At the Moore Thread 2025 MUSA Developer Conference held today, Moore Thread founder, chairman and CEO Zhang Jianzhong announced that a series of chips based on Moore Thread's latest generation GPU architecture "Huagang" - Huashan and Lushan, will be mass-produced and launched next year.
According to Zhang Jianzhong, "Huagang" will use a new generation of instruction sets to support asynchronous programming models and efficient threads; at the same time, the computing power density will be increased by 50% and the energy efficiency will be increased by 10 times. In addition, "Huagang" also supports intelligent computing clusters with a scale of more than 100,000 cards. In order to enhance the utilization of computing power, the chip also invented a new generation of asynchronous programming model.
The "Lushan" high-performance graphics rendering chip based on the Huagang architecture will achieve a 15-fold improvement in 3A game rendering, a 64-fold improvement in AI performance, and a 50-fold improvement in ray tracing performance. In addition to supporting game experience, it also supports all CAD, CAE and other graphic design renderings.
In addition, the GPU chip "Huashan" based on this architecture has achieved many leading and even surpassing the capabilities of international mainstream chips in terms of floating point computing power, memory access bandwidth, memory access capacity and high-speed interconnect bandwidth.
