On March 14th, domestic GPU leader Jingjia Micro announced the successful development of the Jinghong series high-performance AI computing modules and integrated products, which will be soon introduced to the market.
The Jinghong series is designed for applications such as AI training, AI inference, and scientific computing. It supports mixed-precision operations including INT8, FP16, FP32, and FP64, and employs innovative multi-card interconnection technology for computing expansion.
Compatible with mainstream CPUs, operating systems, and server manufacturers both domestically and internationally, the series can support current mainstream computing ecosystems, deep learning frameworks, and algorithm libraries, significantly reducing user adaptation and validation cycles.
With the further development of AI technology, there will be increasing demand for general artificial intelligence and its applications, driving the need for greater computing power.
Currently, China's domestic intelligent computing infrastructure is entering an accelerated construction phase, with major demand coming from sectors such as the Internet, government, and telecommunications operators. Both overseas and domestic AI chips are in high demand.
Previously, domestic chip penetration rates were low. However, with Internet companies increasing their procurement of domestic computing power, led by Huawei's Ascend series, domestic manufacturers will have significant room for development in the future.
