Why is FPGA-GPU Heterogeneity the Best Option for Embedded Deep Neural Networks?
Walther Carballo-HernándezMaxime PelcatFrançois BerryPublished in: CoRR (2021)
Keyphrases
- neural network
- real time
- field programmable gate array
- pattern recognition
- embedded systems
- parallel hardware
- real time image processing
- graphics hardware
- high speed
- fuzzy logic
- parallel computing
- multi layer
- parallel processing
- recurrent neural networks
- genetic algorithm
- gpu accelerated
- artificial neural networks
- hardware design
- graphics processing units
- hardware implementation
- low cost
- back propagation
- parallel implementation
- neural nets
- gpu implementation
- option pricing
- signal processing
- rule extraction
- general purpose
- power consumption
- self organizing maps
- fpga implementation
- hw sw