Processors

Each server chassis supports four (4) Intel® Xeon® Scalable processors:

• – Intel® Xeon® 6788P, 2.0 GHz 86-core 350 W Processor for HPE Compute Scale-up Server 3250
• – Intel® Xeon® 6768P, 2.4 GHz 64-core 330 W Processor for HPE Compute Scale-up Server 3250
• – Intel® Xeon® 6728P, 2.7 GHz 24-core 210 W Processor for HPE Compute Scale-up Server 3250
• – Intel® Xeon® 6724P, 3.6 GHz 16-core 210 W Processor for HPE Compute Scale-up Server 3250
• – Intel® Xeon® 6714P, 4.0 GHz 8-core 165 W Processor for HPE Compute Scale-up Server 3250

Notes:

– All processors within a chassis, partition, and system must be identical

– 6714P 8-core and 6724P 16-core processor support limited to 4 socket configuration

Memory type Registered

• – HPE Compute Scale-up Server 3250 64 GB (1x64GB) Dual Rank x4 DDR5-6400 Registered Standard Memory Kit
• – HPE Compute Scale-up Server 3250 96 GB (1x96GB) Dual Rank x4 DDR5-6400 Registered Standard Memory Kit
• – HPE Compute Scale-up Server 3250 128 GB (1x128GB) Dual Rank x4 DDR5-6400 Registered Standard Memory Kit
• – HPE Compute Scale-up Server 3250 256 GB (1x256GB) Quad Rank x4 DDR5-6400 Registered 3DS Memory Kit

Memory protection

• – DDR5 Error Check and Scrub (ECS) for reduced error counts.
• – DDR5 Post-Package Repair (PPR) for self-healing, including boot time (hPPR) and runtime (sPPR).
• – Advanced ECC and multi-region ADDDC for resiliency.

Operating System

• – Red Hat Enterprise Linux (RHEL)
• – SUSE Linux Enterprise Server (SLES)
• – Oracle Linux
• – Microsoft Windows Server 2025 Standard and Datacenter

Notes:

– HPE Foundation Software is required for all Linux OS environments.

– SLES, RHEL and Oracle Linux certifications include KVM certification.

– Minimum OS versions may be required for certain features and hardware options.

I/O Slots

Chassis support either 16 half-height PCIe slots (8 x16 slots and 8 x8 slots); 12 slots with eight-full height slots (4 x16 slots and 4 x8 slots) + 4 half-height slots (4 x16 slots); 6 slots with four full-height (4 x16) + two half-height (2 x16) or a compute-only configuration (no PCIe slots). The compute-only configuration is only supported with the Expansion Chassis.

Base Chassis Base I/O

• – One OCP Card Access
• – One Reset switch
• – One RMC console (mini USB-B)
• – One 1 GbE LAN on motherboard (LOM) RJ45 network port
• – Two external USB 3.0 ports
• – One DisplayPortTM
• – One serial port (9-pin Dsub)
• – Set of System Indicator LEDs
• – Two iX Industrial® ports for inter-chassis management connection
• – One RJ45 RMC Auxiliary port
• – One RJ45 RMC Management port

Form Factor

5U Base Chassis or Expansion Chassis

High availability-standard server features

• – 2N (N+N) redundant (2400 W) power supplies
• – N+1 fans (or greater depending on the load)
• – Hot-swappable and redundant fans, power supplies
• – HPE Analysis Engine Logging and Reporting
• – Full Support for Fault Resilient boot with indications and deconfiguration capability
• – Enhanced MCA Gen2 recovery
• – Post Package Repair (PPR) with ADDDC memory options
• – ECC, re-tries, and Link Width Reduction on data paths
• – Automatic de-configuration of DIMMs. Processor de-configuration
• – I/O Advanced Error Recovery
• – Enhanced Downstream Port Containment
• – PCIe error Isolation
• – Redundancy and Failover in Fabric Lanes; Automatic rerouting; Error correction
• – Redundant Internal Management LAN
• – F ront Panel Management device

Standard Warranty

Three-year parts, 3-year Labor and 3-year on-site limited global warranty. Protected by HPE Services operational services and a worldwide network of Hewlett Packard Enterprise Authorized Channel Partners.

Hewlett Packard Enterprise branded hardware and options qualified for the HPE Compute Scale-up Server 3250 are covered by a global limited warranty and supported by HPE Services and a worldwide network of HPE Authorized Channel Partners. The HPE branded hardware and options diagnostic support and repair is available for three years from date of purchase, or the warranty length of the server they are attached to, whichever is greater. Additional support may be covered under the warranty or available through additional support packages. Enhancements to warranty services are available through HPE Services operational services or customized service agreements.

Additional information regarding worldwide limited warranty and technical support is available at: https://support.HPE.com/hpsc/doc/public/display?docId=c01865770

Physical and Environmental Information

Systems are comprised of the following components: Base chassis plus Expansion chassis.

Enclosure

The system can be field racked. However, it is strongly recommended that customers order the systems racked from the factory. This provides the customer the benefit of extensive system testing and avoids possible premium service charges for field racking service. Field racking requires the use of an appropriate material lift capable of lifting 125+ lbs.

