orca SCF迷一样的感受

wzkchem5 · 发表于 Post on 2022-3-10 15:32:17

冰释之川发表于 2022-3-10 01:36
下一个版本能否让不支持的泛函自动关闭TRAH，不然用起来不太人性化啊。。

下一个版本应该会让几乎所有泛函都支持TRAH，所以估计不会做自动关闭功能

wzkchem5 · 发表于 Post on 2022-3-10 15:34:00

413 发表于 2022-3-10 02:48
体系为碳氢吸附在金银合金团簇上

这个是orca初猜的一个bug，已经在orca 5.0.3修复了

wzkchem5 · 发表于 Post on 2022-3-10 15:40:18

413 发表于 2022-3-10 05:27
请问正确的格式是再加个ZORA吗？

! pbe angs PRINTBASIS PRINTMOS SARC/J ZORA

应该没问题，不过更方便的写法是先在第一行给所有元素都指定ZORA-def2-SVP，再用NewGTO把Au和Ag的基组改成SARC-ZORA-TZVP，这样也能把DelECP省了

413 · 发表于 Post on 2022-3-10 18:03:27

wzkchem5 发表于 2022-3-10 15:40
应该没问题，不过更方便的写法是先在第一行给所有元素都指定ZORA-def2-SVP，再用NewGTO把Au和Ag的基组改 ...

请问是这样指定吗？

! pbe angs PRINTBASIS PRINTMOS SARC/J ZORA ZORA-def2-SVP
%maxcore 5000
%scf MaxIter 512 Thresh 1e-6 end
%pal nprocs 28 end
%basis
NewGTO Au "SARC-ZORA-TZVP" end #
NewGTO Ag "SARC-ZORA-TZVP" end #
end

wzkchem5 · 发表于 Post on 2022-3-10 19:50:49

413 发表于 2022-3-10 11:03
请问是这样指定吗？

! pbe angs PRINTBASIS PRINTMOS SARC/J ZORA ZORA-def2-SVP

对

413 · 发表于 Post on 2022-3-11 21:49:18

本帖最后由 413 于 2022-3-11 22:00 编辑

wzkchem5 发表于 2022-3-10 19:50
对

搭车请教下，我要算Au、Ag、Pt等过渡元素，想给LANL2DZ指定auxiliary basis set，一直是用autoaux来让orca自动指定。

手册有段话是写了def2/J。但是只能到Kr，不能算Au、Ag、Pt等
The RI-J approximation requires an \auxiliary basis set" in addition to a normal orbital basis set. For
the Karlsruhe basis sets there is the universal auxiliary basis set of Weigend that is called with the
name def2/J (all-electron up to Kr). When scalar relativistic Hamiltonians are used (DKH or ZORA)
along with all-electron basis sets, then a general-purpose auxiliary basis set is the SARC/J that covers
most of the periodic table. Other choices are documented in sections 6.3 and 9.5.

文献里也有，可是手册已经说了def2/J只能用到Kr，还可以给I元素用吗？
“the LANL2DZ pseudopotential and basis set was employed for Cu and I atoms. The def2/J auxiliary basis set was implemented for the resolution-of-identity procedure in the ORCA program”

还想请教下，lanl2dz和zora的全电子基组ZORA-def2-TZVP算Au、Ag、Cu的速度相当吗？我在https://sites.google.com/site/or ... stic-approximations这里还是哪里好像瞟见过速度相当，但是回头找那个网页怎么都找不到了。

wzkchem5 · 发表于 Post on 2022-3-11 22:00:20

413 发表于 2022-3-11 14:49
搭车请教下，我想给LANL2DZ指定auxiliary basis set，一直是用autoaux来让orca自动指定。

查了文献也 ...

我觉得问题不大，原则上讲一个辅助基组X如果可以用于比较大的基组A，另有一个基组B在所有方面（极化函数，弥散函数，价层基函数，内层基函数）普遍小于等于A，那么X也可以用于B。比如适合6-311+G(2df,2pd)的辅助基组（如果有的话）一定可以适用于6-31G(d,p)。LANL2DZ应该是在这些方面普遍小于def2-TZVP的，而def2-TZVP可以结合def2/J，所以我觉得LANL2DZ也可以结合def2/J。但是毕竟LANL2DZ和def2-TZVP不是一个系列的，不排除会被人质疑，所以最好还是能找到文献能证明这样用的精度足够。也就是光查到这么用的文献还不够，还得查查看有没有人用LANL2DZ+def2/J做RI（或RIJCOSX）和不用RI的结果对比，证明两个结果的差别可以忽略。

413 · 发表于 Post on 2022-3-11 22:03:47

wzkchem5 发表于 2022-3-11 22:00
我觉得问题不大，原则上讲一个辅助基组X如果可以用于比较大的基组A，另有一个基组B在所有方面（极化函数 ...

