A network multi-processor for experiments in parallelism
Concurrency: Practice and Experience
Network-based concurrent computing on the PVM system
Concurrency: Practice and Experience
The Linda alternative to message-passing systems
Parallel Computing - Special issue: message passing interfaces
The design of a standard message passing interface for distributed memory concurrent computers
Parallel Computing - Special issue: message passing interfaces
The object-oriented components of the Enterprise parallel programming environment
TOOLS '93 Proceedings of the eleventh international conference on Technology of object-oriented languages and systems
The Enterprise Model for Developing Distributed Applications
IEEE Parallel & Distributed Technology: Systems & Technology
Evaluation of distributed communication systems
CASCON '93 Proceedings of the 1993 conference of the Centre for Advanced Studies on Collaborative research: distributed computing - Volume 2
| Hi-index | 0.00 |
Distributed parallel (DP) programming is a cost effective way of obtaining supercomputer performance. One inherent problem with DP programming is the connection of heterogeneous processor types and file systems. It is becoming increasingly important that the environment transparently handle all communication and I/O connections.This paper looks at some of the obstacles and possible solutions to providing transparent heterogeneity within Enterprise -- an environment for developing distributed memory parallel programs on a network of workstations. Enterprise, using templates and a precompiler, constructs a software layer around the user's application. The user is shielded from the tiresome low-level details of hand crafting the distributed communication portion of the application. In fact, Enterprise programs are transparent to the communication package.One of the tenets of Enterprise is that the user develops programs using only the familiar sequential C programming language. There are no extensions to the language, nor are there any special library calls. However, issues like message and file transparencies are problematic, since they involve user knowledge and dynamic run-time information to create an efficient program. This paper examines the issues of transparently encapsulating and preserving the sequential semantics of the user's code running on a heterogeneous network of workstations and file systems.