Many widely-used aerospace data visualization tools are interactive in nature and are programmed to run on a single processor. While such tools support real-time manipulation of simulation environments, the computations that generate the data are often batch oriented and computation intensive. In many cases, the data generation software is too tuned to a single-processor infrastructure to be readily adapted for emerging parallel and grid computing environments. This paper presents several lessons learned from adapting an aerospace engineering tool to the parallel and grid computing architecture. The architecture provides the ability to perform high-power computing by distributing process execution across many computers connected by a dedicated network or the internet. Some of the challenging tasks involved in the adaptation are (1) to decouple the user interface and display functions from the computational functions, since interaction and graphics are usually unnecessary expenses in parallel and grid computing, (2) to identify and parallelize computationally expensive functions without the drastic modification of the code and data structures, (3) to find a lightweight, yet versatile software solution for a client-server machine interface for remote job execution. The solutions we found for these elaborate tasks are presented and their pros and cons are discussed.
Lessons Learned from Adapting Aerospace Engineering Tools to the Parallel and Grid Computing Environment
2007-03-01
344582 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
NATURAL ENVIRONMENT AND AEROSPACE VEHICLES: SOME LESSONS LEARNED
British Library Conference Proceedings | 2010
|Aerospace Vehicle Wiring, Lessons Learned
SAE Technical Papers | 2020
Aerospace Hydraulics and Actuation Lessons Learned
SAE Technical Papers | 2017
Aerospace Hydraulics and Actuation Lessons Learned
SAE Technical Papers | 2011