还想请教下，lanl2dz和zora的全电子基组ZORA-def2-TZVP算Au、Ag、Cu的速度相当吗？我在https://sites.google.com/site/or ... stic-approximations这里还是哪里好像瞟见过速度相当，但是回头找那个网页怎么都找不到了。

wzkchem5 · 发表于 Post on 2022-3-11 22:10:18

413 发表于 2022-3-11 15:03
还想请教下，lanl2dz和zora的全电子基组ZORA-def2-TZVP算Au、Ag、Cu的速度相当吗？我在https://sites.goo ...

如果非金属原子远多于金属原子，那么速度相当；如果金属原子占总原子数比较大的比例，那么ZORA-def2-TZVP会明显更慢。原因有三个：（1）ZORA本身会占一点时间；（2）LANL2DZ是double zeta基组，ZORA-def2-TZVP是triple zeta基组；（3）LANL2DZ把一部分内层电子赝化掉了。

413 · 发表于 Post on 2022-3-11 22:24:17

本帖最后由 413 于 2022-3-11 22:25 编辑

wzkchem5 发表于 2022-3-11 22:10
如果非金属原子远多于金属原子，那么速度相当；如果金属原子占总原子数比较大的比例，那么ZORA-def2-TZVP ...

还想请教下，这个报错是不是只能用def2/J了？报错信息好像是指定auxiliary basis时的错误"At the moment, SHARK is not equipped to deal with this"
orca版本5.0.3
输入文件：
! pbe lanl2dz angs PRINTBASIS PRINTMOS autoaux
%maxcore 5000
%scf MaxIter 512 end
%pal nprocs 28 end
%basis
NewGTO C "6-31G(d)" end #
DelECP C
NewGTO H "6-31G(d)" end #
DelECP H
end

Calculating one electron integrals       ... done (  1.5 sec)
Calculating ECP integrals                ... done (132.9 sec)
Calculating RI/J V-Matrix + Cholesky decomp.... [file orca_tools/Tool-GTO-Integrals/SHARK/shark.cpp, line 334, Process 31]:  ... sorry, have to bail out
[file orca_tools/Tool-GTO-Integrals/SHARK/shark.cpp, line 334, Process 1]:  ... sorry, have to bail out
......
[file orca_tools/Tool-GTO-Integrals/SHARK/shark.cpp, line 334, Process 28]:  ... sorry, have to bail out

WARNING! Potentially linear dependencies in the auxiliary basis
===>> At the moment, SHARK is not equipped to deal with this
[file orca_tools/Tool-GTO-Integrals/SHARK/shark.cpp, line 334, Process 0]:  ... sorry, have to bail out

wzkchem5 · 发表于 Post on 2022-3-11 23:22:54

413 发表于 2022-3-11 15:24
还想请教下，这个报错是不是只能用def2/J了？报错信息好像是指定auxiliary basis时的错误"At the moment, ...

这个说明autoaux的辅助基组太大了，出现线性相关问题。一般出现这个的话，尽量选取别的辅助基组

413 · 发表于 Post on 2022-3-11 23:25:45

wzkchem5 发表于 2022-3-11 23:22
这个说明autoaux的辅助基组太大了，出现线性相关问题。一般出现这个的话，尽量选取别的辅助基组

哦哦，我试了def2/J，很快就算完了，就是可能需要benchmark下

413 · 发表于 Post on 2022-3-12 12:47:49

wzkchem5 发表于 2022-3-11 23:22
这个说明autoaux的辅助基组太大了，出现线性相关问题。一般出现这个的话，尽量选取别的辅助基组

想请教下，这种的是不是就不太可能收敛了？需要调下结构重新交作业吗？
看起来TRAH有点重复
5.0.3的版本。输入：
! angs PRINTBASIS PRINTMOS BP86 ZORA/RI ZORA-def2-TZVP SARC/J  SlowConv
%scf
Shift Shift 0.1 ErrOff 0.1 end
end
%maxcore 7000
%scf MaxIter 512  end
%pal nprocs 28 end
%basis
NewGTO Au "SARC-ZORA-TZVP" end
NewGTO Ag "SARC-ZORA-TZVP" end
end

