入；systems.yaml：数据集设定；sub.sh：提交脚本；
代码
文本
[60]
! tree 03-DeePKS_Interation
03-DeePKS_Interation ├── glycine │ ├── iter │ │ ├── RECORD │ │ ├── iter.14 │ │ │ ├── 00.scf │ │ │ │ ├── data_train │ │ │ │ │ └── 001 │ │ │ │ │ ├── conv.npy │ │ │ │ │ ├── dm_eig.npy │ │ │ │ │ ├── e_base.npy │ │ │ │ │ ├── e_tot.npy │ │ │ │ │ ├── energy.npy │ │ │ │ │ ├── f_base.npy │ │ │ │ │ ├── f_tot.npy │ │ │ │ │ ├── force.npy │ │ │ │ │ ├── grad_vx.npy │ │ │ │ │ ├── l_e_delta.npy │ │ │ │ │ └── l_f_delta.npy │ │ │ │ ├── log.data │ │ │ │ └── model.ptg │ │ │ └── 01.train │ │ │ ├── log.train │ │ │ └── model.pth │ │ ├── machines.yaml │ │ ├── params.yaml │ │ ├── scf_abacus.yaml │ │ ├── sub.sh │ │ └── systems.yaml │ ├── projector │ │ ├── INPUT │ │ ├── KPT │ │ ├── STRU │ │ ├── jle.orb │ │ └── sub.sh │ └── systems │ ├── 001 │ │ ├── atom.npy │ │ ├── energy.npy │ │ └── force.npy │ └── 002 │ ├── atom.npy │ ├── energy.npy │ └── force.npy ├── orbital_dir │ ├── C_gga_7au_100Ry_2s2p1d.orb │ ├── H_gga_6au_100Ry_2s1p.orb │ ├── N_gga_7au_100Ry_2s2p1d.orb │ └── O_gga_7au_100Ry_2s2p1d.orb └── pseudo_dir ├── C_ONCV_PBE-1.0.upf ├── H_ONCV_PBE-1.0.upf ├── N_ONCV_PBE-1.0.upf └── O_ONCV_PBE-1.0.upf 13 directories, 40 files
代码
文本
3.2 机器配置文件
参考DeePKS-kit文档，以下是笔者的本地机器配置文件machines.yaml示例，使用bohrium平台的机器配置文件待写中...
代码
文本
[61]
! cat 03-DeePKS_Interation/glycine/iter/machines.yaml
# this is only part of input settings.
# should be used together with systems.yaml and params.yaml
scf_machine:
resources:
numb_node: 1 # int; number of nodes; default value is 1
task_per_node: 20 # int; ppn required; default value is 1;
numb_gpu: 0 # int; number of GPUs; default value is 1
time_limit: "100:00:00" # time limit; default value is 1:0:0
mem_limit: 20 # int; memeory limit in GB
partition: "normal2" # string; queue name
module_list: [anaconda/anaconda.2020.02] # e.g., [abacus]
source_list: [activate abacus_3.0.5] # e.g., [/opt/intel/oneapi/setvars.sh; conda activate deepks]
envs: {'OMP_NUM_THREADS': '1'}
group_size: 20
sub_size: 1 # keyword for PySCF; set to 1 for ABACUS SCF jobs
dispatcher:
context: local # "local" to run on local machine, or "ssh" to run on a remote machine
batch: "pbs" # set to shell to run on local machine, you can also use `slurm` or `pbs`
train_machine:
resources:
numb_node: 1 # int; number of nodes; default value is 1
task_per_node: 28 # int; ppn required; default value is 1;
#numb_gpu: # int; number of GPUs; default value is 1
time_limit: "100:00:00" # time limit; default value is 1:0:0
mem_limit: 30 # int; memeory limit in GB
partition: "normal4" # string; queue name
module_list: [anaconda/anaconda.2020.02] # e.g., [abacus]
source_list: [activate deepkskit_0.1] # e.g., [/opt/intel/oneapi/setvars.sh; conda activate deepks]
envs: {'OMP_NUM_THREADS': '1'}
dispatcher:
context: local # "local" to run on local machine, or "ssh" to run on a remote machine
batch: "pbs" # set to shell to run on local machine, you can also use `slurm` or `pbs`
remote_profile: null # use lazy local
# resources are no longer needed, and the task will use gpu automatically if there is one.
python: "python" # use python in path
# other settings (these are default; can be omitted)
cleanup: false # whether to delete slurm and err files
strict: true # do not allow undefined machine parameters
#paras for abacus
use_abacus: true # use abacus in scf calculation代码
文本
3.3 训练参数文件
以下是训练DeePKS模型的参数文件params.yaml示例，本案例只作为练习，只把迭代论述设为2，实际情况要迭代更多次，以及训练超参数可以根据情况拉高。
代码
文本
[62]
! cat 03-DeePKS_Interation/glycine/iter/params.yaml
# this is only part of input settings.
# should be used together with systems.yaml and machines.yaml
# number of iterations to do, can be set to zero for DeePHF training
n_iter: 2
# directory setting (these are default choices, can be omitted)
workdir: "."
share_folder: "share" # folder that stores all other settings
# scf settings, set to false when n_iter = 0 to skip checking
scf_input: false
# train settings for training after init iteration,
# set to false when n_iter = 0 to skip checking
train_input:
# model_args is omitted, which will inherit from init_train
data_args:
batch_size: 1 # training batch size; 16 is recommended
group_batch: 1 # number of batches to be grouped; set to 1 for ABACUS-related training
extra_label: true # set to true to train the model with force, stress, or bandgap labels.
# note that these extra labels will only be included after the init iteration
# only energy label will be included for the init training
conv_filter: true # if set to true (recommended), will read the convergence data from conv_name
# and only use converged datapoints to train; including any unconverged
# datapoints may screw up the training!
conv_name: conv # npy file that records the converged datapoints
preprocess_args:
preshift: false # restarting model already shifted. Will not recompute shift value
prescale: false # same as above
prefit_ridge: 1e1 # the ridge factor used in linear regression
prefit_trainable: false # make the linear regression fixed during the training
train_args:
# start learning rate (lr) will decay a factor of `decay_rate` every `decay_steps` epoches
decay_rate: 0.5
decay_steps: 1000
display_epoch: 100 # show training results every n epoch
force_factor: 1 # the prefactor multiplied infront of the force part of the loss
n_epoch: 5000 # total number of epoch needed in training
start_lr: 0.0001 # the start learning rate, will decay later
# init training settings, these are for DeePHF task
init_model: false # do not use existing model to restart from
init_scf: True # whether to perform init SCF;
init_train: # parameters for init nn training; basically the same as those listed in train_input
model_args:
hidden_sizes: [32, 32, 32] # neurons in hidden layers [100, 100, 100]
output_scale: 100 # the output will be divided by 100 before compare with label
use_resnet: true # skip connection
actv_fn: mygelu # same as gelu, support force calculation
data_args:
batch_size: 1
group_batch: 1
preprocess_args:
preshift: true # shift the descriptor by its mean
prescale: false # scale the descriptor by its variance (can cause convergence problem)
prefit_ridge: 1e1 # do a ridge regression as prefitting
prefit_trainable: false
train_args:
decay_rate: 0.96
decay_steps: 500
display_epoch: 100
n_epoch: 5000
start_lr: 0.0003代码
文本
3.4 SCF迭代参数文件
scf_abacus.yaml文件是为生成ABACUS输入文件做准备，ABACUS需要如下 5 种文件：
INPUT：包含了计算过程中所需的各种参数，定义和控制计算任务，其中我们以PBE泛函为baseline；STRU：结构文件，包含了原子种类、原子位置、晶格常数以及晶格向量等信息；KPT：包含了布里渊区积分所需的k点信息；*.upf：包含了原子的赝势信息；*.orb：包含了原子轨道的数值表示；
代码
文本
[63]
! cat 03-DeePKS_Interation/glycine/iter/scf_abacus.yaml
scf_abacus: ntype: 4 # int; number of different atom species in this calculations, e.g., 2 for H2O nbands: 277 # int; number of bands to be calculated; optional ecutwfc: 100 # real; energy cutoff, unit: Ry scf_thr: 1e-7 # real; SCF convergence threshold for density error; 5e-7 and below is acceptable scf_nmax: 48 # int; maximum SCF iteration steps dft_functional: "pbe" # string; name of the baseline density functional gamma_only: 1 # bool; 1 for gamma-only calculation cal_force: 1 # bool; 1 for force calculation cal_stress: 0 # bool; 1 for stress calculation orb_files: ["H_gga_6au_100Ry_2s1p.orb","O_gga_7au_100Ry_2s2p1d.orb","N_gga_7au_100Ry_2s2p1d.orb","C_gga_7au_100Ry_2s2p1d.orb"] pp_files: ["H_ONCV_PBE-1.0.upf", "O_ONCV_PBE-1.0.upf", "N_ONCV_PBE-1.0.upf", "C_ONCV_PBE-1.0.upf" ] proj_file: ["jle.orb"] lattice_constant: 1.8897259886 lattice_vector: [[12.028,0,0], [0,12.028,0], [0,0,12.028]] coord_type: "Cartesian" run_cmd : "mpirun" abacus_path: "abacus" init_scf_abacus: orb_files: ["H_gga_6au_100Ry_2s1p.orb","O_gga_7au_100Ry_2s2p1d.orb","N_gga_7au_100Ry_2s2p1d.orb","C_gga_7au_100Ry_2s2p1d.orb"] pp_files: ["H_ONCV_PBE-1.0.upf", "O_ONCV_PBE-1.0.upf", "N_ONCV_PBE-1.0.upf", "C_ONCV_PBE-1.0.upf" ] proj_file: ["jle.orb"] ntype: 4 nbands: 277 ecutwfc: 100 scf_thr: 1e-5 scf_nmax: 64 dft_functional: "pbe" gamma_only: 1 cal_force: 0 lattice_constant: 1.8897259886 lattice_vector: [[12.028,0,0], [0,12.028,0], [0,0,12.028]] coord_type: "Cartesian" run_cmd : "mpirun" abacus_path: "abacus"
代码
文本
3.5 投影子轨道的生成
为获得如上图所示的DeePKS描述子，需构造投影密度矩阵，需要先构造DeePKS中使用的投影子，其有如下形式：
为截断半径(推荐5 Bohr)， 的选取需保证，其上限由动能的cutoff决定(推荐与SCF保持一致)，同时的取值也决定了径向球贝塞尔函数的个数。球谐函数决定了角向部分，对于每个球贝塞尔函数，若角量子数，则包含三种角向轨道， 生成个描述子。每个元素总共的描述子个数则为。
代码
文本
[64]
! cat 03-DeePKS_Interation/glycine/projector/INPUT
INPUT_PARAMETERS #Parameters (1.General) suffix abacus pseudo_dir ../../pseudo_dir orbital_dir ../../orbital_dir calculation gen_bessel # calculation type should be gen_bessel ntype 4 nbands 277 symmetry 0 #Parameters (2.Iteration) ecutwfc 100 # kinetic energy cutoff in unit Ry; should be consistent with that set for ABACUS SCF calculation scf_thr 1e-8 scf_nmax 128 #Parameters (3.Basis) basis_type pw gamma_only 1 #Parameters (4.Smearing) smearing_method gaussian smearing_sigma 0.1 #Parameters (5.Mixing) mixing_type pulay mixing_beta 0.4 #Parameters (6. Bessel function) bessel_lmax 2 # maximum angular momentum for projectors; 2 is recommended bessel_rcut 5 # radial cutoff in unit Bohr; 5 or 6 is recommended bessel_tol 1.0e-12
代码
文本
在projector文件夹下，qsub sub.sh提交任务，完成后即可得到投影子轨道文件jle.orb
代码
文本
[65]
! head -n 20 03-DeePKS_Interation/glycine/projector/jle.orb
---------------------------------------------------------------------------
Energy Cutoff(Ry) 100
Radius Cutoff(a.u.) 5
Lmax 2
Number of Sorbitals--> 15
Number of Porbitals--> 15
Number of Dorbitals--> 15
---------------------------------------------------------------------------
SUMMARY END
Mesh 505
dr 0.01
Type L N
0 0 0
1.000000000000e+00 9.999934202767e-01 9.999736812627e-01 9.999407834256e-01
9.998947275446e-01 9.998355147105e-01 9.997631463261e-01 9.996776241054e-01
9.995789500740e-01 9.994671265693e-01 9.993421562398e-01 9.992040420456e-01
9.990527872577e-01 9.988883954587e-01 9.987108705420e-01 9.985202167122e-01
9.983164384844e-01 9.980995406847e-01 9.978695284498e-01 9.976264072266e-01
9.973701827725e-01 9.971008611549e-01 9.968184487513e-01 9.965229522488e-01
代码
文本
3.6 数据集指定文件
systems.yaml文件中指定了训练集和测试集的路径，拟合的目标就是数据集中的M06-2X级别下的能量和受力标签。
代码
文本
[66]
! cat 03-DeePKS_Interation/glycine/iter/systems.yaml
# this is only part of input settings. # should be used together with params.yaml and machines.yaml # training and testing systems systems_train: # can also be files that containing system paths - ../systems/001 # support glob systems_test: # if empty, use the last system of training set - ../systems/002
代码
文本
3.7 运行 DeePKS interation 及其文件输出
在iter文件夹下，执行qsub sub.sh运行作业，工作流思路如下图
代码
文本
输出文件（文件较多只截取了部分作为示例）有：
流程监测文件：iter/RECORD
生成时间点：从程序运行开始直至结束，每一步骤完成均在RECORD内有记录，文件说明：
- (X 0 0): 每个iteration预处理，生成ABACUS SCF所需的输入文件。X=0对应iter.init, X=1对应iter.00，以此类推
- (X 0 1): 运行SCF计算
- (X 0 2): 检查SCF结果并进行当前误差统计log.data
- (X 0): 当前SCF结束
- (X 1 0): 基于SCF计算结果，本轮模型训练开始，学习曲线记录于log.train
- (X 1 1): 本轮模型训练完成
- (X 1): 模型精度测试
- (X): 本轮流程结束
代码
文本
以迭代至iter.14为例，主要的输出文件有：
data_train/*.npy: 当前的训练集，其中grad_vx.npy用于计算受力校正项，因此文件较大（此处的grad_vx.npy并不是真实的，仅作示例）；log.data: 当前结果误差统计；log.train: 模型训练学习曲线文件；model.pth: 模型文件，DeePKS-kit神经网络输出，本轮模型训练完成时生成；model.ptg: 模型文件，传给ABACUS；c++可读，下轮SCF任务开始时生成；
代码
文本
[67]
! cat 03-DeePKS_Interation/glycine/iter/iter.14/00.scf/log.data
Training:
Convergence:
122 / 150 = 0.81333
Energy:
ME: 0.00034515150545378975
MAE: 0.0006003819333447306
MARE: 0.00034084577732682096
Force:
MAE: 0.0006016773367485338
Testing:
Convergence:
22 / 30 = 0.73333
Energy:
ME: 0.001293148748713217
MAE: 0.002438159172519742
MARE: 0.002341818484132394
Force:
MAE: 0.0006044441443446342
代码
文本
值得注意的是并不是所有单点计算都收敛到设定值，说明在哈密顿量中加入神经网络修正后，一定程度上会影响收敛稳定性
代码
文本
[68]
! head -n 15 03-DeePKS_Interation/glycine/iter/iter.14/01.train/log.train
