Sale!

Supermicro 4124GO-NART | AS-4124GO-NART GPU Server 8 x A100 80GB SXM4 NVLink Deep Learning System

$187,187.00

Above price is for informational purposes and includes 2 x AMD EPYC Milan 7763 64 Cores/128 Threads 2.45GHz CPU, 8x Nvidia A100 80 GB GPU with SXM4 NVLink, 16 x 64GB DDR4-3200 2Rx4 Memory (1TB total), 2 x 960GB NVMe PCIe SSD for OS, 8 x 3.84TB NVMe PCIe SSD for data, 2 x 10GbE Ethernet ports.

For best pricing, please fill out the configuration form below. Educational and Inception program discounts available. Compare this to NVIDIA DGX A100.

For other OS or instructions, please specify in comments section below

Supermicro A+ Server AS -4124GO-NART | 4124GO-NART 8 x NVIDIA Ampere A100 SXM4 NVLink GPU AMD EPYC 7002/7003 Deep Learning Server

NVIDIA NGC Pre-Loaded

This Deep Learning server is available with NVIDIA NGC containers that can be preloaded. NGC empowers researchers, data scientists, and developers with performance-engineered containers featuring AI software like TensorFlow, Keras, PyTorch, MXNet, NVIDIA TensorRT™, RAPIDS and more. These pre-integrated containers feature NVIDIA AI stack including NVIDIA® CUDA® Toolkit, NVIDIA deep learning libraries which are easy to upgrade using Docker commands.

Key Features

  • Manufacturer Part Number: AS -4124GO-NART (AS-4124GO-NART)
  • Supports up to 8 NVIDIA® HGX™ A100 GPUs; GPU Communication using NVIDIA® NVLINK™ + NVIDIA® NVSwitch™;
    8 NIC for GPU Direct RDMA (1:1 GPU Ratio)
  • Supports HGX A100 8-GPU 40GB (HBM2) or 80GB (HBM2e)
  • Dual AMD EPYC™ 7003 Series Processors
  • 8TB Registered ECC DDR4 3200MHz SDRAM in 32 DIMMs
  • 8 PCI-E 4.0 x16 via PCI-E switch; 1 PCI-E 4.0 x 16 LP and 1 PCI-E 4.0 x8 LP via CPUs; AIOM Support
  • 6 Hot-swap U.2 NVMe 2.5″ drive bays (4 via PCI-E switch, 2 via CPU )
  • Flexible Networking via AIOM, 1 dedicated IPMI LAN Port
  • 4 Hot-swap heavy-duty cooling fans
  • 2200W Redundant Platinum Level Power Supplies (full redundancy based on configuration and application load); 3000W Titanium Level Redundant Power Supply option coming soon
  • NGC Docker Container for Deep Learning Pre-loaded (Optional)
  • 3 Years Manufacturer’s Warranty

 

 

