Problems in RCCSD(T)

[John Kerkines]

Dear all,

I have been encountering the same problem under similar situations in
RCCSD(T). This time I am performing calculations on a 39 electron system
using Dunning's cc basis sets, in order to reach the CBS limit. But, after
the basis set gets quite large (QZ and up in this case), the RCCSD(T)
calculation stops when calculating the triples perturbative correction. I
get a "Signal too large" message and the Molpro output just reads "ERROR
EXIT". Nothing else.

Now, by resubmitting the job without changing anything at all, the triples
correction is now calculated, but the program stops at the next point in
the potential energy curve (again in the triples calculation).

I have encountered this problem before, but it was

a) at the smallest internuclear  distance (at the most repulsive part of
my calculated curve)

b) when I tried to open core electrons in the RCCSD(T) calculations

Plenty of disk space is available, but it may be a memory problem. Any
ideas?

Best Regards,
John

P.S. I am using Molpro 2000.1