# using seed: 4032076500
# load 5 systems with fields {'lb_e', 'gvx', 'lb_f', 'eig'}
# load 1 systems with fields {'lb_e', 'gvx', 'lb_f', 'eig'}
# working on device: cpu
# epoch trn_err tst_err lr trn_time tst_time
0 1.08e-02 1.25e-02 1.00e-04 0.00 2.07
100 8.46e-04 3.08e-03 1.00e-04 6.52 0.02
200 9.10e-04 2.91e-03 1.00e-04 6.37 0.02
300 8.33e-04 3.09e-03 1.00e-04 6.59 0.02
400 8.53e-04 2.99e-03 1.00e-04 6.42 0.02
500 8.45e-04 3.13e-03 1.00e-04 6.36 0.02
600 8.44e-04 3.06e-03 1.00e-04 6.53 0.02
700 8.52e-04 3.25e-03 1.00e-04 6.50 0.02
800 9.66e-04 2.97e-03 1.00e-04 6.19 0.02
900 8.74e-04 3.11e-03 1.00e-04 6.23 0.02
代码
文本
3.8 模型测试
测试集来自独立的构象，并在M06-2X级别下完成计算。本工作中DeePKS模型的测试误差如下图
代码
文本
代码
文本
4.1 SCF迭代参数文件
在本章节中，我们已经在 04-DeePKS_Labeling 中准备了需要的文件。
输入文件与ABACUS进行SCF计算类似，有
INPUT：包含了计算过程中所需的各种参数，定义和控制计算任务；STRU：结构文件，包含了原子种类、原子位置、晶格常数以及晶格向量等信息；KPT：包含了布里渊区积分所需的k点信息；*.upf：包含了原子的赝势信息；*.orb：包含了原子轨道的数值表示；
代码
文本
[69]
! tree 04-DeePKS_Labeling
04-DeePKS_Labeling ├── INPUT ├── KPT ├── STRU ├── running_scf.log └── sub.sh 0 directories, 5 files
代码
文本
- INPUT 文件。INPUT中的参数设置于章节3中的SCF计算一致，其中我们以PBE泛函为baseline，并指定DeePKS模型路径（其参与到SCF计算）
代码
文本
[70]
! cat 04-DeePKS_Labeling/INPUT
INPUT_PARAMETERS calculation scf ntype 4 ecutwfc 100.000000 scf_thr 1.000000e-07 scf_nmax 18 basis_type lcao dft_functional pbe gamma_only 1 mixing_type pulay mixing_beta 0.400000 symmetry 0 nbands 277 nspin 1 smearing_method fixed smearing_sigma 0.001000 cal_force 1 cal_stress 0 deepks_out_labels 0 deepks_scf 1 deepks_bandgap 0 deepks_model ../03-DeePKS_Interation/glycine/iter/iter.14/00.scf/model.ptg
代码
文本
- STRU 文件。STRU 文件包含了原子种类、原子位置、晶格常数以及晶格向量等信息。并指定了赝势、轨道、投影子的路径：
代码
文本
[71]
! cat 04-DeePKS_Labeling/STRU
ATOMIC_SPECIES H 1.00 ../03-DeePKS_Interation/pseudo_dir/H_ONCV_PBE-1.0.upf O 1.00 ../03-DeePKS_Interation/pseudo_dir/O_ONCV_PBE-1.0.upf N 1.00 ../03-DeePKS_Interation/pseudo_dir/N_ONCV_PBE-1.0.upf C 1.00 ../03-DeePKS_Interation/pseudo_dir/C_ONCV_PBE-1.0.upf LATTICE_CONSTANT 1.8897259886 LATTICE_VECTORS 12.028 0.0 0.0 0.0 12.028 0.0 0.0 0.0 12.028 ATOMIC_POSITIONS Cartesian H 0.0 113 8.477512 11.052771 4.986609 0 0 0 0.713460 6.625312 7.234243 0 0 0 8.379590 10.521278 7.850620 0 0 0 5.126922 0.000242 3.475644 0 0 0 5.780297 10.219974 2.802691 0 0 0 0.476735 10.084497 1.767674 0 0 0 0.458904 9.171061 4.329539 0 0 0 7.225451 2.122528 5.401010 0 0 0 8.518972 6.703912 4.615551 0 0 0 9.529254 11.500578 10.321160 0 0 0 0.237500 10.445055 8.469310 0 0 0 6.736948 11.328984 10.410131 0 0 0 0.479657 9.392328 9.566849 0 0 0 0.743417 6.521008 1.126826 0 0 0 2.709654 7.316965 3.996281 0 0 0 6.454134 9.097445 4.747141 0 0 0 6.707681 7.055103 1.290373 0 0 0 4.720297 0.656670 7.583141 0 0 0 5.875757 0.301229 9.642776 0 0 0 9.305656 3.528874 1.531390 0 0 0 6.066634 7.529630 10.937747 0 0 0 0.347528 5.176702 9.186558 0 0 0 2.696763 10.177609 0.260227 0 0 0 8.680944 8.228310 9.318817 0 0 0 0.579454 0.935519 0.788905 0 0 0 1.143323 5.288025 5.538924 0 0 0 3.929066 5.677605 11.449092 0 0 0 5.886518 2.603033 5.440133 0 0 0 6.054872 4.291443 1.049866 0 0 0 4.555193 3.744698 1.583451 0 0 0 4.802487 11.142155 5.690429 0 0 0 0.507386 7.838729 11.326568 0 0 0 9.800293 8.220575 2.769882 0 0 0 10.356333 10.709876 9.082552 0 0 0 7.517673 1.630213 2.454091 0 0 0 10.937224 5.110278 8.748446 0 0 0 4.358699 3.022382 7.190637 0 0 0 11.152841 9.544364 6.671887 0 0 0 11.204836 10.273087 2.502405 0 0 0 1.540861 11.170513 0.394901 0 0 0 3.320854 9.094880 6.172404 0 0 0 8.280992 9.110674 11.389585 0 0 0 0.466385 7.766830 8.202456 0 0 0 6.651360 2.775823 2.918653 0 0 0 4.795009 4.517778 3.560350 0 0 0 4.635582 7.018159 1.785197 0 0 0 5.946291 1.624960 7.277583 0 0 0 9.156671 4.873720 1.066682 0 0 0 3.819360 5.297258 9.931563 0 0 0 9.659438 0.706990 0.860768 0 0 0 6.913244 11.921111 0.456257 0 0 0 7.156819 7.070028 4.976803 0 0 0 5.016958 1.802284 1.805349 0 0 0 7.764881 5.068147 6.217623 0 0 0 5.111135 7.226218 8.168274 0 0 0 7.180951 6.634584 8.029705 0 0 0 2.935961 7.202505 0.139212 0 0 0 2.983514 4.305543 4.988204 0 0 0 9.496047 6.084994 6.708427 0 0 0 6.869145 9.053115 11.949920 0 0 0 8.755458 10.096845 2.971789 0 0 0 5.283078 8.392680 5.422639 0 0 0 11.491615 5.123042 1.078104 0 0 0 6.983433 11.792354 2.016835 0 0 0 6.863919 8.072224 7.606946 0 0 0 3.944497 2.006696 4.147048 0 0 0 0.339058 6.263888 4.693703 0 0 0 4.748745 5.910080 7.958939 0 0 0 4.941720 10.549195 0.205023 0 0 0 3.581996 4.910788 6.184750 0 0 0 1.250838 7.218697 2.573446 0 0 0 10.184779 7.260062 11.392811 0 0 0 7.382149 11.245316 6.815370 0 0 0 11.864153 7.494446 1.962388 0 0 0 2.774976 4.651631 8.086785 0 0 0 2.521530 9.833419 2.339826 0 0 0 4.377120 8.277041 2.593911 0 0 0 8.887805 9.317310 5.989867 0 0 0 0.705336 8.138920 5.460177 0 0 0 11.688733 7.109435 10.252553 0 0 0 4.678569 2.712175 11.164536 0 0 0 11.146048 9.117496 0.425415 0 0 0 9.706977 9.148640 10.010559 0 0 0 5.219004 10.406819 6.995871 0 0 0 8.807673 0.047341 3.750322 0 0 0 11.927291 10.112556 11.548690 0 0 0 3.789981 0.940396 5.262649 0 0 0 2.726245 6.123123 8.446609 0 0 0 2.837328 5.908180 3.502186 0 0 0 1.127986 0.197139 2.015033 0 0 0 7.214744 4.217977 7.299943 0 0 0 10.784087 6.421259 6.149549 0 0 0 10.973170 12.027130 0.662913 0 0 0 9.772116 8.161264 5.498056 0 0 0 5.211637 1.411018 11.076792 0 0 0 0.056527 5.356343 11.745871 0 0 0 6.370808 8.842716 2.391141 0 0 0 2.997955 7.443466 10.603126 0 0 0 7.471695 6.611041 2.515934 0 0 0 3.515982 1.625137 1.842416 0 0 0 3.978843 9.871054 2.296667 0 0 0 4.775685 7.811378 11.619961 0 0 0 9.480505 7.035846 0.647082 0 0 0 8.868067 10.134663 1.584573 0 0 0 10.735825 7.482301 3.744403 0 0 0 4.797298 3.021638 8.570107 0 0 0 6.479938 0.062927 4.394109 0 0 0 3.069998 7.687982 6.570897 0 0 0 10.811597 8.371383 7.651952 0 0 0 0.835099 3.241213 1.134827 0 0 0 1.501103 3.880456 11.626279 0 0 0 4.961216 6.080938 3.678907 0 0 0 4.561210 11.812593 0.917473 0 0 0 O 0.0 56 2.206114 6.592205 3.589702 0 0 0 10.732100 9.329175 7.464139 0 0 0 2.223431 5.278677 8.587705 0 0 0 7.645067 5.072650 7.182924 0 0 0 9.441138 11.068409 9.449919 0 0 0 4.369172 6.776752 7.936759 0 0 0 3.282936 9.482682 2.791039 0 0 0 6.983782 7.512527 8.446922 0 0 0 10.275126 7.456214 0.335466 0 0 0 6.079891 0.013400 10.521057 0 0 0 5.726767 7.932617 11.787012 0 0 0 6.576653 9.663115 2.952065 0 0 0 0.244163 5.639699 5.458062 0 0 0 6.663630 2.700629 5.933249 0 0 0 5.581856 3.546764 1.450071 0 0 0 0.473027 7.488905 10.421940 0 0 0 4.321913 1.286847 2.236269 0 0 0 10.184948 7.349734 2.937191 0 0 0 4.583616 3.625996 7.904262 0 0 0 5.543302 1.043739 7.917020 0 0 0 0.379637 10.328048 9.422410 0 0 0 9.440982 8.212171 9.931885 0 0 0 4.478494 5.312835 4.033938 0 0 0 7.478191 0.049752 1.204663 0 0 0 3.277196 8.196071 5.791242 0 0 0 7.654345 6.367523 4.523920 0 0 0 10.552967 0.883696 0.564552 0 0 0 7.327133 2.187912 3.213115 0 0 0 2.798415 4.478644 5.937614 0 0 0 5.576144 11.774659 4.268349 0 0 0 4.367718 5.644089 10.609721 0 0 0 11.442246 5.092536 9.586917 0 0 0 1.722208 10.224981 0.290858 0 0 0 7.680344 9.608853 11.908463 0 0 0 0.832246 7.557763 7.383029 0 0 0 4.082949 1.854934 5.180315 0 0 0 0.783357 7.521982 1.750993 0 0 0 8.435534 11.973924 4.652752 0 0 0 9.198317 8.867465 5.160566 0 0 0 3.149756 7.854211 11.453526 0 0 0 6.212763 8.605740 5.521239 0 0 0 0.148318 8.321816 4.692639 0 0 0 11.181329 10.061852 0.143693 0 0 0 9.634097 4.086678 0.898526 0 0 0 1.379823 0.808317 1.324665 0 0 0 7.797373 10.416174 7.147862 0 0 0 4.764720 10.926903 1.060141 0 0 0 10.141476 6.763653 6.761077 0 0 0 5.025996 7.555070 2.435468 0 0 0 4.488856 10.828791 6.602896 0 0 0 9.355994 10.461932 2.326700 0 0 0 4.339259 1.831479 11.135584 0 0 0 0.114986 10.461302 2.601943 0 0 0 7.521582 6.555799 1.539522 0 0 0 3.051578 3.746804 2.105064 0 0 0 2.994047 5.701360 1.057220 0 0 0 N 0.0 1 0.372586 5.305090 0.697343 0 0 0 C 0.0 2 1.266851 4.111790 0.638086 0 0 0 2.533331 4.581331 1.361807 0 0 0 NUMERICAL_ORBITAL ../03-DeePKS_Interation/orbital_dir/H_gga_6au_100Ry_2s1p.orb ../03-DeePKS_Interation/orbital_dir/O_gga_7au_100Ry_2s2p1d.orb ../03-DeePKS_Interation/orbital_dir/N_gga_7au_100Ry_2s2p1d.orb ../03-DeePKS_Interation/orbital_dir/C_gga_7au_100Ry_2s2p1d.orb NUMERICAL_DESCRIPTOR ../03-DeePKS_Interation/glycine/projector/jle.orb
代码
文本
STRU 文件介绍如下：
ATOMIC_SPECIES
H 1.00 ../03-DeePKS_Interation/pseudo_dir/H_ONCV_PBE-1.0.upf # 元素，原子质量，使用的赝势文件
O 1.00 ../03-DeePKS_Interation/pseudo_dir/O_ONCV_PBE-1.0.upf
N 1.00 ../03-DeePKS_Interation/pseudo_dir/N_ONCV_PBE-1.0.upf
C 1.00 ../03-DeePKS_Interation/pseudo_dir/C_ONCV_PBE-1.0.upf
LATTICE_CONSTANT
1.8897259886 # 1.8897259886 Bohr = 1.0 Angstrom
LATTICE_VECTORS # 晶格向量
12.028 0.0 0.0
0.0 12.028 0.0
0.0 0.0 12.028
ATOMIC_POSITIONS
Cartesian # 以笛卡尔坐标表示（Cartesian），单位为晶格常数
H # 元素名称
0.0 # 元素磁性
113 # 原子个数
8.477512 11.052771 4.986609 0 0 0 # 每个原子x，y，z方向的坐标和约束条件（1表示允许在该方向上移动，0表示固定）
...（省略）
NUMERICAL_ORBITAL # 数值轨道
../03-DeePKS_Interation/orbital_dir/H_gga_6au_100Ry_2s1p.orb
../03-DeePKS_Interation/orbital_dir/O_gga_7au_100Ry_2s2p1d.orb
../03-DeePKS_Interation/orbital_dir/N_gga_7au_100Ry_2s2p1d.orb
../03-DeePKS_Interation/orbital_dir/C_gga_7au_100Ry_2s2p1d.orb
NUMERICAL_DESCRIPTOR # 投影子
../03-DeePKS_Interation/glycine/projector/jle.orb
代码
文本
- KPT 文件。KPT 文件包含了SCF 计算的 k 点设置：
代码
文本
[72]
! cat 04-DeePKS_Labeling/KPT
K_POINTS 0 Gamma 1 1 1 0 0 0
代码
文本
*.upf和*.orb文件。各原子的 upf 和 orb 文件可以从 ABACUS 官网下载。
代码
文本
4.2 运行 ABACUS 自洽计算
代码
文本
准备好以上所有输入文件后，我们可以执行甘氨酸溶液体系的 SCF 计算。例如，使用命令行：
- 镜像待安装。笔者在本地机器执行qsub sub.sh即可运行。
- 设置
OMP_NUM_THREADS=1使用单线程进行计算。 abacus为 ABACUS 可执行程序的命令。
运行结束后查看running_scf.log输出文件:
代码
文本
[86]
! cat 04-DeePKS_Labeling/running_scf.log
WELCOME TO ABACUS v3.0
'Atomic-orbital Based Ab-initio Computation at UStc'
Website: http://abacus.ustc.edu.cn/
Version: Parallel, in development
Processor Number is 24
Start Time is Sun Jan 8 00:19:35 2023
------------------------------------------------------------------------------------
READING GENERAL INFORMATION
global_out_dir = OUT.ABACUS/
global_in_card = INPUT
pseudo_dir =
orbital_dir =
DRANK = 1
DSIZE = 24
DCOLOR = 1
GRANK = 1
GSIZE = 1
The esolver type has been set to : ksdft_lcao
>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
| |
| Reading atom information in unitcell: |
| From the input file and the structure file we know the number of |
| different elments in this unitcell, then we list the detail |
| information for each element, especially the zeta and polar atomic |
| orbital number for each element. The total atom number is counted. |
| We calculate the nearest atom distance for each atom and show the |
| Cartesian and Direct coordinates for each atom. We list the file |
| address for atomic orbitals. The volume and the lattice vectors |
| in real and reciprocal space is also shown. |
| |
<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<
READING UNITCELL INFORMATION
ntype = 4
atom label for species 1 = H
atom label for species 2 = O
atom label for species 3 = N
atom label for species 4 = C
lattice constant (Bohr) = 1.88973
lattice constant (Angstrom) = 1
READING ATOM TYPE 1
atom label = H
L=0, number of zeta = 2
L=1, number of zeta = 1
number of atom for this type = 113
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
READING ATOM TYPE 2
atom label = O
L=0, number of zeta = 2
L=1, number of zeta = 2
L=2, number of zeta = 1
number of atom for this type = 56
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
start magnetization = FALSE
READING ATOM TYPE 3
atom label = N
L=0, number of zeta = 2
L=1, number of zeta = 2
L=2, number of zeta = 1
number of atom for this type = 1
start magnetization = FALSE
READING ATOM TYPE 4
atom label = C
L=0, number of zeta = 2
L=1, number of zeta = 2
L=2, number of zeta = 1
number of atom for this type = 2
start magnetization = FALSE
start magnetization = FALSE
TOTAL ATOM NUMBER = 172
CARTESIAN COORDINATES ( UNIT = 1.88973 Bohr ).
