RT-Swap: Addressing GPU Memory Bottlenecks for Real-Time Multi-DNN Inference.
Woo-Sung KangJinkyu LeeYoungmoon LeeSangeun OhKilho LeeHoon Sung ChwaPublished in: RTAS (2024)
Keyphrases
- real time
- graphics hardware
- gpu accelerated
- multi threaded
- gpu implementation
- limited memory
- graph cuts
- control system
- high speed
- vision system
- graphics processing units
- neural network
- memory usage
- parallel processing
- main memory
- data acquisition
- low cost
- bayesian networks
- real time systems
- memory space
- inference process
- quality of service
- programmable graphics hardware