Performance Portability of Programming Strategies for Nearest-Neighbor Communication with GPU-Aware MPI.
James Buford White IIIPublished in: SC Workshops (2023)
Keyphrases
- nearest neighbor
- parallel implementation
- parallel computing
- high dimensional
- knn
- parallel programming
- asynchronous communication
- graphics processing units
- multithreading
- programming language
- programming environment
- high dimensional data
- nearest neighbor search
- parallel algorithm
- communication channels
- real time
- parallelization strategy
- shared memory
- communication networks
- high performance computing
- k nearest neighbor
- general purpose
- data points
- communication systems
- communication cost
- message passing
- parallel processing
- information sharing
- distance function
- nearest neighbor algorithm
- message passing interface
- query processing