atom x y z mag vx vy vz
tauc_H1 8.47751200001 11.052771 4.986609 0 0 0 0
tauc_H2 0.713459999993 6.62531200001 7.234243 0 0 0 0
tauc_H3 8.37958999999 10.521278 7.85061999999 0 0 0 0
tauc_H4 5.12692199999 0.000241999991202 3.475644 0 0 0 0
tauc_H5 5.78029700001 10.219974 2.802691 0 0 0 0
tauc_H6 0.47673500001 10.084497 1.767674 0 0 0 0
tauc_H7 0.458904000007 9.17106099999 4.32953900001 0 0 0 0
tauc_H8 7.22545100001 2.122528 5.40101 0 0 0 0
tauc_H9 8.51897200001 6.70391199999 4.61555099999 0 0 0 0
tauc_H10 9.52925400001 11.500578 10.32116 0 0 0 0
tauc_H11 0.23749999999 10.445055 8.46931000001 0 0 0 0
tauc_H12 6.736948 11.328984 10.410131 0 0 0 0
tauc_H13 0.479656999991 9.39232800001 9.56684900001 0 0 0 0
tauc_H14 0.743417 6.52100800001 1.12682599999 0 0 0 0
tauc_H15 2.709654 7.31696500001 3.996281 0 0 0 0
tauc_H16 6.454134 9.097445 4.74714100001 0 0 0 0
tauc_H17 6.707681 7.05510299999 1.290373 0 0 0 0
tauc_H18 4.720297 0.656670000006 7.583141 0 0 0 0
tauc_H19 5.875757 0.301228999997 9.64277599999 0 0 0 0
tauc_H20 9.30565599999 3.528874 1.53138999999 0 0 0 0
tauc_H21 6.06663400001 7.52963 10.937747 0 0 0 0
tauc_H22 0.347528000003 5.176702 9.186558 0 0 0 0
tauc_H23 2.696763 10.177609 0.260226999996 0 0 0 0
tauc_H24 8.68094400001 8.22831 9.31881699999 0 0 0 0
tauc_H25 0.579453999992 0.935519000004 0.788905000002 0 0 0 0
tauc_H26 1.14332299999 5.288025 5.53892400001 0 0 0 0
tauc_H27 3.92906600001 5.677605 11.449092 0 0 0 0
tauc_H28 5.88651800001 2.60303300001 5.440133 0 0 0 0
tauc_H29 6.054872 4.29144299999 1.049866 0 0 0 0
tauc_H30 4.55519299999 3.74469799999 1.58345099999 0 0 0 0
tauc_H31 4.80248699999 11.142155 5.69042900001 0 0 0 0
tauc_H32 0.507386000007 7.83872900001 11.326568 0 0 0 0
tauc_H33 9.80029300001 8.220575 2.769882 0 0 0 0
tauc_H34 10.356333 10.709876 9.08255199999 0 0 0 0
tauc_H35 7.517673 1.63021300001 2.45409100001 0 0 0 0
tauc_H36 10.937224 5.110278 8.74844599999 0 0 0 0
tauc_H37 4.35869899999 3.022382 7.19063699999 0 0 0 0
tauc_H38 11.152841 9.54436399999 6.671887 0 0 0 0
tauc_H39 11.204836 10.273087 2.502405 0 0 0 0
tauc_H40 1.540861 11.170513 0.394901000004 0 0 0 0
tauc_H41 3.320854 9.09488 6.17240400001 0 0 0 0
tauc_H42 8.28099200001 9.110674 11.389585 0 0 0 0
tauc_H43 0.466384999998 7.76683000001 8.20245599999 0 0 0 0
tauc_H44 6.65135999999 2.77582299999 2.91865300001 0 0 0 0
tauc_H45 4.795009 4.517778 3.56035000001 0 0 0 0
tauc_H46 4.635582 7.01815899999 1.785197 0 0 0 0
tauc_H47 5.94629100001 1.62496 7.27758300001 0 0 0 0
tauc_H48 9.156671 4.87372 1.06668199999 0 0 0 0
tauc_H49 3.81936000001 5.29725800001 9.93156299999 0 0 0 0
tauc_H50 9.659438 0.706990000001 0.860768000008 0 0 0 0
tauc_H51 6.913244 11.921111 0.456256999994 0 0 0 0
tauc_H52 7.156819 7.07002800001 4.97680299999 0 0 0 0
tauc_H53 5.01695800001 1.80228400001 1.80534900001 0 0 0 0
tauc_H54 7.764881 5.068147 6.217623 0 0 0 0
tauc_H55 5.111135 7.22621800001 8.16827399999 0 0 0 0
tauc_H56 7.180951 6.63458399999 8.02970500001 0 0 0 0
tauc_H57 2.93596099999 7.20250499999 0.139211999998 0 0 0 0
tauc_H58 2.983514 4.305543 4.98820399999 0 0 0 0
tauc_H59 9.49604700001 6.08499399999 6.708427 0 0 0 0
tauc_H60 6.86914499999 9.05311500001 11.94992 0 0 0 0
tauc_H61 8.755458 10.096845 2.971789 0 0 0 0
tauc_H62 5.283078 8.39267999999 5.42263899999 0 0 0 0
tauc_H63 11.491615 5.12304199999 1.07810399999 0 0 0 0
tauc_H64 6.98343300001 11.792354 2.01683500001 0 0 0 0
tauc_H65 6.86391899999 8.072224 7.606946 0 0 0 0
tauc_H66 3.944497 2.00669599999 4.14704800001 0 0 0 0
tauc_H67 0.339058000005 6.26388799999 4.69370299999 0 0 0 0
tauc_H68 4.748745 5.91008 7.958939 0 0 0 0
tauc_H69 4.94172000001 10.549195 0.205022999998 0 0 0 0
tauc_H70 3.581996 4.910788 6.18475000001 0 0 0 0
tauc_H71 1.25083799999 7.218697 2.57344599999 0 0 0 0
tauc_H72 10.184779 7.260062 11.392811 0 0 0 0
tauc_H73 7.38214900001 11.245316 6.81537 0 0 0 0
tauc_H74 11.864153 7.49444599999 1.96238799999 0 0 0 0
tauc_H75 2.77497600001 4.65163100001 8.086785 0 0 0 0
tauc_H76 2.52152999999 9.83341899999 2.339826 0 0 0 0
tauc_H77 4.37712000001 8.27704100001 2.59391099999 0 0 0 0
tauc_H78 8.887805 9.31731000001 5.98986699999 0 0 0 0
tauc_H79 0.705336000009 8.13892000001 5.460177 0 0 0 0
tauc_H80 11.688733 7.109435 10.252553 0 0 0 0
tauc_H81 4.678569 2.712175 11.164536 0 0 0 0
tauc_H82 11.146048 9.11749600001 0.425415000002 0 0 0 0
tauc_H83 9.706977 9.14864000001 10.010559 0 0 0 0
tauc_H84 5.21900399999 10.406819 6.99587099999 0 0 0 0
tauc_H85 8.80767300001 0.0473409999954 3.750322 0 0 0 0
tauc_H86 11.927291 10.112556 11.54869 0 0 0 0
tauc_H87 3.78998099999 0.940395999993 5.262649 0 0 0 0
tauc_H88 2.72624500001 6.123123 8.44660900001 0 0 0 0
tauc_H89 2.837328 5.90818 3.502186 0 0 0 0
tauc_H90 1.12798599999 0.197138999997 2.01503300001 0 0 0 0
tauc_H91 7.214744 4.21797699999 7.299943 0 0 0 0
tauc_H92 10.784087 6.42125900001 6.149549 0 0 0 0
tauc_H93 10.97317 12.02713 0.662913000007 0 0 0 0
tauc_H94 9.77211600001 8.16126399999 5.49805600001 0 0 0 0
tauc_H95 5.211637 1.41101800001 11.076792 0 0 0 0
tauc_H96 0.0565269999972 5.35634299999 11.745871 0 0 0 0
tauc_H97 6.37080799999 8.842716 2.391141 0 0 0 0
tauc_H98 2.99795500001 7.443466 10.603126 0 0 0 0
tauc_H99 7.47169499999 6.611041 2.515934 0 0 0 0
tauc_H100 3.51598200001 1.62513699999 1.842416 0 0 0 0
tauc_H101 3.978843 9.87105400001 2.29666700001 0 0 0 0
tauc_H102 4.775685 7.81137799999 11.619961 0 0 0 0
tauc_H103 9.480505 7.035846 0.647082000009 0 0 0 0
tauc_H104 8.86806700001 10.134663 1.58457300001 0 0 0 0
tauc_H105 10.735825 7.482301 3.744403 0 0 0 0
tauc_H106 4.797298 3.021638 8.570107 0 0 0 0
tauc_H107 6.47993800001 0.062927000002 4.39410899999 0 0 0 0
tauc_H108 3.069998 7.68798199999 6.570897 0 0 0 0
tauc_H109 10.811597 8.37138300001 7.651952 0 0 0 0
tauc_H110 0.835098999993 3.241213 1.13482700001 0 0 0 0
tauc_H111 1.50110300001 3.88045600001 11.626279 0 0 0 0
tauc_H112 4.96121599999 6.080938 3.678907 0 0 0 0
tauc_H113 4.56120999999 11.812593 0.917473000008 0 0 0 0
tauc_O1 2.206114 6.59220499999 3.589702 0 0 0 0
tauc_O2 10.7321 9.32917500001 7.464139 0 0 0 0
tauc_O3 2.223431 5.27867700001 8.58770499999 0 0 0 0
tauc_O4 7.645067 5.07265 7.182924 0 0 0 0
tauc_O5 9.441138 11.068409 9.44991900001 0 0 0 0
tauc_O6 4.369172 6.77675200001 7.936759 0 0 0 0
tauc_O7 3.28293599999 9.482682 2.79103900001 0 0 0 0
tauc_O8 6.983782 7.512527 8.446922 0 0 0 0
tauc_O9 10.275126 7.456214 0.33546600001 0 0 0 0
tauc_O10 6.079891 0.0133999999914 10.521057 0 0 0 0
tauc_O11 5.72676700001 7.932617 11.787012 0 0 0 0
tauc_O12 6.576653 9.663115 2.95206499999 0 0 0 0
tauc_O13 0.244163000006 5.63969899999 5.45806200001 0 0 0 0
tauc_O14 6.66363000001 2.70062899999 5.93324900001 0 0 0 0
tauc_O15 5.58185600001 3.54676400001 1.450071 0 0 0 0
tauc_O16 0.473026999994 7.48890499999 10.42194 0 0 0 0
tauc_O17 4.32191300001 1.286847 2.23626899999 0 0 0 0
tauc_O18 10.184948 7.349734 2.937191 0 0 0 0
tauc_O19 4.583616 3.625996 7.904262 0 0 0 0
tauc_O20 5.543302 1.043739 7.91702000001 0 0 0 0
tauc_O21 0.379636999991 10.328048 9.42240999999 0 0 0 0
tauc_O22 9.44098199999 8.212171 9.93188500001 0 0 0 0
tauc_O23 4.478494 5.31283499999 4.03393800001 0 0 0 0
tauc_O24 7.47819100001 0.0497519999956 1.20466299999 0 0 0 0
tauc_O25 3.27719599999 8.19607100001 5.79124199999 0 0 0 0
tauc_O26 7.654345 6.36752299999 4.52392 0 0 0 0
tauc_O27 10.552967 0.883696 0.564552000006 0 0 0 0
tauc_O28 7.327133 2.18791199999 3.21311499999 0 0 0 0
tauc_O29 2.79841500001 4.47864400001 5.937614 0 0 0 0
tauc_O30 5.57614399999 11.774659 4.268349 0 0 0 0
tauc_O31 4.367718 5.644089 10.609721 0 0 0 0
tauc_O32 11.442246 5.092536 9.586917 0 0 0 0
tauc_O33 1.72220799999 10.224981 0.290858000008 0 0 0 0
tauc_O34 7.68034399999 9.60885300001 11.908463 0 0 0 0
tauc_O35 0.832246000008 7.55776299999 7.38302900001 0 0 0 0
tauc_O36 4.08294900001 1.85493400001 5.18031499999 0 0 0 0
tauc_O37 0.783357000006 7.52198200001 1.750993 0 0 0 0
tauc_O38 8.43553399999 11.973924 4.65275200001 0 0 0 0
tauc_O39 9.19831700001 8.867465 5.16056599999 0 0 0 0
tauc_O40 3.149756 7.85421099999 11.453526 0 0 0 0
tauc_O41 6.21276299999 8.60574 5.52123900001 0 0 0 0
tauc_O42 0.148317999998 8.32181600001 4.69263900001 0 0 0 0
tauc_O43 11.181329 10.061852 0.143692999992 0 0 0 0
tauc_O44 9.634097 4.086678 0.898525999991 0 0 0 0
tauc_O45 1.37982299999 0.808317000009 1.324665 0 0 0 0
tauc_O46 7.79737300001 10.416174 7.147862 0 0 0 0
tauc_O47 4.76472000001 10.926903 1.06014099999 0 0 0 0
tauc_O48 10.141476 6.76365300001 6.761077 0 0 0 0
tauc_O49 5.02599600001 7.55507000001 2.43546799999 0 0 0 0
tauc_O50 4.48885600001 10.828791 6.60289599999 0 0 0 0
tauc_O51 9.35599400001 10.461932 2.3267 0 0 0 0
tauc_O52 4.33925899999 1.831479 11.135584 0 0 0 0
tauc_O53 0.114985999997 10.461302 2.601943 0 0 0 0
tauc_O54 7.521582 6.55579899999 1.539522 0 0 0 0
tauc_O55 3.051578 3.74680400001 2.10506400001 0 0 0 0
tauc_O56 2.994047 5.70136000001 1.05722000001 0 0 0 0
