Posted by Per Åke Malmqvist on October 23, 2006 at 09:59:10:
In Reply to: GUGA without symmetry posted by pawel wielgus on June 23, 2006 at 12:34:46:
Dear Pawel,
The GUGA and MRCI codes are very old, and are not actively
maintained. The problem is (apart from cryptic error messages...)
that an array needs to be redefined in an include file for guga,
namely src/niocr.inc. There, a parameter NIOCR is set to 20000.
It can be increased: I have tried with 60000, and it seems to work
fine.
However, there is a similar problem with MRCI. There, it is a parameter limiting
the number of reference CSF's to 1000. You need 4500. I have tried to increase thie
size limit to 5000, but the the program bombs with segmentation error.
I may continue to persue this problem, if I find the time, running MRCI through a
debugger. Meanwhile, consider the fact that the calculation you attempt may be
larger than you anticipated? Perhaps you do not need that large a reference?
( There will be 2434575 CSF's, which is not in itself terribly big.)
Best,
PÅ