IEI Releases the New AI Training System for Deep Learning - GRAND-C422-20D!

Jan 24, 2019

IEI product News
GRAND-C422 ai training server
dot GRAND-C422 dot Key Features dot Applications
dot Specifications dot Packing List  
 

GRAND AI Training Server System

GRAND-C422 ai training server system
GRAND-C422

Intel® Xeon® W family processor supported

PEMUX single board computer supports up to 56-lane PCIe signal

Water cooling system on CPU

Support two U.2 SSD

Support one M.2 SSD M-key slot (NVMe PCIe 3.0 x4)

Support 10GbE network

 
Key Features

Expandable to Suit Your Needs

Support up to 4 GPU card expansion slots (dual-width PCIe x8) and 2 single-width expansion slots (PCIe x4) for max. expansion ability to meet computing needs.
Six backplane slots connected directly to the system host board to minimize latency.
IEI grand c422 expandable server

High-Speed Storage Transmission

The U.2 SSD interface features high-speed, low-latency and low-power NVMe standard protocol, and PCIe 3.0 x4 channel. The theoretical transmission speed is up to 32Gbps, while SATA is only 6Gbps, which is 5 times faster than SATA.

AI training server SSD performance

High Capacity Storage

The GRAND-C422-20D supports M.2 NVMe SSD, U.2 SSD and SATA HDD/SSD. It has a built-in M.2 NVMe port and 20 HDD/SSD bays including two U.2 SDD slots. It supports M.2 solid-state disk which is the next-gen. small-sized form factor introduced by Intel® to offer better performance and power-saving.

AI training sever with High Capacity Storage

Water Cooling System

IEI designs a higher-efficiency, compact water cooling system for the latest 14nm Intel® Xeon® Processor W family to cool the CPU and keep the high performance. The water cooling system can support up to 250W TDP.

server with water cooling system
 
Applications

Hyper Converged Infrastructure

Hyper converged infrastructure (HCI) is scale-out software-defined infrastructure that converges core data services on flash-accelerated, industry-standard servers, delivering flexible and powerful building blocks under unified management.

Efficient, agile, flexible, and integrated, these systems allow for easy scale-out storage, cost-savings, and simplicity to manage your systems. It is the best solution for your data center.

AI training server for Hyper Converged Infrastructure

AI Training System

The GRAND-C442 is dedicated for AI training tasks based on its wide range of slot support for storage expansion,  accelerator cards, video capture, Thunderbolt ™ or PoE add-on cards for unlimited data acquisition possibilities. In order to develop a useful training model, those existing and widely used deep learning training frameworks are recommended.

IEI offers different accelerator cards to facilitate AI inference computing. By combining Mustang-F100-A10, Mustang-V100-MX8 and GPU cards with the GRAND-C422, a complete AI ready solution is easily achieved.

Grand-C422-20D deep learning models

Specifications
Model Name GRAND-C422-20D
Chassis Dimensions (WxHxD) 176.15 mm x 480.94 mm x 644 mm
System Fan 2 x 120 mm, 12V DC
Chassis Construction 4U, Rackmount
System Cooling 2 x Cooling Fans with Smart Fan
Motherboard CPU Support LGA-2066 Intel® Xeon® W family processor
Processor Cooling Water cooling system
Chipset C422
Memory Total slot: 4 x DDR4 ECC RDIMM/LRDIMM
Memory expandable up to:256GB (4 x 64GB)
GRAND-C422-20D-model-naming-convention
Security TPM 1 x TPM 2.0 Pin header
IPMI IPMI Solution IPMI LAN port, IPMI VGA
Storage Hard Drive 12 x 2.5" / 3.5" drive bay
8 x 2.5" drive bay
M.2 1 x M.2 built in on SBC
U.2 2 x U.2 SSD drive bay compatible to SATA
Networking Ethernat IC 1 GbE NIC: Intel® i210-AT with NCSI support
10 GbE NIC: Aquantia AQC107
I/O Interfaces USB 3.0 4
USB 2.0 2
Ethernet 1 x 1GbE RJ45 combo LAN ports / IPMI 1 x 10GbE RJ45 LAN port
Display 1 x IPMI VGA display
Buttons Power button
Internal I/O COM port 2 x RS232 pin header
USB 3.0 2 x USB 3.0 pin header
USB 2.0 1 x USB DOM header
Indicator LEDs 10 GbE, Status, LAN, Storage Expansion Port Status
  LCM LCM, 2 buttons
Expansions PCIe 4 x PCIe 3.0 x8
2 x PCIe 3.0 x4
Power Power Input 110-240 AC,47-63Hz
Power Consumption In Operation: 285W
Type/Watt Redundant Power 1600W
Reliability Operating Temperature 0~40°C
Relative Humidity 5 to 95% non-condensing, wet bulb: 27˚C.
Weight 23.59 kg
Certification CE/FCC
OS Supported OS Windows server 2016
Linux
 
Packing List
Flat head screws (for 2.5” HDD) Flat head screws (for 3.5” HDD)
1 x Cat5e LAN cable 1 x QIG
2 x Power cord 1 x Cat6A LAN cable