      TRAH Step control
---------------------------------
  predicted energy change = -0.000001
  actual energy  change = 0.000007
  energy change ratio    = -9.270889
  old trust radius       = 0.000268
  new trust radius       = 0.000188
  reject step?          =  1
44 -659738.354670954286    1.480198e-01                   0.000 (TRAH MAcro)  Yes
45 -659738.354670953588    5.078232e-03                   0.000 (TRAH MAcro) No
45    dE -4.537582e-10    1.316559e-02 -3.5420e-09    0.000 (TRAH MIcro)
45    dE -4.963770e-07    7.225652e-03 -8.1193e-01    0.000 (TRAH MIcro)
45    dE -7.055799e-07    2.936996e-03 -1.0849e+00    0.001 (TRAH MIcro)
45    dE -7.420193e-07    8.787871e-04 -1.1572e+00    0.001 (TRAH MIcro)
45    dE -7.341458e-07    2.456786e-04 -1.1709e+00    0.001 (TRAH MIcro)
---------------------------------
      TRAH Step control
---------------------------------
  predicted energy change = -0.000001
  actual energy  change = 0.000006
  energy change ratio    = -8.674073
  old trust radius       = 0.000188
  new trust radius       = 0.000131
  reject step?          =  1
46 -659738.354670953588    1.480198e-01                   0.000 (TRAH MAcro)  Yes
47 -659738.354670953820    5.078230e-03                   0.000 (TRAH MAcro) No
47    dE -4.537574e-10    1.316558e-02 -5.5070e-09    0.000 (TRAH MIcro)
47    dE -4.963647e-07    7.225543e-03 -8.1191e-01    0.000 (TRAH MIcro)
47    dE -7.055467e-07    2.937039e-03 -1.0849e+00    0.001 (TRAH MIcro)
47    dE -7.419772e-07    8.789702e-04 -1.1572e+00    0.001 (TRAH MIcro)
47    dE -7.341059e-07    2.457316e-04 -1.1708e+00    0.001 (TRAH MIcro)
---------------------------------
      TRAH Step control
---------------------------------
  predicted energy change = -0.000001
  actual energy  change = 0.000007
  energy change ratio    = -9.113339
  old trust radius       = 0.000131
  new trust radius       = 0.000092
  reject step?          =  1
48 -659738.354670953820    1.480198e-01                   0.000 (TRAH MAcro)  Yes
49 -659738.354670953820    5.078226e-03                   0.000 (TRAH MAcro) No
49    dE -4.537831e-10    1.316568e-02 -2.0089e-09    0.000 (TRAH MIcro)
49    dE -4.963777e-07    7.225663e-03 -8.1193e-01    0.000 (TRAH MIcro)
49    dE -7.055724e-07    2.937045e-03 -1.0849e+00    0.001 (TRAH MIcro)
49    dE -7.420100e-07    8.788316e-04 -1.1572e+00    0.001 (TRAH MIcro)
49    dE -7.341362e-07    2.457352e-04 -1.1709e+00    0.001 (TRAH MIcro)
---------------------------------
      TRAH Step control
---------------------------------
  predicted energy change = -0.000001
  actual energy  change = 0.000007
  energy change ratio    = -9.062694
  old trust radius       = 0.000092
  new trust radius       = 0.000064
  reject step?          =  1
50 -659738.354670953820    1.480198e-01                   0.000 (TRAH MAcro)  Yes
51 -659738.354670955450    5.078229e-03                   0.000 (TRAH MAcro) No
51    dE -4.537694e-10    1.316567e-02 -3.3268e-09    0.000 (TRAH MIcro)
51    dE -4.963800e-07    7.225714e-03 -8.1193e-01    0.000 (TRAH MIcro)
51    dE -7.055883e-07    2.937027e-03 -1.0849e+00    0.001 (TRAH MIcro)
51    dE -7.420348e-07    8.787032e-04 -1.1573e+00    0.001 (TRAH MIcro)
51    dE -7.341602e-07    2.456871e-04 -1.1709e+00    0.001 (TRAH MIcro)
---------------------------------
      TRAH Step control
---------------------------------
  predicted energy change = -0.000001
  actual energy  change = 0.000007
  energy change ratio    = -9.130583
  old trust radius       = 0.000064
  new trust radius       = 0.000045
  reject step?          =  1
52 -659738.354670955450    1.480198e-01                   0.000 (TRAH MAcro)  Yes
53 -659738.354670953122    5.078204e-03                   0.000 (TRAH MAcro) No
53    dE -4.537601e-10    1.316538e-02 -1.4042e-09    0.000 (TRAH MIcro)
53    dE -4.963827e-07    7.225588e-03 -8.1193e-01    0.000 (TRAH MIcro)
53    dE -7.055955e-07    2.936940e-03 -1.0850e+00    0.001 (TRAH MIcro)
53    dE -7.420383e-07    8.787173e-04 -1.1573e+00    0.001 (TRAH MIcro)
53    dE -7.341627e-07    2.457356e-04 -1.1709e+00    0.001 (TRAH MIcro)

