RES AI XR6 4U, 8 Drive 22.5 Inch Rugged GPU Server

Dual Intel® Xeon® Scalable processors, 8 double-wide NVIDIA GPUs, 2TB memory and 8 disk drives

  • slide
  • slide
  • slide
  • slide

Compact 22.5 Inch Deep, Rear I/O Supercomputing Server

Leveraging the latest data center processing and co-processing technologies to accelerate the most demanding workloads, Mercury’s RES AI 4U supercomputing server packs dual Intel® Xeon® Scalable processors, eight double-wide NVIDIA Tesla GPUs, 2TB DDR4 ECC memory and 120TB of storage in a compact 49.5lb server form factor. 

DOWNLOAD RES AI XR6 4U, 8 DRIVE 22.5 INCH DATASHEET

CONFIGURE YOUR SERVER

Features
  • Up to eight NVIDIA® Tesla® GPGPUs or Quadro® GPU
  • Up to two Intel® Xeon® Scalable CPU with up to 28 cores per processor 
  • Up to 2TB DDR4 ECC memory
  • Up to 120TB of storage with eight  2.5" front-access disk drives (SAS3, SATA or NVME)
  • 9 PCIe 3.0 cards
  • MIL-STD: 810G, 901D, 167-1, 1474D, 740-2, 461F
  • Customizable with U.S. designed and manufactured motherboards and advanced security features
  • Manufactured in AS5553 compliant, AS9100D facilities

Patented Technology Options
In addition to memory stabilization, this server features numerous patented technology options including:

  • System control module for remote management
  • Aeroloc Baffle system for low airflow operations
  • Remote battery bypass that enables BIOS battery replacement without system shutdown
Specifications
  • Dual Intel® Xeon® Scalable CPUs with up to 28 cores per processor
  • Bronze, Silver, Gold, or Platinum
  • Up to 8 double-wide NVIDIA Tesla or Quadro GPU accelerators
  • Up to 2TB 2933MHz memory with 12 DIMM slots

Patented Technologies

  • Memory stabilization
  • Aeroloc baffle system
  • System control module for acoustic and remote management

Management and Operating System

  • Windows®, Linux®, VMware® and other hypervisors
  • IPMI v2.0, Redfish option available
  • TPM 1.2 or 2.0 support

Expansion and Modular Maintainability

  • 9 PCIe 3.0 Cards
    • 8 PCIe 3.0 x16 GPU slots + 1 PCIe x16 high speed networking slot

Power Supply Options

  • 2000W 110 VAC power (2+1 redundancy)
  • 3200W 220 VAC power (2+1 redundancy)
  • MIL-STD 461, 704F, 1399-300B

Input/Output Versatility

Front Access

  • Up to 8 removable, hot pluggable, 2.5” SATA/SAS3 drives, U.2 NVME option available
    can configure with up to (4) 15mm or (8) 7mm (SATA) high drives
  • 1 Power/Reset Switch
  • 1 CFM Switch (optional)
  • 1 Power on LED
  • 1 Blu-Ray or DVD/CD ROM drive (optional)
  • 2 USB 3.0
  • 1 Quick Change CMOS battery

Rear Access

  • 2 1GBaseT ethernet ports (RJ45) with 1 port shared with IPMI 2.0
  • 1 Display port with adapter to convert to VGA and 2 USB 2.0 ports

Additional Options

  • Front door filter
  • Slide rails
  • CAC card reader
  • Read/Write switches to prevent accidental rewrite

iot-solutions-alliance-logo.png   

Environmental & Mechanical

MIL-STD / Industrial Specifications

  • MIL-STD 810G
  • Shock: MIL-STD 901D Grade A, IEC 60068-2-27
  • EMI/RFI: MIL-STD 461F, CE102 standard
  • Vibration: MIL-STD 167-1, MIL-STD 810G, IEC 60068-2-64
  • Airborne noise: MIL-STD 1474D
  • Structure borne noise: MIL-STD 740-2
  • Temperature: IEC 60068-2-2 test Bb, 60068-2-1 test Ab

