Pod-racing: bulk-bitwise to floating-point compute in racetrack memory for machine learning at the edge.
Sébastien OllivierXinyi ZhangYue TangChayanika ChoudhuriJingtong HuAlex K. JonesPublished in: IEEE Micro (2022)
Keyphrases
- floating point
- machine learning
- fixed point
- memory bandwidth
- edge detection
- square root
- sparse matrices
- instruction set
- artificial intelligence
- floating point arithmetic
- interval arithmetic
- memory requirements
- random access
- main memory
- image processing
- hamming distance
- memory space
- higher order
- general purpose
- computer vision