AMD Instinct™ GPU Servers for Machine Learning and LLM

Optimize your machine learning and LLM workloads with a high performance AMD Instinct GPU server.

Outstanding performance

HPE Enterprise Bare Metal Servers, powered by advanced AMD Instinct™ MI Series Accelerators are specifically designed to manage even the most challenging AI and HPC workloads. Enjoy outstanding computing power, increased memory density, and improved memory bandwidth.

See configurations

AMD Instinct MI300A APU

Integrated design

The AMD Instinct MI300A APU integrates an AMD Instinct accelerator and an AMD EPYC™ Choose a high-performance server equipped with the AMD Instinct MI300A APU, tailored to meet scalable requirements, boasting an impressive 192GB HBM3 memory capacity.

Cutting-edge technology

Equipped with CDNA3 GPU architecture and Zen 4 CPU cores, the MI300A offers exceptional computational power and facilities rapid data analysis and advanced modeling tasks with remarkable speed.

AMD Instinct MI300A APU bare metal servers powered by the CDNA3 architecture

Exceptional performance

Built on the revolutionary CDNA3 architecture, the MI300A APU excels in delivering unparalleled computing performance, perfect for the most demanding AI and HPC workloads.

High-efficiency

The integration of AMD Instinct accelerators with EPYC™ Zen 4 CPU cores (24 cores) improves efficiency, flexibility and programmability while eliminating data transfer delays.

Scalability

Choose a high-performance server equipped with the AMD Instinct MI300A APU, tailored to meet scalable requirements, boasting an impressive 192GB HBM3 memory capacity.

AMD Instinct™ MI200 series

AMD Instinct MI250X Accelerator

The AMD Instinct MI250X is the idealaccelerator for HPC workloads, specificallyengineered for the exascale computing era.

AMD Instinct MI250 Accelerator

The AMD Instinct MI250 acceleratordelivers unmatched performance for HPCand AI applications, making it an invaluableasset for enterprises, research andacademic institutions.

AMD Instinct MI210 Accelerator

Enhancing HPC and AI capabilities, theAMD Instinct MI210 Accelerator is tailoredfor research, academic and businessenvironments, hosting both single-serverand larger solutions.

AMD Instinct MI200 dedicated server features

Computational power

The AMD Instinct™ MI200 series, powered by 2nd Gen AMD CDNA™ architecture, utilizes amulti-chip design for maximum throughput and power efficiency in demanding HPC and AI workloads.

AI performance

With advanced AI capabilities, your M200 dedicated server accelerate deep learning training and inference, providing powerful solutions for AI-based projects.

CDNA™ architecture

Featuring 2nd CDNA™ architecture and 3rd Infinity architecture, these servers seamlessly integrate CPU and GPU resources, maximizing system efficiency and throughput.

Connectivity

Servers equipped with AMD Instinct MI200 series accelerators offer advanced peer-to-peer connectivity with up to 8 AMD Infinity Fabric™ links, ensuring seamless and efficient data transfer for demanding workloads.