KR6298-X2
6U 8-OAM AI Server for Next-Gen AI, LLM
- 8x fully-connected Habana® Gaudi®2 accelerators
- 2x 4th Gen Intel® Xeon® Scalable
- 32x DDR5, 8TB maximum system memory
- Unprecedented performance to support generative AI, mainstream AI frameworks, development tools and LLM
KR6298V2 with 4th Gen Intel® Xeon® Scalable processors is an advanced AI system supporting 8 Habana® Gaudi®2 OAM accelerators. Leveraging the OAM form factor, it delivers scalable, high-speed performance for next-generation AI workloads and LLM.
Scalable High Performance
- 96GB HBM3, 4.2TB/s converged P2P bandwidth
- Supports AMX/DSA AI engines
- 4000Gbps cross-node communication bandwidth enables pipeline, data parallelism
Optimized Energy Efficiency
- Refined thermal design, intelligent thermal management, 12V/54V dual power input
- Power dynamic matching for high-power workloads, zero power conversion loss lead to 8% system power reduction
Open Ecosystem
- Complies with OAM v1.1, UBB v1.5
- Supports mainstream AI frameworks and AI development tools
- Mature open ecosystem enables cost-efficient AI compute solution development
| Model | KR6288-X2-A0-R0-00 |
|---|---|
| Form Factor | 6U rack server |
| OAM | 8x Habana® Gaudi®2 HL225B, 600W |
| CPU | 2x 4th Gen Intel® Xeon® Scalable processors |
| Chipset | Intel® C740 series |
| Memory | 32x DDR5 DIMMs, up to 4800MT/s |
| Storage | 24x 2.5” SSD, up to 16x NVMe U.2 |
| M.2 | 2x NVMe/SATA M.2 |
| Expansion Slots | 8x low-profile x16 + 2x FHHL x16 |
| I/O | Front: 1x USB 3.0, 1x USB 2.0, 1x VGA |
| Rear: 2x USB 3.0, 1x RJ45, 1x MicroUSB, 1x VGA |
|
| OCP | Supports 1x OCP 3.0, supports NCSI |
| Management | Aspeed 2600 |
| Security | TPM 2.0 (Trusted Platform Module) |
| PSU | 2x 12V 3200W 1+1, 6x 54V 2700W 3+3 |
| Dimensions | 482mm (W) x 263mm (H) x 860mm (D) |