{"title":"DPU","description":null,"products":[{"product_id":"nvidia-bluefield-3-2x-200gbe-b3220-dpu","title":"NVIDIA BlueField-3 2x 200GbE B3220 DPU","description":"\u003cdiv class=\"product-section container jc-between-center is-flex-row\"\u003e\n\u003cdiv class=\"twelve columns medium-down--one-whole-NG product-section-extras quick-specs\"\u003e\n\u003cdiv class=\"one-whole columns quick-specs-data\"\u003e\n\u003cdiv class=\"hfour\"\u003eDPU Specs\u003c\/div\u003e\n\u003cul class=\"margin-empty margin-bottom\"\u003e\n\u003cli\u003e\n\u003cb\u003eType\u003c\/b\u003e: Data Processing Units (DPU)\u003c\/li\u003e\n\u003cli\u003e\n\u003cb\u003eNetworking\u003c\/b\u003e: 2x 200GbE\u003c\/li\u003e\n\u003cli\u003e\n\u003cb\u003eForm Factor\u003c\/b\u003e: FHHL\u003c\/li\u003e\n\u003cli\u003e\n\u003cb\u003eInterface\u003c\/b\u003e: PCIe Gen5 x16\u003c\/li\u003e\n\u003cli\u003e\n\u003cb\u003ePower Consumption\u003c\/b\u003e: 150W\u003c\/li\u003e\n\u003cli\u003e\n\u003cb\u003eCompatible Risers\u003c\/b\u003e: RC 3(Slot 1)\u003c\/li\u003e\n\u003c\/ul\u003e\n\u003c\/div\u003e\n\u003c\/div\u003e\n\u003cdiv class=\"four-NG medium-down--one-whole-NG product-section-extras ideal-for\"\u003e\n\u003cdiv class=\"hfour\"\u003ePart Number:\u003c\/div\u003e\n\u003cul\u003e\n\u003cli\u003e\n\u003c\/li\u003e\n\u003c\/ul\u003e\n\u003c\/div\u003e\n\u003c\/div\u003e\n\u003cdiv class=\"product-section container jc-between-center is-flex-row\"\u003e\n  \u003cdiv class=\"one-half medium-down--one-whole-NG\"\u003e\n  \u003cb\u003eWhat is a DPU?\u003c\/b\u003e\n\u003cp\u003e\nA Data Processing Unit (DPU) is a specialized processor designed to handle data movement, networking, storage, and security tasks in servers and data centers. It functions as a programmable SmartNIC with its own Arm CPU cores, hardware accelerators, and high-speed networking capabilities. DPUs offload heavy infrastructure workloads from the main CPU, freeing it up for application-level work. They are commonly used in modern cloud and AI environments to improve efficiency and performance at scale. Popular examples include NVIDIA’s BlueField series of DPUs.\n\u003c\/p\u003e\n  \u003c\/div\u003e\n    \u003cdiv class=\"one-half medium-down--one-whole-NG\"\u003e\n\n\u003cb\u003eWhy is a DPU different from a GPU?\u003c\/b\u003e\n\u003cp\u003e\nA DPU is different from a GPU because it is optimized for moving, securing, and processing data in transit (networking \u0026amp; storage), whereas a GPU is built for performing massive parallel mathematical computations such as AI training and graphics rendering.\n\u003c\/p\u003e\n    \u003c\/div\u003e\n\u003c\/div\u003e\n\u003chr\u003e\n\u003cb\u003eKey Differences Between DPU and GPU\u003c\/b\u003e\n\u003cdiv class=\"product-section container jc-between-center is-flex-row\"\u003e\n  \u003cdiv class=\"one-half medium-down--one-whole-NG\"\u003e\n\n\u003cul\u003e\n    \u003cli\u003e\n\u003cb\u003ePurpose:\u003c\/b\u003e DPU handles data movement, networking, encryption, and storage offloading;\u003c\/li\u003e\n    \u003cli\u003e\n\u003cb\u003eWorkload:\u003c\/b\u003e DPU excels at infrastructure tasks (packets, security, telemetry); \u003c\/li\u003e\n    \u003cli\u003e\n\u003cb\u003eLocation \u0026amp; Role:\u003c\/b\u003e DPU sits on the data path as a SmartNIC; \u003c\/li\u003e\n    \u003cli\u003e\n\u003cb\u003eEfficiency Goal:\u003c\/b\u003e DPU improves overall server efficiency by freeing CPU cycles; \u003c\/li\u003e\n\u003c\/ul\u003e\n    \u003c\/div\u003e\n    \u003cdiv class=\"one-half medium-down--one-whole-NG\"\u003e\n      \u003cul\u003e\n    \u003cli\u003e\n\u003cb\u003ePurpose:\u003c\/b\u003e GPU focuses on heavy parallel math and compute workloads.\u003c\/li\u003e\n    \u003cli\u003e\n\u003cb\u003eWorkload:\u003c\/b\u003e GPU excels at AI\/ML training, simulations, and rendering.\u003c\/li\u003e\n    \u003cli\u003e\n\u003cb\u003eLocation \u0026amp; Role:\u003c\/b\u003e GPU acts as a compute accelerator.\u003c\/li\u003e\n    \u003cli\u003e\n\u003cb\u003eEfficiency Goal:\u003c\/b\u003e GPU increases raw computational power.