GPU Computing Pipeline Inefficiencies and Optimization Opportunities in Heterogeneous CPU-GPU Processors.
Joel HestnessStephen W. KecklerDavid A. WoodPublished in: IISWC (2015)
Keyphrases
- graphics processing units
- heterogeneous computing
- gpu implementation
- parallel computing
- graphics processors
- parallel processing
- general purpose
- parallel computation
- multithreading
- real time
- parallel algorithm
- graphics hardware
- memory bandwidth
- parallel implementation
- parallel architecture
- parallel programming
- global optimization
- compute unified device architecture
- gpu accelerated
- computing systems
- highly parallel
- optimization problems
- optimization algorithm
- high end
- parallel architectures
- processing units
- data transfer
- grid computing
- parallel computers
- optimization process
- high performance computing
- real time rendering
- compute intensive
- massively parallel
- single instruction multiple data
- constrained optimization