tauc_N1 0.372586000007 5.30509 0.697343 0 0 0 0
tauc_C1 1.266851 4.11179 0.63808600001 0 0 0 0
tauc_C2 2.533331 4.58133100001 1.36180699999 0 0 0 0
READING ORBITAL FILE NAMES FOR LCAO
orbital file: /data/HOME_BACKUP/pengchao/glycine/deepks/glycine/pbe_2/iter/H_gga_6au_100Ry_2s1p.orb
orbital file: /data/HOME_BACKUP/pengchao/glycine/deepks/glycine/pbe_2/iter/O_gga_7au_100Ry_2s2p1d.orb
orbital file: /data/HOME_BACKUP/pengchao/glycine/deepks/glycine/pbe_2/iter/N_gga_7au_100Ry_2s2p1d.orb
orbital file: /data/HOME_BACKUP/pengchao/glycine/deepks/glycine/pbe_2/iter/C_gga_7au_100Ry_2s2p1d.orb
Volume (Bohr^3) = 11742.937938
Volume (A^3) = 1740.12178961
Lattice vectors: (Cartesian coordinate: in unit of a_0)
+12.028 +0 +0
+0 +12.028 +0
+0 +0 +12.028
Reciprocal vectors: (Cartesian coordinate: in unit of 2 pi/a_0)
+0.0831393415364 -0 +0
-0 +0.0831393415364 -0
+0 -0 +0.0831393415364
>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
| |
| Reading pseudopotentials files: |
| The pseudopotential file is in UPF format. The 'NC' indicates that |
| the type of pseudopotential is 'norm conserving'. Functional of |
| exchange and correlation is decided by 4 given parameters in UPF |
| file. We also read in the 'core correction' if there exists. |
| Also we can read the valence electrons number and the maximal |
| angular momentum used in this pseudopotential. We also read in the |
| trail wave function, trail atomic density and local-pseudopotential|
| on logrithmic grid. The non-local pseudopotential projector is also|
| read in if there is any. |
| |
<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<
PAO radial cut off (Bohr) = 15
Read in pseudopotential file is /data/HOME_BACKUP/pengchao/glycine/deepks/glycine/pbe_2/iter/H_ONCV_PBE-1.0.upf
pseudopotential type = NC
exchange-correlation functional = PBE
nonlocal core correction = 0
valence electrons = 1
lmax = 0
number of zeta = 0
number of projectors = 2
L of projector = 0
L of projector = 0
PAO radial cut off (Bohr) = 15
Read in pseudopotential file is /data/HOME_BACKUP/pengchao/glycine/deepks/glycine/pbe_2/iter/O_ONCV_PBE-1.0.upf
pseudopotential type = NC
exchange-correlation functional = PBE
nonlocal core correction = 0
valence electrons = 6
lmax = 1
number of zeta = 0
number of projectors = 4
L of projector = 0
L of projector = 0
L of projector = 1
L of projector = 1
PAO radial cut off (Bohr) = 15
Read in pseudopotential file is /data/HOME_BACKUP/pengchao/glycine/deepks/glycine/pbe_2/iter/N_ONCV_PBE-1.0.upf
pseudopotential type = NC
exchange-correlation functional = PBE
nonlocal core correction = 0
valence electrons = 5
lmax = 1
number of zeta = 0
number of projectors = 4
L of projector = 0
L of projector = 0
L of projector = 1
L of projector = 1
PAO radial cut off (Bohr) = 15
Read in pseudopotential file is /data/HOME_BACKUP/pengchao/glycine/deepks/glycine/pbe_2/iter/C_ONCV_PBE-1.0.upf
pseudopotential type = NC
exchange-correlation functional = PBE
nonlocal core correction = 0
valence electrons = 4
lmax = 1
number of zeta = 0
number of projectors = 4
L of projector = 0
L of projector = 0
L of projector = 1
L of projector = 1
initial pseudo atomic orbital number = 0
NLOCAL = 1332
>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
| |
| Setup plane waves of charge/potential: |
| Use the energy cutoff and the lattice vectors to generate the |
| dimensions of FFT grid. The number of FFT grid on each processor |
| is 'nrxx'. The number of plane wave basis in reciprocal space is |
| different for charege/potential and wave functions. We also set |
| the 'sticks' for the parallel of FFT. The number of plane waves |
| is 'npw' in each processor. |
<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<
SETUP THE PLANE WAVE BASIS
energy cutoff for charge/potential (unit:Ry) = 400
[fft grid for charge/potential] = 150, 150, 150
[fft grid division] = 2, 2, 2
[big fft grid for charge/potential] = 75, 75, 75
nbxx = 22500
nrxx = 180000
SETUP PLANE WAVES FOR CHARGE/POTENTIAL
number of plane waves = 1586405
number of sticks = 16453
PARALLEL PW FOR CHARGE/POTENTIAL
PROC COLUMNS(POT) PW
1 685 66101
2 685 66101
3 685 66101
4 685 66101
5 685 66101
6 685 66101
7 686 66100
8 686 66100
9 686 66100
10 686 66102
11 686 66102
12 686 66100
13 686 66100
14 686 66100
15 686 66100
16 686 66100
17 686 66100
18 686 66100
19 686 66100
20 685 66099
21 685 66099
22 685 66099
23 685 66099
24 685 66099
--------------- sum -------------------
24 16453 1586405
number of |g| = 4218
max |g| = 36.1781936815
min |g| = 0.062209351
SETUP THE ELECTRONS NUMBER
electron number of element H = 1
total electron number of element H = 113
electron number of element O = 6
total electron number of element O = 336
electron number of element N = 5
total electron number of element N = 5
electron number of element C = 4
total electron number of element C = 8
AUTOSET number of electrons: = 462
DONE : SETUP UNITCELL Time : 0.414418935776 (SEC)
>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
| |
| Setup K-points |
| We setup the k-points according to input parameters. |
| The reduced k-points are set according to symmetry operations. |
| We treat the spin as another set of k-points. |
| |
<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<
SETUP K-POINTS
nspin = 1
Input type of k points = Monkhorst-Pack(Gamma)
nkstot = 1
nkstot_ibz = 1
IBZ DirectX DirectY DirectZ Weight ibz2bz
1 0 0 0 1 0
nkstot now = 1
KPOINTS DIRECT_X DIRECT_Y DIRECT_Z WEIGHT
1 0 0 0 1
k-point number in this process = 1
minimum distributed K point number = 1
KPOINTS CARTESIAN_X CARTESIAN_Y CARTESIAN_Z WEIGHT
1 0 0 0 2
KPOINTS DIRECT_X DIRECT_Y DIRECT_Z WEIGHT
1 0 0 0 2
DONE : INIT K-POINTS Time : 0.448012828827 (SEC)
occupied bands = 231
NLOCAL = 1332
NBANDS = 277
NBANDS = 277
>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
| |
| Setup numerical orbitals: |
| This part setup: numerical atomic orbitals, non-local projectors |
| and neutral potential (1D). The atomic orbitals information |
| including the radius, angular momentum and zeta number. |
| The neutral potential is the sum of local part of pseudopotential |
| and potential given by atomic charge, they will cancel out beyond |
| a certain radius cutoff, because the Z/r character. |
| |
<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<
SETUP ONE DIMENSIONAL ORBITALS/POTENTIAL
delta k (1/Bohr) = 0.01
delta r (Bohr) = 0.01
dr_uniform (Bohr) = 0.001
rmax (Bohr) = 30
kmesh = 1005
ORBITAL L N nr dr RCUT CHECK_UNIT NEW_UNIT
1 0 0 601 0.01 6 1 1
2 0 1 601 0.01 6 1 1
3 1 0 601 0.01 6 1 1
ORBITAL L N nr dr RCUT CHECK_UNIT NEW_UNIT
1 0 0 701 0.01 7 1 1
2 0 1 701 0.01 7 1 1
3 1 0 701 0.01 7 1 1
4 1 1 701 0.01 7 1 1
5 2 0 701 0.01 7 1 1
ORBITAL L N nr dr RCUT CHECK_UNIT NEW_UNIT
1 0 0 701 0.01 7 1 1
2 0 1 701 0.01 7 1 1
3 1 0 701 0.01 7 1 1
4 1 1 701 0.01 7 1 1
5 2 0 701 0.01 7 1 1
ORBITAL L N nr dr RCUT CHECK_UNIT NEW_UNIT
1 0 0 701 0.01 7 1 1
2 0 1 701 0.01 7 1 1
3 1 0 701 0.01 7 1 1
4 1 1 701 0.01 7 1 1
5 2 0 701 0.01 7 1 1
DESCRIPTOR L N nr dr RCUT CHECK_UNIT NEW_UNIT
1 0 0 505 0.01 5.04 6.33 1
2 0 1 505 0.01 5.04 1.58 1
3 0 2 505 0.01 5.04 0.704 1
4 0 3 505 0.01 5.04 0.396 1
5 0 4 505 0.01 5.04 0.253 1
6 0 5 505 0.01 5.04 0.176 1
7 0 6 505 0.01 5.04 0.129 1
8 0 7 505 0.01 5.04 0.099 1
9 0 8 505 0.01 5.04 0.0782 1
10 0 9 505 0.01 5.04 0.0633 1
11 0 10 505 0.01 5.04 0.0524 1
12 0 11 505 0.01 5.04 0.044 1
13 0 12 505 0.01 5.04 0.0375 1
14 0 13 505 0.01 5.04 0.0323 1
15 0 14 505 0.01 5.04 0.0282 1
16 1 0 505 0.01 5.04 2.95 1
17 1 1 505 0.01 5.04 1.03 1
18 1 2 505 0.01 5.04 0.521 1
19 1 3 505 0.01 5.04 0.314 1
20 1 4 505 0.01 5.04 0.21 1
21 1 5 505 0.01 5.04 0.15 1
22 1 6 505 0.01 5.04 0.113 1
23 1 7 505 0.01 5.04 0.0878 1
24 1 8 505 0.01 5.04 0.0703 1
25 1 9 505 0.01 5.04 0.0575 1
26 1 10 505 0.01 5.04 0.0479 1
27 1 11 505 0.01 5.04 0.0406 1
28 1 12 505 0.01 5.04 0.0348 1
29 1 13 505 0.01 5.04 0.0302 1
30 1 14 505 0.01 5.04 0.0264 1
31 2 0 505 0.01 5.04 1.71 1
32 2 1 505 0.01 5.04 0.728 1
33 2 2 505 0.01 5.04 0.403 1
34 2 3 505 0.01 5.04 0.256 1
35 2 4 505 0.01 5.04 0.177 1
36 2 5 505 0.01 5.04 0.13 1
37 2 6 505 0.01 5.04 0.0994 1
38 2 7 505 0.01 5.04 0.0785 1
39 2 8 505 0.01 5.04 0.0635 1
40 2 9 505 0.01 5.04 0.0525 1
41 2 10 505 0.01 5.04 0.0441 1
42 2 11 505 0.01 5.04 0.0376 1
43 2 12 505 0.01 5.04 0.0324 1
44 2 13 505 0.01 5.04 0.0282 1
45 2 14 505 0.01 5.04 0.0248 1
SET NONLOCAL PSEUDOPOTENTIAL PROJECTORS
SET NONLOCAL PSEUDOPOTENTIAL PROJECTORS
SET NONLOCAL PSEUDOPOTENTIAL PROJECTORS
SET NONLOCAL PSEUDOPOTENTIAL PROJECTORS
max number of nonlocal projetors among all species is 4
SETUP THE TWO-CENTER INTEGRATION TABLES
SETUP THE DIVISION OF H/S MATRIX
divide the H&S matrix using 2D block algorithms.
nb2d = 64
parameter nb2d is too large: nb2d = 64
reset nb2d to value 1, this set would make the program keep working but maybe get slower during diagonalization.
