This paper presents a design space exploration methodology based on higher-order functions to facilitate the tradeoff between execution time and area usage on FPGAs. Higher-order function are transformed, resulting in parameterized nodes where the amount of parallelism and thereby performance, can be controlled. For composition and scheduling of operations, dataflow principles are used. To show the validity of the approach, a particle filter has been transformed and synthesized for FPGA. The resulting architecture is parameterizable and achieves good performance.
|Title of host publication||Reconfigurable Computing: Architectures, Tools, and Applications|
|Place of Publication||London|
|Number of pages||8|
|Publication status||Published - 2014|
|Name||Lecture Notes in Computer Science|
- Higher-order functions
- Particle filter