Home > Server > G7

Redefining the virtualization cost vs performance ratio, the latest AMD EPYCTM 9004 Series Processors provide powerful computing capabilities that support the scenarios such as virtualization, high-performance computing and big data, etc.

Features
  • Extreme Performance Based on AMD EPYCTM 9004 Series Processors using the latest 5nm manufacturing process, it supports up to 96 cores and maximum of 384MB L3 cache. It has a maximum frequency of 4.4GHZ. Upon the request of high volume of cloud-based cores and high performance per core, the performance increased by 86%* can meet such demands. * Comparison with the previous generation
  • Breakthrough Innovation Introduced new components such as E1.S and E3.S, which are shorter than U.2, allowing for a smaller size, larger storage density, higher heat dissipation efficiency, over 10 million IOPS, reduction in the number of user clusters, reduction in system failures, and improvements in reliability. 2x OCP expansion slots provides users with better network card scalability and resource allocation. In limited space, continuously improve the GPU computing density by 100% for 2U2S server.
  • Reliability The system can operate stably without failure by means of PSU, BIOS/BMC and other core component redundancy. Based on a firmware encryption/digital signature, with the cooperation of security chassis detection and TPM trusted computing module, provides an end-to-end security solution from the underlying hardware to upper applications.
  • Availability Modular design with tool-free disassembly and maintenance. Abundant platform management O&M suite, including infrastructure management platform ISPIM, automatic deployment tool ISIB, automatic O&M tool ISA, single product management software ISQP, and management driver ISMD. Full lifecycle management, asset management, intelligent fault management, large-scale unified batch management, intelligent power consumption management and intelligent automatic configuration of the server is offered, dramatically reducing operation and maintenance costs.
NF5280G7

Flexible Expansion Compute and Storage Server

NF5280-A7-A0-R0-00

- 2x AMD EPYC™ 9004 Series Processors 
- Flexible PCIe expansions with OCP3.0 enabling both on-premise and cloud deployment
- Up to 192 cores and 384MB L3 cache
- Up to 400TB storage and maximum of 4x dual-width GPUs
- 24 x E3.S SSDs deliver high flash storage density and power efficiency
- Tool-free maintenance with modular server design
- Monitoring system via BMC ensures stable operation
NF5180G7

High-Density Compute Server

NF5180-A7-A0-R0-00

- 2x AMD EPYC™ 9004 Series Processors
- Up to 96 cores, TDP up to 400W, 384MB L3 cache
- Memory frequency up to 4800 MHz greatly improves operation efficiency
- 2x E1.S front SSDs deliver greater storage capacity and easier maintenance
- Supports complex workloads like virtualization and high-performance computing
NF5688G7

Industry-Proven AI Training Server

NF5688-A7-A0-R0-00
- 2x AMD EPYC™ 9004 Series Processors and 8x 700W GPUs in 6U
- Flexible PCIe expansions with CX7, OCP3.0, and multiple SmartNICs enabling both on-premise and cloud deployment
- Supports demanding AI computing tasks like trillion-parameter Transformer model training, massive recommender systems, and AI + Graphics workloads

NF5468G7

Versatile AI Server

NF5468-A7-A0-R0-00
- 2x AMD EPYC™ 9004 Series Processors & up to 10 FHFL Dual-slot high-performance GPU and ASIC cards in 4U
- Flexible I/O design and self-developed trouble shooting module offer easy maintenance and versatile build options
- Support specialized AI tasks including training, inference, AI4Science, graphics and video codec processing, making it the ideal choice for both datacenter and enterprise applications

i24G7

High-Density Optimized Multi-Node Server

i24-A7-A0-R0-00
- 2U 4-node, 1x AMD EPYC™ 9004 Series processor per node
- High-density 2U delivers high performance and reliability as well as space-, energy-, and cost-efficiency
- Suitable for high-performance computing, virtualization, high-density data centers, and other computation-intensive applications