Orbital record saving

[Nathan DeYonker]

Hey everyone,

I'm having a bit of trouble saving and starting from the appropriate
configuration set when running an MRCI job.

I'm trying to calculate an excited state potential energy surfaces of
HBO-BOH isomerization.

What I want to do is optimize the HB/BO bond distances at a range of
angles. Here is how the main part of my input is set up:

do k=1,#iang

ang1(k)=ang1e+stpr(iang(k)+1)

geometry={z;angstrom
           h
           b,h,hb
           o,b,bo,h,ang1(k)}

rhf;closed,6,1;wf,14,1,0;

multi;core,2;wf,14,1,0;wf,14,2,0;noextra;expec2,lzz
multi;closed,2;wf,14,2,0;state,4;wf,14,1,0;state,4;noextra;expec2,lzz;


kk=0
label
kk=kk+1

multi;closed,2;save,ref=6500.2+kk;wf,14,1,0;state,4;wf,14,2,0;state,4;noextra;expec2,lzz;
ci;core,2;maxiter,30;select,6500.2+kk,,0.01;wf,14,1,0;state,4;
edavid=energd

opt,startcmd=label;inactive,ang1;numerical,hb=0.01,bo=0.01;variable,edavid;
kk=0
enddo
---

I have the do loop running over various H-B-O angles from 0 to 180
degrees.

The problem is this, upon finishing the first optimization (say at 180
degrees), the second optimization (at 170 degrees) is not "overwriting"
the orbital records. 

These records quickly get to 200 with numerical derivatives, whereupon the
program crashes for have too many records... 

 More than           200  records on file             2
 ? Error
 ? Too many records
 ? The problem occurs in reservem

 ERROR EXIT
 CURRENT STACK:      MAIN

When the first optimization is completed with 11 optimization steps and 5
points per opt (5 x 11), the last MCSCF states the following correctly:

A'
 Orbital configurations saved on  6555.2
A''
 Orbital configurations saved on  6655.2

Then the MRCI:

 144 Reference configurations selected from 6555.2 for reference states 1
- 4 with threshold  0.010000

However, at the beginning of the second optimization, with KK = 0,
shouldn't the save,ref=6500.2+kk command renew the orbital records at
6500/6600 ??

This isn't what happens,

A'
 Orbital configurations saved on  6556.2
A''
 Orbital configurations saved on  6656.2


MRCI picks up the wrong set of configurations...

 158 Reference configurations selected from 6501.2 for reference states 1
- 4 with threshold  0.010000

Job crashes before my PES is even nearly mapped out. 

Is this a bug? Are there any workarounds other than re-compiling with a
higher maximum number of orbital records? 

 --------------------------------------------------------------------
  Nathan DeYonker              | phone: 706 542-2067
  Ctr. for Comp. Quantum Chem. | fax:   706 542-0406
  University of Georgia        | e-mail: nate at ccqc.uga.edu
  Athens, GA  30602-2556       | http://zopyros.ccqc.uga.edu
 --------------------------------------------------------------------