EET Taiwan Article: Choose AI Hardware Acceleration Scheme For Different Applications

From cloud-based big data processing to edge-side keyword recognition and image analysis, the explosive growth of artificial intelligence (AI) applications has prompted experts to develop optimal architectures to accelerate machine learning (ML) algorithms. A variety of emerging solutions highlight the importance of designers clearly defining applications and their needs before choosing a hardware platform.
To help reduce the power consumption of this edge AI image processing, GreenWaves Technologies‘ GAP8 processor integrates nine RISC-V cores; one core is responsible for hardware and I/O control functions, and the remaining eight cores are around shared data and the instruction memory forms a cluster. This structure forms the CNN inference engine accelerator with additional RISC-V ISA instructions to enhance DSP-type operations.
 
To read more, please visit: https://www.eettaiwan.com/news/article/20181012NT51-AI-hardware-acceleration-needs-careful-requirements-planning. Please note the article is in Chinese.