image image image image image

On This Page

The NVIDIA® BlueField®-3 data processing unit (DPU) is the 3rd-generation data center infrastructure-on-a-chip that enables organizations to build software-defined, hardware-accelerated IT infrastructures from cloud to core data center to edge. With 400Gb/s Ethernet or NDR 400Gb/s InfiniBand network connectivity, BlueField-3 DPU offloads, accelerates, and isolates software-defined networking, storage, security, and management functions in ways that profoundly improve data center performance, efficiency, and security. Providing powerful computing, and a broad range of programmable acceleration engines in the I/O path, BlueField-3 is perfectly positioned to address the infrastructure needs of the most demanding applications, while delivering full software backward compatibility through the NVIDIA DOCA™ software framework. 
BlueField-3 DPUs transform traditional computing environments into secure and accelerated virtual private clouds, allowing organizations to run application workloads in secure, multi-tenant environments. Decoupling data center infrastructure from business applications, BlueField-3 enhances data center security, streamlines operations, and reduces the total cost of ownership. Featuring NVIDIA’s in-network computing technology, BlueField-3 enables the next generation of supercomputing platforms, delivering optimal bare-metal performance and native support for multi-node tenant isolation.

System Requirements

PCI Express slot

In PCIe x16 Configuration
PCIe Gen 5.0 (32GT/s) through x16 edge connector.

In PCIe x16 Extension Option - Switch DSP (Data Stream Port)

  • PCIe Gen 5.0 SERDES @32GT/s through edge connector
  • PCIe Gen 5.0 SERDES @32GT/s through PCIe Auxiliary Connection Card
System Power SupplyMinimum 75W or greater system power supply for all cards.

P-Series DPUs with PCIe Gen 5.0 x16 require a supplementary 8-pin ATX power supply connectivity available through the external power supply connector.

NOTE: The power supply harness is not included in the package. 

To power-up the DPU, power the ATX power supply and the PCIe golden fingers simultaneously. Failure to do so may harm the DPU.

Operating SystemBlueField-3 DPU is shipped with Ubuntu – a Linux commercial operating system – which includes the NVIDIA OFED stack (MLNX_OFED), and is capable of running all customer-based Linux applications seamlessly. For more information, please refer to the DOCA SDK documentation or NVIDIA BlueField-3 Software User Manual. 
  • Interoperable with 1/10/25/40/50/100/200/400 Gb/s Ethernet switches and SDR/DDR/EDR/HDR100/HDR/NDR200/NDR InfiniBand switches
  • Passive copper cable with ESD protection
  • Powered connectors for optical and active cable support

For detailed information, see Specifications.

Package Contents  

Prior to unpacking your DPU, it is important to make sure your server meets all the system requirements listed above for a smooth installation. Be sure to inspect each piece of equipment shipped in the packing box. If anything is missing or damaged, contact your reseller.

Card Package

For FHHL P-Series DPUs, you need an 8-pin PCIe external power cable to activate the card. The cable is not included in the package. For further details, please refer to External PCIe Power Supply Connector.

Card1x BlueField-3 DPU

1x tall bracket (shipped assembled on the card)

Accessories Kit 

This is an optional accessories kit used for debugging purposes and can be ordered separately. 

Kit OPNContents

4-pin USB to female USB Type-A cable
20-pin shrouded connector to USB Type-A cable

PCIe Auxiliary Card Package 

This is an optional kit which applies to following OPNs: 900-9D3B6-00CV-AA0, 900-9D3B6-00SV-AA0, 900-9D3B6-00CC-AA0, 900-9D3B6-00SC-AA0, 900-9D3B6-00CN-AB0 and 900-9D3B6-00SN-AB0.

The PCIe auxiliary kit can be purchased separately to operate selected DPUs in a dual-socket server. For package contents, refer to PCIe Auxiliary Card Kit.

Features and Benefits

This section describes hardware features and capabilities. Please refer to the relevant driver and/or firmware release notes for feature availability.


InfiniBand Architecture Specification v1.5 compliant

BlueField-3 DPU delivers low latency, high bandwidth, and computing efficiency for high-performance computing (HPC), artificial intelligence (AI), and hyperscale cloud data centers applications.
BlueField-3 DPU is InfiniBand Architecture Specification v1.5 compliant.

