Cascading Divide-and-Conquer (Classic Reprint): A Technique for Designing Parallel Algorithms - Couverture souple
Synopsis
Unlock parallel speed with proven techniques for designing fast algorithms. This book offers a practical framework for solving complex problems in parallel, using a structured approach that scales with many processors.
This work presents general methods for divide-and-conquer that apply beyond simple merges. It shows how to model problems as trees, manage ordered data, and keep operations efficient even when combining subresults. Readers will see how to apply these ideas to geometric problems and data structures, always with an emphasis on time and processor efficiency.
- General cascading merges that run in optimal time using many processors
- Parallel fractional cascading and its adaptation to new data structures
- Techniques for planar geometry problems like trapezoidal decomposition and point location
- Applications to 3D maxima, dominance counting, and visibility problems
Ideal for readers who want concrete methods to design parallel algorithms and understand how to achieve near-optimal performance in practice.
Les informations fournies dans la section « Synopsis » peuvent faire référence à une autre édition de ce titre.
