Intel Data Center GPU Max 1350

The Intel Data Center GPU Max 1350 is a high-performance discrete GPU built on Intel's Ponte Vecchio architecture, fabricated using a multi-tile, multi-chiplet design that combines Intel 7 process nodes with advanced packaging technology. It delivers substantial compute throughput across FP64, FP32, FP16, and BF16 precisions, making it well-suited for the demanding numerical workloads found in modern high-performance computing (HPC) and AI training environments. The GPU integrates High Bandwidth Memory (HBM2e), providing the memory bandwidth essential for data-intensive scientific simulations and large-scale deep learning operations.

The Max 1350 is the lower-TDP member of the GPU Max 1350 product line, offering a configurable TDP that enables deployment in a broader range of server and rack configurations where thermal and power envelope constraints are a consideration. It exposes the full Xe-HPC microarchitecture feature set, including support for Intel's oneAPI unified programming model, which allows developers to target the GPU using standard SYCL, OpenMP offload, and familiar HPC frameworks without requiring proprietary vendor-locked toolchains. The architecture includes Xe Matrix Extensions (XMX) engines that accelerate matrix multiply-accumulate operations central to both AI inference and training.

Designed for enterprise data centers, cloud service providers, and national laboratories, the Intel Data Center GPU Max 1350 integrates into PCIe form-factor server slots and is supported across major server platforms. Software ecosystem support spans the Intel oneAPI HPC Toolkit, Intel AI Analytics Toolkit, PyTorch, TensorFlow, and OpenMP, providing a broad foundation for organizations standardizing on heterogeneous compute infrastructure across UAE, GCC, EMEA, and APAC regions.

Ideal for

• Large-scale scientific simulation and computational fluid dynamics workloads requiring high FP64 throughput and substantial memory bandwidth
• AI and deep learning model training leveraging XMX matrix acceleration engines across BF16 and FP16 mixed-precision workflows
• Seismic processing and geophysical modeling in energy-sector HPC clusters where dense linear algebra performance is critical
• Molecular dynamics and quantum chemistry simulations in life sciences and research computing environments
• Financial risk modeling and Monte Carlo simulations demanding consistent double-precision floating-point accuracy at scale
• Government and academic HPC center deployments requiring open-standard oneAPI and OpenMP offload programming compatibility across heterogeneous compute nodes

Technical specifications

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