XML Schema for Monte Carlo simulations
The complexity of developing XML schemas for physics
simulations forces us to split the project into smaller pieces and
to first identify those areas where common ground between many applications
can be found.
Once we get to higher level structures, such as the input or output
of whole programs, individual applications
often tend to differ too much for a common schema to make sense. However
for a limited class of applications, such as, e.g. Monte Carlo simulations,
or even more specific, such as Monte Carlo simulations on lattice models,
a common data format can make much sense, as it allows common evaluation
or archiving solutions to be created. Here we thus present a simple first
draft of a schema for Monte Carlo simulations, based on the parameter , random
number generator , lattice, and Monte Carlo averages schemas presented before.
The main structure of the proposed schema is:
<SIMULATION>
<PARAMETERS>
<PARAMETER name="lattice">square</PARAMETER>
<PARAMETER name="L">10<>/PARAMETER>
</PARAMETERS>
<AVERAGES>
...
</AVERAGES>
<MCRUN>
...
</MCRUN>
<MCRUN>
...
</MCRUN>
</SIMULATION>
The PARAMETERS element contains a list of parameters , and the AVERAGES section
results of Monte Carlo measurements , as
discussed in the other sections.
The main new element which needs some discussion is MCRUN. Monte
Carlo simulations can trivially be parallelized by starting them on different
nodes with different random seeds. Each MCRUN element describes
a single of these
parallel Monte carlo runs:
<MCRUN>
<RNG>
...
</RNG>
<EXECUTED>
<FROM>Wed Aug 22 18:53:39 2001</FROM>
<TO>Wed Aug 22 18:53:39 2001</TO>
<MACHINE><NAME>ogdoad.ethz.ch</NAME></MACHINE>
</EXECUTED>
<EXECUTED>
<FROM>Wed Aug 22 18:53:40 2001</FROM>
<TO>Wed Aug 22 18:53:40 2001</TO>
<MACHINE><NAME>ogdoad.ethz.ch</NAME></MACHINE>
</EXECUTED>
<AVERAGES>
...
</AVERAGES>
</MCRUN>
The RNG element describes the random seed used for this particular
run, the EXECUTED elements information about when and where the
simulation was performed, and the AVERAGES element the Monte Carlo
measurements for this particular run.
For a full example of a real simulation see this example .
The above schema is the current draft which we
use to archive our results. We expect it to change and grow as we interact
with others to find a common useful format. We encourage your comments and ideas
.
Back to XML in Compuational Physics