Scalable and custom-precision floating-point hardware convolution core for using in AI edge processors.
Mahdi ShafieiHassan DaryanavardAhmad HatamPublished in: J. Real Time Image Process. (2023)
Keyphrases
- floating point
- instruction set
- floating point arithmetic
- fast fourier transform
- graphics processing units
- commodity hardware
- memory bandwidth
- fixed point
- square root
- high end
- application specific
- artificial intelligence
- mesh connected
- sparse matrices
- parallel processing
- image processing
- parallel architectures
- parallel algorithm
- edge detection
- parallel computing
- floating point unit
- parallel processors
- multithreading
- processing units
- massively parallel
- multi core processors
- low cost
- interval arithmetic
- level parallelism
- parallel programming
- embedded processors
- computer systems