[Octopus-users] Problems with TD and Casida

[Xavier Andrade]

Dear Juzar,

In my experience calculations should not take that long, specially if you 
are using large spacings. Of course it depends on your systems (hardware, 
compilers, BLAS, etc.), can you give us some details about it? How much 
does it take each step of time propagation?

Cheers,

Xavier

On Tue, 23 Dec 2008, juzar thingna wrote:

> Dear all,
> We are running octopus code for a 65 atom cluster (Si29H36). We are
> intersted in
> doing casida and td calculations. There are a large number of unocc
> states....the casida and td calculations are running on 4/8 CPUs for
> almost one week and still going on.....should this take that long....
> please advise us....The input files are as follows
> 
> ---------------------------------------------------------------------------
> -------------------
> Casida Input file.
> ---------------------------------------------------------------------------
> -------------------
> CalculationMode = casida
> Units = eV_Angstrom
> 
> radius = 4.0
> spacing = 0.32
> 
> XYZCoordinates = "si29h36.xyz"
> 
> NumberUnoccStates = 150
> ---------------------------------------------------------------------------
> --------------------
> TD Input file
> ---------------------------------------------------------------------------
> --------------------
> CalculationMode = td
> Units = eV_Angstrom
> fromScratch = yes
> 
> radius = 4.0
> spacing = 0.30
> 
> XYZCoordinates = "si29h36.xyz"
> 
> TDDeltaStrength = 0.01
> TDPolarizationDirection = 1
> TDPolarizationEquivAxis = 3
> 
> tmax = 30
> TDEvolutionMethod = aetrs
> TDTimeStep = 0.002
> TDMaximumIter = tmax/TDTimeStep
> SpecEnergyStep = 0.001
> ---------------------------------------------------------------------------
> ----------------------------
> 
> 
> --
> "At the end of all our searching we will arrive at the place we began and
> know it for the first time."
> 
>