ORCA在计算scf以及hessian时停止

Theor_Comp · 发表于 Post on 2019-9-12 11:42:03

在分别进行1. 结构优化（opt）以及 2.(opt=calcfc)中，

1.
SCF ITERATIONS
--------------
ITER    Energy       Delta-E       Max-DP    RMS-DP    [F,P]    Damp
  0  -4431.7082633400 0.00000000000085.64289961  0.19976724  1.6913454 0.7000
  1  -4420.5417743541  11.166488985949206.08165123  0.50838331  1.1232059 0.7000
  2  -4396.2212765560  24.3204977980391192.21655076  2.10510263  0.8221542 0.7000
  3  -4357.5517797277  38.669496828273353.96299444  0.98517323  0.7391678 0.7000
  4  -4340.8265020596  16.7252776681361242.05437141  2.66444139  0.9487423 0.7000
  5  -4338.8921942219 1.934307837671681.72232588  1.51495368  0.6363816 0.7000
  6  -4308.4780988843  30.414095337638284.61510885  0.84830087  0.6623655 0.7000
  7  -4317.0134749989  -8.535376114562785.53625500  1.65676255  0.4510681 0.7000
  8  -4321.4094876839  -4.396012685037389.49149155  1.09120564  0.4585606 0.7000
  9  -4333.4127495059 -12.003261821975362.57953659  0.76560399  0.3619837 0.7000
10  -4340.9440966898  -7.531347183956724.59742824  1.35813254  0.3024128 0.7000
11  -4327.2914719072  13.652624782616765.29678012  1.25914317  0.4977518 0.7000
                           ***Turning on DIIS***
                  *** Initiating the SOSCF procedure ***
                        *** Shutting down DIIS ***
                  *** Re-Reading the Fockian ***
                  *** Removing any level shift ***
ITER    Energy    Delta-E       Grad    Rot    Max-DP RMS-DP
12  -4317.02122564  10.2702462636  0.546213  0.546213 709.660759  1.758232
13  -4303.37206165  13.6491639894  0.410047  5.612014 2268.246212  5.293211

  ************************************************************
      *                   SERIOUS PROBLEM IN SOSCF          *
      *                   ------------------------          *
      *          HUGE, UNRELIABLE STEP WAS ABOUT TO BE TAKEN *
      *                      ABORTING THE RUN                *
      ************************************************************
[file orca_scf/scf2.cpp, line 1377, Process 0]:
[file orca_scf/scf2.cpp, line 1377, Process 1]:
[file orca_scf/scf2.cpp, line 1377, Process 2]:
[file orca_scf/scf2.cpp, line 1377, Process 3]:
[file orca_scf/scf2.cpp, line 1377, Process 4]:
[file orca_scf/scf2.cpp, line 1377, Process 5]:
[file orca_scf/scf2.cpp, line 1377, Process 7]:
[file orca_scf/scf2.cpp, line 1377, Process 8]:
[file orca_scf/scf2.cpp, line 1377, Process 9]:
[file orca_scf/scf2.cpp, line 1377, Process 10]:
[file orca_scf/scf2.cpp, line 1377, Process 11]:
[file orca_scf/scf2.cpp, line 1377, Process 6]:
-------------------------------------------------------
Primary job  terminated normally, but 1 process returned
a non-zero exit code.. Per user-direction, the job has been aborted.
前面有已经两个SUCCESS  SCF CONVERGED AFTER  17 CYCLES  ，但是到这第三个SCF ITERATIONS的时候出错了，不知为何？

2. 由于加了calcfc，所以会计算hessian，该情况下scf收敛，但是计算hessian时出错，
Adding the E*S(x)*S(y) terms to the Hessian    ... done (    22.1 sec)
Calculating energy weighted overlap derivatives  ... done (    1.5 sec)
Two electron integral derivatives (RI)          ... done ( 6826.2 sec)
Exchange-correlation integral derivatives       ... done ( 2166.3 sec)
tr(F(y)Q(x)) contribution to the Hessian       ... done (    39.8 sec)

