Improving Power of DSP and CNN Hardware Accelerators Using Approximate Floating-point Multipliers.
Vasileios LeonTheodora PaparouniEvangelos PetrongonasDimitrios SoudrisKiamal Z. PekmestziPublished in: ACM Trans. Embed. Comput. Syst. (2021)
Keyphrases
- floating point
- fixed point
- cellular neural networks
- sparse matrices
- square root
- floating point arithmetic
- fast fourier transform
- digital signal processing
- instruction set
- interval arithmetic
- power consumption
- signal processing
- convolutional neural network
- floating point unit
- database management systems
- parallel algorithm
- data processing
- low cost
- higher order
- digital signal processor
- probabilistic model
- search space