计算化学公社

标题: 求助ORCA计算高精度单点能时出现错误 [打印本页]

作者
Author: Caroline2022 时间: 2024-2-26 10:10
标题: 求助ORCA计算高精度单点能时出现错误
在仿照Sob老师帖子http://bbs.keinsci.com/thread-16255-1-1.html中做构象搜索，前面做了几次都正常运行并结束，再运算时出现报错。
用Molclus调用Gaussian做几何优化和振动分析都提示正确结束，没有虚频。调用Shermo也正常结束。但运行ORCA时报错如下：
Now run single point task based on the template_SP.inp in current folder
Running ORCA: "/sob/orca503/orca" orcaSP.inp  > orcaSP.out

[file orca_tools/Tool-IntegralList/qcmatstore.cpp, line 2174, Process 13]: Error TMatrixContainers::AddMatrix - Failed to add 65 0 to orcaSP.IAV_aa.tmp.proc13

[file orca_tools/Tool-IntegralList/qcmatstore.cpp, line 2174, Process 28]: Error TMatrixContainers::AddMatrix - Failed to add 95 0 to orcaSP.IAV_aa.tmp.proc28

[file orca_tools/Tool-IntegralList/qcmatstore.cpp, line 2174, Process 29]: Error TMatrixContainers::AddMatrix - Failed to add 97 0 to orcaSP.IAV_aa.tmp.proc29

[file orca_tools/Tool-IntegralList/qcmatstore.cpp, line 2174, Process 31]: Error TMatrixContainers::AddMatrix - Failed to add 101 0 to orcaSP.IAV_aa.tmp.proc31

[file orca_tools/Tool-IntegralList/qcmatstore.cpp, line 2174, Process 32]: Error TMatrixContainers::AddMatrix - Failed to add 103 0 to orcaSP.IAV_aa.tmp.proc32

[file orca_tools/Tool-IntegralList/qcmatstore.cpp, line 2174, Process 33]: Error TMatrixContainers::AddMatrix - Failed to add 105 0 to orcaSP.IAV_aa.tmp.proc33

[file orca_tools/Tool-IntegralList/qcmatstore.cpp, line 2174, Process 2]: Error TMatrixContainers::AddMatrix - Failed to add 43 0 to orcaSP.IAV_aa.tmp.proc2

[file orca_tools/Tool-IntegralList/qcmatstore.cpp, line 2174, Process 5]: Error TMatrixContainers::AddMatrix - Failed to add 49 0 to orcaSP.IAV_aa.tmp.proc5

[file orca_tools/Tool-IntegralList/qcmatstore.cpp, line 2174, Process 8]: Error TMatrixContainers::AddMatrix - Failed to add 55 0 to orcaSP.IAV_aa.tmp.proc8

[file orca_tools/Tool-IntegralList/qcmatstore.cpp, line 2174, Process 9]: Error TMatrixContainers::AddMatrix - Failed to add 57 0 to orcaSP.IAV_aa.tmp.proc9

[file orca_tools/Tool-IntegralList/qcmatstore.cpp, line 2174, Process 11]: Error TMatrixContainers::AddMatrix - Failed to add 61 0 to orcaSP.IAV_aa.tmp.proc11

[file orca_tools/Tool-IntegralList/qcmatstore.cpp, line 2174, Process 15]: Error TMatrixContainers::AddMatrix - Failed to add 69 0 to orcaSP.IAV_aa.tmp.proc15

[file orca_tools/Tool-IntegralList/qcmatstore.cpp, line 2174, Process 17]: Error TMatrixContainers::AddMatrix - Failed to add 73 0 to orcaSP.IAV_aa.tmp.proc17

[file orca_tools/Tool-IntegralList/qcmatstore.cpp, line 2174, Process 22]: Error TMatrixContainers::AddMatrix - Failed to add 84 0 to orcaSP.IAV_aa.tmp.proc22

[file orca_tools/Tool-Parallel/pal1.cpp, line 2835, Process 23]: Error in PAL_fseek

[file orca_tools/Tool-IntegralList/qcmatstore.cpp, line 2174, Process 25]: Error TMatrixContainers::AddMatrix - Failed to add 89 0 to orcaSP.IAV_aa.tmp.proc25

