HPE ProLiant DL380 Gen11 Intel Gaudi 3 Accelerator Option

The HPE ProLiant DL380 Gen11 Intel Gaudi 3 Accelerator Option (P65396-B21) integrates Intel's Gaudi 3 AI accelerator architecture directly into the HPE ProLiant DL380 Gen11 platform, delivering purpose-built deep learning training and inference performance for enterprise data center environments. Gaudi 3 is built on a 5nm process node and features 64 Tensor Processor Cores alongside 8 matrix multiplication engines, with 128 GB of high-bandwidth memory (HBM2e) per accelerator card providing the memory capacity required for large-scale model training and serving. The architecture incorporates 24 integrated 200 GbE RDMA-capable network ports per accelerator, enabling high-throughput, low-latency scale-out communication without reliance on a separate networking fabric component.

This option kit is engineered specifically for compatibility and validated operation within the HPE ProLiant DL380 Gen11 server ecosystem, ensuring full integration with HPE's system management tooling, including iLO 6 and HPE OneView, for unified monitoring and lifecycle management of accelerator resources alongside host compute infrastructure. The Gaudi 3 accelerator supports the Intel Gaudi software stack, which provides compatibility with PyTorch and TensorFlow frameworks through the SynapseAI SDK, allowing enterprises to migrate existing AI workloads without significant code refactoring.

Targeted at organizations running large language model (LLM) training, generative AI inference, and high-performance deep learning pipelines, the DL380 Gen11 Intel Gaudi 3 Accelerator Option positions enterprise IT teams to deploy scalable AI infrastructure using an open, standards-aligned software ecosystem. Its fit within the widely deployed DL380 Gen11 2U form factor makes it operationally consistent with existing data center rack layouts, power delivery infrastructure, and HPE support structures across UAE, GCC, EMEA, and APAC deployments.

Ideal for

• Large language model (LLM) training and fine-tuning for enterprise AI development teams requiring high-bandwidth memory and scale-out networking within a managed server platform
• Generative AI inference serving for production deployments of transformer-based models where consistent throughput and low latency are required at data center scale
• Computer vision model training pipelines in manufacturing, healthcare imaging, and retail analytics where GPU-class tensor compute is integrated into existing HPE ProLiant infrastructure
• High-performance deep learning research environments in universities, national laboratories, and R&D organizations that require validated, enterprise-supported accelerator hardware
• AI-accelerated data processing and recommendation system training for large-scale e-commerce, financial services, and telecom operators across EMEA and APAC regions
• Hybrid cloud AI infrastructure buildouts where on-premises accelerated compute must integrate with HPE management tooling and align with standardized rack and power specifications

Technical specifications

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