Demo Run

Programming, Compiling and Deploying nBody Code with the POLCA toolbox

The POLCA tool box offers the full support to compile (annotated) C code to different target platforms, ranging from CPUs to FPGAs. With its annotation model, legacy code can be easily extended and readied for POLCA so that existing code becomes more portable and easier to maintain.

Since POLCA generates enriched source code, standard compilers can be used for the final conversion step. At the time of writing this, POCLA can generate OpenMP, OpenCL, MaxJ, FlexaWare and VHDL code. In the following we will demonstrate how POLCA supports development and compilation of a simple n-Body code:

Step 0: Preparing the Demo

Describes how to install the tools to set up a demo that you can run on your own machine.

Step 1: Annotating the Code

In this step we will show you how to use the tools to write and annotate a simple n-body simulation.

Step 2: Assessing Performance, Analysing the Graph

POLCA offers you the right tools to assess the performance of a code without necessitating its compilation and deployment.

Step 3: Transformation Rules

What are "transformations"? And what does POLCA do with the annotated code?

Step 4: Code Adaptation According to the Transformation Rules

This step describes how you can generate code for CPUs, GPUs or FPGAs using the POLCA tools.

Step 5: Measuring Performance and Improving the Results

Testing the resulting code is easy with these POLCA steps.