[file orca_tools/Tool-IntegralList/qcmatstore.cpp, line 2174, Process 26]: Error TMatrixContainers::AddMatrix - Failed to add 92 0 to orcaSP.IAV_aa.tmp.proc26

[file orca_tools/Tool-IntegralList/qcmatstore.cpp, line 2174, Process 27]: Error TMatrixContainers::AddMatrix - Failed to add 93 0 to orcaSP.IAV_aa.tmp.proc27

[file orca_tools/Tool-IntegralList/qcmatstore.cpp, line 2174, Process 7]: Error TMatrixContainers::AddMatrix - Failed to add 53 0 to orcaSP.IAV_aa.tmp.proc7

[file orca_tools/Tool-IntegralList/qcmatstore.cpp, line 2174, Process 10]: Error TMatrixContainers::AddMatrix - Failed to add 59 0 to orcaSP.IAV_aa.tmp.proc10

[file orca_tools/Tool-IntegralList/qcmatstore.cpp, line 2174, Process 12]: Error TMatrixContainers::AddMatrix - Failed to add 64 0 to orcaSP.IAV_aa.tmp.proc12

[file orca_tools/Tool-IntegralList/qcmatstore.cpp, line 2174, Process 14]: Error TMatrixContainers::AddMatrix - Failed to add 68 0 to orcaSP.IAV_aa.tmp.proc14

[file orca_tools/Tool-IntegralList/qcmatstore.cpp, line 2174, Process 30]: Error TMatrixContainers::AddMatrix - Failed to add 100 0 to orcaSP.IAV_aa.tmp.proc30

[file orca_tools/Tool-IntegralList/qcmatstore.cpp, line 2174, Process 24]: Error TMatrixContainers::AddMatrix - Failed to add 87 0 to orcaSP.IAV_aa.tmp.proc24

[file orca_tools/Tool-IntegralList/qcmatstore.cpp, line 2174, Process 19]: Error TMatrixContainers::AddMatrix - Failed to add 77 0 to orcaSP.IAV_aa.tmp.proc19

[file orca_tools/Tool-Parallel/pal1.cpp, line 2835, Process 1]: Error in PAL_fseek

[file orca_tools/Tool-IntegralList/qcmatstore.cpp, line 2174, Process 4]: Error TMatrixContainers::AddMatrix - Failed to add 47 0 to orcaSP.IAV_aa.tmp.proc4

[file orca_tools/Tool-IntegralList/qcmatstore.cpp, line 2174, Process 16]: Error TMatrixContainers::AddMatrix - Failed to add 71 0 to orcaSP.IAV_aa.tmp.proc16

[file orca_tools/Tool-IntegralList/qcmatstore.cpp, line 2174, Process 20]: Error TMatrixContainers::AddMatrix - Failed to add 79 0 to orcaSP.IAV_aa.tmp.proc20

[file orca_tools/Tool-IntegralList/qcmatstore.cpp, line 2174, Process 6]: Error TMatrixContainers::AddMatrix - Failed to add 51 0 to orcaSP.IAV_aa.tmp.proc6

--------------------------------------------------------------------------
Primary job  terminated normally, but 1 process returned
a non-zero exit code. Per user-direction, the job has been aborted.
--------------------------------------------------------------------------
--------------------------------------------------------------------------
mpirun detected that one or more processes exited with non-zero status, thus causing
the job to be terminated. The first process to do so was:

  Process name: [[21135,1],28]
  Exit code: 64
--------------------------------------------------------------------------
[file orca_tools/qcmsg.cpp, line 465]:
  .... aborting the run

Error: Single point calculation did not normally terminated! The task is regarded as failed
The orcaSP.out has been backed up to orcaSP00020.out
Deleting orcaSP.* orcaSP_*
The gau.out has been backed up to gau00020.out
The gau.chk has been backed up to gau00020.chk
Wall clock time elapsed for calculating this configuration: 1716 s
Deleting gau.gjf gau.out gau.chk

Running finished!
All calculations totally took up wall clock time    41391 s (  11.50 hours)

Gibbs free energy (G), thermal correction to G (Gcorr), numbers of imaginary frequencies (Nimag) and geometry of all successful configurations have been collectively outputted to isomers.xyz in current folder.
In this file, the value after "Frame:" is the frame index in the isomers.xyz, "Step:" is the step number in this time of molclus execution, "Geom:" is geometry index in the inputted trajectory file

