VASP on-the-fly method issue

[stefani_setiono]

Dear all, I hope you are well. I am new to VASP and also on-the-fly
simulation methods, I was trying to run my simulation however, i have two
issues. firstly, I am unable to run the simulation at a higher number of
core, I had initially assume this was due to lack of memory issue but that
did not solve the issue, the only way I was able to get it to start running
was by decreasing the simulation to 1 core. Secondly, whilst running the
simulation, i came across these issues as well:

DAV: 24 -0.941385358407E+04 -0.25654E-04 -0.70247E-06 7336
0.438E-03 0.410E-03
DAV: 25 -0.941385359527E+04 -0.11196E-04 -0.51861E-06 6784
0.270E-03 0.213E-03
DAV: 26 -0.941385360015E+04 -0.48841E-05 -0.19369E-06 5432
0.166E-03 0.249E-03
DAV: 27 -0.941385360242E+04 -0.22684E-05 -0.75264E-07 4776
0.142E-03 0.124E-03
DAV: 28 -0.941385360348E+04 -0.10589E-05 -0.69777E-07 4800
0.105E-03 0.139E-03
DAV: 29 -0.941385360398E+04 -0.50099E-06 -0.47129E-07 4784
0.856E-04

-----------------------------------------------------------------------------
|
|
| EEEEEEE RRRRRR RRRRRR OOOOOOO RRRRRR ### ### ###
|
| E R R R R O O R R ### ### ###
|
| E R R R R O O R R ### ### ###
|
| EEEEE RRRRRR RRRRRR O O RRRRRR # # #
|
| E R R R R O O R R
|
| E R R R R O O R R ### ### ###
|
| EEEEEEE R R R R OOOOOOO R R ### ### ###
|
|
|
| Unsupported xc functional, please specify VDW_A1, VDW_A2, and
|
| VDW_S8 in the INCAR file.
|
|
|
| ----> I REFUSE TO CONTINUE WITH THIS SICK JOB ... BYE!!! <----
|
|
|

-----------------------------------------------------------------------------

i am not sure where to move on from here and I'd be happy to share you my
input files.

Thank you so much in advance for your help!

[alexey.tal]

Dear stefani_setiono,

It is rather difficult to give recommendations without seeing what kind of calculation you are trying to run. These two issues don't seem to be related, so perhaps you could provide input/output files for each corresponding issue. Please see the guidelines on how to provide the files.

Best wishes,
Alexey

[stefani_setiono]

Dear Alexey,
thank you so much for replying. The main issue I am having right now is that I am unable to run the simulation without it stopping suddenly. I think that this might be due to memory issue but I was unable to fix this.

attached are all my input files and the important output files.

Many thanks,
Stefani

[alexey.tal]

Thank you for sending the files. It does look like this job is running out of memory. As far as I see, you are trying to run it on 8-core machine with 64 Gb? Could you run it on a machine with more memory?

[stefani_setiono]

Hi Alexey,
thanks for the reply! I have tried but the job took too long in the queue to start... I was wondering if there is a way to gauge how much memory i might need... Other than that, could it perhaps be that my structure with 1980 atoms is too big for this method?

[alexey.tal]

You can find how the required memory is estimated on our wiki.
In your ML_LOGFILE I see that the estimated memory is around 58 Gb per MPI rank. How much memory do you have on your machine?

[alexey.tal]

I tried to run your job and found out that it requires over 1Tb of memory. However, I was able to reduce it to just under 1Tb by building VASP with the shared memory support. You need to add the following lines in your makefile.include:
-Duse_shmem
and
OBJECTS_LIB+= getshmem.o

However, training on such a large structure will be a long calculation. So you should try to reduce the structure for training the force-field. Your structure looks like some type of MOF system. Maybe there is a way to reduce the structure to just one unit with a few hundred atoms. Furthermore, right now you are using the default 1500 reference configurations, which is most likely not going to be enough for an accurate fore-field. You should consider that you will need over 5000 configurations.