wzkchem5 · 发表于 Post on 2022-3-12 16:03:51

413 发表于 2022-3-12 05:47
想请教下，这种的是不是就不太可能收敛了？需要调下结构重新交作业吗？
看起来TRAH有点重复
5.0.3的版 ...

可能格点不够，数值噪音导致一直不收敛。不太可能是调结构就能解决的问题，但是值得检查一下结构是否合理。

413 · 发表于 Post on 2022-3-12 21:48:29

wzkchem5 发表于 2022-3-12 16:03
可能格点不够，数值噪音导致一直不收敛。不太可能是调结构就能解决的问题，但是值得检查一下结构是否合理 ...

想请教下，对于过渡金属的相对论效应计算，第一圈的自洽场迭代基本徘徊在10^-4左右。我用了较高密度的grid。也试了soscf，还是会徘徊在10^-4左右，手册上有这么一段说明，说是可以进一步调节these parameters，请问有哪些parameters可以调节呢？
orca5.0.3
|  1> ! angs PRINTBASIS PRINTMOS BP86 ZORA/RI ZORA-def2-TZVP SARC/J  SlowConv  defgrid3
|  2> %scf
|  3> AutoTRAH true
|  4> AutoTRAHTOl 1.125 #Threshold to determine when TRAH should be activated. Default value 1.125
|  5> AutoTRAHIter 20 #Number of iterations before interpolation is used
|  6> AutoTRAHNInter 10 #Number of iterations used in interpolation. Default 20.
|  7> Shift Shift 0.1 ErrOff 0.1 end
|  8> end
|  9> %maxcore 7000
| 10> %scf MaxIter 512  end
| 11> %pal nprocs 28 end
| 12> %basis
| 13> NewGTO Au "SARC-ZORA-TZVP" end
| 14> NewGTO Ag "SARC-ZORA-TZVP" end
| 15> end

The integration grids used in DFT should be viewed together with the basis set. If large basis set
calculations are converged to high accuracy it is advisable to also use large DFT integration grids
(like ! DEFGRID3). For \unlimited" accuracy (i.e. benchmark calculations) it is probably best to use
product grids (Grid=0) with a large value for IntAcc (perhaps around 6.0). The default grids have
been chosen such that they provide adequate accuracy at the lowest possible computational cost, but
for all-electron calculations on heavy elements in conjunction with scalar relativistic Hamiltonians
you should examine the grid dependency very carefully and adjust these parameters accordingly to
minimize errors.

SCF ITERATIONS
--------------
ITER    Energy       Delta-E       Max-DP    RMS-DP    [F,P]    Damp
            ***  Starting incremental Fock matrix formation  ***
  0 -659738.3540527759 0.000000000000 0.21391546  0.00019557  0.0017077 0.8500
  1 -659738.3540408637 0.000011912198 0.20527759  0.00018164  0.0016423 0.8500
                           ***Turning on DIIS***
  2 -659738.3540321524 0.000008711242 1.23562683  0.00107020  0.0013794 0.0000
  3 -659738.3543270219  -0.000294869416 0.11405034  0.00015842  0.0000800 0.0000
  4 -659738.3542744694 0.000052552437 0.05903348  0.00004354  0.0004660 0.0000
  5 -659738.3544131238  -0.000138654374 0.02848066  0.00004887  0.0004301 0.0000
  6 -659738.3544615146  -0.000048390822 0.01669514  0.00001511  0.0001651 0.0000
  7 -659738.3544918646  -0.000030349940 0.00674935  0.00001112  0.0003387 0.0000
  8 -659738.3545418921  -0.000050027505 0.12783734  0.00013369  0.0003635 0.0000
  9 -659738.3547533691  -0.000211477047 0.04098745  0.00003925  0.0007819 0.0000
10 -659738.3546819927 0.000071376446 0.04363270  0.00007961  0.0006423 0.0000
11 -659738.3548215146  -0.000139521901 0.18179230  0.00013426  0.0009463 0.0000
12 -659738.3548966042  -0.000075089629 0.22489512  0.00022523  0.0010666 0.0000
13 -659738.3548787254 0.000017878832 0.21901004  0.00022447  0.0008757 0.0000
14 -659738.3548678381 0.000010887277 0.22634525  0.00017091  0.0007922 0.0000
15 -659738.3548004244 0.000067413668 0.50208747  0.00035035  0.0008152 0.0000
16 -659738.3546156262 0.000184798264 0.08646566  0.00011077  0.0010997 0.0000
17 -659738.3545748875 0.000040738611 0.19839879  0.00022266  0.0003713 0.0000
18 -659738.3548108202  -0.000235932646 0.11255784  0.00011107  0.0011895 0.0000
19 -659738.3547273217 0.000083498540 0.13255248  0.00007975  0.0011443 0.0000
            *** Restarting incremental Fock matrix formation ***
                                 *** Resetting DIIS ***

