The ParaNut Processor
The goal of the ParaNut project is to develop an open, scalable and practically usable multi-core processor
architecture for embedded systems. Scalability is provided by a special focus on parallelism at thread level and data level.
A ParaNut consists of an arbitrary number of simple processing cores, each of which is capable of performing 32 bit RISC
The ParaNut introduces a new concept for SIMD (single instruction, multiple data) vectorization. Whereas SIMD
extentsions for workstation processors or embbeded systems frequently contain specialized instructions leading to an
inherently bad compiler support, SIMD code for the ParaNut can be programmed in a high-level language according to
a paradigm similar to thread programming.
The instruction set is kept compatible to the OpenRISC 1000 specification. Hence, the OpenRISC GCC tool chain and
libraries/operation systems (newlib, Linux with some necessary extensions) can be used with the ParaNut architecture