Systems are supported in the HPE 600 mm wide and 800 mm wide racks, and the HPE D-rack. Rack availability is dependent on size of system complex.

Other products may be placed in the same rack as the system. Placement of these other products must not result in moving the server chassis. All racks in the same order must be the same height and width.

The system is supported in the HPE 600 mm and 800 mm series racks and the HPE D-rack with a standard rack door.

Configuration Rules

The chassis is the basic building block.

• – A single system can be supported in 1-Chassis to 4-Chassis configurations. An embedded Rack Management Controller is included with Base chassis.
• – Each system starts with one (1) Base Chassis. Up to three (3) Expansion Chassis can be added to scale the system.
• – There are single phase and three phase power distribution options.
• – All processors within a chassis, partition or system must be of the same processor type
• – Default boot slots dependent on what I/O selection is made and slot is populated with a bootable PCIe card

  • 6-slot: 1, 2, 6
  • 12-slot:3,1,11
  • 16-slot:3,1,11
• – One (1) Additional/Customizable boot slot available for a total of four (4) boot slots

Racking Choices

HPE Compute Scale-up Server 3250 can be racked in many of the HPE G2 Enterprise Series and Advanced Series racks, and the HPE D-Rack. Complete ordering rules can be found in the HPE Compute Scale-up Server 3250 menu and in the ordering and configuration tools.

The HPE Compute Scale-up Server 3250 can also be rack mounted in 3rd party rack. Specific rules and guidelines for this are available here:

HPE Compute Scale-up Server 3250 Rack Planning Checklist

HPE Compute Scale-up Server 3250 Rackmounting Guide

The following racks are supported with HPE Compute Scale-up Server 3250—refer to the server menu for ordering and configuration rules:

• – HPE 22U 600x1075mm Adv G2 Shock Rack
• – HPE 36U 600x1075mm Adv G2 Kit Shock Rack
• – HPE 42U 600x1200mm Adv G2 Kit Shock Rack
• – HPE 42U 600x1075mm Adv G2 Kit Shock Rack
• – HPE 42U 600x1075 Ent G2 Shock Rack
• – HPE 42U 600x1200 Ent G2 Shock Rack
• – HPE 42U 800x1075 Ent G2 Shock Rack
• – HPE 42U 800x1200 Ent G2 Shock Rack
• – HPE 48U 600x1075 Ent G2 Shock Rack
• – HPE 48U 600x1200 Ent G2 Shock Rack
• – HPE 48U 800x1200 Ent G2 Shock Rack
• – HPE 42U 610 mm x 1156 mm D-Rack
• – HPE D-Rack 42U 610 mm x 1156 mm Extended

The default assumption is that chassis are loaded in the rack at the bottom. It is recommended that 1U is left below the bottom of the compute enclosure in the 42U rack to provide PDU and cabling exit space.

Supported configurations have the system located at the bottom of the rack with peripherals located above all chassis.

HPE G2 Enterprise Series Racks QuickSpecs:

https://www.HPE.com/h20195/v2/GetDocument.aspx?docname=a00002907enw

HPE G2 Advanced Series Racks QuickSpecs:

https://www.HPE.com/h20195/v2/GetDocument.aspx?docname=c05324689

Enlogic by nVent G3 Enterprise Power Distribution Units for HPE QuickSpecs:

https://h20195.www2.HPE.com/v2/GetDocument.aspx?docname=c05324691

HPE G2 Metered Power Distribution Units QuickSpecs:

https://www.HPE.com/psnow/doc/a00002908enw.pdf?jumpid=in_pdp-psnow-qs

HPE Destination Rack (D-Rack)

The HPE D-Rack is available for HPE Compute Scale-up Server 3250 in two models:

• – HPE 42U 610 mm x 1156 mm D-Rack (H7C27A)
• – HPE D-Rack 42U 610 mm x 1156 mm Extended (Q2T97A) The extended rack includes a 2U extension for a total of 44 rack units (44U).

HPE Power Advisor

The HPE Power Advisor is a tool provided by Hewlett-Packard to assist in the estimation of power consumption at a system, rack, and multi-rack level.

Available at: https://poweradvisorext.it.HPE.com/?Page=Index

Processor Support

HPE Compute Scale-up Server 3250 systems support Intel® Xeon® 6th Generation specified in the following table.

Notes:

– Maximum memory per socket includes DDR5 memory within the system

– Maximum memory is 4 TB per socket

– Minimum OS levels are required for support of some processors and options

Processor Mixing Support

Governing rules for mixing processors are as follows:

• – No mixing of processor types within the same system

DDR5 Memory Support

Systems will use DDR5 DIMM technology.

The following DIMMs are supported in the chassis:

• – 64 GB 2Rx4 DDR5-6400R Registered DIMM
• – 96 GB 2Rx4 DDR5-6400R Registered DIMM
• – 128 GB 2Rx4 DDR5-6400R Registered DIMM
• – 256 GB 4Rx4 DDR5-6400R Registered DIMM

Notes:

– Only DIMMs that Hewlett Packard Enterprise has qualified are supported.

