H3C UniServer R5300 G6, a next-generation 4U dual-processor AI server, is designed to deliver significant upgrades in processor and GPU specifications compared to the previous generation. With outstanding performance, flexible architecture, strong expansion capabilities, and high reliability, this server supports diverse workloads such as deep learning model training, inference, high performance computing, and data analytics. Its integrated computing and storage design simplifies management and deployment, making it a versatile computing engine for the AI era.
H3C UniServer R5300 G6 supports various types of AI accelerator cards to meet diverse heterogeneous computing demands in AI scenarios. The system supports processor-to-GPU ratios of 1:4 and 1:8, along with various topology configurations such as four-card or eight-card direct connect. Users can flexibly switch configurations based on application models with one-click operations. The system also accommodates diverse AI acceleration cards and smart network adapters to meet different AI scenario requirements.
H3C UniServer R5300 G6 supports two Intel® Xeon® 4th/5th Gen Scalable processors, with up to 64 cores per processor. With the eight-channel 5600MT/s DDR5 memory technology, the system can deliver up to 12TB memory expansion and a 50% bandwidth increase. The I/O supports PCIe 5.0 bandwidth, doubling data bandwidth compared to the previous generation. The R5300 G6 supports up to 24 2.5/3.5-inch drives, including eight NVMe drives, meeting the high memory space demands of AI for massive data processing. The R5300 G6 supports up to 20 PCIe slots, including one dedicated storage controller slot and one OCP 3.0 dedicated network adapter slot. It also accommodates 400GE network adapters to meet the high-speed bandwidth demands of AI applications.
The following contents are complex, and it is recommended to browse on PC.

Enter c.h3c.com.cn on the PC browser and operate according to the page to synchronize to the PC and continue browsing.
Continue by mobile