NVIDIA RTX 4000 Ada Graphics Accelerator for HPE

Introduction

Do you require higher performance for artificial intelligence (AI) training and inference, high-performance computing (HPC) or graphics? NVIDIA Accelerators for HPE help solve the world’s most important scientific, industrial, and business challenges with AI and HPC. Visualize complex content to create cutting-edge products, tell immersive stories, and reimagine cities of the future. Extract new insights from massive datasets. Hewlett Packard Enterprise servers with NVIDIA accelerators are designed for the age of elastic computing, providing unmatched acceleration at every scale.

What is new supporting point:
With 96 GB of ultra-fast GDDR7 memory, the NVIDIA RTX PRO 6000 Blackwell accelerates a range of use cases from agentic AI, physical AI, and scientific computing to rendering, 3D graphics, and video.
Built on the NVIDIA Blackwell architecture, the NVIDIA RTX PRO™ 6000 Blackwell Server Edition delivers a powerful combination of AI and visual computing to accelerate enterprise data center workloads.
NVIDIA H200 NVL comes with a five-year NVIDIA AI Enterprise subscription and simplifies the way you build an enterprise AI-ready platform.
Multi-Instance GPU (MIG) expands the performance and value of RTX PRO 6000 Blackwell by enabling the creation of up to four (4) fully isolated instances.
With up to four GPUs connected by NVIDIA NVLink™ and a 1.5x memory increase, LLM inference can be accelerated up to 1.7x and HPC up to 1.3x on H200 NVL compared to H100 NVL.
NVIDIA H200 NVL is ideal for lower-power, air-cooled enterprise rack designs that require flexible configurations, delivering acceleration for AI and HPC workloads.

Processor

CUDA

Yes

CUDA cores

6144

Graphics processor family

NVIDIA

Graphics processor

RTX 4000 Ada

Memory

Discrete graphics card memory

20 GB

Graphics card memory type

GDDR6

Memory bus

160 bit

Memory bandwidth (max)

360 GB/s

Ports & interfaces

Interface type

PCI Express x16 4.0

DisplayPorts quantity

4

DisplayPort version

1.4a

Performance

DirectX version

12.0

Shader model version

6.7

OpenGL version

4.6

OpenCL version

3.0

Design

Cooling type

Active

Power

Power consumption (typical)

130 W

System requirements

Linux operating systems supported

Yes

Server operating systems supported

Yes

Other features

Weight (imperial)

1.02 kg (2.25 lbs)

Height (imperial)

11.1 cm (4.37")

Compatible products

HPE Mainstream Compute Platforms

Weight & dimensions

Weight

1.02 kg

Length

111 mm

Depth

241.3 mm

Height

17.3 mm

Width

17.3 mm