Hi,
I couldn't restart a VSCF calculation, and I think I've found the problem.
If I understood the logic of the vscf.F file, in the RDCOUP subroutine, lines 2067 and 2082 (version 6.1.1 of NWChem) should read
do i=1, NM-1
instead of
do i=1, NM
and more important, in lines 2096-2097
9019 if ((nstart(2).eq.nm).and.(nstart(3).eq.n).and.(nstart(4).eq.nm)
& .and.(nstart(5).eq.n)) nstart(1)=2
should be replaced by
9019 if ((nstart(2).eq.(nm-1)).and.(nstart(3).eq.n)
& .and.(nstart(4).eq.nm).and.(nstart(5).eq.n)) nstart(1)=2
Otherwise, VSCF will always fail to restart a calculation with "coupling pair".
Maybe a similar change will be necessary in RDTRIP:
do i=1, NM-2
do j=i+1, NM-1
do k=j+1, NM
instead of
do i=1, NM
do j=i+1, NM
do k=j+1, NM
and the correponding in lines 2153-2155:
9024 if ((nstart(2).eq.nm).and.(nstart(3).eq.n).and.(nstart(4).eq.nm)
& .and.(nstart(5).eq.n).and.(nstart(6).eq.nm)
& .and.(nstart(7).eq.n)) nstart(1)=3
Javier
|