Title :
Design and implementation of a NoC supporting priority-based communications for many-core SoCs
Author :
Chang, Kuei-Chung ; Liao, Ming ; Shiu, Bo-Yi
Author_Institution :
Dept. of Inf. Eng. & Comput. Sci., Feng-Chia Univ., Taichung, Taiwan
Abstract :
As technology scaling enables the integration of billions of transistors on a chip, economies of scale are prompting the move toward parallel chip architectures with application-specific systems-on-a-chip (SoC) leveraging multiple cores on a single chip for better performance at manageable design costs. The demand for communicating capability of many-core SoC will definitely increase because the traffic flow between cores, memory, and cache will be massive and complicated. As these parallel chip architectures scale in size, on-chip networks (NoC) have become the main communication architecture, replacing dedicated interconnections and shared buses. These NoCs have been adopted to mitigate wire delay in specific domains such as Nonuniform Cache Architectures (NUCA). The paper presents a NoC design which provides priority-based communications coordinating with wormhole-switching as well as XY routing algorithm. In the proposed NoC, the important packets will be transmitted first with the higher priority to save the waiting time. The experimental results show the proposed priority-based NoC can increase the transmission efficiency in NUCA for many-core SoCs.
Keywords :
cache storage; economies of scale; interconnections; microprocessor chips; network routing; network-on-chip; parallel architectures; NUCA; XY routing algorithm; application specific systems-on-chip leveraging multiple core; communication architecture; economies of scale; manageable design cost; many core SoC; nonuniform cache architecture; on-chip network; parallel chip architecture; priority based communication; shared bus; technology scaling; traffic flow; transmission efficiency; wire delay; wormhole switching; Channel allocation; Computer architecture; Protocols; Resource management; Routing; Switches; System-on-a-chip; many-core systems; nonuniform cache architecture; on-chip interconnection;
Conference_Titel :
Computer Symposium (ICS), 2010 International
Conference_Location :
Tainan
Print_ISBN :
978-1-4244-7639-8
DOI :
10.1109/COMPSYM.2010.5685465