Loading...
Please wait, while we are loading the content...
Similar Documents
Unify: A scalable, loosely-coupled, distributed shared memory multicomputer (1993)
| Content Provider | CiteSeerX |
|---|---|
| Author | Griffioen, James Yavatkar, Rajendra Finkel, Raphael |
| Abstract | Unify is a scalable approach for designing distributed multicomputers. It uses high-speed local and wide area networks as a distributed backplane to interconnect hundreds or thousands of workstations and supercomputers into a single massively parallel multicomputer. From a programmer's point of view, Unify's memory model provides a high-performance shared memory multiprocessor, making it easy to program massively parallel applications. Unify achieves scalability beyond that of existing distributed shared memory approaches by providing scalable synchronization primitives, multiple grades of consistency for shared memory, lazy evaluation for kernel data structures, optimizations for local sharing, and efficient multicast protocols. This paper describes the Unify design and discusses the scalability issues involved in constructing a large scale distributed shared memory multicomputer. This research was supported in part by the National Science Foundation Grant No. NCR-9111323 and Grant ... |
| File Format | |
| Language | English |
| Publisher Date | 1993-01-01 |
| Access Restriction | Open |
| Subject Keyword | Memory Multicomputer Scalability Issue Unify Design Multiple Grade Shared Memory Multicomputer Lazy Evaluation Memory Multiprocessor Wide Area Network Local Sharing Shared Memory Approach Efficient Multicast Protocol Shared Memory National Science Foundation Grant Memory Model Parallel Multicomputer Distributed Multicomputers Parallel Application Scalable Synchronization Primitive Large Scale Scalable Approach Kernel Data Structure Distributed Backplane |
| Content Type | Text |
| Resource Type | Technical Report |
