Extended Cyclostatic Dataflow Program Compilation and Execution for an Integrated Manycore Processor

Extended Cyclostatic Dataflow Program Compilation and Execution for an Integrated Manycore Processor

The ever-growing number of cores in embedded chips emphasizes more than ever the complexity inherent to parallel programming. To solve these programmability issues, there is a renewed interest in the dataflow paradigm. In this context, we present a compilation toolchain for the ΣC language, which allows the hierarchical construction of stream applications and automatic mapping of this applicati...

A Dynamically Reconfigurable Processor for Dataflow Graph Execution

IFPEC: Integrated Framework for Processor Extension Generation and Application Compilation

This paper presents our integrated design framework, IFPEC, used for automatic selection and synthesis of processor’s extensions as well as efficient application compilation for these newly generated extensions. Our approach is based on constraint programming (CP) and is able to solve hard design problems. Processor extensions, in our approach, are defined as computational patterns and represen...

A Curry-Howard Isomorphism for Compilation and Program Execution

This paper establishes a Curry-Howard isomorphism for compilation and program execution by showing the following facts. (1) The set of A-normal forms, which is often used as an intermediate language for compilation, corresponds to a subsystem of Kleene’s contraction-free variant of Gentzen’s intuitionistic sequent calculus. (2) Compiling the lambda terms to the set of A-normal forms corresponds...

Toward a Core Design to Distribute an Execution on a Manycore Processor

This paper presents a parallel execution model and a manycore processor design to run C programs in parallel. The model automatically builds parallel sections of machine instructions from the run trace. It parallelizes instructions fetches, renamings, executions and retirements. Predictor based fetch is replaced by a fetch-decode-and-partlyexecute stage able to compute in-order most of the cont...