home.html

<center>
  <img src="logo.png" />
</center>

<h1>Peano - a Framework for Solvers on Spacetree Grids</h1>
<h2>www.peano-framework.org</h2>


<p>
Peano is an open source C++ solver framework. 
It is based upon the fact that spacetrees, a generalisation of the classical 
octree concept, yield a cascade of adaptive Cartesian grids. 
Consequently, any spacetree traversal is equivalent to an element-wise 
traversal of the hierarchy of the adaptive Cartesian grids. 
The software Peano realises such a grid traversal and storage algorithm, and 
it provides hook-in points for applications performing per-element, per-vertex, 
and so forth operations on the grid. 
It also provides interfaces for dynamic load balancing, sophisticated 
geometry representations, and other features.
Some properties are enlisted below.
</p>

<p>
Peano is currently available in its third generation.
The development of the original set of Peano codes started around 2002.
2005-2009, we merged these codes into one Peano kernel (2nd generation).
2009, I started a complete reimplementation of the kernel with special emphasis on
reusability, application-independent design and the support for rapid prototyping.
Currently, HPC challenges and extensions dominate the development.  
All data you can download today are stable, i.e. the interfaces will not change.
Updates to the kernel (in particular to speed up simulations) will be hidden from any application or
will be made available due to additional, non-mandatory APIs.
</p>

<p>
Peano is plain C++ code and depends only on MPI and Intel's TBB or OpenMP if you want 
to run it with distributed or shared memory support.
C++ 11 is used.
There are no further dependencies or libraries required.
If you intend to use Peano, we provide a small Java tool to facilitate rapid 
prototyping and to get rid of writing glue code.
This Peano Development Toolkit (PDT) is pure Java and uses <a href="http://www5.in.tum.de/dastgen/">DaStGen</a>.
While we provide the PDT's sources, there's also a jar file available that comprises all required 
Java libraries and runs stand alone. 
</p>



<table  border="1" style="background-color:#dedeee">
  <tr>
    <td> <b>Features</b> </td>
    <td> <b>Parallelisation</b> </td>
    <td> <b>Extensions</b> </td>
    <td> <b>Applications</b> </td>
  </tr>
  <tr>
    <td> Arbitrary dynamic adaptivity </td>
    <td> MPI support (domain decomposition) </td>
    <td> Tailored HDF5 output </td>
    <td> Peano is one of the building <br />blocks of the <a href="http://www5.in.tum.de/forschung/exahype/" target="_blank">ExaHyPE</a> project </td>
  </tr>
  <tr>
    <td> Arbitrary spatial dimensions <br /> or space-time grids</td>
    <td> OpenMP and TBB support </td>
    <td> Multiscale data compression <br &> with reduced floating point <br /> accuracy </td>
    <td> Interface to <a href="http://www.mcs.anl.gov/petsc/" target="_blank">PETSc</a> </td> 
  </tr>
  <tr>
    <td> Inherent geometric multiscale <br /> representation of the computational <br /> domain </td>
    <td> Plug-in points for hybrid <br /> load balancing plus autotuning <br />prototypes </td>
    <td> Support for Particle-in-Cell (PIC) particle <br />
         administration, algebraic-geometric multigrid <br />
         (additive, BPX, multiplicative), and patch-based <br />
         block-structured solvers with automated <br />
         adjacency tracking </td>
    <td> Interface to <a href="http://numerics.kaust.edu.sa/pyclaw/" target="_blank">PyClaw/ClawPack</a> <br /> due to <a href="http://www.peano-framework.org/download.shtml">PeanoClaw</a> 
  </tr>
  <tr>
    <td> Low memory requirements </td>
    <!--
    <td> Resiliency support due to <br /><a href="http://www.exascience.com/cobra/" target="_blank">Cobra</a> </td>
    -->
    <td> Tailored performance analysis outputs and <br />
         extensions for Scalasca/Score-P </td>  
  </tr>
  <tr>
    <td> Good cache and memory bus <br /> behaviour </td> 
  </tr>
  <tr>
    <td> Complex-valued degrees of <br /> freedom </td>
  </tr>
</table>


<h2>Citing Peano</h2>
<p>
If you use Peano, please cite it as 
</p>
<table  border="1"><tr><td>
<pre>
@Misc{Software:Peano,
  author = {Tobias Weinzierl and others},
  title  = {Peano---a {F}ramework for {PDE} {S}olvers on {S}pacetree {G}rids},
  url    = {http://www.peano-framework.org},
  year   = {201x},   % adopt to your paper
  note   = {http://www.peano-framework.org}
}
</pre>
</td></tr></table>