trace_loc_row dimension = 1332
trace_loc_col dimension = 1332
nloc = 73926
Initialize the descriptor index for DeePKS (lcao line)
lmax of descriptor = 2
nmax of descriptor= 15
total basis (all atoms) for descriptor=
Type 1 number_of_atoms 113
Type 2 number_of_atoms 56
Type 3 number_of_atoms 1
Type 4 number_of_atoms 2
descriptors_per_atom 135
total_descriptors 23220
-------------------------------------------
SELF-CONSISTENT
-------------------------------------------
>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
| |
| Search adjacent atoms: |
| Set the adjacent atoms for each atom and set the periodic boundary |
| condition for the atoms on real space FFT grid. For k-dependent |
| algorithm, we also need to set the sparse H and S matrix element |
| for each atom. |
| |
<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<
SETUP SEARCHING RADIUS FOR PROGRAM TO SEARCH ADJACENT ATOMS
longest orb rcut (Bohr) = 7
longest nonlocal projector rcut (Bohr) = 1.64
searching radius is (Bohr)) = 14
searching radius unit is (Bohr)) = 1.89
SETUP EXTENDED REAL SPACE GRID FOR GRID INTEGRATION
[real space grid] = 150, 150, 150
[big cell numbers in grid] = 75, 75, 75
[meshcell numbers in big cell] = 2, 2, 2
[extended fft grid] = 24, 24, 24
[dimension of extened grid] = 124, 124, 124
UnitCellTotal = 27
Atom number in sub-FFT-grid = 117
Local orbitals number in sub-FFT-grid = 905
Successful completion of Potential's registration : local
Successful completion of Potential's registration : hartree
Successful completion of Potential's registration : xc
Successful completion of Potential's registration : tddft
lgd_last = 0
lgd_now = 905
allocate DM , the dimension is 905
enter setAlltoallvParameter, nblk = 1
checkpoint 2
pnum = 0
prow = 0
pcol = 0
nRow_in_proc = 229
nCol_in_proc = 149
pnum = 1
prow = 0
pcol = 1
nRow_in_proc = 229
nCol_in_proc = 149
pnum = 2
prow = 0
pcol = 2
nRow_in_proc = 229
nCol_in_proc = 148
pnum = 3
prow = 0
pcol = 3
nRow_in_proc = 229
nCol_in_proc = 151
pnum = 4
prow = 0
pcol = 4
nRow_in_proc = 229
nCol_in_proc = 154
pnum = 5
prow = 0
pcol = 5
nRow_in_proc = 229
nCol_in_proc = 154
pnum = 6
prow = 1
pcol = 0
nRow_in_proc = 226
nCol_in_proc = 149
pnum = 7
prow = 1
pcol = 1
nRow_in_proc = 226
nCol_in_proc = 149
pnum = 8
prow = 1
pcol = 2
nRow_in_proc = 226
nCol_in_proc = 148
pnum = 9
prow = 1
pcol = 3
nRow_in_proc = 226
nCol_in_proc = 151
pnum = 10
prow = 1
pcol = 4
nRow_in_proc = 226
nCol_in_proc = 154
pnum = 11
prow = 1
pcol = 5
nRow_in_proc = 226
nCol_in_proc = 154
pnum = 12
prow = 2
pcol = 0
nRow_in_proc = 222
nCol_in_proc = 149
pnum = 13
prow = 2
pcol = 1
nRow_in_proc = 222
nCol_in_proc = 149
pnum = 14
prow = 2
pcol = 2
nRow_in_proc = 222
nCol_in_proc = 148
pnum = 15
prow = 2
pcol = 3
nRow_in_proc = 222
nCol_in_proc = 151
pnum = 16
prow = 2
pcol = 4
nRow_in_proc = 222
nCol_in_proc = 154
pnum = 17
prow = 2
pcol = 5
nRow_in_proc = 222
nCol_in_proc = 154
pnum = 18
prow = 3
pcol = 0
nRow_in_proc = 228
nCol_in_proc = 149
pnum = 19
prow = 3
pcol = 1
nRow_in_proc = 228
nCol_in_proc = 149
pnum = 20
prow = 3
pcol = 2
nRow_in_proc = 228
nCol_in_proc = 148
pnum = 21
prow = 3
pcol = 3
nRow_in_proc = 228
nCol_in_proc = 151
pnum = 22
prow = 3
pcol = 4
nRow_in_proc = 228
nCol_in_proc = 154
pnum = 23
prow = 3
pcol = 5
nRow_in_proc = 228
nCol_in_proc = 154
receiver_size is 819025 ; receiver_size of each process is:
34121 34121 33892 34579 35266 35266 33674 33674 33448 34126 34804 34804 33078 33078 32856 33522 34188 34188 33972 33972 33744 34428 35112 35112
sender_size is 826434 ; sender_size of each process is:
34121 36036 40590 41168 40016 42926 43010 45232 43172 37120 33592 30595 28560 25021 24768 24130 26136 27470 30738 31524 32256 36261 36190 35802
init_chg = atomic
DONE : INIT SCF Time : 16.878 (SEC)
LCAO ALGORITHM --------------- ION= 1 ELEC= 1--------------------------------
vlocal exchange index is built
buffer size(M): = 12
buffer index size(M): = 12
vlocal data are put in sender_buffer, size(M): = 6
vlocal data are exchanged, received size(M): = 6
K-S equation was solved by genelpa2
eigenvalues were copied to ekb
cal_dk_gamma_from_2D, NSPIN = 1
number of non-zero elements in sender_buffer = 826434
sender_size = 826434
last sender_buffer = -0.00404168
number of non-zero elements in receiver_buffer = 819025
receiver_size = 819025
last receiver_buffer = 0.00622052
Density error is 0.227314000648
Energy Rydberg eV
E_KohnSham -1976.27439151 -26888.592536
E_Harris -1988.47857191 -27054.6389289
E_Fermi -0.419967773079 -5.71395469025
LCAO ALGORITHM --------------- ION= 1 ELEC= 2--------------------------------
vlocal exchange index is built
buffer size(M): = 12
buffer index size(M): = 12
vlocal data are put in sender_buffer, size(M): = 6
vlocal data are exchanged, received size(M): = 6
K-S equation was solved by genelpa2
eigenvalues were copied to ekb
cal_dk_gamma_from_2D, NSPIN = 1
number of non-zero elements in sender_buffer = 826434
sender_size = 826434
last sender_buffer = -0.00364001815946
number of non-zero elements in receiver_buffer = 819025
receiver_size = 819025
last receiver_buffer = 0.00635725846671
Density error is 0.0814572657435
Energy Rydberg eV
E_KohnSham -1977.63064707 -26907.0453396
E_Harris -1977.76500213 -26908.873334
E_Fermi -0.191615043833 -2.60705641865
LCAO ALGORITHM --------------- ION= 1 ELEC= 3--------------------------------
vlocal exchange index is built
buffer size(M): = 12
buffer index size(M): = 12
vlocal data are put in sender_buffer, size(M): = 6
vlocal data are exchanged, received size(M): = 6
K-S equation was solved by genelpa2
eigenvalues were copied to ekb
cal_dk_gamma_from_2D, NSPIN = 1
number of non-zero elements in sender_buffer = 826434
sender_size = 826434
last sender_buffer = -0.00341992685964
number of non-zero elements in receiver_buffer = 819025
receiver_size = 819025
last receiver_buffer = 0.00586206564199
Density error is 0.0498116059971
Energy Rydberg eV
E_KohnSham -1976.13026302 -26886.6315674
E_Harris -1976.20532493 -26887.652837
E_Fermi -0.154651923865 -2.10414737123
LCAO ALGORITHM --------------- ION= 1 ELEC= 4--------------------------------
vlocal exchange index is built
buffer size(M): = 12
buffer index size(M): = 12
vlocal data are put in sender_buffer, size(M): = 6
vlocal data are exchanged, received size(M): = 6
K-S equation was solved by genelpa2
eigenvalues were copied to ekb
cal_dk_gamma_from_2D, NSPIN = 1
number of non-zero elements in sender_buffer = 826434
sender_size = 826434
last sender_buffer = -0.00337862138725
number of non-zero elements in receiver_buffer = 819025
receiver_size = 819025
last receiver_buffer = 0.00570424464891
Density error is 0.00621204948294
Energy Rydberg eV
E_KohnSham -1975.94394593 -26884.0965933
E_Harris -1975.95007152 -26884.1799362
E_Fermi -0.128938410702 -1.7542970766
LCAO ALGORITHM --------------- ION= 1 ELEC= 5--------------------------------
vlocal exchange index is built
buffer size(M): = 12
buffer index size(M): = 12
vlocal data are put in sender_buffer, size(M): = 6
vlocal data are exchanged, received size(M): = 6
K-S equation was solved by genelpa2
eigenvalues were copied to ekb
cal_dk_gamma_from_2D, NSPIN = 1
number of non-zero elements in sender_buffer = 826434
sender_size = 826434
last sender_buffer = -0.00343248496684
number of non-zero elements in receiver_buffer = 819025
receiver_size = 819025
last receiver_buffer = 0.00574741905578
Density error is 0.00213983594865
Energy Rydberg eV
E_KohnSham -1975.98943873 -26884.7155546
E_Harris -1975.98972332 -26884.7194265
E_Fermi -0.128890193412 -1.75364104673
LCAO ALGORITHM --------------- ION= 1 ELEC= 6--------------------------------
vlocal exchange index is built
buffer size(M): = 12
buffer index size(M): = 12
vlocal data are put in sender_buffer, size(M): = 6
vlocal data are exchanged, received size(M): = 6
K-S equation was solved by genelpa2
eigenvalues were copied to ekb
cal_dk_gamma_from_2D, NSPIN = 1
number of non-zero elements in sender_buffer = 826434
sender_size = 826434
last sender_buffer = -0.00345884007165
number of non-zero elements in receiver_buffer = 819025
receiver_size = 819025
last receiver_buffer = 0.00580476537254
Density error is 0.00105747326241
Energy Rydberg eV
E_KohnSham -1975.96141636 -26884.3342906
E_Harris -1975.96146289 -26884.3349238
E_Fermi -0.130394966584 -1.77411453607
LCAO ALGORITHM --------------- ION= 1 ELEC= 7--------------------------------
vlocal exchange index is built
buffer size(M): = 12
buffer index size(M): = 12
vlocal data are put in sender_buffer, size(M): = 6
vlocal data are exchanged, received size(M): = 6
K-S equation was solved by genelpa2
eigenvalues were copied to ekb
cal_dk_gamma_from_2D, NSPIN = 1
number of non-zero elements in sender_buffer = 826434
sender_size = 826434
last sender_buffer = -0.00347396918484
number of non-zero elements in receiver_buffer = 819025
receiver_size = 819025
last receiver_buffer = 0.00580968228732
Density error is 0.00023116392174
Energy Rydberg eV
E_KohnSham -1975.96138028 -26884.3337998
E_Harris -1975.96139028 -26884.3339357
E_Fermi -0.12987438085 -1.76703160378
LCAO ALGORITHM --------------- ION= 1 ELEC= 8--------------------------------
vlocal exchange index is built
buffer size(M): = 12
buffer index size(M): = 12
vlocal data are put in sender_buffer, size(M): = 6
vlocal data are exchanged, received size(M): = 6
K-S equation was solved by genelpa2
eigenvalues were copied to ekb
cal_dk_gamma_from_2D, NSPIN = 1
number of non-zero elements in sender_buffer = 826434
sender_size = 826434
last sender_buffer = -0.00347061200901
number of non-zero elements in receiver_buffer = 819025
receiver_size = 819025
last receiver_buffer = 0.0058137461561
Density error is 8.84520165868e-05
Energy Rydberg eV
E_KohnSham -1975.96095129 -26884.3279631
E_Harris -1975.96095208 -26884.3279737
E_Fermi -0.130417549668 -1.77442179468
LCAO ALGORITHM --------------- ION= 1 ELEC= 9--------------------------------
vlocal exchange index is built
buffer size(M): = 12
buffer index size(M): = 12
vlocal data are put in sender_buffer, size(M): = 6
vlocal data are exchanged, received size(M): = 6
K-S equation was solved by genelpa2
eigenvalues were copied to ekb
cal_dk_gamma_from_2D, NSPIN = 1
number of non-zero elements in sender_buffer = 826434
sender_size = 826434
last sender_buffer = -0.00347291587297
number of non-zero elements in receiver_buffer = 819025
receiver_size = 819025
last receiver_buffer = 0.0058131406283
Density error is 2.73897806268e-05
Energy Rydberg eV
E_KohnSham -1975.96116376 -26884.3308538
E_Harris -1975.96116385 -26884.330855
E_Fermi -0.130231381083 -1.77188884114
LCAO ALGORITHM --------------- ION= 1 ELEC= 10--------------------------------
vlocal exchange index is built
buffer size(M): = 12
buffer index size(M): = 12
vlocal data are put in sender_buffer, size(M): = 6
vlocal data are exchanged, received size(M): = 6
K-S equation was solved by genelpa2
eigenvalues were copied to ekb
cal_dk_gamma_from_2D, NSPIN = 1
number of non-zero elements in sender_buffer = 826434
sender_size = 826434
last sender_buffer = -0.00347338182217
number of non-zero elements in receiver_buffer = 819025
receiver_size = 819025
last receiver_buffer = 0.00581435951942
Density error is 7.63084372372e-06
Energy Rydberg eV
E_KohnSham -1975.96128205 -26884.3324633
E_Harris -1975.96128206 -26884.3324634
E_Fermi -0.130263216266 -1.77232198103
LCAO ALGORITHM --------------- ION= 1 ELEC= 11--------------------------------
vlocal exchange index is built
buffer size(M): = 12
buffer index size(M): = 12
vlocal data are put in sender_buffer, size(M): = 6
vlocal data are exchanged, received size(M): = 6
K-S equation was solved by genelpa2
eigenvalues were copied to ekb
cal_dk_gamma_from_2D, NSPIN = 1
number of non-zero elements in sender_buffer = 826434
sender_size = 826434
last sender_buffer = -0.00347368912615
number of non-zero elements in receiver_buffer = 819025
receiver_size = 819025
last receiver_buffer = 0.00581451675197
Density error is 3.11716499145e-06
Energy Rydberg eV
E_KohnSham -1975.96116371 -26884.3308531
E_Harris -1975.96116371 -26884.3308531
