Cambricon
About
About Cambricon
Cambricon’s MLU370 accelerator cards target server and data-centre workloads, while MLU220 modules bring inference into compact edge systems. The hardware is supported by the NeuWare software environment and MagicMind optimisation tools, which help developers deploy trained models across Cambricon processors.
In a robot or autonomous machine, this stack can sit between sensors and application software, accelerating vision, recognition and other neural-network workloads locally. Cambricon supplies the processing architecture rather than navigation, motion control or a complete robot, leaving system integrators to connect its compute with cameras, actuators and operational software.
Financials
- 688256.SS - SSE - CNY
Timeline
Events
- 2019-11-14 - Product Launch - Cambricon releases Siyuan 220 edge AI chip
Cambricon released its Siyuan 220 edge AI chip and M.2 accelerator card at the 21st China Hi-Tech Fair. The release extended Cambricon's portfolio from cloud AI accelerators into edge computing hardware. The chip used the MLUv02 architecture and targeted low-power AI inference at the edge.
Source