InfiniBand Network Protocols and Rates: 

ProtocolStandardRate (Gb/s)Comments
4x Port
(4 Lanes)
2x Ports
(2 Lanes)
NDR/NDR200IBTA Vol2 1.5425212.5PAM4 256b/257b encoding and RS-FEC
HDR/HDR100IBTA Vol2 1.4212.5106.25PAM4 256b/257b encoding and RS-FEC
EDRIBTA Vol2 1.3.1103.12551.5625NRZ 64b/66b encoding
FDRIBTA Vol2 1.256.25N/ANRZ 64b/66b encoding

Up to 400 Gigabit Ethernet

BlueField-3 DPU complies with the following IEEE 802.3 standards:
400GbE / 200GbE / 100GbE / 50GbE / 40GbE / 25GbE / 10GbE


MAC Rate

IEEE802.3ck400/200/100 Gigabit Ethernet
(Include ETC enhancement)


400/200/100 Gigabit Ethernet
(Include ETC enhancement)
IEEE 802.3bj
IEEE 802.3bm
100 Gigabit Ethernet
IEEE 802.3by
Ethernet Consortium25
50/25 Gigabit Ethernet
IEEE 802.3ba40 Gigabit Ethernet
IEEE 802.3ae10 Gigabit Ethernet
IEEE 802.3cb2.5/5 Gigabit Ethernet
(For 2.5: support only 2.5 x1000BASE-X)
IEEE 802.3ap Based on auto-negotiation and KR startup
IEEE 802.3ad
IEEE 802.1AX 
Link Aggregation
IEEE 802.1Q
IEEE 802.1P VLAN tags and priority

IEEE 802.1Qau (QCN)
Congestion Notification
IEEE 802.1Qaz (ETS)
EEE 802.1Qbb (PFC)
IEEE 802.1Qbg
IEEE 1588v2
IEEE 802.1AE (MACSec) 
Jumbo frame support (9.6KB)

On-board Memory
  • Quad SPI NOR FLASH - includes 256Mbit for Firmware image.
  • UVPS EEPROM - includes 2Mbit.

  • FRU EEPROM - Stores the parameters and personality of the card. The EEPROM capacity is 128Kbit. FRU I2C address is (0x50) and is accessible through the PCIe SMBus.

  • DPU_BMC Flashes: 

    • 2x 64MByte for BMC Image
    • 512MByte for Config Data
  • eMMC pSLC 40GB with 30K Write Cycles eMMC for SoC BIOS and OS.

  • SSD (onboard BGA) 128GByte for user SoC OS, logs and application SW.

  • DDR5 SDRAM - 16GB/32GB @5600MT/s single/dual-channel DDR5 SDRAM memory. Solder down on-board. 128bit + 16bit ECC

BlueField-3 ICThe NVIDIA BlueField-3 DPU integrates x8 / x16 Armv8.2+ A78 Hercules cores (64-bit) is interconnected by a coherent mesh network, one DRAM controller, an RDMA intelligent network adapter supporting up to 400Gb/s, an embedded PCIe switch with endpoint and root complex functionality, and up to 32 lanes of PCIe Gen 5.0.
Overlay NetworksIn order to better scale their networks, data center operators often create overlay networks that carry traffic from individual virtual machines over logical tunnels in encapsulated formats such as NVGRE and VXLAN. While this solves network scalability issues, it hides the TCP packet from the hardware offloading engines, placing higher loads on the host CPU. NVIDIA DPU effectively addresses this by providing advanced NVGRE and VXLAN hardware offloading engines that encapsulate and de-capsulate the overlay protocol.
RDMA and RDMA over Converged InfiniBand/Ethernet (RoCE)NVIDIA DPU, utilizing IBTA RDMA (Remote Data Memory Access) and RoCE (RDMA over Converged InfiniBand/Ethernet) technology, delivers low-latency and high-performance over InfiniBand/Ethernet networks. Leveraging data center bridging (DCB) capabilities as well as advanced congestion control hardware mechanisms, RoCE provides efficient low-latency RDMA services over Layer 2 and Layer 3 networks.
Quality of Service (QoS)Support for port-based Quality of Service enabling various application requirements for latency and SLA.
Storage Acceleration
  • A consolidated compute and storage network achieves significant cost-performance advantages over multi-fabric networks. Standard block and file access protocols can leverage RDMA for high-performance storage access: NVMe over Fabric offloads for the target machine
  • BlueField-3 DPU may operate as a co-processor offloading specific storage tasks from the host, isolating part of the storage media from the host, or enabling abstraction of software-defined storage logic using the NVIDIA BlueField-3 Arm cores. On the storage initiator side, NVIDIA BlueField-3 DPU can prove an efficient solution for hyper-converged systems to enable the host CPU to focus on compute while all the storage interface is handled through the Arm cores.
NVMe-oFNon-volatile Memory Express (NVMe) over Fabrics is a protocol for communicating block storage IO requests over RDMA to transfer data between a host computer and a target solid-state storage device or system over a network. NVIDIA BlueField-3 DPU may operate as a co-processor offloading specific storage tasks from the host using its powerful NVMe over Fabrics Offload accelerator.

