Leading Architecture Design
Based on the 4th Gen Intel®Xeon®Scalable Processor, up to 350W TDP
Support DDR5 technology +PCIe Gen5 latest technology
PCIe passthrough configuration with low latency creates ultimate performance
Front and rear IO maintenance compatible design, suitable for new and old server rooms
Flexible Topology
CPU-GPU supports passthrough and Switch solutions, up to 10 DW 350W GPU
CPU-Switch can choose X16 or X32 lane interconnection to achieve high-bandwidth communication
Support different CPU-GPU interconnection topologies such as common and performance, enabling broader application scenarios
Extensive Partner Ecosystem
Support single and double-wide GPU cards
Support domestic and foreign brand GPU cards
Support customized development needs with our dedicated team
| Name | Taihang H8230 AI Server |
| CPU | 1/2 x 4th Gen Intel®Xeon®Scalable Processor, TDP 350W |
| Memory | Up to 32x RDIMM & LRDIMMs, 8 channels per CPU |
| Storage | Up to 12x 3.5 SATA/SAS/NVMe drives Up to 24x 2.5 SATA/SAS/NVMe drives+16x NVMe drives 1x M.2 SSD |
| PCIe Extension | Rear PCIe Slot: 8x FHFL DW GPU+5x FHHL + 1x OCP Front PCIe Slot: 3x FHHL + 1x OCP |
| Management | BMC: AST2600, 1x dedicated RJ45 MGMT port |
| IO | 3x USB, Type-A 1x PWR BTN with LED 1x Status LED 1x UART PORT, Type C 1x Front UID BTN with LED + 1x Rear UID LED 1x Front VGA |
| Thermal | 12 x High-performance hot-swappable fans, N+1 Redundancy Optional cold plate cooling |
| Temperature | 5ºC - 35ºC |
| PSU | 4x CRPS power supplies, 3+1 Redundancy |
| Dimension | 175mm (H) x 447mm (W) x 900mm (D) |