WARNING: the maximum gradient error descreased on average only by a factor 1.0
      during the last 20 iterations

                  *** Initiating the TRAH-SCF procedure ***

TRAH GRID
---------

General Integration Accuracy    IntAcc    ... 3.467
Radial Grid Type                RadialGrid  ... OptM3 with GC (2021)
Angular Grid (max. ang.)       AngularGrid ... 3 (Lebedev-194)
Angular grid pruning method    GridPruning ... 4 (adaptive)
Weight generation scheme       WeightScheme... Becke
Basis function cutoff          BFCut    ... 1.0000e-10
Integration weight cutoff       WCut       ... 1.0000e-14
Angular grids for H and He will be reduced by one unit
Partially contracted basis set             ... off
Rotationally invariant grid construction    ... off

Total number of grid points                ... 540385
Total number of batches                   ...    8498
Average number of points per batch          ...    63
Average number of grid points per atom    ...    4740

--------------------------------------------------------------------------------------------
Iter.       energy          ||Error||_2       Shift    TRadius  Mac/Mic       Rej.
--------------------------------------------------------------------------------------------
0 -659738.354224959272    5.584408e-02                   0.400 (TRAH MAcro) No
0    dE -7.843961e-06    3.634366e-01 -1.5688e-05    0.000 (TRAH MIcro)
0    dE -1.029186e-04    8.373204e-02 -2.0583e-04    0.007 (TRAH MIcro)
0    dE -1.246020e-04    1.508290e-02 -2.4919e-04    0.007 (TRAH MIcro)
0    dE -1.279635e-04    2.716917e-03 -2.5591e-04    0.008 (TRAH MIcro)
1 -659738.354418403003    1.257599e-02                   0.480 (TRAH MAcro) No
1    dE -1.440615e-07    2.239469e-01 -2.8977e-07    0.000 (TRAH MIcro)
1    dE -5.015682e-06    5.965489e-02 -1.0031e-05    0.002 (TRAH MIcro)
1    dE -6.553713e-06    6.244602e-03 -1.3107e-05    0.002 (TRAH MIcro)
1    dE -6.727213e-06    7.022422e-04 -1.3454e-05    0.002 (TRAH MIcro)
2 -659738.354433833621    2.481142e-03                   0.576 (TRAH MAcro) No
2    dE -1.036866e-08    5.408486e-02 -2.3343e-08    0.000 (TRAH MIcro)
2    dE -1.591271e-07    2.159381e-02 -3.1826e-07    0.000 (TRAH MIcro)
2    dE -2.254930e-07    4.984132e-03 -4.5099e-07    0.000 (TRAH MIcro)
2    dE -2.396939e-07    9.108237e-04 -4.7939e-07    0.000 (TRAH MIcro)
2    dE -2.425997e-07    7.104017e-05 -4.8520e-07    0.000 (TRAH MIcro)
3 -659738.354435423738    1.449313e-03                   0.691 (TRAH MAcro) No
3    dE -4.618169e-09    3.007258e-02 -1.1313e-08    0.000 (TRAH MIcro)
3    dE -7.989533e-08    2.966794e-04 -1.5979e-07    0.000 (TRAH MIcro)
3    dE -8.060404e-08    4.900243e-05 -1.6121e-07    0.000 (TRAH MIcro)
4 -659738.354438695707    1.972488e-03                   0.829 (TRAH MAcro) No
4    dE -1.318298e-08    2.573613e-02 -2.7176e-08    0.000 (TRAH MIcro)
4    dE -1.554553e-07    1.482226e-04 -3.1091e-07    0.000 (TRAH MIcro)

		自动登录 Automatic login	找回密码 Forget password
密码 Password			注册 Register

[ORCA] orca SCF迷一样的感受

浏览过的版块