参考帖子http://bbs.keinsci.com/thread-34880-1-1.html中也出现错误“Error in PAL_fseek”，不知道是不是也是类似的因为硬盘空间问题，还是有其它方面的问题。
另外，如何查看ORCA在每次计算中产生了多少临时文件？哪些文件夹或者字母开头的是属于临时文件？（因为是ORCA小白，不敢胡乱删除，担心把有用的文件误删了）
请各位老师不吝指教，多谢多谢

作者
Author: Satoru 时间: 2024-2-26 10:25
本帖最后由 Satoru 于 2024-2-26 10:27 编辑

看起来是盘写满了，可以df -h检查一下剩余空间。ORCA的临时文件存在提交任务的目录下

作者
Author: Satoru 时间: 2024-2-26 10:25
本帖最后由 Satoru 于 2024-2-26 10:27 编辑

ORCA的临时文件一般文件名都是*.tmp或*.tmp.*

作者
Author: Caroline2022 时间: 2024-2-26 15:48

Satoru 发表于 2024-2-26 10:25
看起来是盘写满了，可以df -h检查一下剩余空间。ORCA的临时文件存在提交任务的目录下

老师，我用df -h命令查看磁盘空间后结果如下：
[root@localhost ~]# df -h
Filesystem             Size  Used Avail Use% Mounted on
devtmpfs                126G    0  126G 0% /dev
tmpfs                   126G    0  126G 0% /dev/shm
tmpfs                   126G 44M  126G 1% /run
tmpfs                   126G    0  126G 0% /sys/fs/cgroup
/dev/mapper/centos-root 50G 38G 13G  76% /
/dev/nvme0n1p2       1014M  259M  756M  26% /boot
/dev/nvme0n1p1          200M 12M  189M 6% /boot/efi
/dev/mapper/centos-home  1.8T 20G  1.8T 2% /home
tmpfs                   26G 48K 26G 1% /run/user/0
/dev/sdb2             7.3T 93M  6.9T 1% /mnt/home
[root@localhost ~]#
似乎没有占用特别大空间的（/dev/mapper/centos-root 这个是centos的系统吧）。

orca计算单点能时是通过Molclus程序调用的，我仿照了sob老师帖子中的settings.ini，其中关于ORCA的控制参数如下：
--- Below for ORCA ---
  orca_path= "/sob/orca503/orca"  // Command for invoking ORCA
  ibkgbw= 0  // The same as ibkout, but for .gbw file
  ibktrj= 0  // 1: Backup optimization trajectory (*_trj.xyz) with step number in the file name, 0: Don't backup
  mpioption= none  // Special MPI option for running ORCA. e.g. --allow-run-as-root. If no option is needed, this option should be set to "none"
输入文件为template_SP.inp，内容如下：
! PWPB95 D3 def2-QZVPP def2/J def2-QZVPP/C RIJCOSX defgrid3 tightSCF noautostart miniprint nopop
%maxcore 6000
%pal nprocs 56 end
* xyz 0 1
[GEOMETRY]
*
ORCA程序是安装在/sob/orca503下面的，我在其中没有找到含tmp的文件，在Molclus文件夹中也没有找到含tmp的文件（照片见附件）。
另外附件中还附上了第20个构象计算的输出文件，其中的报错提示和前文一样。
例如其中：[file orca_tools/Tool-IntegralList/qcmatstore.cpp, line 2174, Process 13]: Error TMatrixContainers::AddMatrix - Failed to add 65 0 to orcaSP.IAV_aa.tmp.proc13
我又用语句find / -name 'orca*tmp*' 尝试在根目录（所有硬盘）下查找文件名中含orca和tmp的文件，也没有任何符合条件的文件。

请问老师这些临时文件应该在哪里呢？
另外，在输出文件中有提示：
IA-Transformation
Memory available                      ... 5703 MB

是否是因为内存不足引起的报错呢？

作者
Author: sai77 时间: 2024-2-26 21:31
本帖最后由 sai77 于 2024-2-26 21:50 编辑

Molclus输出信息里写了 “Deleting orcaSP.* orcaSP_*”，临时文件都被删除了。
临时文件在输入文件同目录下，不在orca安装目录。
不清楚报错原因，不过你这50多个原子，给了56核*6000的资源，服务器内存多大啊？28*2000感觉足够了

作者
Author: Caroline2022 时间: 2024-2-28 11:35

sai77 发表于 2024-2-26 21:31
Molclus输出信息里写了 “Deleting orcaSP.* orcaSP_*”，临时文件都被删除了。
临时文件在输入文件同目录 ...

