Position Paper: Validity of the Static Dataflow Approach for Exascale Computing Challenges
We present CHIARA, the programming language used to program CODACS (a general purpose dataflow architecture exploiting FPGA technology), and describe the compiling strategies leading from CHIARA programs to the CODACS dataflow graphs. CHIARA is a functional programming language based on Backus'FP. We designed both the language and the compiling tools in such a way that CHIARA programs can be efficiently compiled and run onto the overall architecture and CODACS platform-processors being, for the latter, also the low-level (assembly) programming language. Some preliminary experimental results are discussed, demonstrating that the CHIARA approach to CODACS programming is feasible and promising.