NVIDIA A30 24GB PCIe Gen4 Passive GPU for HPE ProLiant DL380

The NVIDIA A30 24GB PCIe Gen4 Passive GPU for HPE ProLiant DL380 (P38788-B21) is a data center-grade accelerator built on NVIDIA's Ampere architecture, engineered specifically for enterprise AI inference, HPC, and data analytics workloads. It delivers 165 TFLOPS of INT8 performance and 10.3 TFLOPS of FP64 tensor performance, supported by 24GB of High Bandwidth Memory 2 (HBM2) with 933 GB/s of memory bandwidth. The passive cooling design integrates seamlessly into the airflow-managed environment of HPE ProLiant DL380 rack servers, making it well-suited for high-density datacenter deployments.

As an HPE factory-integrated option, the A30 is validated and qualified against HPE's ProLiant firmware and system management stack, enabling full visibility through HPE iLO and Integrated Smart Array management. The GPU supports PCIe Gen4 x16 interconnect, providing the bandwidth necessary for latency-sensitive inference pipelines and large-scale model serving. Multi-Instance GPU (MIG) technology allows the A30 to be partitioned into up to four isolated GPU instances, enabling multiple workloads or tenants to share a single physical device with guaranteed quality-of-service.

The A30 supports NVIDIA's full software ecosystem including CUDA, cuDNN, TensorRT, and the NVIDIA AI Enterprise suite, making it a versatile platform for organizations deploying machine learning inference at scale, scientific simulation, and financial modeling. Its passive thermal design and enterprise-grade MTBF characteristics align with the continuous-operation requirements of production datacenter environments across industries including healthcare, financial services, telecommunications, and government.

Ideal for

• Large-scale AI inference serving for natural language processing and computer vision models in production datacenter environments
• High-performance computing (HPC) simulations in scientific research, computational fluid dynamics, and molecular dynamics applications
• Multi-tenant GPU virtualization using MIG partitioning to isolate workloads across teams or business units within a shared ProLiant DL380 infrastructure
• Real-time data analytics and in-database acceleration for financial risk modeling, fraud detection, and algorithmic trading platforms
• Training and fine-tuning of mid-scale deep learning models within HPE ProLiant-based on-premises AI clusters
• Medical imaging processing and genomics workloads requiring high memory bandwidth and FP64 precision in regulated, on-premises environments

Technical specifications

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