A method for documenting architectural solutions of computing platforms

The article describes a method for documenting the principles of functioning and internal organization of computing platforms, including reconfigurable computing systems and non-standard processor architectures. The novelty is in using of unified tools to describe: the design process and the computing process, hardware, software and tools, computing components of different granularity. The proposed approach is to describe computing platform as an ideal model that represents abstract algorithms for fulfilling functional requirements without specifying of how to implement it. Then the iterative model refinement follows including selection of physical implementation options, specifying the technological stacks, and additional mechanisms that provide the specified system qualities. A feature of the method is a kernel used for structuring information, classifying and describing the computational mechanisms the system consists of. The kernel includes elements common for different systems and is based on the analysis of a large number of computing architectures. The method describes the principles of the organization of platforms which are usually not considered together. These are: generalized processors with classical architecture which is an evolution of the von Neumann principles; systems based on microcontrollers; operating systems; large- and small-granular reconfigurable systems; specialized processors and accelerators; artificial neural networks. The proposed method can be used to structure information in both traditional and rapidly developing areas: reconfigurable systems and specialized processors. Based on the method, it is possible to create a common database of computing mechanisms suitable for use in different functional units of the system and at different levels of granularity. The results of the work can be useful for system architects to describe complex computing mechanisms consisting of software, hardware and dynamically generated adaptive “intelligent” components which will simplify their reuse and can be used to generate new architectural solutions. Also, the proposed method can be used in the process of training specialists, for a visual demonstration of the basic principles of computer technology.