====== Advanced settings ======

==== Parallelization ====

=== Number of threads ===


The number of threads can be controlled using the command <Maximum_Number_of_Threads>
<code>
<Simulation_Parameter>
  ...
  <Maximum_Number_of_Threads>12</Maximum_Number_of_Threads>
  ...
</Simulation_Parameter>
</code>

For an automatic setting, use 0 or do not specify this command. In this case, the number of threads will be set to half of the number of physical cores.

In any case, we recommended that <Maximum_Number_of_Threads> does not exceed the number of **physical** cores (note that when hyperthreading is activated, the number of physical cores is half the one on logical processsors).

=== Parallelization in case of combined Temperature-Voltage sweep ===
In the case where a combined Temperature-Voltage sweep is used, there are two levels of parallelization. The first is at the level of the command <Threads> in the <SweepParameters> section.

<code>
<SweepParameters>   
  <SweepType>Temperature-Voltage</SweepType>
  ...
 <Threads>6</Threads> <!-- Parallelization for Temperature-Voltage sweep -->
 </SweepParameters>
</code>

The second level of parallelization occurs in the <Simulation_Parameter> section:
<code>
<Simulation_Parameter>
  ...
  <Maximum_Number_of_Threads>2</Maximum_Number_of_Threads>
  ...
</Simulation_Parameter>
</code>

In this example, the total number of threads will be <Threads>* <Maximum_Number_of_Threads> = 6*2 = 12.


==== Definition of graded alloy and graded interfaces ====

There are two possibilities to define graded alloy profile. 

=== Linearly graded alloys ===
To define a linear graded alloy profile, two materials first need to be defined in the ''<Material>'' section
For example,
<code>
    <Material>
      <Name>GaAs</Name>                           <!-- Binary material -->
      <Alias>mat1</Alias>
      <Effective_mass_from_kp_parameters>yes</Effective_mass_from_kp_parameters>
    </Material>

    <Material>
      <Name>Al(x)Ga(1-x)As</Name>                 <!-- Ternary material -->
      <Alloy_Composition>0.15</Alloy_Composition> <!-- alloy composition x -->
      <Alias>mat2</Alias>
      <Effective_mass_from_kp_parameters>yes</Effective_mass_from_kp_parameters>
    </Material>
</code>

Then, in the layer definition (named ''<Superlattice>''), the following can be used to define a layer with a linear alloy profile, with composition starting from the material "mat1" and ending with the material "mat2" composition.
<code>
<Layer> <!-- all the material parameters will be linearily interpolated between <Material1> and <Material2> -->
    <Material1>mat1</Material1>
    <Material2>mat2</Material2>
    <Thickness unit="nm">5.0</Thickness>
  </Layer>
</code>


=== Graded interfaces ===

To specify a grading for all the interfaces of the structure, the following command ''<InterfaceWidth>'' has to be included in the ''<Interface_Roughness>'' section.
<code>
<Interface_Roughness>
      <InterfaceWidth unit="nm">0.8</InterfaceWidth>
      ...
   </Interface_Roughness>
</code>
In this case, the alloy profile of the all strucure is convoluted by a Gaussian. For a well defined interface, this results in an error function profile of the form:
$$
c(z) = c_0 + d_0 \text{erf} \left[2
\sqrt{ln(2)}(z − z_0)/L \right]
$$ 
where $L$ is the quantity specified in ''<InterfaceWidth>''.

==== Scattering processes ====
=== Contributions of individual scattering processes ===
The following command allow to display self-energies and spectral functions for some specific mechanisms, like e.g. scattering due to optical phonons:
<code>
<Scattering>
  ...
  <Separate_Scattering>yes</Separate_Scattering>
  ...
</Scattering>
</code>
This command is only for analysis purpose, and has no influence on the other calculated quantities. It only involves an additionnal calculation of self-energies and Green's functions.
In each basis folder, a folder with the name of the scattering process will appear, containing the retarded self-energy and the spectral function arising only from the specific scattering mechanism. Hence this spectral function gives the contribution to broadening of an individual scattering process.

=== Homogeneous Coulomb scattering ===
To speed up the calculation, it is possible to consider the assumption of homogeneously distributed Coulomb scatterers (ionized impurities and other charge carriers) using the following command:
<code>
<Scattering>
  ...
  <Homogeneous_Coulomb>yes</Homogeneous_Coulomb>
  ...
</Scattering>
</code>


/*==== Asymetric interface roughness ====
It is possible to specify asymetric interface roughness using the following command (note: this feature is currently not working due to syntax changes).
<code>
  <Interface_Roughness>

   <InterfaceAutoCorrelationType> 0 </InterfaceAutoCorrelationType>
   <Correlation_Length_in_XY unit="nm"> 8 </Correlation_Length_in_XY>

   <Asymmetric_Interfaces> yes </Asymmetric_Interfaces>
   <Amplitude_in_Z_Left>  0.1 </Amplitude_in_Z_Left>
   <Amplitude_in_Z_Right> 0.2 </Amplitude_in_Z_Right>

  </Interface_Roughness>
</code>*/

==== Output format for 2D plots ====
By default, 2D plots are output in a [[http://www.vtk.org|VTK format]] (.vtr extension).
 [[http://www.gnuplot.info/|Gnuplot]] files (.plt extension) are also generated (in a file explorer, double click on the file to generate the gnuplot figure).

In addition, to output 2D plots in a [[http://www.avs.com|AVS/Express format]] (.fld extension), the following command "<FLD_format>" can be added in the <Output> section of the input file.

The following commands can be used to control the output formats:

<code>
<Output>
...
<VTK_format>yes</VTK_format>
<FLD_format>yes</FLD_format>
<Gnuplot_format>yes</Gnuplot_format>
...
</Output>
</code>



==== Scaling 1D wavefunctions ====
In order to scale the output of the square of the wavefunctions (1D plots), the following command can be used:
<code>
<Output>
...
<ScaleWaveFunction>0.1</ScaleWaveFunction>
...
</Output>
</code>