Environmental*

Operating

  • Temperature: 0°C to 50°C
  • Extended Temperature: -20°C to 55°C
  • Humidity: 5% to 95% (non-condensing)
  • Shock: 3 axis, 35g, 25ms
  • Vibration: 4.76Grms, 4Hz to 2000 Hz (SSD)
  • Altitude: 10,000 ASL

Non-Operating

  • Temperature: -40°C to 80°C
  • Humidity: 5% to 95% (non-condensing)
  • Altitude: 40,000 ASL
  • Conformal Coating: IPC-CC 830 (optional)
 

Mechanical

  • Height: 4U or 7 inches (177mm)
  • Width: 17 inches (431.8mm)
  • Depth: 22.5 inches (571.5mm)
  • Weight (Typical)*: 49.5lbs (22.5kg)
  • Dip brazed welded aluminum chassis
  • Cooling: Internal fan-cooled (rear vent) front to rear
  • 19” rackmountable

* Mercury Systems designs all products to meet or exceed listed data sheet specifications. Some specifications including I/O profiles, weight, and thermal profiles are configuration dependent. Contact Mercury for information specific to your desired configuration requirements.

Configurable to your HPC Field Application

Equipped with 9 PCIe 3.0 slots that accommodate a mix of GPUs, FPGA accelerators and other expansion cards, RES AI 4U maximizes scalability and minimizes the overall computing footprint to accelerate an array of high performance computing (HPC) workloads in the field. Eight hot-pluggable, front-access drives support the latest 2.5", 15mm high and 7mm high SAS3, SATA and NVME drives and can be configured for redundancy. 

To ensure sustained optimal performance in almost any environment, our 4U server minimizes airborne and structural noise and is certified to multiple International Electrotechnical Commission (IEC) and military standard specifications including MIL-STD 810G, 901D, 167-1, 1474D, 740-2 and 461F.  Patented memory stabilization technology prevents disconnect during system shock and vibration while 2+1 redundant AC power supplies enhance availability. 

Maximize Throughput with Parallel Computing

Powered by eight of the latest NVIDIA® Volta architecture GPUs, RES AI 4U harnesses parallel processing to maximize throughput, boost productivity and push the boundaries of compute-heavy mission-critical applications such as signal intelligence (SIGINT), cryptography, deep learning, artificial intelligence (AI), surveillance, sensor fusion, visualization, image processing, tracking and big data analytics. Other GPU options are available. 

Scale Supercomputing Applications from the Cloud to the Edge

Compatible with multiple operating systems, applications and software, RES AI adheres to open standards and integrates the latest commercial technologies to support supercomputing from the cloud to the edge. 

The RES AI 4U is designed in the U.S. and manufactured in AS5553 compliant, AS9100D and ISO9001 facilities to maintain stringent quality standards.

DISCOVER MORE

This is not a drill: Preparing for a severe cyber storm

Designing post-quantum cryptography (PQC) is much more than just implementing an algorithm. You need...

SEE MORE

Article: Optimizing the Edge Through Distributed Disaggregation

Disaggregating processing is now enabling low-latency, network-attached everything at the edge with...

SEE MORE

White Paper: Enabling big data processing and AI-powered everything, everywhere

Learn how Mercury and Intel collaborate to scale and deploy composable data center capabilities...

SEE MORE

Webinar: Using NVIDIA-Based Distributed Processing to Speed Mission-Critical AI at the Edge

For the first time in the market, there is an optimized network-attached rugged distributed GPU...

SEE MORE

Vodcast: Enabling the Army to get higher performance and SWaP optimization in a smaller footprint

Live from the International Armored Vehicle Show in Austin, Texas, discover how Mercury is enabling...

SEE MORE

White Paper: Massive Disaggregated Processing for Sensors at the Edge

Optimized networked-attached GPU distributed processing architecture delivers NVIDIA A100 GPU...

SEE MORE

Tech Brief: Expedite Critical Wideband Signal Recorders in Defense Applications

Recording solutions accelerate system development, performance and deployment. Capturing RF signals...

SEE MORE

Case Study: Mercury + U.S. Coast Guard, Defending Borders and Protecting the Maritime Environment

Learn all about how Mercury helped the U.S. Coast Guard perform their maritime duties with solutions...

SEE MORE