Re: Block Length

Search:

Re: Block Length

From: Stan van Gisbergen <vangisbergen_at_email.domain.hidden>
Date: Tue, 25 Nov 2003 11:11:35 +0100

Dear Juan,

The program needs more memory in case of a geometry optimization.
If ADF estimates it will not have sufficient memory for a certain job type,
it will adjust the vectorlength automatically, so the calculation will
fit in
the available memory.

If you increase MAXMEMORYUSAGE, this effect should disappear and
the default vectorlength will be used.

By the way, the results should be (almost) independent of the vectorlength,
so I would be surprised if you can get the SCF to converge by only changing
that parameter in the input file.

If you have reasons to believe you have encountered a bug in the software,
please send a report with your input and output files to support_at_scm.com.
Thank you.

Best regards,
Stan van Gisbergen

Juan María García Lastra wrote:

> Dear all,
>
>
>
> I perform a SP calculation and I do not have any problem with the
> convergence.
>
> When I perform a geometry optimization, whose initial geometry is the
> same than in the previous SP calculation, the first geometry cycle
> does not converge (so geometry optimization stops). I have seen in the
> output that the only difference is in the Block Length (64 for SP and
> 41 for geometry optimization). I have tried to set the same Block
> Length with several options: vectorlength, fullscf and fullfock.
> However Block Length for geometry optimization still remains in 41, so
> it does not converge.
>
>
>
> Do you know anything about this?
>
>
>
> Thank you
>
>
>

-- 
Dr. S.J.A. van Gisbergen
Scientific Computing & Modelling NV
TC, FEW, Vrije Universiteit
De Boelelaan 1083
1081 HV Amsterdam 
The Netherlands 
T +31-20-4447626
F +31-20-4447629
vangisbergen_at_scm.com
www.scm.com
Received on 2003-11-25 11:12:59

This archive was generated by hypermail 2.2.0 : 2006-11-02 07:00:02 CET