# CCL:G: Frequency Calculation error

*From*: "Soumya Lipsa Rath"
<soumyalipsa*gmail.com>
*Subject*: CCL:G: Frequency Calculation error
*Date*: Mon, 22 Feb 2016 22:53:34 -0500

Sent to CCL by: "Soumya Lipsa Rath" [soumyalipsa(_)gmail.com]
Dear all,
I am trying to do a frequency calculation from a geometry optimized structure
using the following:
%chk=hess.chk
%nproc=6
%mem=96GB
# MP2/6-31G* Geom=(AllCheck,ModRedundant) Freq NoSymm IOp(7/33=1) SCF=Tight
Guess=Read
B * * K
A * * * K
L * * * K
D * * * * K
B 2 1 A
B 3 2 A
B 6 2 A
B 4 3 A
B 31 3 A
B 5 4 A
B 8 4 A
B 7 6 A
B 32 6 A
........................................................
However, I encountered the following lines during calculation after which the
calculation stops.
Solved reduced A of dimension 1114 with 153 vectors.
End of Minotr F.D. properties file 721 does not exist.
End of Minotr F.D. properties file 722 does not exist.
End of Minotr F.D. properties file 788 does not exist.
MDV= 12884901888.
Form MO integral derivatives with frozen-active canonical
formalism.
Discarding MO integrals.
Reordered first order wavefunction length = 4657669128
In DefCFB: NBatch= 1 ICI=126 ICA=383 LFMax= 64
Large arrays: LIAPS= 50010923592 LIARS= 5158297620 words.
Assuming memory error, I ran the same locally, where it stops again after the
following lines
Disk-based method using ON**2 memory for 79 occupieds at a time.
Permanent disk used for amplitudes= 2904914439 words.
Estimated scratch disk usage= 17599594491 words.
Actual scratch disk usage= 16841782779 words.
JobTyp=1 Pass 1: I= 48 to 126 NPSUse= 6 ParTrn=T ParDer=T
DoDerP=T.
(rs|ai) integrals will be sorted in core.
Surprisingly if I use HF instead of MP2 in my gaussian input file, it runs and
terminates normally. I wanted to know, how much different is the frequency
calculation using MP2 and HF? If its very different could you please suggest how
to run the MP2 based frequency calculation without encountering the above
errors?
Thanks,
Soumya