Response fock operator R(S(x)) (RIJCOSX)       ...
IBatch  1 (of  3):                                  -----------------------------------------------------------
---------------
mpirun noticed that process rank 0 with PID 4495 on node cnx exited on signal 9 (Killed).
--------------------------------------------------------------------------

ORCA finished by error termination in SCF Hessian
Calling Command: mpirun -np 12  /orca/4.0.1/orca_scfhess_mpi mol.scfhess.inp
[file orca_tools/qcmsg.cpp, line 432]:
  .... aborting the run

下面是它的输入文件：
================================================================================
                                    INPUT FILE
================================================================================
NAME = mol.inp
|  1> ! B3LYP D3 def2-TZVPD def2/J RIJCOSX tightscf grid4 engrad freq BOHRS nopop
|  2> %scf soscfmaxit 12
|  3>    directresetfreq 1
|  4> end
|  5> %pal nprocs 12 end
|  6> %maxcore 2700
|  7> * xyz -2 1
|  8>       6    -1.341630032048    -5.503571899534    -2.417394088660
|  9>       6    -3.414532230268    -4.063042608199    -1.927924349513
| 10>       6    -3.253676412689    -1.600151117810    -0.841552600061
| 11>       6    -0.819863464019    -0.695510491387    -0.266793964698
| 12>       6    1.328187226006    -2.134231081939    -0.755271876029

麻烦各位及Sob老师了。

chrinide · 发表于 Post on 2019-9-12 14:01:45

本帖最后由 chrinide 于 2019-9-12 14:09 编辑

这么小的体系(末尾是给的完整输入文件吧?)不要用这么多核心
你没有附上完整的输入文件吧？？？

Theor_Comp · 发表于 Post on 2019-9-12 14:06:51

chrinide 发表于 2019-9-12 14:01
这么小的体系(末尾是给的完整输入文件吧?)不要用这么多核心

不是，一共59个原子

chrinide · 发表于 Post on 2019-9-12 14:13:13

Theor_Comp 发表于 2019-9-12 14:06
不是，一共59个原子

发一下完整输入文件

mutron · 发表于 Post on 2019-9-12 18:56:03

本帖最后由 mutron 于 2019-9-12 20:11 编辑

soscf没必要强制12步迭代后开启（it is a lot less robust than DIIS），不过这里应该不是soscf问题。当使用体积很大或含有弥散函数的基组时（比如你的def2-TZVPD），遇到Max-DP数值突然暴涨到甚至和前一个数值黏在一起的情况，很有可能是碰到linear dependencies问题，检查一下Smallest eigenvalue看看是不是约等于或小于e-08。参考：https://sites.google.com/site/or ... -convergence-issues

如果确认是以上问题，解决方法：
%scf
thresh 1e-12
tcut 1e-14
end

或者更推荐使用不带弥散函数的def2-TZVP优化结构，正常开启DIIS（去掉soscfmaxit 12）。如果还是不收敛尝试用更小的基组计算单点再moread读取初猜、NRSCF

347lixin · 发表于 Post on 2022-10-22 08:57:13

安装完orca测试成功，当用我的inp文件做时一直显示失败，也不知道为什么，用的i7-12700H，核心14，和处理器有关吗，还是因为什么原因，毕竟sob大大说过12代cpu不适合做计算化学

