HPE Cray XD670 AI Server (NVIDIA H100 NVL)

The HPE Cray XD670 is a purpose-built AI server engineered for large-scale machine learning, deep learning, and high-performance computing workloads. Each node features up to eight NVIDIA H100 NVL GPUs delivering 141.1 teraflops of FP8 performance per GPU, with 141GB of HBM3e memory per accelerator optimized for generative AI and LLM inference at scale.

The system architecture incorporates a 2-socket 4th generation Intel Xeon Scalable processor foundation with up to 60 cores per socket, paired with 12-channel DDR5 memory and PCIe Gen5 connectivity. The H100 NVL configuration emphasizes inference acceleration and batch processing, with unified memory addressing and NVIDIA NVLink-C2C fabric supporting full-duplex 900GB/s inter-GPU communication.

Ideal for

• Large language model (LLM) inference serving and prompt optimization
• Generative AI model fine-tuning and continued pre-training
• Data center-scale machine learning feature engineering and ETL pipelines
• Climate modeling, molecular dynamics, and scientific computing simulations
• Telecommunications network optimization and 5G/6G workload acceleration
• Financial risk modeling and computational trading analytics

Technical specifications

Available from Omnixon Global. Submit an RFQ and our team will confirm configuration and availability for your order.