Co-Investigator(Kenkyū-buntansha) |
ABE Masato Iwate University, Professor, 工学部情報工学科, 教授 (00159443)
ASO Hirotomo Tohoku University, Professor, 大学院工学研究科, 教授 (10005522)
VIJAY Jain 南フロリダ大学, 電子工学科, 教授
FRANK Hsu フォーダム大学, 計算科学科, 教授
FANRIZIO Lom ノースイースタン大学, 計算科学科, 教授
JAIN Vijay University of South Florida, Professor
LOMBARDI Fanrizio Northeastern University, Professor
HSU Frank Fordham University, Professor
|
Research Abstract |
With the continuing advances in VLSI technology, multiprocessor systems with tens to hundreds of processing elements are expected to rival today's supercomputers in the next decade. Systems consisting of a few tens to a million processing elements have indeed been described in the literatures. The development of such large scale systems will make numerous applications such as multi-media, computer vision, and modeling of physical phenomena to be feasible on desk-top workstations. However, a major issue in designing large scale multiprocessor systems is the construction of a flexible interconnection network to provide efficient inter-processor communication and fault tolerance techniques to achieve both higher reliability and higher production yield. This research report explores new interconnection networks for massively parallel computers and distributed systems. First, the research report a new interconnection network, "Tori connected meshes (TESH)", consisting of k-ary n-cube connect
… More
ion of supemodes that consist of meshes of lower level nodes. Its key features are the following : it is hierarchical, thus allowing exploitation of computation locality as well as easy expansion (up to a million processors). Second, the performance of TESH network are evaluated and TESH network appears to be well suited for 3-D VLSI implementation, for it requires far fewer number of vertical wires than almost all known multi-computer networks. To solve a restriction of vertical links increases the diameter of TESH networks and makes the network feature worse for more than several thousand processing elements. Thus, we propose new types of three dimensional hierarchical interconnection networks : 3D Hierarchical Torus network, called 3DH-torus and 3D Hierarchical Mesh network, called 3DH-mesh. These 3D hierarchical interconnection networks decrease the number of vertical links in 3D stacked implementation by keeping good network feature. Finally, advanced applications of the network to massively parallel computing are presented. Specifically, we discuss the mapping on to TESH network of stack filtering, a hardware oriented technique for order statistic image filtering. The mapping of advanced anolications : sortinn and EFT to the 3D Hierarchical networks. Less
|