1
$\begingroup$

I started optimising a system containing amino acid residue, water molecule and a toxic metal ion, in proving a supposed reaction steps that might take place in inhibiting enzyme activity. The suggestion is the toxic metal ion potentially hinder the interactions of ASP and water molecule by interacting with ASP. So considering an input file for the reactants containing around 15 atoms (input coordinates are given below) I started optimising the input using DFT (B3LYP) LANL2DZ basis set. The calculation started but at one point I cannot see any progress in the calculations and now it is almost around for 2 days. Gaussian does not show any error also.What can be the reason? I have attached the input and last few lines of the output file of another system in which the same issue in processing was observed. The new system is the optimization of aspartic acid with a water molecule

**%nprocshared=4
%mem=1GB
%chk=C:\Users\Chemistry\Desktop\asp_opt[i].chk
# opt b3lyp/cc-pvtz geom=connectivity

asp

-1 1
 C                  1.19957000   -0.49798300    0.13741800
 C                  0.55937200    0.74926400   -0.44764700
 H                  0.44339200    0.59076300   -1.51903000
 N                  1.55336100    1.83916500   -0.25779900
 H                  1.10438900    2.71239400   -0.51456000
 H                  1.73266800    1.90850600    0.74017500
 C                 -0.80611000    1.00817800    0.18465200
 H                 -1.23792900    1.92271700   -0.23130400
 H                 -0.67292000    1.17051900    1.25931800
 O                  1.45896000   -0.63706000    1.31192900
 O                  1.65241300   -1.36767700   -0.79729900
 H                  2.01459900   -2.11079700   -0.29521300
 O                 -1.24221200   -1.29130000   -0.17155100
 C                 -1.80218000   -0.17703500   -0.00346500
 O                 -3.01436500    0.10068700    0.04935500
 O                 -0.08458867   -1.88429493    1.50213956
 H                  0.67747839   -1.30667491    1.58704485
 H                 -0.52974695   -1.70729416    0.67021118

