An error with CCSD(T)/CCPVQZ Single point energy through TCE


Just Got Here
Hi,

Please I need this forum to help me overcome this error. I have been trying to run a single point energy calculation for Pyrrolidine-BH3 but everytime I was not successful. Find below my input:

echo

start PYRBH3

title "Pyrrolidine-borane CCSDT calculation"
charge 0

geometry units angstroms print xyz autosym
zmatrix
C
C 1 r2
C 1 r3 2 a3
H 1 r4 2 a4 3 d4
H 1 r5 2 a5 3 d5
C 2 r6 1 a6 3 d6
H 2 r7 1 a7 3 d7
H 2 r8 1 a8 3 d8
N 3 r9 1 a9 2 d9
H 3 r10 1 a10 2 d10
H 3 r11 1 a11 2 d11
H 6 r12 2 a12 1 d12
H 6 r13 2 a13 1 d13
B 9 r14 3 a14 1 d14
H 14 r15 9 a15 3 d15
H 14 r16 9 a16 3 d16
H 14 r17 9 a17 3 d17
H 9 r18 3 a18 1 d18
variables
r2 1.54621
r3 1.52602
a3 103.51863
r4 1.09526
a4 110.71891
d4 242.31675
r5 1.09250
a5 112.82688
d5 121.89968
r6 1.54531
a6 104.86526
d6 18.29287
r7 1.09256
a7 111.01127
d7 259.33736
r8 1.09246
a8 111.67801
d8 139.70715
r9 1.48452
a9 104.07549
d9 323.09556
r10 1.09117
a10 115.46467
d10 202.98876
r11 1.09345
a11 112.21655
d11 77.75787
r12 1.09143
a12 112.26288
d12 250.81587
r13 1.09098
a13 114.64820
d13 127.05214
r14 1.62902
a14 115.40258
d14 167.96900
r15 1.21697
a15 105.06416
d15 300.58824
r16 1.21616
a16 105.63946
d16 180.74166
r17 1.21617
a17 105.72104
d17 60.74564
r18 1.02091
a18 107.60238
d18 286.68647
end
end

memory heap 1000 mb stack 4000 mb global 10000 mb noverify

basis spherical
 * library cc-pVQZ
end

scf
singlet
thresh 1e-10
rohf
end

tce
freeze atomic
ccsd(t)
maxiter 200
diis 10
tilesize 18
thresh 1e-7
io ga
2EORB
2EMET 13
end

task tce energy


and also my batch file is:

  1. !/bin/bash -e
  2. SBATCH --job-name CCSDT_CBS
  3. SBATCH --account ****
  4. SBATCH --time 2-00:00:00
  5. SBATCH --tasks 10
  6. SBATCH --cpus-per-task 10
  7. SBATCH --tasks-per-node 2
  8. SBATCH --mem-per-cpu 4G
  9. SBATCH --output PYRBH3_CCSDT_VQZ.out # Include the job ID in the names
  10. SBATCH --error PYRBH3_CCSDT_VQZ.err # of the output and error files
  11. Since --output is set, there is no need to redirect output to a file
  12. on the NWChem command line.
export ARMCI_DEFAULT_SHMMAX=4096
module load NWChem/6.8.revision47-gimkl-2017a-2017-12-14-Python-2.7.14
srun nwchem PYRBH3_CCSDT_VQZ.nw


The end of my output file is:


Global array virtual files algorithm will be used

Parallel file system coherency ......... OK

Fock matrix recomputed
1-e file size = 467856
1-e file name = ./PYRBH3.f1
Cpu & wall time / sec 120.9 117.3
4-electron integrals stored in orbital form
available GA memory 10484980712 bytes
------------------------------------------------------------------------
createfile: failed ga_create size/nproc bytes 23087245708
------------------------------------------------------------------------
------------------------------------------------------------------------
current input line :
available GA memory 10484971712 bytes
createfile: failed ga_create size/nproc bytes 23087245708
------------------------------------------------------------------------
------------------------------------------------------------------------
current input line :
104: task tce energy
------------------------------------------------------------------------
------------------------------------------------------------------------
------------------------------------------------------------------------
For more information see the NWChem manual at http://nwchemgit.github.io/index.php/NWChem_Documentation


For further details see manual section:                                                                                                                                                                                                                                                                
0:0:createfile: failed ga_create size/nproc bytes:: 1612409228
    0: 


I am running on 5 nodes on a cluster that has a kernel.shmmax = 4294967296

I hope you will come to my aid.

Cheers

Forum Vet
Your doing is confusing.
It is impossible to finishi the huge task CCSDT here with your current setting, pehaps your current facility, because the physical memory should be large enough. It might be very slow also; thus I think you might resort to supercomputers installing official NWCHEM in US national labs if possible.


Forum >> NWChem's corner >> Running NWChem