Difference between revisions of "Tutorials:Parameters"

From ALPS
Jump to: navigation, search
(Additional parameters for exact diagonalization)
 
(4 intermediate revisions by 2 users not shown)
Line 1: Line 1:
 +
{{Languages|Tutorials:Parameters}}
 +
 +
 
The following input parameters are common to most of the ALPS applications
 
The following input parameters are common to most of the ALPS applications
  
Line 78: Line 81:
 
| align="center"| 0
 
| align="center"| 0
 
|| The random number seed used in the next run. After using a seed in the creation of a Monte Carlo run, this value gets incremented by one.
 
|| The random number seed used in the next run. After using a seed in the creation of a Monte Carlo run, this value gets incremented by one.
 +
|-
 +
|| RNG
 +
| align="center"| "mt19937"
 +
|| The pseudo-random number generator to be used. Allowed values are [http://www.boost.org/doc/libs/1_35_0/libs/random/random-generators.html#lagged_fibonacci_spec "lagged_fibonacci607"] and [http://www.boost.org/doc/libs/1_35_0/libs/random/random-generators.html#mt19937 "mt19937"].
 
|-
 
|-
 
|| WORK_FACTOR
 
|| WORK_FACTOR
Line 91: Line 98:
 
||Number of Monte Carlo sweeps for thermalization
 
||Number of Monte Carlo sweeps for thermalization
 
|}
 
|}
 +
 
==Additional parameters for exact diagonalization ==
 
==Additional parameters for exact diagonalization ==
  
Line 109: Line 117:
 
||TRANSLATION_SYMMETRY
 
||TRANSLATION_SYMMETRY
 
||true
 
||true
||<tt>sparsediag</tt> exploits translational symmetry and classifies eigenstates by their momentum quantum numbers when possible. This symmetry reductions can be switched off with TRANSLATION_SYMMETRY=false.
+
||<tt>fulldiag</tt> and <tt>sparsediag</tt> exploit translational symmetry and classify eigenstates by their momentum quantum numbers when possible. This symmetry reductions can be switched off with TRANSLATION_SYMMETRY=false.
 
|-
 
|-
 
||TOTAL_MOMENTUM
 
||TOTAL_MOMENTUM
 
||
 
||
 
|| fixes the value of the total momentum. Further explanations can be found below.
 
|| fixes the value of the total momentum. Further explanations can be found below.
 +
|-
 +
||MEASURE_ENERGY
 +
||false
 +
||if no measurements are explicitly specified, <tt>fulldiag</tt> and <tt>sparsediag</tt> do not store any information on eigenstates by default. Of course, the energy can always be computed for any eigenstate. If you wish to have this in the output and no other measurements are specified, you can specify MEASURE_ENERGY=true.
 
|}
 
|}
 +
'''Note:''' Instead of true and false, you can also specify 1 and 0, respectively.
  
  
Line 122: Line 135:
  
 
TOTAL_MOMENTUM takes the momentum quantum numbers as a vector,
 
TOTAL_MOMENTUM takes the momentum quantum numbers as a vector,
<i>i.e.</i> a space-separated list of numbers. Typically, each momentum
+
''i.e.'' a space-separated list of numbers. Typically, each momentum
 
quantum number <i>k<sub>i</sub></i> will be of the form
 
quantum number <i>k<sub>i</sub></i> will be of the form
<br>
+
 
 
<center><font size=+1>
 
<center><font size=+1>
 
<i>k<sub>i</sub></i> =
 
<i>k<sub>i</sub></i> =
 
2 &pi; <i>n<sub>i</sub></i>/<i>L<sub>i</sub></i> ,
 
2 &pi; <i>n<sub>i</sub></i>/<i>L<sub>i</sub></i> ,
 
</font></center>
 
</font></center>
<br>
+
 
 
where <i>n<sub>i</sub></i> is an integer and <i>L<sub>i</sub></i> the
 
where <i>n<sub>i</sub></i> is an integer and <i>L<sub>i</sub></i> the
 
linear extent in the corresponding direction. It is possible to specify
 
linear extent in the corresponding direction. It is possible to specify
 
a symbolic number such as <tt>2*Pi/5</tt> if you put the values in
 
a symbolic number such as <tt>2*Pi/5</tt> if you put the values in
 
quotation marks, e.g. <tt>TOTAL_MOMENTUM="2*Pi/5 0"</tt>.
 
