On November 4th, the AMD 2023 Annual Academic Meeting for Educational and Research Experts, along with a seminar on advancing new productivity in scientific computing, was successfully held in Beijing. Gooxi, as an ecosystem partner of AMD, was honored to participate. The conference revolved around the meteoric rise of ChatGPT by the end of 2022 and explored cutting-edge topics such as the integration of AI with traditional scientific computing and whether CPUs and GPUs can collaborate effectively. Gooxi joined hands with experts and academic leaders from across the country in the field of education and research to discuss how to promote the practical implementation of "new productivity" in the education and research sector.

On the path of digital intelligence, Gooxi has unwaveringly partnered with AMD. In November of last year, when AMD unveiled its fourth-generation EPYC processors, Gooxi simultaneously introduced a series of servers based on these new processors. This series caters to various computational needs, covering AI, edge computing, and high-performance computing. It addresses a range of computational requirements, including general computing, small-scale inference scenarios, and large-scale training scenarios. Notably, the Gooxi AMD Genoa dual-processor high-end server is designed around the next-generation AMD EPYC 9004 series high-end processors. With an upgrade from DDR4 to DDR5, offering a 50% increase in speed and incorporating PCIe 5.0 technology, it meets the demands of high computational performance, high network bandwidth, and large memory capacity. It effectively mitigates high concurrency issues and delivers faster response times. The modular design supports various hard drive configurations, making it suitable for high-performance computing, cloud computing, big data processing, virtualization, and more.

With the advent of ChatGPT, the relationship between AI and computational power has become even more intertwined. In response to the evolving demands of the AI industry, Gooxi has made a comprehensive and profound effort to establish a high-quality computational capacity that combines open and diverse services with server solutions. This initiative helps alleviate the bottlenecks in AI computational power implementation. Gooxi AMD Milan dual-socket 4U 10-GPU server, compacted within a 4U form factor, accommodates two high-performance AMD general-purpose processors and ten double-width, full-height, and full-length GPU cards (compatible with 10x RTX 3090, 8x RTX 4090, 8x A800). It utilizes high-speed interconnection via a switch to meet the extreme performance demands of AI training and model computation, with high computational power and interconnectivity. Additionally, it offers 20 PCIe expansion slots, facilitating effective performance expansion. This solution will enable customers to swiftly deploy high-performance AI computational systems, expediting innovation and development in generative AI industries.

The wise adapt to the times, and the knowledgeable adapt to circumstances. As a peer in AMD's digital transformation journey, Gooxi will continue to maintain close communication and cooperation with AMD. We will collaboratively develop a range of digital intelligence innovations, striving to advance the development of new productivity and accelerate the pace of digital transformation and upgrade across various industries.