感谢您的回复
我的服务器内存是256G，按照56*6000=336G，确实是超了。我调整为28*2000后算了一个构象，但是依然报错了。
我发现.out文件中不再出现“Error in PAL_fseek”的报错信息。但是仍然提示无法写入临时文件，并且可用内存量不知为什么又降低到了1704MB，我又尝试了将%maxcore修改为4000，可用内存量又变为了3704MB。
IA-Transformation
Memory available                      ... 1704 MB
Max. # MO pairs treated in a batch    ... 3
# of internal orbitals                ... 68
# batches for internal orbitals       ... 1
[file orca_tools/Tool-IntegralList/qcmatstore.cpp, line 2174, Process 19]: Error TMatrixContainers::AddMatrix - Failed to add 98 0 to orcaSP.IAV_aa.tmp.proc19

[file orca_tools/Tool-IntegralList/qcmatstore.cpp, line 2174, Process 7]: Error TMatrixContainers::AddMatrix - Failed to add 62 0 to orcaSP.IAV_aa.tmp.proc7

[file orca_tools/Tool-IntegralList/qcmatstore.cpp, line 2174, Process 3]: Error TMatrixContainers::AddMatrix - Failed to add 50 0 to orcaSP.IAV_aa.tmp.proc3

ORCA finished by error termination in MP2
Calling Command: mpirun -np 28  /sob/orca503/orca_mp2_mpi orcaSP.mp2inp.tmp orcaSP
[file orca_tools/qcmsg.cpp, line 465]:
  .... aborting the run

这个可用内存量似乎是随着设置的%maxcore数值降低而降低的，并且总是达不到需要的数值。请问报错的原因是因为内存不足么？又应该如何调整呢？
附件中是输出文件

作者
Author: 啊不错的飞过海 时间: 2024-2-28 15:57
/home只剩20个G了，orca稍微写写就满了。清理一下吧。

作者
Author: Caroline2022 时间: 2024-2-29 20:11

啊不错的飞过海发表于 2024-2-28 15:57
/home只剩20个G了，orca稍微写写就满了。清理一下吧。

老师，您是说这里么？/dev/mapper/centos-home 1.8T 20G 1.8T 2% /home 这应该是使用了20G，还有1.8T剩余吧？

作者
Author: hebrewsnabla 时间: 2024-2-29 20:23
你是在哪个目录下提交计算的？

作者
Author: Caroline2022 时间: 2024-3-1 09:42

hebrewsnabla 发表于 2024-2-29 20:23
你是在哪个目录下提交计算的？

是通过Molclus调用的ORCA，Molclus在/sob目录下

作者
Author: hebrewsnabla 时间: 2024-3-1 14:17

Caroline2022 发表于 2024-3-1 09:42
是通过Molclus调用的ORCA，Molclus在/sob目录下

在哪个目录下提交计算是指在什么目录下执行的命令，输入文件在哪个目录。

作者
Author: Caroline2022 时间: 2024-3-1 17:05

hebrewsnabla 发表于 2024-3-1 14:17
在哪个目录下提交计算是指在什么目录下执行的命令，输入文件在哪个目录。

/sob/molclus_1.9.9.9_Linux

作者
Author: hebrewsnabla 时间: 2024-3-1 17:25

Caroline2022 发表于 2024-3-1 17:05
/sob/molclus_1.9.9.9_Linux

从4L贴的东西来看，这个目录所在分区只剩13G，肯定没法做计算的。

作者
Author: Caroline2022 时间: 2024-3-4 11:03

hebrewsnabla 发表于 2024-3-1 17:25
从4L贴的东西来看，这个目录所在分区只剩13G，肯定没法做计算的。

谢谢老师，我转移了一部分程序，现在可以计算了

欢迎光临计算化学公社 (http://bbs.keinsci.com/)