Group Members

	Marek Tudruj, Ph.D., Associate Professor, Head of the Group
	Eryk Laskowski, Ph.D.
	Łukasz Masko, Ph.D.

	Program execution control in systems based on dynamic processor clusters with shared memory.
	MNiSW grant No 3 T11C 025 29
	Coordinator: Marek Tudruj
	Methods and support tools for fine-grained parallelism in computer systems.
	KBN Project No. 4 T11C 007 22
	Coordinator: Marek Tudruj
	Structuring parallel programs for computer systems with redundant communication resources.
	KBN grant No 7 T11C 015 20
	Coordinator: Eryk Laskowski

Research Domain
Currently, the Group is conducting research on new data communication solutions for parallel systems based on multicore processors and on modern implementation technologies, such as "Networks on Chip" and "Systems on Chip". The research concerns methodology for dynamic reconfiguration of parallel systems structure for efficient communication by message passing and shared memory enabling operation time transparency features. The results are a reduction of the program execution time and better use of system resources. The mechanisms under research enable adjusting hardware structures to the executed application programs, which corresponds to the functionality of parallel embedded systems.

The Group's research also concerns optimization algorithms for programs meant for efficient execution in parallel and distributed systems. This includes task scheduling algorithms in programs for systems with novel data communication architectures, such as those needed for systems of globally interconnected multicore processor modules. The research is conducted using simulation-based verification of hypotheses on system architecture for representative program classes. For simulation purposes, application programs and system behaviour are modelled with the use of graph representations. Special attention is paid to program and system graph representation methods which provide short simulation time.