2U 2-Node 8-Bay GPU Server

SKU: ASA2223-GH-R

NVIDIA GH200 Grace Hopper^™ Superchip
Up to 480GB CPU LPDDR5X ECC memory per node
8 x 2.5" Gen4 NVMe hot-swappable bays
4 x M.2 slots
2 x PCIe x16 (Gen5 x16) FHHL dual slots
2 x PCIe x16 (Gen5 x16) FHHL single slots
2 x OCP 3.0 slots with PCIe Gen5 x16 bandwidth
Triple 3000W (240V) 80 PLUS Titanium redundant power supply

Key Applications:

High Performance Computing
AI/Deep Learning Training and Inference
Large Language Model (LLM) and Generative AI

Starting Configuration: (Contact Us)

Processor

CPU

NVIDIA GH200 Grace Hopper^™ Superchip

1 x Grace CPU

1 x Hopper H100 GPU

900GB/s NVLink-C2C Interconnect

CPU+GPU design for giant-scale AI and HPC

TDP up to 1000W (CPU + GPU + memory)

Motherboard

MV13-HD0

GPU

Supported GPU

NVIDIA: GH200 Grace Hopper^™ Superchip

CPU-GPU Interconnect

900GB/s NVLink-C2C Interconnect

System Memory

Grace CPU

Up to 480GB of LPDDR5X memory with ECC (per node)

Memory bandwidth up to 512GB/s (per node)

Hopper H100 GPU

Up to 96GB HBM3 or 144GB HBM3e (per node)

Memory bandwidth up to 4TB/s (or 4.9TB/s) (per node)

Drive Bays

Hot Swap

8 x 2.5" Gen5 NVMe bays

Expansion Slots

PCI-Express

2 x PCIe x16 (Gen5 x16) FHHL slots (Dual slot)

2 x PCIe x16 (Gen5 x16) FHHL slots (Single slot)

2 x OCP 3.0 slots with PCIe Gen5 x16 bandwidth

M.2 Slots

M-key

PCIe Gen5 x4, from PEX89048

Support 2280/22110 cards

Input/Output

LAN

4 x RJ45

2 x MLAN

USB

4 x USB 3.2 Gen1

Video

Integrated in Aspeed^® AST2600

2D Video Graphic Adapter with PCIe bus interface

1920x1200@60Hz 32bpp

System Cooling

Fans

4 x 80x80x80mm (16,500rpm)

Power Supply

2+1 3000W 80 PLUS Titanium redundant power supplies

Dimensions and Weight

Height

3.45" (87.5mm)

Width

17.33" (440mm)

Depth

33.46" (850mm)

Packaging

12.05" (H) x 27.8" (W) x 46.46" (D)

Operating Environment

Operating temperature

10�C(50�F) ~ 35�C(95�F)

Non-operating temperature

-40�C(14�F) ~ 60�C(140�F)

Operating Relative Humidity

8%~80% non-condensing

Non-operating Relative Humidity

20%~95% non-condensing

GPU Servers