Loading...
Please wait, while we are loading the content...
Similar Documents
Efficient Implementation of High-Level Parallel Programs (1991)
| Content Provider | CiteSeerX |
|---|---|
| Author | Bagrodia, Rajive Mathur, Sharad Hall, F. Boelter |
| Description | The efficiency of a parallel program is related to the implementation of its data structures on the distributed (or shared) memory of a specific architecture. This paper describes a declarative approach that may be used to modify the mapping of the program data on a specific architecture. The ideas are developed in the context of a new language called UC and its implementation on the Connection Machine. The paper also contains measurements on sample programs to illustrate the effectiveness of data mappings in improving the execution efficiency of example programs. 1 Introduction In designing parallel languages, software engineering concerns dictate that the language abstract out architectural details and present a simple virtual machine to the programmer. However, experience with programming parallel architectures has indicated that optimal performance can be extracted only by exploiting the communication topology of the architecture such that communication and synchronization costs ... In Fourth International Conference on Architectural Support for Programming Languages and Operating Systems |
| File Format | |
| Language | English |
| Publisher Date | 1991-01-01 |
| Access Restriction | Open |
| Subject Keyword | Simple Virtual Machine Optimal Performance Declarative Approach Efficient Implementation Synchronization Cost Data Structure Specific Architecture Data Mapping Execution Efficiency Parallel Language Example Program Connection Machine Architectural Detail High-level Parallel Program Parallel Architecture Software Engineering Concern New Language Sample Program Program Data Language Abstract Communication Topology Parallel Program |
| Content Type | Text |
| Resource Type | Article |