 1 2 1.0 10 2.0 11 1.0
 2 3 1.0 4 1.0 7 1.0
 3
 4 5 1.0 6 1.0
 5
 6
 7 8 1.0 9 1.0 14 1.0
 8
 9
 10
 11 12 1.0
 12
 13 14 1.0
 14 15 2.0
 15
 16 17 1.0 18 1.0
 17
 18**
**(A) (A) (A) (A) (A) (A) (A) (A) (A) (A) (A) (A)
 Requested convergence on RMS density matrix=1.00D-08 within 128 cycles.
 Requested convergence on MAX density matrix=1.00D-06.
 Requested convergence on             energy=1.00D-06.
 No special actions if energy rises.
 SCF Done:  E(RB3LYP) =  -588.437660556     A.U. after    9 cycles
             Convg  =    0.6247D-08             -V/T =  2.0058
 Calling FoFJK, ICntrl=      2127 FMM=F ISym2X=0 I1Cent= 0 IOpClX= 0 NMat=1 NMatS=1 NMatT=0.
 ***** Axes restored to original set *****
 -------------------------------------------------------------------
 Center     Atomic                   Forces (Hartrees/Bohr)
 Number     Number              X              Y              Z
 -------------------------------------------------------------------
      1        6          -0.003297899    0.000968992   -0.001205417
      2        6           0.001174672    0.000030078   -0.002058134
      3        1           0.000120836   -0.000724917    0.000428308
      4        7           0.000179028   -0.000406405    0.000490257
      5        1          -0.000094934    0.000238362   -0.000114970
      6        1           0.000060431   -0.000078420    0.000048740
      7        6          -0.002031199    0.003200366    0.004191686
      8        1           0.000368507   -0.000184638    0.000325103
      9        1           0.000963855   -0.000094190    0.000676509
     10        8           0.005990754    0.007413835   -0.001731561
     11        8           0.000065729   -0.002001216   -0.001545829
     12        1           0.000236666    0.000364910    0.000031038
     13        8           0.000381196    0.004747433    0.001670893
     14        6           0.003841548   -0.002121460   -0.011415519
     15        8          -0.001868448   -0.000631479    0.002953022
     16        8          -0.005940381    0.005686767   -0.013553137
     17        1           0.012654684   -0.001668370    0.014136764
     18        1          -0.012805045   -0.014739648    0.006672248
 -------------------------------------------------------------------
 Cartesian Forces:  Max     0.014739648 RMS     0.005058324**
```
$\endgroup$
6
  • 1
    $\begingroup$ Given the nature of the question, there are too many unknowns. For all that you might know, it could be that you don't have enough memory, or you have too many ERIs to compute. Have you considered, first optimizing the structure with UFF or any other force fields? $\endgroup$
    – Fracton
    Commented Feb 4 at 5:20
  • $\begingroup$ @Fracton Yes I tried optimising the structure using UFF. But then it gave the error ''error termination request processed by link 9999. Error termination via lnkle in c:\g09W\19999.exe'' . Then as explained in a video I tried using the keywords geom=allcheck and opt=calcfc for the mentioned error. That also didnt work with the error ''error termination in NtrErr: NtrErr Called from FileIO.'' I couldn't find any solution for this and hoping for a help.. $\endgroup$
    – PriZarah
    Commented Feb 16 at 15:28
  • $\begingroup$ @PriZarah I'm afraid this question isn't answerable at present. What do you mean by 'I cannot see any progress in the calculations', do you mean there's no output in the log file? Are you running on a server or on PC? $\endgroup$
    – leeman
    Commented Mar 2 at 17:14
  • $\begingroup$ @leeman I am doing my calculations on my PC. Actually what I meant in my question is, during the optimization though it generated output file, at a point onwards nothing was written newly in the output and no.of cycles of processing didnt change. It seemed like the optimization has got stuck at a point. I guess it is due to the basis set I used and the lack of processing speed because when I used a lower basis set for the calculation the processing completed. But I am not sure with my assumption whether it is correct or not. I would appreciate if anyone can suggest a way for my calculations. $\endgroup$
    – PriZarah
    Commented Mar 4 at 15:53
  • $\begingroup$ Hi @PriZarah, yes you might be right, LANL2DZ is fairly large (especially for metals) and might struggle on a personal PC. If you edit your question to include the last 10 or so lines from the stuck output file then we may be able to help. Also, how many CPUs are you using? $\endgroup$
    – leeman
    Commented Mar 4 at 18:15

1 Answer 1

1
$\begingroup$

Thanks for updating the question PriZarah, I can now at least suggest what might be causing the problem.

It's probably just very, very slow

In your recent example, you are calculating a reasonably large system (~20 atoms) with a large basis set (triple-zeta) with only 4 CPUs (%nprocshared=4) and very little memory (%mem=1GB). For a calculation of this complexity you need to allocate a lot more resources to have it complete in a reasonable amount of time. Most calculations like this are performed on high-performing computing (HPC) servers using many CPUs in parallel (~16-32 wouldn't be uncommon) and very large amount of memory (~20-80 GB perhaps).

To run this sort of calculation on a personal PC you need to allocate as many CPUs and as much memory as possible. Open task manager and see how much free memory your system has at idle, and allocate all of this (for example if you see you have 16GB total and 8GB free, allocate %mem=8GB). Likewise determine how many CPU cores you have and allocate all but one of these (so if you have 16 cores total, allocate %nprocshared=15). Start with a much smaller basis set and increase sequentially until you get the results you want, or you run out of patience. An optimisation with a triplet-zeta basis set should take hours for a reasonably large system. I would recommend:

  • First STO-3G, to make sure everything is working
  • Then a small split valence set, maybe 3-21G
  • Then consider cc-pVDZ
  • If you're still going, then try cc-pVTZ

Why does it get stuck here?

I'm not sure exactly which part of the calculation Gaussian get's stuck at, but this looks like the start of the first optimisation cycle. I would guess it's calculating the gradient, which is often a very slow operation. You can always check whether Gaussian is working hard in task manager. If you see lots of Gaussian processes consuming CPU, it's doing its thing.

$\endgroup$
0

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .