求助Gaussian计算中出现的问题，停在某一步很久了

agone997

在gaussian优化结构的时候停在一步很长时间，作业显示仍在计算，以下是log输出文件结尾部分重复部分用....................代替

  G2DrvN:  IPasSy=1 do centers  201 through  201.
Calling FoFCou, ICntrl=  3107 FMM=T I1Cent=   0 AccDes= 0.00D+00.
CoulSu:  requested number of processors reduced to:   4 ShMem   1 Linda.
CalDSu:  requested number of processors reduced to:   6 ShMem   1 Linda.
End of G2Drv F.D. properties file   721 does not exist.
End of G2Drv F.D. properties file   722 does not exist.
End of G2Drv F.D. properties file   788 does not exist.
Leave Link 1110 at Sun Dec  8 22:18:08 2024, MaxMem=   104857600 cpu:          200086.7 elap:           35872.0
(Enter /public/home/acg3c246oo/apprepo/gaussian/16C01-none/app/g16/l1002.exe)
Minotr:  Closed shell wavefunction.
          IDoAtm=11111111111111111111111111111111111111111111111111
          IDoAtm=11111111111111111111111111111111111111111111111111
          IDoAtm=11111111111111111111111111111111111111111111111111
          IDoAtm=11111111111111111111111111111111111111111111111111
          IDoAtm=1
          Direct CPHF calculation.
          Differentiating once with respect to electric field.
                with respect to dipole field.
          Differentiating once with respect to nuclear coordinates.
          Requested convergence is 1.0D-08 RMS, and 1.0D-07 maximum.
          Secondary convergence is 1.0D-12 RMS, and 1.0D-12 maximum.
          NewPWx=T KeepS1=F KeepF1=F KeepIn=T MapXYZ=F SortEE=F KeepMc=T.
        63637 words used for storage of precomputed grid.
Two-electron integrals replicated using symmetry.
          MDV=     104857600 using IRadAn=       1.
          Solving linear equations simultaneously, MaxMat=      84.
Estimated number of processors is:    4
FoF2E skips out because all densities are zero.
CalDSu exits because no D1Ps are significant.
FoFJK:  IHMeth= 1 ICntrl=       0 DoSepK=F KAlg= 0 I1Cent=           0 FoldK=F
IRaf= 320000000 NMat=       3 IRICut=       1 DoRegI=T DoRafI=F ISym2E= 2 IDoP0=0 IntGTp=1.
FoFCou: FMM=T IPFlag=           0 FMFlag=      100000 FMFlg1=        2001
         NFxFlg=           0 DoJE=F BraDBF=F KetDBF=F FulRan=T
         wScrn=  0.000000 ICntrl=         0 IOpCl=  0 I1Cent=           0 NGrid=           0
         NMat0=    3 NMatS0=      3 NMatT0=    0 NMatD0=    3 NMtDS0=    0 NMtDT0=    0
Symmetry not used in FoFCou.
FMM levels:  10  Number of levels for PrismC:   9
CoulSu:  requested number of processors reduced to:   6 ShMem   1 Linda.

...........................


CalDSu:  requested number of processors reduced to:   7 ShMem   1 Linda.
CoulSu:  requested number of processors reduced to:   7 ShMem   1 Linda.
CalDSu:  requested number of processors reduced to:   7 ShMem   1 Linda.
    606 vectors produced by pass  3 Test12= 7.65D-13 1.00D-09 XBig12= 9.71D-02 2.89D-02.
Estimated number of processors is:    4
CoulSu:  requested number of processors reduced to:   6 ShMem   1 Linda.
CalDSu:  requested number of processors reduced to:   7 ShMem   1 Linda.
CoulSu:  requested number of processors reduced to:   6 ShMem   1 Linda.

...........................


CalDSu:  requested number of processors reduced to:   7 ShMem   1 Linda.
CoulSu:  requested number of processors reduced to:   7 ShMem   1 Linda.
CalDSu:  requested number of processors reduced to:   7 ShMem   1 Linda.
CoulSu:  requested number of processors reduced to:   7 ShMem   1 Linda.
CalDSu:  requested number of processors reduced to:   7 ShMem   1 Linda.

wal

看着没问题 等

agone997

wal 发表于 2024-12-10 15:21
看着没问题 等

好嘞，谢谢您

Freeman

内存给少了，所以高斯少用了几个线程以降低内存需求，于是CPSCF就慢了许多。你可以继续等着，也可以杀掉这个任务，调大内存再提交。

agone997

Freeman 发表于 2024-12-10 16:06
内存给少了，所以高斯少用了几个线程以降低内存需求，于是CPSCF就慢了许多。你可以继续等着，也可以杀掉这 ...

28核，内存我写的70GB

sobereva

甭自己手写[新手求助]这种标签，给你去了，以后注意！

大体系振动分析28核才给70GB太抠门了，本来振动分析就是耗内存高的任务。该怎么设内存仔细看
Gaussian的安装方法及运行时的相关问题
http://sobereva.com/439（http://bbs.keinsci.com/thread-10814-1-1.html）
注意当前不是在做优化，而是之后振动分析时迭代求解CPHF方程

agone997

sobereva 发表于 2024-12-11 02:56
甭自己手写[新手求助]这种标签，给你去了，以后注意！

大体系振动分析28核才给70GB太抠门了，本来振动分 ...

谢谢sob老师，我学习一下。