FileMan error: End of file reached prematurely reading FILE_OVERLAP_MATRIX

I am running a single point HSE calculation on a fairly large (2736 atoms) system, and it crashes when reaching the overlap matrix.
Relevant segments of the input and output are:

$molecule
0 1
(In,P,H. 2736 atoms)
$end

$rem
method HSE-HJS
BASIS LANL2DZ
ECP fit-LANL2DZ
scf_algorithm RCA
max_scf_cycles 3000
mem_static 250000
mem_total 500000
sym_ignore true
scf_final_print 1
thresh 10
scf_convergence 6
xc_grid 1
KS_gap_print true
BASISPROJTYPE OVPROJECTION
$end

$external_charges
(on the H’s, used to implement pseudo-hydrogens)
$end

Q-Chem warning in module libmdc/get_mega.C, line 171:

Allowed MEM_STATIC has been capped at MEM_STATIC=8191.
Reduce the value of MEM_STATIC to silence this warning.

Total QAlloc Memory Limit 500000 MB
Mega-Array Size 8007 MB
MEM_STATIC part 8191 MB
A cutoff of 1.0D-10 yielded2002662 shell pairs
There are 8280882 function pairs
Requested basis set is LANL2DZ
Compound shells will be simplified
There are 9792 shells and 18432 basis functions
A cutoff of 1.0D-10 yielded2002662 shell pairs
There are 8280882 function pairs

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

FileMan error: End of file reached prematurely reading (570425360) bytes in file FILE_OVERLAP_MATRIX
Path: /scratch/nuritman/qchem//qchem54.9972959.pbs/320.1

Please submit a crash report at Q-Chem Crash Reporter


The calculation is run on a 500 gb computer node (I get the same error while working with other nodes). I note that the same calculation on similar, smaller systems works fine.

I realize this is a memory issue, but I do not see how to solve it…

Let’s see if others have suggestions but you may simply have reached an upper limit of Q-Chem’s SCF code. 18,432 basis functions is about 3x larger than anything I have done personally. My largest calculation [M06-2X/6-311+G(2df,2p) on ~200 atoms and ~6,100 basis functions] certainly isn’t an upper limit, but your system is extremely large even by SCF standards.

I understand. Thank you.

@nurit Could you please send the full input file to support@q-chem.com for us to take a look? Thank you.

Any particular reason the SCF code is hard coded with a 18,432 basis function limit?

There’s no hard-coded limit, I just suspect that some kind of resource is being maxed out.

Code resource or system resource? What is defining this limit?

I just did. Thank you.