– Each chassis supports up to 64 DIMMs. This breaks down to sixteen DIMMs per socket, 8 memory channels. General memory configuration rules:

• For best performance, the amount of memory on each chassis within the system should be the same.
• Must use the same amount of memory on each processor module within a partition.

Notes:

– * Indicates one or two tethered super-caps are mounted in center DVD bay

– # If more than one controller, this indicates the primary boot controller.

Notes:

– The speeds above are based on storage interface data transfer rates and are NOT measured performance.

– When multiple controller cards are used, there will be multiple RAID groups because drives on different controllers can't be in the same RAID group.

– When multiple controller cards are used, the first controller listed is the default boot controller.

– RAID 0, 1, 10, 5, 50, 6, and 60 supported

Security

HPE Compute Scale-up Server 3250 provides important architectural security features not found in other industry standard servers. HPE Compute Scale-up Server 3250 Mission Critical focus provides unique customer benefits without exposure to many vulnerabilities found in common server firmware.

• – UEFI Secure Boot and Secure Start support, Enabled by default
• – Immutable HPE iLO ASIC Silicon Root of Trust
• – PQC (Post Quantum Cryptography) compliant root of trust
• – IDevID - Factory provisioned server identity
• – Auto secure FW recovery. RMC & PDHC watchdogs are fully enabled allowing RMC and PDHC to automatically reboot to recover from unexpected hangs.
• – Tamper-free updates – components digitally signed and verified
• – Components are dual-signed with RSA and PQC cryptos
• – Secure Recovery – recover critical firmware to known a good state on detection of compromised firmware (per NIST SP800-193)
• – Firmware rollback protection for critical vulnerability fixes
• – System TPM (Trusted Platform Module) 2.0 tamper-resistant and soldered down for all systems
• – Manageability TPM for RMC
• – Manageability Flash Encryption
• – Chassis Intrusion detection standard
• – FW update protected by RMC Admin authentication
• – Firewalled Manageability Network
• – Secure Out-of-Box passwords
• – RMC uses CNSA 1.0 by default
• – Directory access control (LDAP/Active Directory)
• – Alternatives to PXE (Directed LAN Boot, HTTP Boot)
• – Ability to disable HOST and RMC USB ports
• – Ability to enable Total Memory Encryption (TME)

Platform Management

The HPE Compute Scale-up Server 3250 has a built-in and always available platform management system. By integrating the management into the server platform, Hewlett Packard Enterprise ensures that every HPE Compute Scale-up Server 3250 comes with the full set of management features and simplifies the task of integrating HPE Compute Scale-up Server 3250 into the data center. The purpose of the HPE Compute Scale-up Server 3250 management system is to:

• – Provide built-in tools to manage hardware and provide mission-critical system availability (inventory, monitor, diagnose, configure, maintain, and self-healing)
• – Make it easier for users and applications to manage the system (inventory, start, stop, connect console, and so on)

The HPE Compute Scale-up Server 3250 manageability system provides a very powerful control point for the system, and the embedded Rack Management Controller (RMC) makes managing the HPE Compute Scale-up Server 3250 much easier by centralizing the control and building the management into the hardware and firmware of the system. It provides the following features:

• – Web GUI for RMC – Redfish driven. Covers:

  • System inventory, health, configuration
  • Launch vMedia, vKVM (HTML5)
  • Configure UEFI reboot, power
  • RMC security settings
  • RMC LAN settings
• – CLI for easy access to all RMC functions, providing potential scripting and power user convenience
• – Console, and console logs
• – Built-in Error Analysis Engine constantly monitors all system hardware, analyzes log and telemetry data, and determines corrective actions for highest system uptime (often performing corrective actions automatically)
• – HPE Compute Scale-up Server 3250 RMC will interface directly with the HPE Remote Support software for data center wide fault management visibility and tie-in to HPE support services, such as the Insight Online portal – Front Panel Management device functionality that provides Power Button, System status Indicator LEDs including Power, UID and RMC indicator.
• – The RMC accepts all valid update requests. Firmware updates will be done immediately if possible.

The HPE Compute Scale-up Server 3250 delivers system administration, control, and platform management both via a Web GUI, a programmable Redfish API, or a comprehensive and concise command-line interface. The Redfish® API can be used in many ways including:

• – Directly in simple scripts to obtain inventory and monitoring information
• – With HPE OneView for a graphical user interface, as well as manage multiple HPE systems concurrently in the datacenter at once
• – With Openstack Ironic for Provisioning the OS

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.

Ordering and Configuration

This section lists some of the steps required to configure a Factory Integrated Model. To ensure only valid configurations are ordered, Hewlett Packard Enterprise recommends the use of an HPE approved configurator. Contact your local sales representative for information on CTO product offerings and requirements.

Notes:

– Configure-to-order servers must start with a CTO Chassis

– FIO indicates that this option is a Factory Installable Option.

Additional Technical documentation may be found at: http://www.HPE.com/support/computescaleup3200-docs

Regulatory model numbers:

• − Chassis (S4S70A, S4S71A) RMN: CHPF-074