PROJECT DESIGN FOR COMPUTER ARCHITECTURE PRACTICAL SESSIONS BASED ON FIELD-PROGRAMMABLE GATE ARRAY

Field-programmable logic arrays are often used in courses on computer architecture. The student must describe the processor with the external components necessary for its operation in the specified HDL (hardware description language) language according to the provided specification during a certain number of projects. The weakness of this approach is that the basis of such projects is a processor of one specific architecture, so the lecturer faces the issue of individualization of projects. This article proposes a solution based on dedicated processors instead of one programmable processor of a specific architecture. It’s shown here that the issue of project individualization is easier solvable in the proposed way, and it does not deviate from the theory of computer architecture, because the programmable processor is a generalization of a dedicated processor. The article describes project design ideas based on dedicated processors and gives some examples. Represented different instance than was applied during practical sessions of Computer Architecture that are held at the Department of Electronic Systems within VILNIUS TECH, i.e. certain modifications, and additions were applied.

Implementation of asymmetric processing on multi core processors to implement IOT applications on GNU/Linux framework

Internet of Things brought in a bigger computing challenges where there came a need for running tasks in a multi-sensor and large data processing is involved. In order to implement this requirement multiprocessors are being used for implementation of IoT Gateways. There comes a need for specific tasks having a resource dedicated for its job. To fulfill this we face a hurdle in choosing dedicated processor or shared processor in a Symmetric Processing Architecture. Dedicated processor are the one in which all the tasks are being processed on a single core where as in fair share processors specific processes are assigned to specific cores. Symmetric processing makes use of dedicated processors where as Asymmetric processor makes use of shared processors. Asymmetric Multi Processing can be used in real time applications in order to solve real time problems, one such platform is IOT. In this paper we have evaluated Asymmetric processing on GNU/Linux Platform to test multiple threads running on different multi-core processors architectures to realize the same for running IOT applications having higher computational requirements in the future.