On latency in GPU throughput microarchitectures.
Michael AnderschJan LucasMauricio Alvarez MesaBen H. H. JuurlinkPublished in: ISPASS (2015)
Keyphrases
- memory bandwidth
- low latency
- response time
- heterogeneous computing
- real time
- resource utilization
- level parallelism
- floating point
- processing power
- parallel processing
- prefetching
- data transfer
- parallel programming
- gpu implementation
- graphics processing units
- highly efficient
- graphics hardware
- commodity hardware
- graphics processors
- clock frequency
- high speed
- parallel computing
- high throughput
- gpu accelerated
- virtual machine
- processing units
- memory access
- end to end
- stream processing
- higher throughput