Efficient Processor Allocation Strategies for Mesh Multicomputers : Performance Evaluation of Processor Allocation Strategies for Mesh Interconnection Networks

ISBN-13: 9783639171525 / Angielski / Miękka / 2009 / 208 str.

Saad Bani Mohammad

Opis

We have identified the limitations of the existing §work in the area of processor allocation in mesh-§connected multicomputers. These limitations form the §motives behind undertaking such a research. We can §summarise the main contributions of this research §into three main points. The first part of this §research presents an efficient contiguous processor §allocation algorithm, referred to as Turning Busy §List (or TBL for short), for 3D mesh-connected §multicomputers that overcome the limitations of the §existing strategies suggested previously for the 3D §mesh networks. The second part of this research §proposes a new non-contiguous processor allocation §algorithm, referred to as Greedy Available Busy List §(or GABL for short), for the 2D mesh-connected §multicomputer. And finally, we study the impact of §heavy-tailed job execution times on the performance §of the existing contiguous allocation strategies for §3D mesh-connected multicomputers, including the ones §proposed in this research.

O autorze

Saad Bani Mohammad received the BSc degree in computer science
from Yarmouk University, Jordan in 1994, the MSc degree in
computer science from Al al-Bayt university, Jordan in 2002, and
the PhD degree in computer science from University of Glasgow,
U.K., in 2008.