E_Fermi -0.130261732168 -1.77230178883
LCAO ALGORITHM --------------- ION= 1 ELEC= 12--------------------------------
vlocal exchange index is built
buffer size(M): = 12
buffer index size(M): = 12
vlocal data are put in sender_buffer, size(M): = 6
vlocal data are exchanged, received size(M): = 6
K-S equation was solved by genelpa2
eigenvalues were copied to ekb
cal_dk_gamma_from_2D, NSPIN = 1
number of non-zero elements in sender_buffer = 826434
sender_size = 826434
last sender_buffer = -0.00347381684749
number of non-zero elements in receiver_buffer = 819025
receiver_size = 819025
last receiver_buffer = 0.00581473774185
Density error is 9.93014256052e-07
Energy Rydberg eV
E_KohnSham -1975.96118534 -26884.3311475
E_Harris -1975.96118534 -26884.3311475
E_Fermi -0.130260787559 -1.77228893677
LCAO ALGORITHM --------------- ION= 1 ELEC= 13--------------------------------
vlocal exchange index is built
buffer size(M): = 12
buffer index size(M): = 12
vlocal data are put in sender_buffer, size(M): = 6
vlocal data are exchanged, received size(M): = 6
K-S equation was solved by genelpa2
eigenvalues were copied to ekb
cal_dk_gamma_from_2D, NSPIN = 1
number of non-zero elements in sender_buffer = 826434
sender_size = 826434
last sender_buffer = -0.00347385093602
number of non-zero elements in receiver_buffer = 819025
receiver_size = 819025
last receiver_buffer = 0.00581476393895
Density error is 2.68495121111e-07
Energy Rydberg eV
E_KohnSham -1975.9611737 -26884.3309891
E_Harris -1975.9611737 -26884.3309891
E_Fermi -0.130261539369 -1.77229916567
LCAO ALGORITHM --------------- ION= 1 ELEC= 14--------------------------------
vlocal exchange index is built
buffer size(M): = 12
buffer index size(M): = 12
vlocal data are put in sender_buffer, size(M): = 6
vlocal data are exchanged, received size(M): = 6
K-S equation was solved by genelpa2
eigenvalues were copied to ekb
cal_dk_gamma_from_2D, NSPIN = 1
number of non-zero elements in sender_buffer = 826434
sender_size = 826434
last sender_buffer = -0.00347386253503
number of non-zero elements in receiver_buffer = 819025
receiver_size = 819025
last receiver_buffer = 0.00581477661417
Density error is 8.74652341678e-08
Energy Rydberg eV
E_KohnSham -1975.9611767 -26884.3310299
E_Harris -1975.9611767 -26884.3310299
E_band -315.293237188 -4289.78456662
E_one_elec -2526.32621042 -34372.4314685
E_Hartree +1424.85196768 +19386.105567
E_xc -492.328005796 -6698.4661638
E_Ewald -367.312466806 -4997.54249499
E_demet +0 +0
E_descf +0 +0
E_exx +0 +0
E_DeePKS -10.2075189363 -138.880419977
E_Fermi -0.130261482713 -1.77229839482
charge density convergence is achieved
final etot is -26884.3310299 eV
STATE ENERGY(eV) AND OCCUPATIONS NSPIN == 1
1/1 kpoint (Cartesian) = 0 0 0 (0 pws)
1 -21.5207 2.00000
2 -21.5049 2.00000
3 -21.2301 2.00000
4 -21.1716 2.00000
5 -21.0795 2.00000
6 -21.0265 2.00000
7 -20.9430 2.00000
8 -20.8734 2.00000
9 -20.7968 2.00000
10 -20.7920 2.00000
11 -20.7387 2.00000
12 -20.6533 2.00000
13 -20.5900 2.00000
14 -20.5729 2.00000
15 -20.5107 2.00000
16 -20.4876 2.00000
17 -20.4210 2.00000
18 -20.3643 2.00000
19 -20.3520 2.00000
20 -20.3076 2.00000
21 -20.3048 2.00000
22 -20.2978 2.00000
23 -20.2927 2.00000
24 -20.2753 2.00000
25 -20.2635 2.00000
26 -20.2557 2.00000
27 -20.2172 2.00000
28 -20.2052 2.00000
29 -20.1655 2.00000
30 -20.1623 2.00000
31 -20.1384 2.00000
32 -20.1350 2.00000
33 -20.1296 2.00000
34 -20.1171 2.00000
35 -20.0741 2.00000
36 -20.0706 2.00000
37 -20.0625 2.00000
38 -20.0609 2.00000
39 -20.0250 2.00000
40 -19.9960 2.00000
41 -19.9531 2.00000
42 -19.9445 2.00000
43 -19.9185 2.00000
44 -19.9105 2.00000
45 -19.8880 2.00000
46 -19.8415 2.00000
47 -19.8144 2.00000
48 -19.7917 2.00000
49 -19.6701 2.00000
50 -19.6281 2.00000
51 -19.6047 2.00000
52 -19.5113 2.00000
53 -19.4784 2.00000
54 -19.4072 2.00000
55 -19.3661 2.00000
56 -19.2763 2.00000
57 -17.0095 2.00000
58 -12.3776 2.00000
59 -10.2042 2.00000
60 -10.1740 2.00000
61 -10.1230 2.00000
62 -10.0389 2.00000
63 -9.88033 2.00000
64 -9.69702 2.00000
65 -9.63540 2.00000
66 -9.46562 2.00000
67 -9.42815 2.00000
68 -9.41445 2.00000
69 -9.33401 2.00000
70 -9.31343 2.00000
71 -9.30406 2.00000
72 -9.27733 2.00000
73 -9.24322 2.00000
74 -9.21329 2.00000
75 -9.20078 2.00000
76 -9.16802 2.00000
77 -9.15671 2.00000
78 -9.13828 2.00000
79 -9.09078 2.00000
80 -9.05675 2.00000
81 -9.03949 2.00000
82 -9.00280 2.00000
83 -8.97253 2.00000
84 -8.93545 2.00000
85 -8.86454 2.00000
86 -8.84675 2.00000
87 -8.84025 2.00000
88 -8.82007 2.00000
89 -8.78470 2.00000
90 -8.75765 2.00000
91 -8.73006 2.00000
92 -8.68736 2.00000
93 -8.64042 2.00000
94 -8.62125 2.00000
95 -8.59262 2.00000
96 -8.56856 2.00000
97 -8.50508 2.00000
98 -8.47330 2.00000
99 -8.43895 2.00000
100 -8.40773 2.00000
101 -8.39712 2.00000
102 -8.35720 2.00000
103 -8.32913 2.00000
104 -8.28455 2.00000
105 -8.23797 2.00000
106 -8.23426 2.00000
107 -8.17393 2.00000
108 -8.16099 2.00000
109 -8.11939 2.00000
110 -8.10206 2.00000
111 -7.99835 2.00000
112 -7.97738 2.00000
113 -7.87437 2.00000
114 -7.80247 2.00000
115 -7.77870 2.00000
116 -7.59997 2.00000
117 -6.99779 2.00000
118 -6.76582 2.00000
119 -6.59047 2.00000
120 -6.44710 2.00000
121 -6.21954 2.00000
122 -6.10992 2.00000
123 -6.09983 2.00000
124 -6.04022 2.00000
125 -6.00559 2.00000
126 -5.92126 2.00000
127 -5.90712 2.00000
128 -5.87413 2.00000
129 -5.82055 2.00000
130 -5.78688 2.00000
131 -5.77889 2.00000
132 -5.76072 2.00000
133 -5.66270 2.00000
134 -5.59001 2.00000
135 -5.57352 2.00000
136 -5.50792 2.00000
137 -5.47762 2.00000
138 -5.45039 2.00000
139 -5.40025 2.00000
140 -5.28850 2.00000
141 -5.27204 2.00000
142 -5.24963 2.00000
143 -5.20924 2.00000
144 -5.18637 2.00000
145 -5.13202 2.00000
146 -5.11968 2.00000
147 -5.05162 2.00000
148 -5.02708 2.00000
149 -5.01026 2.00000
150 -4.99454 2.00000
151 -4.96277 2.00000
152 -4.92088 2.00000
153 -4.91477 2.00000
154 -4.86606 2.00000
155 -4.83963 2.00000
156 -4.81681 2.00000
157 -4.79178 2.00000
158 -4.75537 2.00000
159 -4.73558 2.00000
160 -4.70838 2.00000
161 -4.65602 2.00000
162 -4.64186 2.00000
163 -4.59755 2.00000
164 -4.57465 2.00000
165 -4.54004 2.00000
166 -4.49979 2.00000
167 -4.47810 2.00000
168 -4.44861 2.00000
169 -4.38479 2.00000
170 -4.30018 2.00000
171 -4.24634 2.00000
172 -4.22749 2.00000
173 -4.13705 2.00000
174 -4.04407 2.00000
175 -3.86148 2.00000
176 -3.66326 2.00000
177 -3.54996 2.00000
178 -3.52747 2.00000
179 -3.51617 2.00000
180 -3.48918 2.00000
181 -3.48535 2.00000
182 -3.45750 2.00000
183 -3.41080 2.00000
184 -3.32766 2.00000
185 -3.31196 2.00000
186 -3.27210 2.00000
187 -3.23508 2.00000
188 -3.21352 2.00000
189 -3.19785 2.00000
190 -3.17308 2.00000
191 -3.15281 2.00000
192 -3.13738 2.00000
193 -3.12953 2.00000
194 -3.10737 2.00000
195 -3.07806 2.00000
196 -3.06556 2.00000
197 -3.06221 2.00000
198 -3.04004 2.00000
199 -3.03753 2.00000
200 -3.01146 2.00000
201 -3.00087 2.00000
202 -2.97725 2.00000
203 -2.96947 2.00000
204 -2.95033 2.00000
205 -2.90961 2.00000
206 -2.89446 2.00000
207 -2.86585 2.00000
208 -2.81464 2.00000
209 -2.80289 2.00000
210 -2.75306 2.00000
211 -2.74202 2.00000
212 -2.70760 2.00000
213 -2.69705 2.00000
214 -2.68068 2.00000
215 -2.66161 2.00000
216 -2.60501 2.00000
217 -2.57592 2.00000
218 -2.55472 2.00000
219 -2.53332 2.00000
220 -2.50601 2.00000
221 -2.49766 2.00000
222 -2.47761 2.00000
223 -2.45365 2.00000
224 -2.34536 2.00000
225 -2.29944 2.00000
226 -2.28225 2.00000
227 -2.22395 2.00000
228 -2.14925 2.00000
229 -2.11448 2.00000
230 -2.04941 2.00000
231 -1.77230 2.00000
232 3.47015 0.00000
233 4.05965 0.00000
234 4.76907 0.00000
235 4.92119 0.00000
236 5.06944 0.00000
237 5.23686 0.00000
238 5.35402 0.00000
239 5.38800 0.00000
240 5.67504 0.00000
241 5.77959 0.00000
242 5.82633 0.00000
243 5.92846 0.00000
244 6.02261 0.00000
245 6.08244 0.00000
246 6.12751 0.00000
247 6.23592 0.00000
248 6.30846 0.00000
249 6.36624 0.00000
250 6.52017 0.00000
251 6.58951 0.00000
252 6.66717 0.00000
253 6.69861 0.00000
254 6.77893 0.00000
255 6.89973 0.00000
256 7.05175 0.00000
257 7.08273 0.00000
258 7.11536 0.00000
259 7.15347 0.00000
260 7.20289 0.00000
261 7.29817 0.00000
262 7.36034 0.00000
263 7.39300 0.00000
264 7.49937 0.00000
265 7.59276 0.00000
266 7.63609 0.00000
267 7.68085 0.00000
268 7.72190 0.00000
269 7.79853 0.00000
270 7.83316 0.00000
271 7.92673 0.00000
272 7.98688 0.00000
273 8.03352 0.00000
274 8.12809 0.00000
275 8.14448 0.00000
276 8.25088 0.00000
277 8.25988 0.00000
EFERMI = -1.772298394824647 eV
OUT.ABACUS/ final etot is -26884.33102989164 eV
correction force for each atom along direction 1 is 1.658097903011684e-05
correction force for each atom along direction 2 is 1.227943257710597e-06
correction force for each atom along direction 3 is 1.017309399449144e-05
><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><
TOTAL-FORCE (eV/Angstrom)
><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><
atom x y z
H1 +0.48096941 +0.70969173 -0.80047080
H2 -0.050144161 -0.63341955 -0.11815875
H3 +1.6718354 +1.2422693 +2.2965521
H4 -0.10531528 +1.0704930 -0.75041216
H5 +1.1722744 -0.46078862 +0.35698735
H6 -0.17961370 +0.30874855 +0.64808057
H7 -0.32177274 -0.33048919 -0.31259074
H8 +0.74204574 +0.18787296 +0.060485623
H9 +1.8193521 +0.59574426 -0.34309619
H10 +0.51329726 -0.32866201 -0.72031480
H11 +0.20257213 -0.23053974 +0.59267615
H12 -0.18161689 +0.35043753 +0.19838510
H13 +0.15615839 -0.67696930 +0.078838059
H14 +0.044719171 -1.3615400 -0.24073049
H15 -0.32600390 -0.0040231262 +0.26119144
H16 +0.35990326 +0.61877230 -1.0533055
H17 +0.93586322 -0.56652308 -0.41105650
H18 +0.50047561 -0.064276966 -0.19761125
H19 -0.26117389 +0.49420229 -1.3602747
H20 -0.99637960 -2.5472231 +2.3478563
H21 -0.12620387 +0.48392667 +1.3258744
H22 -1.4116302 -0.33529229 +1.2415928
H23 -0.78045742 -0.21018792 -0.31072496
H24 +0.21081430 -0.40679010 +0.094374212
H25 +0.21273934 +0.43304328 +0.44822157
H26 +0.13966085 -0.33271136 +0.32566487
H27 -0.67144688 -0.44632074 +0.54446162
H28 -2.4795441 +0.042028522 -0.96977937
H29 -0.37181512 -0.17200000 +0.15431077
H30 +2.1039065 +0.032493009 -0.12318493
H31 -0.22350382 -0.72292775 +1.4511636
H32 +0.0084130824 -0.16507758 -0.49155835
H33 -0.042645148 +0.0066157095 +0.085590759
H34 -2.4452487 +1.4426908 +1.5450391
H35 +0.030286173 -0.044138952 -0.36921607
H36 -0.19133150 -0.17266336 +1.0210716
H37 -0.29468743 +0.28404747 -0.55361197
H38 +0.92116311 -0.011844577 -2.1407033
H39 -0.66189699 +0.37420710 +0.19559932
H40 +0.23263600 +0.24369132 -0.12104779
H41 +0.18547225 -0.093857539 -0.57971970
H42 +1.0309600 -0.80162192 -0.76516760
H43 -0.58668349 +0.34734919 +2.2044837
H44 -0.70323973 +0.89472119 -0.59651644
H45 -0.11515728 +0.86080926 +0.13871329
H46 -1.0402475 -1.4064839 -1.4634598
H47 +0.33946064 +0.36828985 -0.78818830
H48 -0.89728210 +1.4581762 +0.040119533
H49 -1.1385098 -0.60685986 -0.49415299
H50 -0.46481241 +0.045666816 +0.087652999
H51 -0.37291954 -0.14392557 -0.39361549
H52 -0.48699041 +0.081754540 -0.0078706809
H53 +0.13001021 +0.17204237 -0.0093401225
H54 -0.33480308 +0.37959297 +0.0069879031
H55 +3.4088559 +3.0039729 +1.3409105
H56 +0.11172661 +0.36552038 +0.61509664
H57 -0.30188568 +0.30788882 -0.87222712
H58 -0.28603654 -0.069939159 +1.0196279
H59 -1.4667942 -0.86109198 +0.59856950
H60 +0.47219138 +0.39798042 +0.067008671
H61 -0.14806512 -0.018952402 +1.1824826
H62 -0.54633078 -0.12133952 +0.069234003
H63 -0.58817222 -0.69780657 +1.0305047
H64 +0.79079274 -0.096391710 -1.5250127
H65 +0.39952918 -0.58916586 +1.6453130
H66 +0.50286347 -0.44258700 +2.9506375
H67 +0.067668462 -0.82306456 +0.94933880
H68 -0.036723745 -1.4204486 -0.15364696
H69 +0.38467434 -0.31250961 -0.60991807
H70 +1.6803802 +1.1714888 +0.50085412
H71 +0.81592543 -0.13055147 +0.23325986
H72 -0.012991837 +0.78478957 +1.1221611
H73 +0.89513694 -0.79516402 +1.0886582
H74 -0.70206870 +0.46043332 +0.31023773
H75 -0.57744550 -0.24299025 +0.25864459
H76 -0.90566013 -0.11050710 -0.22374568
H77 -0.026894771 -0.97427662 +0.099259953
H78 +0.14188597 -0.63657267 -0.83647998
H79 +0.050244896 -0.046109183 +0.24053702
H80 -2.8271437 -1.5689191 -0.42398789
