Nvidia DGX h100 Datasheet Nvidia Us Web

Datasheet
NVIDIA DGX H100

The gold standard for AI infrastructure.
Artificial intelligence has become the go-to approach for solving

Specifications
difficult business challenges. Whether improving customer service,
GPU 8x NVIDIA H100 Tensor Core GPUs
optimizing supply chains, extracting business intelligence, or
designing leading-edge products and services with generative AI GPU memory 640GB total
and other transformer models, AI gives organizations across nearly Performance 32 petaFLOPS FP8
every industry the mechanism to realize innovation. And as a pioneer
NVIDIA® NVSwitch™ 4x
in AI infrastructure, NVIDIA DGX™ provides the most powerful and
complete AI platform for bringing these essential ideas to fruition. System power 10.2kW max
usage
NVIDIA DGX H100 powers business innovation and optimization.
CPU Dual Intel® Xeon® Platinum 8480C Processors
Part of the DGX platform and the latest iteration of NVIDIA’s 112 Cores total, 2.00 GHz (Base),
legendary DGX systems, DGX H100 is the AI powerhouse that’s the 3.80 GHz (Max Boost)
foundation of NVIDIA DGX SuperPOD™, accelerated by the System memory 2TB

groundbreaking performance of the NVIDIA H100 Tensor Core GPU.
Networking 4x OSFP ports serving 8x single-port NVIDIA
The system is designed to maximize AI throughput, providing ConnectX-7 VPI
enterprises with a highly refined, systemized, and scalable platform > Up to 400Gb/s InfiniBand/Ethernet
to help them achieve breakthroughs in natural language 2x dual-port QSFP112 NVIDIA ConnectX-7 VPI
> Up to 400Gb/s InfiniBand/Ethernet
processing, recommender systems, data analytics, and much more.
Available on-premises and through a wide variety of access and Management 10Gb/s onboard NIC with RJ45
network 100Gb/s Ethernet NIC
deployment options, DGX H100 delivers the performance needed
Host baseboard management controller
for enterprises to solve the biggest challenges with AI.
(BMC) with RJ45
Storage OS: 2x 1.92TB NVMe M.2

The Cornerstone of Your AI Center of Excellence
Internal storage: 8x 3.84TB NVMe U.2
AI has bridged the gap between science and business. No longer
the domain of experimentation, AI is used day in and day out by Software NVIDIA AI Enterprise – Optimized AI software
companies large and small to fuel their innovation and optimize NVIDIA Base Command – Orchestration,
scheduling, and cluster management
their business. As the fourth generation of the world’s first
DGX OS / Ubuntu / Red Hat Enterprise Linux /
purpose-built AI infrastructure, DGX H100 is designed to be the Rocky – Operating System
centerpiece of an enterprise AI center of excellence. It’s a fully
Support Comes with 3-year business-standard
optimized hardware and software platform that includes full hardware and software support
support for the new range of NVIDIA AI software solutions, a rich
System weight 287.6lbs (130.45kgs)
ecosystem of third-party support, and access to expert advice
from NVIDIA professional services. DGX H100 offers proven Packaged system 376lbs (170.45kgs)
weight
reliability, with the DGX platform being used by thousands of
customers around the world spanning nearly every industry. System dimensions Height: 14.0in (356mm)
Width: 19.0in (482.2mm)
Length: 35.3in (897.1mm)
Break Through the Barriers to AI at Scale Operating 5–30°C (41–86°F)
As the world’s first system with the NVIDIA H100 Tensor Core temperature range
GPU, NVIDIA DGX H100 breaks the limits of AI scale and

performance. It features 9X more performance, 2X faster networking with NVIDIA
ConnectX®-7 smart network interface cards (SmartNICs), and high-speed scalability
for NVIDIA DGX SuperPOD. The next-generation architecture is supercharged for
the largest, most complex AI jobs, such as generative AI, natural language
processing and deep learning recommendation models.
Powered by NVIDIA Base Command

NVIDIA Base Command powers the DGX platform, enabling organizations to leverage
the best of NVIDIA software innovation. Enterprises can unleash the full potential
of their DGX infrastructure with a proven platform that includes enterprise-grade
orchestration and cluster management, libraries that accelerate compute, storage
and network infrastructure, and an operating system optimized for AI workloads.
Additionally, DGX infrastructure includes NVIDIA AI Enterprise, offering a suite of
software optimized to streamline AI development and deployment.
Leadership-Class Infrastructure on Your Terms

AI for business is about more than performance and capabilities. It’s also about
fitting neatly into an organization’s IT envelope and practices. DGX H100 can be
installed on premises or accessed through a variety of managed and colocation
options. And with the DGX-Ready Lifecycle Management program, organizations
get a predictable financial model to keep their deployment at the leading edge. This
makes DGX H100 as easy to use and acquire as traditional IT infrastructure, with no
additional burden on busy IT staff—which lets organizations leverage AI for their
businesses today instead of waiting for tomorrow.
Ready to Get Started?

To learn more about NVIDIA DGX H100, visit
nvidia.com/DGX-H100
© 2023 NVIDIA Corporation and affiliates. All rights reserved. NVIDIA, the NVIDIA logo, Base Command, ConnectX,
DGX, DGX SuperPOD, and NVSwitch are trademarks and/or registered trademarks of NVIDIA Corporation and
affiliates in the U.S. and other countries. All other trademarks and copyrights are the property of their respective
owners. Specifications are subject to change without notice. 2795800. May23

Nvidia DGX h100 Datasheet Nvidia Us Web

Uploaded by

Copyright:

Available Formats

Nvidia DGX h100 Datasheet Nvidia Us Web

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Nvidia DGX h100 Datasheet Nvidia Us Web

Uploaded by

Copyright:

Available Formats

Datasheet

NVIDIA DGX H100

Artificial intelligence has become the go-to approach for solving

foundation of NVIDIA DGX SuperPOD™, accelerated by the System memory 2TB

Storage OS: 2x 1.92TB NVMe M.2

GPU, NVIDIA DGX H100 breaks the limits of AI scale and

Powered by NVIDIA Base Command

Leadership-Class Infrastructure on Your Terms

Ready to Get Started?

You might also like