347lixin · 发表于 Post on 2022-10-22 08:59:05

! B97-3c opt freq tightSCF noautostart miniprint nopop
%maxcore  1000
%pal nprocs  14 end
* xyz 0 1
C -42.98600000  -27.18200000 14.67200000
C -42.96200000  -28.66200000 14.35500000
C -42.43800000  -31.05600000 8.53900000
C -42.61200000  -30.28800000 12.37000000
C -42.87800000  -30.31300000 10.88000000
C -40.88200000  -21.66200000 23.21300000
O -41.06400000  -21.52900000 24.50900000
C -42.08800000  -22.14900000 22.46300000
O -39.84100000  -21.40500000 22.66200000
C -41.80100000  -23.23000000 21.43100000
C -43.05400000  -23.97800000 20.98100000
C -43.72100000  -23.39900000 19.74600000
C -43.12600000  -23.84700000 18.42100000
C -43.82700000  -25.04000000 17.80900000
C -43.34700000  -25.37700000 16.41000000
C -43.53600000  -26.83800000 16.03600000
C -42.80700000  -28.86700000 12.86400000
C -42.00000000  -31.17200000 10.00400000
H -43.58970000  -26.69320000 13.93610000
H -41.96730000  -26.85480000 14.66540000
H -43.87750000  -29.10960000 14.68140000
H -42.13630000  -29.11720000 14.86080000
H -41.80820000  -31.66920000 7.92880000
H -42.35840000  -30.03700000 8.22260000
H -43.45300000  -31.38090000 8.44340000
H -41.60830000  -30.60410000 12.56380000
H -43.28340000  -30.94990000 12.87590000
H -42.78830000  -29.30700000 10.52670000
H -43.84700000  -30.75900000 10.79540000
H -42.52610000  -21.31370000 21.95780000
H -42.73830000  -22.58970000 23.18960000
H -39.76150000  -21.51660000 21.73510000
H -41.11850000  -23.93510000 21.85750000
H -41.39170000  -22.74650000 20.56870000
H -43.76150000  -23.95870000 21.78350000
H -42.74210000  -24.96860000 20.72340000
H -43.64730000  -22.33280000 19.79870000
H -44.72900000  -23.75790000 19.75620000
H -42.10000000  -24.10460000 18.58180000
H -43.24140000  -23.03280000 17.73640000
H -44.87530000  -24.83010000 17.76520000
H -43.60550000  -25.88440000 18.42770000
H -42.30480000  -25.14340000 16.34480000
H -43.93730000  -24.80040000 15.72890000
H -43.03800000  -27.44230000 16.76520000
H -44.58930000  -27.02430000 16.00900000
H -41.95720000  -28.29840000 12.54870000
H -43.74320000  -28.55910000 12.44720000
H -40.98430000  -30.84830000 10.09590000
H -42.08790000  -32.19200000 10.31490000
*
这是我的inp文件，有大佬可以帮忙测试一下吗，是我的inp文件出现问题了吗？

347lixin · 发表于 Post on 2022-10-22 09:08:54

100 #选择其他功能的part 1
2 #文件转换功能
12 #生成 ORCA 输入文件
your path #输入你想保存的路径和名称，如果按回车就会保存在当前文件夹
#如果你使用的是 4.x 版本的 ORCA，请进行下一步操作；如果不是跳过就好
-11 #选择 ORCA 版本
1 #选择 5.0及更新版本
0 #更改任务类型，因为我们要进行键的优化嘛
4 #优化
1 #B97-3c

inp文件我是根据这个步骤用multiwfn文件做出来的，求大佬帮忙测试一下文件是否有问题

wzkchem5 · 发表于 Post on 2022-10-22 16:01:10

347lixin 发表于 2022-10-22 01:57
安装完orca测试成功，当用我的inp文件做时一直显示失败，也不知道为什么，用的i7-12700H，核心14，和处理器 ...

大概率是内存不够，maxcore给太少了

347lixin · 发表于 Post on 2022-10-22 22:08:38

wzkchem5 发表于 2022-10-22 16:01
大概率是内存不够，maxcore给太少了

如果是这样的话，我就不瞎折腾了，

347lixin · 发表于 Post on 2022-10-22 22:09:00

wzkchem5 发表于 2022-10-22 16:01
大概率是内存不够，maxcore给太少了

谢谢你

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

[ORCA] ORCA在计算scf以及hessian时停止

浏览过的版块