H81 -0.50376483 +1.4514235 +0.24058608
H82 -0.49613437 +0.94304377 -0.35471552
H83 -0.072637706 -0.33295239 -0.033056303
H84 +1.5154769 -0.87623818 +0.68086866
H85 -0.088554318 -0.61365317 +0.94330032
H86 -0.44372153 +0.048423757 +0.16606378
H87 +0.18530683 +0.18998161 +0.16687474
H88 -0.87464318 -0.77272128 +0.20973740
H89 +1.3213285 -1.3849067 -0.35821524
H90 -0.50767165 -0.19359016 +0.59529347
H91 -0.063774990 -0.071014214 +0.25453841
H92 +1.4088989 -0.33172651 -0.30769636
H93 -0.70056017 +0.49455152 -0.13561924
H94 -0.22537704 +0.48701095 +0.079435412
H95 -0.51655129 -0.79909047 +0.083592739
H96 +0.11880465 +0.77352392 +0.21239347
H97 +0.55613189 +1.5265920 +1.1326383
H98 -0.18808910 -0.24362210 -0.027714438
H99 +0.49314200 -0.40552617 -0.17208726
H100 -0.74798206 +0.13216779 -0.20696615
H101 +0.96368255 +1.2314292 -0.88055314
H102 -0.87370024 +0.40038029 +0.21392108
H103 -0.81114626 -0.015788460 +0.99725590
H104 -0.36061179 +0.17980255 -2.0979670
H105 -0.50303333 +0.20722373 -0.77152027
H106 +0.62051876 -1.3121257 +2.0017023
H107 +0.20246228 +0.29654036 -0.58049802
H108 -0.015288230 -0.72698435 +0.28027882
H109 -0.28436894 +0.88824120 -0.70877742
H110 +0.073473722 +0.29743915 -0.12727667
H111 +0.074624273 -0.018605861 +0.12087532
H112 -0.036643578 -0.26315033 +0.011559984
H113 -0.78044870 +2.8032608 -0.16478120
O1 -1.2030135 +1.8529469 +0.62913957
O2 -1.2294966 -0.27371504 +2.7722233
O3 +0.89742016 +0.76119989 -0.17066901
O4 +0.19500594 -0.14567018 -0.52851637
O5 +2.0742425 -1.1129994 -0.85874927
O6 -3.3492941 -1.8162295 -1.1881180
O7 -0.60227955 -0.59311412 +1.1578565
O8 -0.10003688 +0.54995962 -2.4440918
O9 +0.61783870 -0.92605359 -1.8625995
O10 +0.95977547 -1.4667366 +0.99881869
O11 +0.96775365 -1.1283806 -1.7955552
O12 -0.98724816 -1.1533591 -1.3403954
O13 -0.040098638 +0.76120053 -1.2798908
O14 +1.9593693 -0.35180574 +0.89376792
O15 -1.5439376 +0.48688615 +0.28962972
O16 +3.1376818 +1.9667084 +0.62257771
O17 +0.35839609 -0.37821726 +0.25570825
O18 +0.88821889 -0.11349089 +0.94994077
O19 -0.59163402 +1.4678182 -1.3862988
O20 -0.47989375 -0.72177139 +1.1408573
O21 -0.14994357 +0.95720664 -0.35640587
O22 -0.19357644 +0.48278816 -0.39833142
O23 +0.69343308 -0.62650398 -0.27457126
O24 -0.23314929 +0.75739736 +1.9468758
O25 -0.067054515 +1.0907924 -0.10950753
O26 -1.4415151 -0.56061807 +0.13375480
O27 +0.81845933 -1.0535754 +0.017110284
O28 +0.80541175 -0.77930787 +0.68293165
O29 -1.6987951 -1.0345474 -1.6677007
O30 -0.10505782 -1.1369805 +1.2825300
O31 +2.0576844 +0.18651866 -0.027230476
O32 +1.7274135 +0.52748537 -2.2309477
O33 +1.0831526 -0.0058532472 +0.36328692
O34 -1.0882906 +0.80577323 +0.77050671
O35 +1.3300959 +0.20325474 -1.9016910
O36 -0.86765572 +0.60838443 -3.1603118
O37 -0.56823424 +0.23220642 -0.64405670
O38 -0.45562609 -0.10438344 +0.028381676
O39 -0.41184215 +0.54302659 +0.64143245
O40 -0.12228119 +0.48309189 +1.1512981
O41 -0.051972356 -0.89480105 +1.1655228
O42 +0.24917478 +0.54401172 +0.010129432
O43 +0.17146260 -0.80955112 -0.57769640
O44 +1.7842386 +1.2940435 -2.1883350
O45 +0.13621901 +0.020541096 -0.79706504
O46 -2.4798129 -0.17776721 -3.1365447
O47 +0.38380110 -2.6662810 +0.37702562
O48 +0.14913148 +0.61274148 +0.0079436920
O49 +0.66260731 +1.3061527 +1.5532722
O50 -1.4717108 +1.3660259 -2.1956946
O51 +0.48707268 -0.57747006 +0.94384937
O52 +0.74950365 -0.22493944 -0.049001827
O53 +1.1643262 -0.46584961 -0.25855320
O54 -0.94057355 +0.57023197 +0.42474692
O55 +1.2901126 -0.87000379 +1.2287939
O56 +0.56245948 +0.50269089 +0.068679204
N1 -0.39319043 -0.081417097 -2.6905335
C1 -0.29531506 +1.2203175 +1.1982769
C2 -0.81565481 -0.68525539 -1.5946904
--------------------------------------------
!FINAL_ETOT_IS -26884.33102989164 eV
--------------------------------------------
|CLASS_NAME---------|NAME---------------|TIME(Sec)-----|CALLS----|AVG------|PER%-------
total 274.00 7 39. 1.0e+02%
Driver driver_line 273.99 1 2.7e+02 1.e+02%
PW_Basis setuptransform 0.36106 1 0.36 0.13%
PW_Basis setup_struc_factor 0.99831 1 1.0 0.36%
ORB_control read_orb_first 0.43508 1 0.44 0.16%
LCAO_Orbitals Read_Orbitals 0.43508 1 0.44 0.16%
NOrbital_Lm extra_uniform 0.26745 63 0.0042 0.098%
Mathzone_Add1 Uni_Deriv_Phi 0.23874 63 0.0038 0.087%
ORB_control set_orb_tables 11.402 1 11. 4.2%
ORB_gen_tables gen_tables 11.402 1 11. 4.2%
ORB_table_phi init_Table 3.0108 1 3.0 1.1%
ORB_table_phi cal_ST_Phi12_R 2.9945 548 0.0055 1.1%
ORB_table_beta init_Table_Beta 1.2610 1 1.3 0.46%
ORB_table_beta VNL_PhiBeta_R 1.2545 312 0.0040 0.46%
ORB_table_alpha init_Table_Alpha 6.8116 1 6.8 2.5%
ORB_table_alpha S_PhiAlpha_R 6.7816 1155 0.0059 2.5%
ppcell_vl init_vloc 0.32731 1 0.33 0.12%
ESolver_KS_LCAO Run 237.49 1 2.4e+02 87.%
ESolver_KS_LCAO beforescf 3.0821 1 3.1 1.1%
ESolver_KS_LCAO beforesolver 2.0786 1 2.1 0.76%
ESolver_KS_LCAO set_matrix_grid 0.22531 1 0.23 0.082%
Grid_Technique init 0.18953 1 0.19 0.069%
LCAO_Deepks build_psialpha 1.8341 1 1.8 0.67%
ORB_gen_tables snap_psialpha_half 1.7062 44699 3.8e-05 0.62%
Charge atomic_rho 0.79093 1 0.79 0.29%
PW_Basis recip2real 1.0473 77 0.014 0.38%
PW_Basis gathers_scatterp 0.69031 77 0.0090 0.25%
Potential init_pot 0.19203 1 0.19 0.070%
Potential update_from_charge 2.7668 15 0.18 1.0%
Potential cal_v_eff 2.7569 15 0.18 1.0%
H_Hartree_pw v_hartree 0.27649 15 0.018 0.10%
PW_Basis real2recip 0.80642 120 0.0067 0.29%
PW_Basis gatherp_scatters 0.32097 120 0.0027 0.12%
PotXC cal_v_eff 2.4671 15 0.16 0.90%
XC_Functional v_xc 2.4592 15 0.16 0.90%
HSolverLCAO solve 225.38 14 16. 82.%
HamiltLCAO updateHk 195.17 14 14. 71.%
OperatorLCAO init 169.62 42 4.0 62.%
Nonlocal<OperatorLCAO> contributeHR 0.12831 1 0.13 0.047%
LCAO_gen_fixedHbuild_Nonlocal_beta_new 0.12825 1 0.13 0.047%
Veff contributeHk 25.557 14 1.8 9.3%
Gint_interface cal_gint 49.163 29 1.7 18.%
Gint_interface cal_gint_vlocal 22.905 14 1.6 8.4%
Gint_Gamma distri_vl 2.6080 14 0.19 0.95%
Gint_Gamma distri_vl_index 2.2035 14 0.16 0.80%
Gint_Gamma distri_vl_value 0.40449 14 0.029 0.15%
DeePKS contributeHR 169.42 14 12. 62.%
LCAO_Deepks cal_projected_DM 164.29 16 10. 60.%
LCAO_gen_fixedH add_v_delta 116.86 7 17. 43.%
LCAO_DESCRIPTOR add_v_delta 116.60 7 17. 43.%
HSolverLCAO hamiltSolvePsiK 6.3403 14 0.45 2.3%
DiagoElpa elpa_solve 1.8431 14 0.13 0.67%
ElecStateLCAO psiToRho 23.870 14 1.7 8.7%
elecstate cal_dm 0.17092 15 0.011 0.062%
psiMulPsiMpi pdgemm 0.16168 15 0.011 0.059%
LCAO_Charge dm_2dTOgrid 0.29234 14 0.021 0.11%
Gint_interface cal_gint_rho 21.613 14 1.5 7.9%
Charge mix_rho 0.26330 13 0.020 0.096%
Force_Stress_LCAO getForceStress 22.647 1 23. 8.3%
Forces cal_force_loc 0.95109 1 0.95 0.35%
Forces cal_force_ew 0.83219 1 0.83 0.30%
Forces cal_force_scc 1.2014 1 1.2 0.44%
Force_LCAO_gamma ftable_gamma 19.662 1 20. 7.2%
Force_LCAO_gamma allocate_gamma 0.54057 1 0.54 0.20%
LCAO_gen_fixedHbuild_Nonlocal_mu_new 0.35481 1 0.35 0.13%
Force_LCAO_gamma cal_foverlap 0.13804 1 0.14 0.050%
Force_LCAO_gamma cal_edm_2d 0.13528 1 0.14 0.049%
Force_LCAO_gamma cal_fvnl_dbeta_new 0.22399 1 0.22 0.082%
Force_LCAO_gamma cal_fvl_dphi 4.6455 1 4.6 1.7%
Gint_interface cal_gint_force 4.6455 1 4.6 1.7%
----------------------------------------------------------------------------------------
CLASS_NAME---------|NAME---------------|MEMORY(MB)--------
7.620e+04
ORB_table_phi Jl(x) 138.2
ORB_table_alpha Table_DSR 16.60
Charge_Mixing Rrho 10.99
ORB_table_phi Table_SR&TR 10.72
Charge_Mixing dRrho 9.613
Charge_Mixing drho 9.613
Grid_Meshcell index2normal 7.273
Grid_Meshcell index2ucell 7.273
LocalOrbital_Charge Density_Kernal 6.249
ORB_table_beta Table_NR 4.136
PW_Basis struc_fac 4.034
atoms_on_grid which_atom 2.377
atoms_on_grid which_bigcell 2.377
atoms_on_grid which_unitcell 2.377
force_lo dS 1.692
force_lo dTVNL 1.692
init_meshball meshball_pos 1.426
Charge rho 1.373
Charge rho_save 1.373
Charge rho_core 1.373
Potential v_effective_fixed 1.373
Potential vr_eff 1.373
Charge_Mixing rho_save2 1.373
----------------------------------------------------------
Start Time : Sun Jan 8 00:19:35 2023
Finish Time : Sun Jan 8 00:24:09 2023
Total Time : 0 h 4 mins 34 secs
代码
文本
在这里，我们关注：
...（省略）
LCAO ALGORITHM --------------- ION= 1 ELEC= 1--------------------------------
vlocal exchange index is built
buffer size(M): = 12
buffer index size(M): = 12
vlocal data are put in sender_buffer, size(M): = 6
vlocal data are exchanged, received size(M): = 6
K-S equation was solved by genelpa2
eigenvalues were copied to ekb
cal_dk_gamma_from_2D, NSPIN = 1
number of non-zero elements in sender_buffer = 826434
sender_size = 826434
last sender_buffer = -0.00404168
number of non-zero elements in receiver_buffer = 819025
receiver_size = 819025
last receiver_buffer = 0.00622052
Density error is 0.227314000648
Energy Rydberg eV
E_KohnSham -1976.27439151 -26888.592536
E_Harris -1988.47857191 -27054.6389289
E_Fermi -0.419967773079 -5.71395469025
...（省略）
LCAO ALGORITHM --------------- ION= 1 ELEC= 14--------------------------------
vlocal exchange index is built
buffer size(M): = 12
buffer index size(M): = 12
vlocal data are put in sender_buffer, size(M): = 6
vlocal data are exchanged, received size(M): = 6
K-S equation was solved by genelpa2
eigenvalues were copied to ekb
cal_dk_gamma_from_2D, NSPIN = 1
number of non-zero elements in sender_buffer = 826434
sender_size = 826434
last sender_buffer = -0.00347386253503
number of non-zero elements in receiver_buffer = 819025
receiver_size = 819025
last receiver_buffer = 0.00581477661417
Density error is 8.74652341678e-08
Energy Rydberg eV
E_KohnSham -1975.9611767 -26884.3310299
E_Harris -1975.9611767 -26884.3310299
E_band -315.293237188 -4289.78456662
E_one_elec -2526.32621042 -34372.4314685
E_Hartree +1424.85196768 +19386.105567
E_xc -492.328005796 -6698.4661638
E_Ewald -367.312466806 -4997.54249499
E_demet +0 +0
E_descf +0 +0
E_exx +0 +0
E_DeePKS -10.2075189363 -138.880419977
E_Fermi -0.130261482713 -1.77229839482
charge density convergence is achieved
final etot is -26884.3310299 eV
...（省略）
><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><
TOTAL-FORCE (eV/Angstrom)
><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><
atom x y z
H1 +0.48096941 +0.70969173 -0.80047080
H2 -0.050144161 -0.63341955 -0.11815875
H3 +1.6718354 +1.2422693 +2.2965521
...（省略）
可以看到，经过 14 次迭代后，电荷密度收敛，密度误差达到 8.74652341678e-08，最终总能量为 -26884.3310299 eV。
输出文件中也打印了受力，最后收集坐标、能量和受力以制作DP的数据集。
代码
文本
已有溶液体系的反应力场并不全面，这里我们使用 DeePMD-kit 训练一个DP模型。在这个示例中，训练集来自章节4。
代码
文本
5.1 下载教程资源
我们已经在 05-DP_Training 中准备了需要的文件。
代码
文本
[74]
! tree 05-DP_Training
05-DP_Training ├── lcurve.out ├── partial_dataset │ └── glycine │ └── 001 │ ├── set.000 │ │ ├── box.npy │ │ ├── coord.npy │ │ ├── energy.npy │ │ └── force.npy │ ├── type.raw │ └── type_map.raw ├── run.json └── sub.sh 4 directories, 9 files
代码
文本
在 05-DP_Training 文件夹下输入文件有
- partial_dataset 文件夹用于存放训练和测试数据，
- run.json 包含使用 DeePMD-kit 训练模型的参数设置，
代码
文本
5.2 准备训练数据
调用 dpdata 的工具，将 ABACUS生成的数据转换为 DeePMD-kit 训练所需的数据格式（NumPy数组）。
代码
文本
在开始训练之前，我们可以先检查数据集组成
代码
文本
[75]
! tree 05-DP_Training/partial_dataset/glycine/001
05-DP_Training/partial_dataset/glycine/001 ├── set.000 │ ├── box.npy │ ├── coord.npy │ ├── energy.npy │ └── force.npy ├── type.raw └── type_map.raw 1 directory, 6 files
代码
文本
这些文件的作用如下：
set.000：是一个目录，包含压缩格式的数据（NumPy压缩数组）。所有训练数据应首先转换为此格式，然后在 DeePMD-kit 中使用。该数据格式在 DeePMD-kit 手册中有详细解释，可以在 DeePMD-kit Data Introduction 中找到。type.raw：是一个文件，包含原子的类型（以整数表示）。type_map.raw：是一个文件，包含原子的类型名称。
让我们来看一下这些文件：
代码
文本
[76]
! cat 05-DP_Training/partial_dataset/glycine/001/type.raw
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 3 3
代码
文本
这告诉我们这个例子中有 172 个原子，其中4个原子类型由"0、1、2、3"表示。映射可以通过文件type_map.raw给出。
由于系统中的所有帧都具有相同的原子类型和原子序号，因此我们只需为整个系统指定一次类型信息。
代码
文本
[77]
! cat 05-DP_Training/partial_dataset/glycine/001/type_map.raw
H O N C
代码
文本
其中原子H、O、N、C分别被赋予类型"0、1、2、3"
代码
文本
5.3 准备输入脚本
训练数据准备完成后，接下来就可以进行训练。让我们进入训练目录看一下运行所需的脚本：
代码
文本
[78]
! cat 05-DP_Training/run.json
{
"model": {
"type_map": ["H","O","N","C"],
"descriptor": {
"_comment": "if type = se_a_tpe: can’t apply compression method while using atom type embedding",
"type": "se_e2_a",
"_comment": "sel: [16 32 32] means maximal number of neighbors = 16H 32O 32N",
