Neural acceleration for GPU throughput processors.
Amir YazdanbakhshJongse ParkHardik SharmaPejman Lotfi-KamranHadi EsmaeilzadehPublished in: MICRO (2015)
Keyphrases
- parallel processing
- parallel computation
- parallel computing
- memory bandwidth
- graphics processing units
- parallel algorithm
- network architecture
- parallel programming
- response time
- neural network
- parallel architectures
- real time
- multithreading
- biologically plausible
- parallel implementation
- single instruction multiple data
- neural model
- heterogeneous computing
- bio inspired
- graphics hardware
- floating point
- associative memory
- high end
- parallel processors
- biologically inspired
- highly parallel
- high performance computing
- resource utilization
- level parallelism
- neural fuzzy
- instruction set
- massively parallel
- processor array
- higher throughput
- message passing interface
- graphics processors
- gpu implementation
- buffer size
- spike trains
- single processor
- processing units
- general purpose
- end to end