Description

AS -4124GO-NART Technical Specifications

Product SKUs
AS -4124GO-NART
  • A+ Server 4124GO-NART
Motherboard
Super H12DGO-6
Processor/Cache
CPU
  • Dual AMD EPYC™ 7002 Series Processors
  • Socket SP3
  • Supports CPU TDP up to 280W*
Cores
  • Up to 128 Cores (64 per CPU)
Note * Certain CPUs with high TDP may be supported only under specific conditions. Please contact Supermicro Technical Support for additional information about specialized system optimization
GPU Support
  • Supports HGX A100 8-GPU 40GB (HBM2) or 80GB(HBM2e) with NVLink/NVSwitch GPU interconnect and PCI-E Gen4 host CPUs
System Memory
Memory Capacity
  • 32 DIMM slots
  • Up to 8TB 3DS ECC DDR4-3200MH RDIMM/LRDIMM
Memory Type
  • 3200MHz ECC DDR4 RDIMM/LRDIMM
On-Board Devices
Chipset
  • System on Chip (SoC)
SATA
  • SATA3 (6Gbps)
Network Controllers
  • Provided via AIOM
IPMI
  • Support for Intelligent Platform Management Interface v.2.0
  • IPMI 2.0 with virtual media over LAN and KVM-over-LAN support
Graphics
  • ASPEED AST2600 BMC
Input / Output
SATA
  • 4 SATA3 (6Gbps) ports
LAN
  • Provided by AIOM
  • 1 RJ45 Dedicated IPMI LAN port
USB
  • 2 USB 3.0 ports (front)
Video
  • 1 VGA Connector (front)
COM Port
  • 1 COM port (header)
System BIOS
BIOS Type
  • AMI 256Mb SPI Flash ROM
Management
Software
PC Health Monitoring
CPU
  • Monitors for CPU Cores, Chipset Voltages, Memory.
  • 4+1 Phase-switching voltage regulator
FAN
  • Fans with tachometer monitoring
  • Status monitor for speed control
  • Pulse Width Modulated (PWM) fan connectors
Temperature
  • Monitoring for CPU and chassis environment
  • Thermal Control for fan connectors
Chassis
Form Factor
  • 4U Rackmountable
  • Rackmount Kit (MCP-290-00057-0N)
Model
  • CSE-438G
Dimensions
Height
  • 6.9″ (174mm)
Width
  • 17.6″ (446mm)
Depth
  • 35.4″ (900mm)
Package
  • 29.5″ (750mm) H x 27.4″ (695mm) W x 44.9″ (1140mm) D
Weight
  • Net Weight: 166.0 lbs (75.3 kg)
  • Gross Weight: 225.0 lbs (102.1 kg)
Drive Bays / Storage
Hot-swap
  • 6 hot-swap U.2 NVMe 2.5″ drive bays (4 via PCI-E switch, 2 via CPU, SATA/NVMe Hybrid or SAS with optional HBA)
    (up to 10 hot-swap U.2 NVMe 2.5″ available)
Expansion Slots
PCI-Express
  • 8 PCI-E 4.0 x16 via PCI-E switch
  • 1 PCI-E 4.0 x16 LP and 1 PCI-E 4.0 x8 LP via CPUs
System Cooling
Fans
  • 4 x Hot-swap 11.5K RPM heavy duty fans
Power Supply
2200W Redundant Platinum Level Power Supplies
Total Output Power
  • 1000W with Input 100 – 127Vac
  • 2200W with Input 220 – 240Vac (for UL/cUL only)
  • 2090W with Input 230 – 240Vdc (for CQC only)
Dimension
(W x H x L)
  • 106.5 x 82.4 x 203.5 mm
Input
  • 1000W: 100-127 Vac / 12-9.5 A / 50-60 Hz
  • 1800W: 200-220 Vac / 10-9.5 A / 50-60 Hz
  • 1980W: 220-230 Vac / 10-9.5 A / 50-60 Hz
  • 2090W: 230-240 Vac / 10-9.8 A / 50-60 Hz
  • 2200W: 220-240 Vac / 11.8-9.6 A / 50-60 Hz (for UL/cUL only)
  • 2090W: 230-240 Vdc / 10-9.8 A / 50-60 Hz (for CQC only)
+12V
  • Max: 41.67A / Min: 0A (500W)
+54V
  • Max: 18.5A / Min: 0A (1000W)
  • Max: 33.33A / Min: 0A (1800W)
  • Max: 36.6A / Min: 0A (1980W)
  • Max: 38.7A / Min: 0A (2090W)
  • Max: 40.74A / Min: 0A (2200W)
12Vsb
  • Max: 3A / Min: 0A
Output Type
  • Gold Finger (connector on M/P)
Certification Platinum Level Certified95%+
Platinum Certified
Test Report ]
Operating Environment
RoHS
  • RoHS Compliant
Environmental Spec.
  • Operating Temperature:
    10°C ~ 35°C (50°F ~ 95°F)
  • Non-operating Temperature:
    -40°C to 60°C (-40°F to 140°F)
  • Operating Relative Humidity:
    8% to 90% (non-condensing)
  • Non-operating Relative Humidity:
    5% to 95% (non-condensing)

Additional information

Weight 225 lbs
Dimensions 35.4 × 17.6 × 6.9 in
GPU Compatibility

NVIDIA A100 80GB NVLink SXM4 (HGX)

Max GPU

8

Reviews

There are no reviews yet.

Be the first to review “Supermicro 4124GO-NART | AS-4124GO-NART GPU Server 8 x A100 80GB SXM4 NVLink Deep Learning System”

Your email address will not be published.

You may also like…