"_comment": "sel: auto:1.1 means automatically counts maximal number of neighbors*1.1",
"sel": "auto:1.1",
"rcut_smth": 0.5,
"rcut": 6.0,
"neuron": [25,50,100],
"activation_function": "tanh",
"resnet_dt": false,
"_comment": "axis_neuron: Size of the submatrix of G (embedding matrix)",
"axis_neuron": 16,
"seed": 20230129,
"_comment": "descriptor that's all--------------------------------------------------------------"
},
"fitting_net": {
"_comment": "other types: dipole or polar",
"type": "ener",
"neuron": [240,240,240],
"activation_function": "tanh",
"resnet_dt": true,
"seed": 20230129,
"_comment": "fitting_net that's all-------------------------------------------------------------"
},
"_comment": " model that's all------------------------------------------------------------------------------"
},
"loss": {
"_comment": "loss = pref_e * loss_e + pref_f * loss_f + pref_v * loss_v",
"_comment": "pref_f(t) = start_pref_f * ( lr(t) / start_lr ) + limit_pref_f * ( 1 - lr(t) / start_lr )",
"start_pref_e": 0.02,
"limit_pref_e": 1.0,
"start_pref_f": 1000,
"limit_pref_f": 1.0,
"start_pref_v": 0.0,
"limit_pref_v": 0.0,
"_comment": " loss that's all-----------------------------------------------------------------------------"
},
"learning_rate": {
"_comment": "lr(t) = start_lr * decay_rate ^ ( training step / decay_steps )",
"_comment": "decay_rate and decay_steps are automatically determined by start_lr, stop_lr and training step)",
"type": "exp",
"_comment": "When parallel training or batch size scaled, how to alter learning rate. Valid values are linear`(default), `sqrt or none.",
"_scale_by_worker": "sqrt",
"start_lr": 1e-03,
"stop_lr": 1e-08,
"_comment": " learning_rate that's all--------------------------------------------------------------------"
},
"training": {
"numb_steps": 100000,
"seed": 20230129,
"disp_file": "lcurve.out",
"disp_freq": 1000,
"numb_test": 10,
"save_freq": 10000,
"save_ckpt": "model.ckpt",
"disp_training": true,
"time_training": true,
"profiling": false,
"profiling_file": "timeline.json",
"training_data": {
"systems": [
"partial_dataset/glycine/001"
],
"batch_size": "auto",
"_auto_prob": "prob_sys_size; 0:14:0.25; 14:56:0.75",
"_comment": " training_data that's all--------------------------------------------------------------------"
},
"_comment": " training that's all-------------------------------------------------------------------------"
}
}
代码
文本
input.json 包含了 DP 模型训练过程中所需的各种参数，定义和控制训练任务。这些参数在 DeePMD-kit 手册中有详细的解释，所以这里只做简单介绍。
在 model 模块, 指定嵌入和拟合网络的参数。
"model":{
"type_map": ["H","O","N","C"], # 元素名称
"descriptor":{
"type": "se_e2_a", # 描述符类型
"sel": "auto:1.1", # 原子的选定邻居数
"rcut_smth": 0.50, # 光滑截止半径
"rcut": 6.00, # 截止半径
"neuron": [25,50,100], # 嵌入网络尺寸
"resnet_dt": false,
"axis_neuron": 16, # 嵌入子网络横向尺寸
},
"fitting_net":{
"neuron": [240,240,240], # 拟合网络尺寸
"resnet_dt": true,
},
},
描述符 se_e2_a 用于 DP 模型的训练。将嵌入和拟合神经网络的大小分别设置为 [25,50,100] 和 [240,240,240]。 里的成分会从 0.5 到 6 Å 平滑地趋于 0。
下面的参数指定学习效率和损失函数：
"learning_rate" :{
"type": "exp",
"start_lr": 1e-03, # 起始学习率
"stop_lr": 1e-08, # 结束学习率
},
"loss" :{
"type": "ener",
"start_pref_e": 0.02, # 能量起始权重
"limit_pref_e": 1, # 能量最终权重
"start_pref_f": 1000, # 力起始权重
"limit_pref_f": 1, # 力最终权重
"start_pref_v": 0, # 维里
"limit_pref_v": 0,
},
在损失函数中， 从 0.02 逐渐增加到 1 ，而 f 从 1000 逐渐减小到 1 ，这意味着力项在开始时占主导地位，而能量项和维里项在结束时变得重要。这种策略非常有效，并且减少了总训练时间。 设为 0 ，这表明训练过程中不包含任何维里数据。将起始学习率、停止学习率分别设置为 1e-3，1e-8。模型训练步数为 。
训练参数如下：
"training" : {
"training_data": {
"systems": ["partial_dataset/glycine/001"], # 训练数据路径
"batch_size": "auto", # 自动确定，natoms*batch_size≥32
},
"numb_steps": 100000, # 训练步数
"disp_file": "lcurve.out", # 写入学习曲线到文件
"disp_freq": 1000, # 写入学习曲线的频率
"save_freq": 10000, # 保存模型相关文件频率
},
代码
文本
5.4 模型训练
准备好训练脚本后，我们可以用 DeePMD-kit 开始训练，本地机器只需运行qsub sub.sh:
代码
文本
[79]
! head -n 5 05-DP_Training/lcurve.out
! tail -n 4 05-DP_Training/lcurve.out
# step rmse_trn rmse_e_trn rmse_f_trn lr
0 2.93e+01 4.04e-01 9.25e-01 1.0e-03
1000 8.96e+00 3.79e-02 2.83e-01 1.0e-03
2000 7.41e+00 4.41e-02 2.34e-01 1.0e-03
3000 6.39e+00 1.03e-01 2.02e-01 1.0e-03
9997000 6.88e-02 4.36e-04 6.82e-02 1.0e-08
9998000 4.95e-02 6.52e-04 4.86e-02 1.0e-08
9999000 5.79e-02 4.61e-04 5.73e-02 1.0e-08
10000000 5.72e-02 4.55e-05 5.69e-02 1.0e-08
代码
文本
第2、3、4列分别介绍了总loss、能量和力的训练误差。 本工作经过 10000000 步训练，训练误差可以通过简单的 Python 脚本对该文件进行可视化：
代码
文本
[80]
import numpy as np
import matplotlib.pyplot as plt
data = np.genfromtxt("./05-DP_Training/lcurve.out", names=True)
for name in data.dtype.names[1:-1]:
plt.plot(data['step'], data[name], label=name)
plt.legend()
plt.xlabel('Step')
plt.ylabel('Loss')
plt.yscale('log')
plt.show()
代码
文本
当训练过程异常停止时，我们可以从提供的检查点重新开始训练，只需运行
代码
文本
[ ]
! cd ./05-DP_Training && dp train --restart model.ckpt input.json
代码
文本
需要注意的是 input.json 需要和上一个保持一致。
代码
文本
5.5 冻结和压缩模型
在训练结束时，保存在 TensorFlow 的 checkpoint 文件中的模型参数通常需要冻结为一个以扩展名 .pb 结尾的模型文件。 当训练结束时，在sub.sh脚本中会自动执行dp freeze，将输出一个名为frozen_model.pb模型文件。紧接着，压缩frozen_model.pb模型通常会将基于 DP 的计算速度提高一个数量级，并且消耗更少的内存。 通过dp compress压缩，将输出一个名为frozen_model_cpmpressed.pb模型文件。
代码
文本
代码
文本
[82]
! cat 05-DP_Training/sub.sh
#!/bin/bash
#PBS -l select=1:ncpus=4:mpiprocs=1:ngpus=1:ompthreads=4
#PBS -l walltime=24:00:00
#PBS -q gpu_a100
#PBS -o dptrain.out
#PBS -j oe
#PBS -N train
# change dir submission dir
cd $PBS_O_WORKDIR
# source env
export PATH=/work/pzhang/app/miniconda/3/bin:$PATH
source activate deepmdkit_2.1.5
# run job
if [ ! -f tag_0_finished ] ;then
{ if [ ! -f model.ckpt.index ]; then
CUDA_VISIBLE_DEVICES=0 horovodrun -np 1
dp train --mpi-log=master run.json ;
else
CUDA_VISIBLE_DEVICES=0 horovodrun -np 1
dp train --mpi-log=master run.json --restart model.ckpt; fi }
1>> train.log 2>> train.log
{ if test $? -ne 0; then touch tag_0_failure;
else
dp freeze && dp compress && touch tag_0_finished; fi }
fi
# Run the application - the line below is just a random examp
#dp train run.json
#dp train run.json --restart model.ckpt
#dp freeze
#dp compress
代码
文本
5.6 模型测试
我们可以通过运行dp test命令检查训练模型的质量，测试集来自独立的构象，并在M06-2X级别下完成计算。本工作的测试误差如下图（因为描述子更简单，在构象数数量比DeePKS数据集高两个数量级下，DP模型测试误差相对比DeePKS模型大一些，但精度足够用以跑DPMD）
代码
文本
6.1 下载教程资源
我们已经在 06-DPMD+Enhanced_sampling 中准备了需要的文件。
代码
文本
[83]
! tree 06-DPMD+Enhanced_sampling
06-DPMD+Enhanced_sampling ├── COLVAR ├── Glycine128H2O_opt_atomic.data ├── geo.lammpstrj ├── in.lammps └── input.plumed 0 directories, 5 files
代码
文本
in.lammps和Glycine128H2O_opt_atomic.data是 LAMMPS 的输入文件input.plumed是 PLUMED 的输入文件
本教程采用 DeePMD-kit(2.1.5)软件包中预置的 LAMMSP 程序完成，PLUMED需要手动修改CV代码后，借助plumed-feedstock安装在DeePMD-kit环境下。
代码
文本
6.2 LAMMPS输入文件
溶液体系 LAMMPS NVT-MD 的控制文件如下：
代码
文本
[84]
! cat 06-DPMD+Enhanced_sampling/in.lammps
# Variables: Specifies variables and their values to be used in the script.
variable NSTEPS equal 500000
variable THERMO_FREQ equal 10
variable DUMP_FREQ equal 10
variable TEMP equal 300.0
variable TAU_T equal 0.040
variable SEED1 equal 20230216
variable SEED2 equal 20230216
# Prints the following lines to the screen
echo screen
units metal # Specifies the units used in the simulation (metal units).
boundary p p p # Sets the boundary conditions for the simulation (periodic in all three directions).
atom_style atomic # Sets the atom style used to define atom properties (atomic style).
read_data Glycine128H2O_opt_atomic.data # Reads the data file that contains the initial atomic positions and properties of the system.
#read_restart ../5000000.comp
neighbor 1.0 bin # Defines the neighbor list settings for pairwise interactions (bin method with a cutoff of 1.0).
pair_style deepmd frozen_model.pb # Specifies the potential model used for pairwise interactions (Deep Potential Molecular Dynamics).
pair_coeff * * # indicates that the potential coefficients are assigned to all pairs of atom types.
thermo_style custom step temp density pe ke etotal vol press # Specifies the output style for thermodynamic properties.
thermo ${THERMO_FREQ} # is the frequency at which thermodynamic output should be printed.
velocity all create ${TEMP} ${SEED2} dist gaussian # Assigns velocities to all atoms.
# a fix using the PLUMED package with an input and output file specified.
fix dpgen_plm all plumed plumedfile input.plumed outfile output.plumed
# a thermostat fix that controls the temperature using a stochastic velocity rescaling algorithm.
fix 1 all nve
fix 2 all temp/csvr ${TEMP} ${TEMP} ${TAU_T} ${SEED1}
# Specifies the dump output for atomic positions.
dump glycine1 all custom ${DUMP_FREQ} geo.lammpstrj id type x y z
dump_modify glycine1 sort id
restart 100000 restart.*.cont # is the frequency at which restart files will be written.
timestep 0.001 # ps, Specifies the length of each time step in the simulation.
run ${NSTEPS} # Runs the simulation for the specified number of steps (${NSTEPS}).
write_restart *.comp # Writes a restart file for the simulation.
write_data *.data # Writes the final atomic configuration to a data file.代码
文本
6.3 PLUMED输入文件
与章节1中的采样设置类似，我们只额外加了一些用于后处理分析的输出参数，溶液体系 OPES Voronoi CVs 的控制文件如下：
代码
文本
[85]
! cat 06-DPMD+Enhanced_sampling/input.plumed
# this is using Angstrom UNITS LENGTH=A # set atom groups GlyN: GROUP ATOMS=385 GlyOhasH: GROUP ATOMS=391 GlyOnotH: GROUP ATOMS=393 GlyH: GROUP ATOMS=386,392,394 WaterO: GROUP ATOMS=1-382:3 WaterH: GROUP ATOMS=2-383:3,3-384:3 # define CVs cp: VORONOIC1P GROUPA=WaterO,GlyN,GlyOhasH,GlyOnotH GROUPB=WaterH,GlyH NRX=3 LAMBDA=-30 D_0=2 D_1=2 D_2=0.5 D_3=0.5 NLIST NL_CUTOFF=2.4 NL_STRIDE=1 cm: VORONOIC1M GROUPA=WaterO,GlyN,GlyOhasH,GlyOnotH GROUPB=WaterH,GlyH NRX=3 LAMBDA=-30 D_0=2 D_1=2 D_2=0.5 D_3=0.5 NLIST NL_CUTOFF=2.4 NL_STRIDE=1 d05: VORONOID2 GROUPA=WaterO,GlyN,GlyOhasH,GlyOnotH GROUPB=WaterH,GlyH NRX=3 LAMBDA=-5 D_0=2 D_1=2 D_2=0.5 D_3=0.5 NLIST NL_CUTOFF=2.4 NL_STRIDE=1 s05: VORONOIS1 GROUPA=WaterO,GlyN,GlyOhasH,GlyOnotH GROUPB=WaterH,GlyH NRX=3 LAMBDA=-5 D_0=256 D_1=3 NLIST NL_CUTOFF=2.4 NL_STRIDE=1 c05: VORONOIC0 GROUPA=WaterO,GlyN,GlyOhasH,GlyOnotH GROUPB=WaterH,GlyH NRX=3 LAMBDA=-5 D_0=2 D_1=2 D_2=0.5 D_3=0.5 NLIST NL_CUTOFF=2.4 NL_STRIDE=1 d08: VORONOID2 GROUPA=WaterO,GlyN,GlyOhasH,GlyOnotH GROUPB=WaterH,GlyH NRX=3 LAMBDA=-8 D_0=2 D_1=2 D_2=0.5 D_3=0.5 NLIST NL_CUTOFF=2.4 NL_STRIDE=1 s08: VORONOIS1 GROUPA=WaterO,GlyN,GlyOhasH,GlyOnotH GROUPB=WaterH,GlyH NRX=3 LAMBDA=-8 D_0=256 D_1=3 NLIST NL_CUTOFF=2.4 NL_STRIDE=1 c08: VORONOIC0 GROUPA=WaterO,GlyN,GlyOhasH,GlyOnotH GROUPB=WaterH,GlyH NRX=3 LAMBDA=-8 D_0=2 D_1=2 D_2=0.5 D_3=0.5 NLIST NL_CUTOFF=2.4 NL_STRIDE=1 d100: VORONOID2 GROUPA=WaterO,GlyN,GlyOhasH,GlyOnotH GROUPB=WaterH,GlyH NRX=3 LAMBDA=-100 D_0=2 D_1=2 D_2=0.5 D_3=0.5 NLIST NL_CUTOFF=2.4 NL_STRIDE=1 s100: VORONOIS1 GROUPA=WaterO,GlyN,GlyOhasH,GlyOnotH GROUPB=WaterH,GlyH NRX=3 LAMBDA=-100 D_0=256 D_1=3 NLIST NL_CUTOFF=2.4 NL_STRIDE=1 c100: VORONOIC0 GROUPA=WaterO,GlyN,GlyOhasH,GlyOnotH GROUPB=WaterH,GlyH NRX=3 LAMBDA=-100 D_0=2 D_1=2 D_2=0.5 D_3=0.5 NLIST NL_CUTOFF=2.4 NL_STRIDE=1 d_N_O1: DISTANCE ATOMS=GlyN,GlyOhasH d_N_O2: DISTANCE ATOMS=GlyN,GlyOnotH ene: ENERGY # define opes OPES_METAD ... RESTART=NO LABEL=opes ARG=s05,d05 FILE=HILLS TEMP=300 PACE=500 BARRIER=35 STATE_WFILE=STATE STATE_WSTRIDE=5000 STORE_STATES WALKERS_MPI ... OPES_METAD # print CVs and bias FLUSH STRIDE=1000 PRINT ... ARG=s05,d05,c05,s08,d08,c08,s100,d100,c100,cp,cm,d_N_O1,d_N_O2,opes.*,ene STRIDE=1 FILE=COLVAR RESTART=NO ... PRINT ENDPLUMED
代码
文本
6.4 结果简析
本工作进行了30ns的充分增强采样，CVs与bias随时间的变化如下图
最终我们可以计算得到的反应自由能面，基本包含了所有的构象转变
正好对应3条质子转移路径
代码
文本
结语
在这项工作中，我们通过如下工作流（主要借助DeePKS、DeePMD、OPES、Voronoi CVs）完成了甘氨酸在水溶液中的多种质子转移通道的探索，相信该方法也能适用于多种场景。
代码
文本
参考文献
如果您对本notebook中介绍的具体案例感兴趣，请看
Intramolecular and water mediated tautomerism of solvated glycine JCIM arXiv
Cooresponding notebooks：
OPES(on-the-fly probability enhanced sampling)
Voronoi CVs for enhanced sampling autoionization and tautomerism
代码
文本
已赞9
本文被以下合集收录
机器学习与DFT精华帖
gtang
更新于 2024-09-10
38 篇21 人关注
打通DPA2-ABACUS-DeepKS-DeepMD的基础与案例
weiqingzhou@whu.edu.cn
更新于 2024-05-21
1 篇5 人关注
推荐阅读
公开
从 DFT 先去 DeePKS 再到 DeePMD | DeePKS案例篇 + 增强采样副本山妖
发布于 2024-05-21
公开
从 DFT 先去 DeePKS 再到 DeePMD | DeePKS基础篇zhangpengchao@dp.tech
发布于 2023-09-28
10 赞14 转存文件
评论
zhangwei2011@hotmail.com
zhangpengchao@dp.tech