NVIDIA GB200 NVL4 by HPE SKUs

Below is a list of SKUs to configure the NVL4 system.

Additional software and support SKUs will be added soon.

Platform Information

Form Factor

• – 48U Rack-scale System

System

• – Compute tray: 2x Grace CPUs
• – 4x Blackwell GPUs
• – PCIe slots: x2

  • Primary slot will have a BlueField-3 3240 DPU by default
  • Primary slot can be replaced with a 10/25G NIC
  • If two NICs are needed, the second slot can be installed with a 10/25GbE NIC
• – Storage: SEDs – Opal/TCP Capable

  • 4x E1.S – Local data
  • 1x M.2 – OS/Boot Drive
• – Up to 8x 33 kW Power Shelves
• – 50 V DC Busbar
• – Optional: 2x SN2201 Top of Rack Switches
• – Rack Liquid Manifold
• – Front IO:
• – 1x USB 3.0
• – 1x USB 2.0 for UART
• – 1x Mini DisplayPort
• – 1x Management 1 GbE Port
• – 3x Status LEDs

Processors – Grace ARM processor

The GB200 NVL4 compute tray contains 2x GB200 boards. Each board has 1 Grace ARM processor.

On System Management Chipset

Based on NVIDIA OpenBMC

ASPEED 2600 management controller

Memory

Soldered on 512 GB LPDDR5x – 480 GB usable of LPDDR5x

Memory ECC – Supported and enabled by default

Notes: Memory is not configurable

Front Panel Connections

System Unit Specifications

Dimensions (Height x Width x Depth)

• – Rack

2303 mm (H) x 600 mm (W) x 1151 mm (D)

Weight (approximate)

• – Fully Loaded (with PGW):
• 4161 lbs

Internal Storage

Hot-pluggable and hot-swappable

Power Delivery and Requirements

Server Power: The NVL4 tray receives 54 V DC power through the rear bus bar clip.

Rack Power: NVL4 will share the same rack as NVL72, also known as the MGX rack. The MGX rack is based an “OCP-like” form factor having

• 1 Power Shelf contains
• – 6x 5.5 kW PSU – Power Supply Units
• – 1x PMC – Power shelf management controller
• – 1x Power Shelf Chassis

CDU (Cooling Distribution Unit) For NVIDIA GB200 NVL4 by HPE

The cooling distribution unit (CDU) is a liquid-to-liquid heat exchanger that is used to remove heat from HPE GB200 NVL4 by HPE.

The CDU uses a secondary loop to circulate a heat transfer liquid to the cold plates. The heat captured in the secondary loop is transferred to the facility’s primary loop via a liquid-to-liquid heat exchanger. The CDU is designed to circulate and control the heat transfer fluid to the manifolds that are in each rack.

The in-row CDU is rated for 1.3 MW of cooling. One CDU supports a maximum of 4 fully configured NVL4 racks.

HPE supports overhead (top feed) and underfloor (bottom feed) secondary loop configurations. A spacer manifold can be used for racks that do not need liquid cooling.

Below are some images showing overhead and underfloor plumbing configurations:

Operating Systems and Virtualization Software Support for HPE Servers

NVIDIA GB200 NVL4 by HPE will support Ubuntu 24.04 and RHEL 9.6

Notes: NVIDIA GB200 NVL4 by HPE supports Red Hat Enterprise Linux (RHEL) version 9.6 only. Access to ongoing 9.6 operating system updates requires the purchase of Red Hat Enterprise Linux Extended Update Support (EUS). Red Hat Enterprise Linux (RHEL) 9.6 does not support Secure Boot. Red Hat Enterprise Linux (RHEL) 9.6 does not support Secure Boot.

HPE Server UEFI (under development – subject to change)

Unified Extensible Firmware Interface (UEFI) is an industry standard that provides better manageability and more secured configuration than the legacy ROM while interacting with your server at boot time. The NVIDIA GB200 NVL4 by HPE has a UEFI Class 3 implementation and supports UEFI Mode only.

