Energy-efficient stencil computations on distributed GPUs using dynamic parallelism and GPU-controlled communication.
Lena OdenBenjamin KlenkHolger FröningPublished in: E2SC@SC (2014)
Keyphrases
- energy efficient
- parallel processing
- sensor networks
- wireless sensor networks
- commodity hardware
- parallel computation
- communication cost
- energy consumption
- graphics processing units
- data dissemination
- graphics hardware
- parallel computing
- data gathering
- information dissemination
- parallel architectures
- base station
- real time
- parallel programming
- computer networks
- general purpose
- energy efficiency
- computational power
- routing protocol
- shared memory
- distributed environment
- sensor nodes
- mobile communication
- relay nodes
- single instruction multiple data
- multi hop
- mobile agents
- parallel implementation
- parallel algorithm
- cost effective
- multithreading