POSTER: Accelerate GPU Concurrent Kernel Execution by Mitigating Memory Pipeline Stalls.
Hongwen DaiZhen LinChao LiChen ZhaoFei WangNanning ZhengHuiyang ZhouPublished in: PACT (2017)
Keyphrases
- concurrent execution
- memory management
- distributed shared memory
- multi threaded
- multithreading
- feature space
- real time
- kernel methods
- resource consumption
- kernel function
- graphics hardware
- memory usage
- data transfer
- computing power
- parallel computation
- support vector
- data flow
- dynamically created
- parallel computing
- parallel architecture
- memory bandwidth
- memory size
- parallel processing
- operating system
- concurrent processes
- convolution kernel
- main memory
- gpu accelerated
- processing pipeline
- random access
- computational power
- parallel execution
- reproducing kernel hilbert space
- kernel machines
- memory space