quotation marks, e.g. <tt>TOTAL_MOMENTUM="2*Pi/5 0"</tt>.
<br>
+
 
<br>
+
 
<font color="red"><b>Warning:</b></font> An illegal value of
+
<font color="red">'''Warning:'''</font> An illegal value of TOTAL_MOMENTUM may lead to incorrect results without any further error message.
TOTAL_MOMENTUM may lead to incorrect results without any further error message.
 

Latest revision as of 07:59, 2 March 2012


The following input parameters are common to most of the ALPS applications


Lattice definition

ALPS applications on lattices specify the lattice with the following three parameters

Parameter Default Meaning
LATTICE_LIBRARY lattices.xml path to a file containing lattice descriptions
LATTICE name of the lattice, specified by dimensionality, extent and unit cell.
GRAPH as an alternative to a lattice, also a specific arbitrary graph defined in the lattice library can be specified.

In addition, the lattice description can require further parameters (e.g. L or W) as specified in the lattice description file.

Model definition

ALPS quantum lattice models can be specified using the following parameters

Parameter Default Meaning
MODEL_LIBRARY models.xml path to a file containing model descriptions</td>
MODEL name of the model (for example "spin" or "boson")

The model description can also require further parameters (e.g. S=1/2 or S=1, h=0.5 for spin models, t=1.5 or mu=0.5 for boson models) as specified in the model description file.

Parameters for finite temperature simulations

Parameter Meaning
T the temperature
BETA inverse of temperature (if temperature is not given

Additional parameters for Monte Carlo simulations

Parameter Default Meaning
SEED 0 The random number seed used in the next run. After using a seed in the creation of a Monte Carlo run, this value gets incremented by one.
RNG "mt19937" The pseudo-random number generator to be used. Allowed values are "lagged_fibonacci607" and "mt19937".
WORK_FACTOR 1 A factor by which the work that needs to be done for a simulation is multiplied in load balancing.
SWEEPS number of Monte Carlo steps (after thermalization)
THERMALIZATION Number of Monte Carlo sweeps for thermalization

Additional parameters for exact diagonalization

Parameter Default Meaning
CONSERVED_QUANTUMNUMBERS specifies conserved global quantum numbers which are used to split the computation into smaller computations for the different sectors. If more than one quantum number is conserved, the quantum numbers are listed in double quotes and separated by commas as in CONSERVED_QUANTUMNUMBERS="N,Sz"
N_total, Sz_total, ... and similar parameters might be defined for your model through a constraint in your model definition. These constraints will be used if these parameters are specified and the quantumnumber is listed in CONSERVED_QUANTUMNUMBERS.
TRANSLATION_SYMMETRY true fulldiag and sparsediag exploit translational symmetry and classify eigenstates by their momentum quantum numbers when possible. This symmetry reductions can be switched off with TRANSLATION_SYMMETRY=false.
TOTAL_MOMENTUM fixes the value of the total momentum. Further explanations can be found below.
MEASURE_ENERGY false if no measurements are explicitly specified, fulldiag and sparsediag do not store any information on eigenstates by default. Of course, the energy can always be computed for any eigenstate. If you wish to have this in the output and no other measurements are specified, you can specify MEASURE_ENERGY=true.

Note: Instead of true and false, you can also specify 1 and 0, respectively.


If the lattice supports translation symmetries, you can specify the total momentum quantum numbers, but you should be quite careful in doing so.

TOTAL_MOMENTUM takes the momentum quantum numbers as a vector, i.e. a space-separated list of numbers. Typically, each momentum quantum number ki will be of the form

ki = 2 π ni/Li ,

where ni is an integer and Li the linear extent in the corresponding direction. It is possible to specify a symbolic number such as 2*Pi/5 if you put the values in quotation marks, e.g. TOTAL_MOMENTUM="2*Pi/5 0".


Warning: An illegal value of TOTAL_MOMENTUM may lead to incorrect results without any further error message.