Parallel applications can be executed using the idle computing capacity of workstation clusters. However, it remains unclear how to schedule the processors among different applications most effectively. Processor scheduling algorithms that were successful for shared-memory machines have proven to be
Parallel FFT on ATM‐based networks of workstations
✍ Scribed by Suresh Chalasani; Parameswaran Ramanathan
- Book ID
- 110380847
- Publisher
- Springer US
- Year
- 1998
- Tongue
- English
- Weight
- 289 KB
- Volume
- 1
- Category
- Article
- ISSN
- 1386-7857
No coin nor oath required. For personal study only.
📜 SIMILAR VOLUMES
The network of workstations (NOW) we consider for scheduling is heterogeneous and nondedicated, where computing power varies among the workstations and local and parallel jobs may interact with each other in execution. An effective NOW scheduling scheme needs sufficient information about system hete
Cluster algorithms have application in diverse areas, including statistical mechanics of polymer solutions, spin models in physics, and the study of ecological systems. Most parallel cluster labeling algorithms are designed for SIMD and MIMD multiprocessors and based on relaxation methods. We presen