NVIDIA DPU SR-IOV technology provides dedicated adapter resources and guaranteed isolation and protection for virtual machines (VM) within the server.

High-Performance Accelerations
  • Tag Matching and Rendezvous Offloads
  • Adaptive Routing on Reliable Transport
  • Burst Buffer Offloads for Background Checkpointing
GPU Direct

GPUDirect RDMA is a technology that provides a direct P2P (Peer-to-Peer) data path between the GPU Memory directly to/from the NVIDIA HCA devices. This provides a significant decrease in GPU-GPU communication latency and completely offloads the CPU, removing it from all GPU-GPU communications across the network. NVIDIA DPU uses high-speed DMA transfers to copy data between P2P devices resulting in more efficient system applications

IsolationBlueField-3 DPU functions as a “computer-in-front-of-a-computer,” unlocking unlimited opportunities for custom security applications on its Arm processors, fully isolated from the host’s CPU. In the event of a compromised host, BlueField-3 may detect/block malicious activities in real-time and at wire speed to prevent the attack from spreading further.
Cryptography Accelerations

From IPsec and TLS data-in-motion inline encryption to AES-XTS block-level data-at-rest encryption and public key acceleration, BlueField-3 DPU hardware-based accelerations offload the crypto operations and free up the CPU, reducing latency and enabling scalable crypto solutions. BlueField-3 “host-unaware” solutions may transmit and receive data, while BlueField-3 acts as a bump-in-the-wire for crypto.

Securing Workloads BlueField-3 DPU accelerates connection tracking with its ASAP2 technology to enable stateful filtering on a per-connection basis. Moreover, BlueField-3 includes a Titan IC regular expression (RXP) acceleration engine supported by IDS/IPS tools to detect host introspection and Application Recognition (AR) in real-time.
Security Accelerators

A consolidated compute and network solution based on DPU achieves significant advantages over a centralized security server solution. Standard encryption protocols and security applications can leverage NVIDIA BlueField-3 compute capabilities and network offloads for security application solutions such as Layer4 Statefull Firewall.

Virtualized Cloud By leveraging BlueField-3 DPU virtualization offloads, data center administrators can benefit from better server utilization, allowing more virtual machines and more tenants on the same hardware, while reducing the TCO and power consumption
Out-of-Band ManagementThe NVIDIA BlueField-3 DPU incorporates a 1GbE RJ45 out-of-band port that allows the network operator to establish trust boundaries in accessing the management function to apply it to network resources. It can also be used to ensure management connectivity (including the ability to determine the status of any network component) independent of the status of other in-band network components.

Some BlueField-3 DPUs incorporate local NIC BMC (Baseboard Management Controller) hardware on the board. The BMC SoC (system on a chip) can utilize either shared or dedicated NICs for remote access. The BMC node enables remote power cycling, board environment monitoring, BlueField-3 chip temperature monitoring, board power and consumption monitoring, and individual interface resets. The BMC also supports the ability to push a bootstream to BlueField-3.
Having a trusted on-board BMC that is fully isolated for the host server ensures highest security for the DPU boards.