WHAT IT IS
Ascend is the AI processor and computing architecture Huawei has offered since 2018, based on the Da Vinci architecture with the CANN software stack. Its chips, such as the Ascend 910 and 310 series, cover training and inference. It addresses providing a self-developed AI chip and software stack beyond the established compute landscape.
Why it matters
It is one of China’s main self-developed AI compute routes, with its CANN stack occupying roughly the place CUDA holds in the ecosystem.
Key specs
- Architecture
- Da Vinci
- Software stack
- CANN
- Representative chips
- Ascend 910 / 310 series
Related concepts
Training & Inference Infrastructure
Memory decides how large a model you can train, communication how long it takes — raw compute is rarely the bottleneck
Inference Optimization & Serving
Training happens once; inference happens a billion times a day — and serving is torn between fast first tokens and high throughput, which usually pull against each other