Strange results with cc-pcvtz vs. aug-cc-pcvtz

Dear all,

I'm computing an all electron CCSD(T) energy with cc-pcvtz and aug-cc-pcvtz. I'm getting very strange energies. I've gotten similar results as the aug-cc-pcvtz with other programs, and other programs with the cc-pcvtz basis give a result closer to aug-cc-pcvtz.

I'm using the April 12, 2017 version compiled with intel 2015 and openmp.

The memory I'm using is overkill and it can be ran with way less.

Ecc-pcvtz= -4583.172882759474
Eaug-cc-pcvtz=-190.081261151800

basis=cc-pcvtz
calc=CCSD(T)
scftyp=RHF
mem=40000MB
scftol=10
cctol=8
charge=-1
core=0
ovirt=MP2
eps=0.9875
geom=xyz
6

C -1.1708030505 0.0784958569 0.0000000000
H -1.9777124093 -0.6733043675 0.0000000000
H -1.2763081940 0.7232314698 0.8883690097
H -1.2763081940 0.7232314698 -0.8883690097
O 0.0370131332 -0.5926235454 0.0000000000
O 1.0591571866 0.4809243702 0.0000000000

I cannot reproduce it, at least with another compiler version. It may be some compiler-related issue. You should try with the precompiled binaries.

That seems to fix the issue. Also, the binary is ever so slightly faster than the version I compiled so that's a major relief for future work with it.