Hello,

How much memory have you specified in the input file. If you're new to Gaussian both memory (mem) and number of processors (nprocshared) can be specified in the top part of the input file (Known as Link0 commands http://gaussian.com/link0/). e.g:

%nprocshared=18

%mem=6GB

%chk=test.chk

# opt freq b3lyp/3-21g geom=connectivity

.........

If you do not specify any, Gaussian will use default value which can be much lower compared to what is required for your system. The same values can also be set in Gaussview (If you're using it) under Calculate> Gaussian Calculation Setup > Link0 section.

Thank you for your response Uttam Pal , Anwesh Pandey

This was the real situation,

First i uploaded an input file in cluster and also on my computer without specific processor and memory. So, the calculation get terminated with this error and recommendation in the log file

" Out-of-memory error in routine ShPair-CalcS1 (IEnd=113078 MxCore= 3981)

Use %mem=33MW to provide the minimum amount of memory required to complete this step.

Error termination via Lnk1e in /home/software/g09/g09/l508.exe

Job cpu time: 0 days 13 hours 26 minutes 26.0 seconds.

File lengths (MBytes): RWF= 1495 Int= 0 D2E= 0 Chk= 54 Scr= 1

forrtl: severe (174): SIGSEGV, segmentation fault occurred"

So, I updated my input file with the given recommendation. The calculation terminated once again with this error :

" 34577563 words are not enough for AlAXAO.

Error termination via Lnk1e in /home/software/g09/g09/l1002.exe

Job cpu time: 0 days 4 hours 1 minutes 54.0 seconds.

File lengths (MBytes): RWF= 2005 Int= 0 D2E= 0 Chk= 54 Scr= 1

forrtl: severe (174): SIGSEGV, segmentation fault occurred "

So, how to know the memory available in cluster and computer and Is there any ways to know the required memory for this particular calculation?

Thank you so much in advance.

Hello,

I'm afraid 33MW is too low for any reasonable calculations. The rule of Thumb is 100MW = 800 MB (or 1 MW = 8 MB) so, 33 MW = 264 MB(approx) . You can check the conversion here: https://deviceanalytics.com/memcalc/

If you do not know the available memory in your cluster. I'll highly suggest you should contact your system admin to get information about available mem and processors in your cluster. This are crucial parameters for Gaussian calculations.

Usually the command 'free -mh' (without quotes) works for memory and 'nproc' for processors in linux systems. May be different for your case though, so without knowing the architecture of your cluster/computer I should not comment. Please check the following link:

https://www.binarytides.com/linux-command-check-memory-usage/

https://www.binarytides.com/linux-cpu-information/

Properly configured Gaussian on clusters should ignore %mem and similar keywords and to use maximal available memory on given node (that is identified by the queuing system).

I would point attention that error came in 13 or 4 hours after the work was started. This is definetly far after the SCF process starts, so why did memory usage grows suddenly? Isn't it already frequency calculation? It would be pretty usefull to see the whole output, preferably if it was runned with #P

Hello everyone, ( Uttam Pal , Anwesh Pandey

with all your valuable suggestion, now I managed to run the files for pcm calculation.

Thank you so much and much appreciated.

Greetings from the other side,

Now I am required to run an AIM analysis. Does anyone know how to generate an .wfn file using server?

I did tried the same way as generating via PC, but it's not performing.

Below, are the input file settings for your reference,

"

%Nprocshared=16

%mem=4GB

%chk=/home/AIM/Aim_in.chk

# sp b3lyp/6-31+g(d,p) scrf=pcm density=current out=wfn

Agg_gg_basic

0 1

H 14.29404400 2.12096900 -1.55465400

H 13.87052800 3.48036400 -0.50323000

H 15.13622800 2.34363400 -0.01344600

/home/AIM/Aim_in.wfn

"

Thank you so much in advance.

Hello,

About the AIM calculation:

It seems that in the above input the spin multiplicity (2S+1 = 1) is incorrect. I'm not sure whether you've edited the structure (Delete some atoms) and forgot to change the spin multiplicity. Since each H has one electron, spin multiplicity =1 is not possible for neutral state. If you run, you'll get the following error,

"The combination of multiplicity 1 and 3 electrons is impossible."

Try with the attached input.

Hello Sir Turbasu Sengupta ,

Thank you for your suggestion, the problem have been solved.

After complete generating wfn file for few file, now I am getting this error message, where i didn't do any amendments in between.

"Error parsing secondary structure:

QPErr --- A syntax error was detected in the input line.

/home/AIM/tg_gg/

'

Last state="Top" "

Why is it so sir?

Thank you so much in advance.

Hi, Can you please upload one of your input file for which you're getting the error. It'll help to understand the error.

Hello Sir Turbasu Sengupta ,

attached is my input file together with the error output.

Thank you Sir

How did you run it that OUTPUT is written into wfn file??? It has to create normal output (.out or .log) and ADDITIONALLY .wfn file (it is not binary but has nothing similar to output).

However my guess about mistake is that you forgot 2 empty lines after the last line.

Hello,

Please follow Dr. Sharapa's suggestion. Just add empty lines at the very end of the file (after /home/ramesh/631_calculation/AIM/tg_gg/Atg_gg_aim.wfn) line, It will run fine.

Also, it is out of topic, and may be was also done, but i would like to mention:

Molecule looks very flexible, so conformational search recommended befor DFT, and order of conformers might be significantly changed by solvation, also by changing continuum solvent to real one on the part of hydrogen bonds with cyclic part.

Here, minimum value should be set up %mem=100MW.