[Pw_forum] cholesky decomposition

[Liping Yu]

Hi, Paolo,
I have been met this kind of error messages many times when doing
structure relaxation with 'bfgs' using espresso-4.0. This error ever
occurred on IMB sp4/5 and Sun cluster comprised of Opteron processors.
There is no such error message if I change 'bfgs' to 'damp', or use the
older code version (< 4.0). I didn't check it in v.4.0.1.
Liping

Paolo Giannozzi wrote:
> On Jul 11, 2008, at 23:03 , Axel Kohlmeyer wrote:
>
>   
>> AC>       from  pzpotrf  : error #       452
>> AC>       problems computing cholesky decomposition
>>
>> the same or very similar problem has been reported very recently
>>     
>
> but nobody has ever provided a test job, and not even details on
> code version, hardware/software. It might be a real bug; it might
> be a numerical problem, or a problem with mathematical libraries,
> or some other weirdness, like this one, fixed in v.4 for real
> diagonalization (Gamma case):
>     http://www.democritos.it:8888/O-sesame/chngview?cn=5777
> and in v.4.0.1 for complex diagonalization (generic k-point):
>     http://www.democritos.it:8888/O-sesame/chngview?cn=5932
> The subspace diagonalization algorithm is the same used in CP
> for iterative orthonormalization, and has been tested on O(1000)
> atoms.
>
> Paolo
> ---
> Paolo Giannozzi, Dept of Physics, University of Udine
> via delle Scienze 208, 33100 Udine, Italy
> Phone +39-0432-558216, fax +39-0432-558222
>
>
>
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://www.democritos.it/mailman/listinfo/pw_forum
>   


-- 
Liping YU, PhD Candidate
Department of Physics & Center for High Performance Simulation
North Carolina State University, Raleigh, NC 27695, USA
tel: 1-919-5157228