Moore Threads Unveils Huagang GPU Architecture at MUSA Conference
Chinese chipmaker Moore Threads has officially launched its next-generation GPU architecture, named Huagang, during its inaugural MUSA Developer Conference. This strategic move aims to bolster the company's standing in the artificial intelligence and high-performance computing sectors. The new Huagang architecture features support for full-precision computing ranging from FP4 to FP64. According to the company, it delivers a 50% increase in compute density while significantly improving energy efficiency by tenfold. Leveraging Moore Threads' proprietary MTLink high-speed interconnect technology, the architecture is designed to scale effectively, supporting intelligent computing clusters of up to 100,000 GPUs. Alongside the architectural announcement, the company introduced two new chips based on this technology: Huashan, an integrated processor optimized for AI training and inference tasks, and Lushan, a chip dedicated to graphics rendering. This launch highlights Moore Threads' continued efforts to innovate in the semiconductor industry and compete in the growing market for advanced AI hardware solutions.
Wire timeline
Moore Threads Unveils Huagang GPU Architecture at MUSA Conference
Chinese chipmaker Moore Threads has officially launched its next-generation GPU architecture, named Huagang, during its inaugural MUSA Developer Conference. This strategic move aims to bolster the company's standing in the artificial intelligence and high-performance computing sectors. The new Huagang architecture features support for full-precision computing ranging from FP4 to FP64. According to the company, it delivers a 50% increase in compute density while significantly improving energy efficiency by tenfold. Leveraging Moore Threads' proprietary MTLink high-speed interconnect technology, the architecture is designed to scale effectively, supporting intelligent computing clusters of up to 100,000 GPUs. Alongside the architectural announcement, the company introduced two new chips based on this technology: Huashan, an integrated processor optimized for AI training and inference tasks, and Lushan, a chip dedicated to graphics rendering. This launch highlights Moore Threads' continued efforts to innovate in the semiconductor industry and compete in the growing market for advanced AI hardware solutions.
TechNode