New Subscribers Exclusive! 3% Off + Free Shipping on Your First System Order
Looking for Something? 24 Hour Live Chat Tech Support
Need Help? Click Here for More Support Resources!

Supermicro AS -3015MR-H5TNR 3U 5-Node Rackmount H13 UP MicroCloud A+ Server

  • Key Features/Applications: Generative AI, AI Inference, Model Training, 3D Graphics and Video Rendering, Cloud Gaming
  • CPU: Single AMD EPYC™ 4005/4004 Series Processor Per Node, up to 16 Cores
  • Chassis: 3U / 5 Nodes
  • Drive: Up to 2x 2.5" Internal Fixed PCIe 5.0 NVMe + 2x 2.5" Internal Fixed SATA Drive Bays Per Node
  • RAM: Up to 192GB of DDR5 ECC/non-ECC UDIMM in 4 DIMM Slots Per Node
  • GPU: Up to 2 GPUs Per Node
  • Network Ports: Flexible Networking Options with PCIe 5.0 MicroLP Interfaces
  • Power Supply: 4x 2000W Redundant (3 + 1) Titanium Level Shared Power Supply

* Complete systems only with all 5 nodes populated.

SKU
AS -3015MR-H5TNR

* Required Fields

AS-3015MR-H5TNR Front

3U MicroCloud with 5 AMD GPU Nodes

The AS -3015MR-H5TNR is a 3U 5-Node MicroCloud with AMD EPYC processors and double-width GPU support in every node, ideal for Cloud Computing and CDN.

High-Density Multi-Node System for AI and 3D Graphics

The demands of generative AI, training and inference, 3D graphics rendering, and video transcoding are shifting the need for computing resources more heavily onto GPU accelerators. Whether in the enterprise or cloud environment, dedicated, bare-metal servers are in demand to deliver the utmost in performance. Powered by AMD EPYC™ 4005/4004 Series processors, the 3U, multi-node AS -3015MR-H5TNR MicroCloud system easily meets these demands. You reap the benefits of shared power and cooling infrastructure, and affordable pricing that is favorable compared to the same number of 1U servers. With hot-swappable nodes, the system is easy to maintain and also easy to upgrade with future nodes.

Multi-Node System for GPU-Accelerated Hosting

This 5-node Supermicro MicroCloud system is designed for applications that need a large number of GPU-accelerated discrete servers, virtualized or nonvirtualized, including:

  • Generative AI that needs GPU power to create responses to customer requests in real time
  • AI inference executes models in production using GPUs to reduce response time and to meet the need for real-time engagement with customers
  • Model training is the most intensive phase of the AI lifecycle, and GPU acceleration is a requirement for iterative training of large AI models
  • 3D graphics and video rendering requires massive amounts of vector processing to help creative teams iterate designs more quickly.
  • Cloud gaming powered by the EPYC 4005/4004 Series, this ultimate gaming processor boasts boost frequencies up to 5.7 GHz

Your Choice of GPU Acceleration

To deliver the acceleration capabilities that match workload requirements, you can choose graphics processors from AMD and NVIDIA, in full-length and either full- or double-width form factors with up to 350W total power consumption. The GPU slot is provided with 16 lanes of PCIe 5.0 connectivity to a single device, or 8 lanes per device for two accelerators.

AS-3015MR-H5TNR Node

Powered AMD EPYC 4005/4004 Series Processors

The AMD EPYC 4005/4004 Series delivers high performance at an affordable price. They leverage AMD’s hybrid, multi-chip architecture to deliver a series with from 4 to 16 cores, and a thermal design that enables boost frequencies up to 5.7 GHz. With one or two eight-core ‘Zen 4’ CPU dies, this processor features a matching I/O die tailored to the needs of small businesses, offering balanced performance at the entry level, while not sacrificing essential features such as the built-in AMD Secure Processor or the choice of AMD 3D V-Cache™ technology for L3 cache sizes up to 128 MB per CPU. Whether you prize overall system price or cost per core, the processor’s attractive price is designed to help you deploy high-value, essential dedicated servers in hosting environments. Choosing the EPYC 4005/4004 Series gives you a fast and practical solution that is tested and validated with the leading server operating systems, so your systems can run the same software as AMD’s high-end server processors.

Flexible Node Configurations

Each node incorporates a single Supermicro low-profile interface slot for network connectivity up to 10 Gigabit Ethernet. Internal I/O includes four 2.5" drive slots can host up to two NVMe and two SATA drives, an M.2 boot-drive slot, and a single 16-lane lowprofile or two 8-lane PCIe 5.0 low-profile slots.

Open Management

Regardless of your data center’s management approach, our open management APIs and tools are ready to support you. In addition to a dedicated IPMI port, and a Web IPMI interface per node, Supermicro® SuperCloud Composer software helps you configure, maintain, and monitor all of your systems using single-pane-of-glass management. Supermicro Server Manager (SSM) and Supermicro Update Manager (SUM) can be used to handle basic administration tasks. If your DevOps teams prefer to use their own tools, industry-standard Redfish® APIs provide access to higher-level tools and scripting languages.

Key Features

  • 3U 5-Node MicroCloud chassis with 30.51" (775mm) depth
  • Single AMD EPYC™ 4005/4004 Processor per node
  • Up to 192GB DDR5-5600MT/s memory per node
  • Single dual-slot GPU or two single-slot GPUs
  • Four internal 2.5" bays for up to two NVMe and two SATA drives
  • One M.2 NVMe drive
  • Flexible networking options with PCIe 5.0 MicroLP interfaces
  • 4x Redundant 2000W Titanium Level Shared Power Supplies
More Information
System Starting Price€11,952.43
CPUAMD EPYC™ 4000 Series, AMD Ryzen™ 7000 Series, AMD Ryzen™ 9000 Series
Specification
We're sorry, an error has occurred while generating this content.