Culler, D. NOW: Towards Everyday Supercomputing on a Network of Workstations. EECS Tech. Rep. UC Berkeley, to app.  Home/Search   Document Not in Database   Summary   Related Articles   Check

.... over a network (network multicomputers) or of share nothing processors with a local storage linked through a high speed network or bus (switched multicomputers) T95] It is well known that multicomputers offer best price performance ratio [T95] M96] Research on multicomputers becomes popular [C94], G96] The Scalable Distributed Data Structures (SDDSs) like LH [LNS93] are new data structures designed for multicomputer files. An SDDS gracefully scales up with inserts over available distributed storage, the distributed RAM storage preferably. One problem that a designer of an SDDS may ....

Culler, D. NOW: Towards Everyday Supercomputing on a Network of Workstations. EECS Tech. Rep. UC Berkeley.

....thousands of nodes, whose distributed RAM reaches dozens of GBytes, and disks reach TBytes. Multicomputers offer best price performance ratio and are potentially capable of computational performance superior to that of supercomputers [T95] M96] Research on multicomputers is increasingly popular [C94], G96] Multicomputers need new data structures where scalability is vital. The Scalable Distributed Data Structures (SDDSs) LNS93] are intended specifically for multicomputers; other avenues to scalability, especially for parallel supercomputers, are also being explored, SPW90] JK93] An ....

Culler, D. NOW: Towards Everyday Supercomputing on a Network of Workstations. EECS Tech. Rep. UC Berkeley.

.... (network multicomputers) or of share nothing processors with a local storage linked through a high speed network or bus (switched multicomputers) T95] Multicomputers offer best price performance ratio and are potentially capable of computational performance superior to that of supercomputers [C94], G96] They require new data structures where scalability is vital. The Scalable Distributed Data Structures (SDDSs) LNS93] are specifically for multicomputers. An SDDS scales up through splits of its buckets stored over the servers of a distributed storage, in particular RAM storage 3 . ....

Culler, D. NOW: Towards Everyday Supercomputing on a Network of Workstations. EECS Tech. Rep. UC Berkeley.

....of the overhead incurred in distributed processing under the SDDS with live optimization. We want to show the following: ffl Optimal Size of a Distributed Partition ffl Overhead Added by SDDSs 5. Implementation and Performance Study 8 ffl Performance Scalability We use a network multicomputer [Cul94] Tan95] in our case a number of Silicon Graphics O 2 s, running IRIX6.3, each having 64 MBytes of memory. For communication the office network is used. This network is a mix of ATM switches and Ethernet. Each workstation has a 180 MHz, R5000 MIPS CPU. All measures are given in milli seconds ....

D. Culler. NOW: Towards Everyday Supercomputing on a Network of Workstations. Technical report, EECS Technical Reports UC Berkeley, 1994.

....LH lh opens a new perspective for high performance applications, especially for the database management of new types of data and in real time environments. 1 Introduction New applications of databases require increased performance. One way is to use parallel and distributed architectures [17][2]. The multicomputers, i.e. networks of multiple CPUs with local storage become a popular hardware platform for this purpose[17] 2] 21] In particular, multicomputer files need to be able to scale to large sizes over the distributed storage, especially the RAM. The Scalable Distributed Data ....

....in real time environments. 1 Introduction New applications of databases require increased performance. One way is to use parallel and distributed architectures [17] 2] The multicomputers, i.e. networks of multiple CPUs with local storage become a popular hardware platform for this purpose[17][2][21] In particular, multicomputer files need to be able to scale to large sizes over the distributed storage, especially the RAM. The Scalable Distributed Data Structures (SDDSs) 15] is an approach towards this goal. An SDDS file can gracefully expand with the inserts from a single storage site ....

D. Culler. NOW: Towards Everyday Supercomputing on a Network of Workstations. Technical report, EECS Tech. Rep. UC Berkeley, 1994.

....We investigate several different client updating schemes, enabling better access load distribution for many slow clients. Keywords Scalable Distributed Data Structure, Spatial Point Index, Ordered Files, Multicomputers 1 Introduction Research is increasingly focusing on using multicomputers [1] [20] 24] Multicomputers are built from mass produced PCs and Workstations, often having special high bandwidth networks. Such inThe 5th International Conference on Foundations of Data Organization (FODO 98) Kobe, Japan, November 1998. frastructures have emerged and many organizations have ....

D. Culler. NOW: Towards Everyday Supercomputing on a Network of Workstations. Technical report, EECS Technical Reports UC Berkeley, 1994.

No context found.

Culler, D. NOW: Towards Everyday Supercomputing on a Network of Workstations. EECS Tech. Rep. UC Berkeley, to app.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC