Memory Problems with ADC2 calculations

Questions
#1

Hi everyone, I am a newbie with this awesome software and I have the following problem. I am trying to do excited-state calculations including point charges using ADC2. the calculation crashes when starting the ADC2 part due to insufficient disk memory in the Scratch directory. I have checked the quota of the directory where Q-Chem scratch is and I still have space:

            Use%    Used    Free    Quota   Limit    Use%    Used           Free     Quota        Limit
           98.62%  1.97T   28.28G  2.00T  2.20T  42.26%  1394647  1905353  3000000  3300000

I also include the relevant input section and the error message.

Input:

$rem
jobtype SP
method adc(2)
basis 6-311++G**
MEM_TOTAL 45000
CC_MEMORY 44000
MEM_STATIC 512
CC_SYMMETRY false
EE_SINGLETS 6
ADC_DO_DIIS false
ADC_PROP_ES true
ADC_PROP_ES2ES true
ADC_PROP_TPA false
make_cube_files true
$end

$plots
grid information and request to plot 6 densities
50 -6.0 6.0
50 -6.0 6.0
50 -6.0 6.0
0 0 6 0
1 2 3 4 5 6
$end

Error message:

Q-Chem fatal error occurred in module libmdc/newfileman.C, line 384:

FileMan error: Error writing 315392 bytes to file FILE_HALF_TRNS1

exported envs: QC QCAUX QCSCRATCH QCRUNNAME QCFILEPREF QCPROG QCPROG_S GUIFILE
FileMan error: Error writing 315392 bytes to file FILE_HALF_TRNS1

Could you help me understand what is the problem here?

Thank you very much in advance

Federico

#2

The FileMan error indicates that Q-Chem fails to write to disk most likely because it’s out of free space.

From the quota printout it looks like you have about 30GB of free disk space available before starting Q-Chem. With MEM_TOTAL=45000 the absolute minimum of free disk space is around 50GB. In ADC calculations and depending on the size of the molecule, disk requirements could easily go to 500GB or more.

#3

Thank you very much! I though that MEM_TOTAL was just for RAM memory, but what you are saying makes total sense. I will try setting a lower amount of memory. My system has 44 atoms, only C-H-N.

Federico

#4

The MEM_TOTAL keyword is for RAM, but in ADC calculations all available RAM is shadowed by a larger disk file, which is used for data spillover (everything that doesn’t fit in RAM, that is). Your 44-atom molecule with a triple-zeta basis set is very big by ADC standards, probably somewhere close to a thousand basis functions, which is way beyond the capabilities of the hardware you seem to be using.

#5

Thank you. I apologize for my late reply.

F