UEFI enables numerous new capabilities such as

• – Secure Boot and Secure Start enabled for enhanced security
• – Embedded UEFI Shell
• – Operating system specific functionality
• – Remote settings configuration using Redfish
• – Support for greater than 2.2 TB (using GPT) boot drives
• – PXE boot support for IPv6 networks
• – USB 3.0 Stack
• – Automatic drive unlocking for self-encrypting drives
• – TPM 2.0 Support
• – iSCSI Software Initiator Support.
• – NVMe Boot Support
• – HTTP/HTTPs Boot support as a PXE alternative.
• – Boot support for option cards that only support a UEFI option ROM

Notes: For UEFI Boot Mode, boot environment and OS image installations should be configured properly to support UEFI.

Industry Standard Compliance (under development – subject to change)

• – ACPI 6.3 Compliant
• – IPMI 2.0
• – PCIe 4.0 Compliant
• – PCIe 5.0 Compliant
• – PXE Support
• – Redfish API
• – TPM 2.0 Support
• – UEFI (Unified Extensible Firmware Interface Forum) 2.6
• – USB 2.0 Compliant
• – USB 3.0 Compliant

Embedded Management (under development – subject to change)

Based off of NVIDIA’s Reference Firmware with HPE Enhancements such as

• – Security Features

  • HWRoT Secure Boot of BMC
  • Recovery BMC firmware image support
  • Ultimaco ESKM appliance support Remote Key Management for self-encrypting devices (OPAL compliance)
• – Audit Logging
• – Restricted SSH Access
• – Serviceability Features

  • AHS logging limited to sensors and events
  • CLI Interface
• – Branding

  • HPE Branded GUI

Server Management

NVIDIA GB200 NVL4 by HPE Baseboard Management Controller Embedded, in-depth server-level monitoring and management technology offering system management, service alerting, reporting and remote management including remote console and virtual media mount.

Industry Standard Redfish

The NVIDIA GB200 NVL4 by HPE supports industry standard DMTF Redfish that provides API conformance and offers simplified server management automation such as configuration and maintenance tasks based on modern industry standards. Learn more at: https://dmtf.org/standards/redfish.

Security (under development – subject to change)

• – Secure Start – only boot Signed FW
• – UEFI Secure Boot
• – NIST SP800-193 Protect, Detect, Recover for BIOS and BMC Tamper-free updates – components digitally signed and verified using CNSA strength RSA4096 signing key.
• – Secure Recovery – recover critical firmware (BIOS, GPU FW) to known good state on detection of compromised firmware. Event recorded in the event log.
• – UEFI TPM 2.0 (Trusted Platform Module 2.0) - supports OpenSSL 3.0 to provide secure remote connectivity sessions to the system
• – Support for multiple Redfish role-based access control (RBAC) accounts ensuring that passwords need not be shared and can provide separation of duties.

For the most up-to-date information on HPE Services, please refer to the HPE Services – Supplemental QuickSpecs, which provides a comprehensive and regularly updated overview of available services.

Emissions Classification (EMC) – Regulatory Information

To view the regulatory information for your product, view the Safety and Compliance Information for Server, Storage, Power, Networking, and Rack Products, available at the Hewlett Packard Enterprise Support Center:

http://www.hpe.com/support/Safety-Compliance-EnterpriseProducts

Environmentally friendly Products and Approach End-of-life Management and Recycling

Hewlett Packard Enterprise offers end-of-life product return, trade-in, and recycling programs, in many geographic areas, for our products. Products returned to Hewlett Packard Enterprise will be recycled, recovered or disposed of in a responsible manner.

The European Union Waste Electrical and Electronic Equipment Directive [EU WEEE] (2002/95/EC) requires manufacturers to provide treatment information for each product type for use by treatment facilities. This information (product disassembly instructions) is posted on the Hewlett Packard Enterprise website. These instructions may be used by recyclers and other WEEE treatment facilities as well as Hewlett Packard Enterprise OEM customers who integrate and re-sell Hewlett Packard Enterprise equipment.