\u003c\/li\u003e\n\u003c\/ul\u003e\n  \u003c\/div\u003e\n   \u003c\/div\u003e","brand":"Nvidia","offers":[{"title":"Default Title","offer_id":52394935615800,"sku":null,"price":0.0,"currency_code":"USD","in_stock":false}]},{"product_id":"nvidia-bluefield-3-1x400-gbe-b3140h","title":"NVIDIA BlueField-3 1x 400 GbE B3140H","description":"\u003cdiv class=\"product-section container jc-between-center is-flex-row\"\u003e\n\u003cdiv class=\"twelve columns medium-down--one-whole-NG product-section-extras quick-specs\"\u003e\n\u003cdiv class=\"one-whole columns quick-specs-data\"\u003e\n\u003cdiv class=\"hfour\"\u003eDPU Specs\u003c\/div\u003e\n\u003cul class=\"margin-empty margin-bottom\"\u003e\n\u003cli\u003e\n\u003cb\u003eType\u003c\/b\u003e: Data Processing Units (DPU)\u003c\/li\u003e\n\u003cli\u003e\n\u003cb\u003eNetworking\u003c\/b\u003e: 1x 400GbE\u003c\/li\u003e\n\u003cli\u003e\n\u003cb\u003eForm Factor\u003c\/b\u003e: FHHL\u003c\/li\u003e\n\u003cli\u003e\n\u003cb\u003eInterface\u003c\/b\u003e: PCIe Gen5 x16\u003c\/li\u003e\n\u003cli\u003e\n\u003cb\u003ePower Consumption\u003c\/b\u003e: 75W\u003c\/li\u003e\n\u003cli\u003e\n\u003cb\u003eCompatible Risers\u003c\/b\u003e: N\/A\u003c\/li\u003e\n\u003c\/ul\u003e\n\u003c\/div\u003e\n\u003c\/div\u003e\n\u003cdiv class=\"four-NG medium-down--one-whole-NG product-section-extras ideal-for\"\u003e\n\u003cdiv class=\"hfour\"\u003ePart Number:\u003c\/div\u003e\n\u003cul\u003e\n\u003cli\u003e\n\u003c\/li\u003e\n\u003c\/ul\u003e\n\u003c\/div\u003e\n\u003c\/div\u003e\n\u003cdiv class=\"product-section container jc-between-center is-flex-row\"\u003e\n\u003cdiv class=\"one-half medium-down--one-whole-NG\"\u003e\n\u003cb\u003eWhat is a DPU?\u003c\/b\u003e\n\u003cp\u003eA Data Processing Unit (DPU) is a specialized processor designed to handle data movement, networking, storage, and security tasks in servers and data centers. It functions as a programmable SmartNIC with its own Arm CPU cores, hardware accelerators, and high-speed networking capabilities. DPUs offload heavy infrastructure workloads from the main CPU, freeing it up for application-level work. They are commonly used in modern cloud and AI environments to improve efficiency and performance at scale. Popular examples include NVIDIA’s BlueField series of DPUs.\u003c\/p\u003e\n\u003c\/div\u003e\n\u003cdiv class=\"one-half medium-down--one-whole-NG\"\u003e\n\u003cb\u003eWhy is a DPU different from a GPU?\u003c\/b\u003e\n\u003cp\u003eA DPU is different from a GPU because it is optimized for moving, securing, and processing data in transit (networking \u0026amp; storage), whereas a GPU is built for performing massive parallel mathematical computations such as AI training and graphics rendering.\u003c\/p\u003e\n\u003c\/div\u003e\n\u003c\/div\u003e\n\u003chr\u003e\n\u003cp\u003e\u003cb\u003eKey Differences Between DPU and GPU\u003c\/b\u003e\u003c\/p\u003e\n\u003cdiv class=\"product-section container jc-between-center is-flex-row\"\u003e\n\u003cdiv class=\"one-half medium-down--one-whole-NG\"\u003e\n\u003cul\u003e\n\u003cli\u003e\n\u003cb\u003ePurpose:\u003c\/b\u003e DPU handles data movement, networking, encryption, and storage offloading;\u003c\/li\u003e\n\u003cli\u003e\n\u003cb\u003eWorkload:\u003c\/b\u003e DPU excels at infrastructure tasks (packets, security, telemetry);\u003c\/li\u003e\n\u003cli\u003e\n\u003cb\u003eLocation \u0026amp; Role:\u003c\/b\u003e DPU sits on the data path as a SmartNIC;\u003c\/li\u003e\n\u003cli\u003e\n\u003cb\u003eEfficiency Goal:\u003c\/b\u003e DPU improves overall server efficiency by freeing CPU cycles;\u003c\/li\u003e\n\u003c\/ul\u003e\n\u003c\/div\u003e\n\u003cdiv class=\"one-half medium-down--one-whole-NG\"\u003e\n\u003cul\u003e\n\u003cli\u003e\n\u003cb\u003ePurpose:\u003c\/b\u003e GPU focuses on heavy parallel math and compute workloads.\u003c\/li\u003e\n\u003cli\u003e\n\u003cb\u003eWorkload:\u003c\/b\u003e GPU excels at AI\/ML training, simulations, and rendering.\u003c\/li\u003e\n\u003cli\u003e\n\u003cb\u003eLocation \u0026amp; Role:\u003c\/b\u003e GPU acts as a compute accelerator.\u003c\/li\u003e\n\u003cli\u003e\n\u003cb\u003eEfficiency Goal:\u003c\/b\u003e GPU increases raw computational power.\u003c\/li\u003e\n\u003c\/ul\u003e\n\u003c\/div\u003e\n\u003c\/div\u003e","brand":"Nvidia","offers":[{"title":"Default Title","offer_id":52394946199864,"sku":null,"price":0.0,"currency_code":"USD","in_stock":false}]},{"product_id":"del-nvidia-bluefield-3-b3220-dpu","title":"NVIDIA BlueField-3 2x 200GbE B3220 DPU","description":"\u003cdiv class=\"eight medium-down--one-whole-NG \"\u003e\n\u003cb\u003eCTO Item\u003c\/b\u003e\u003cp\u003eThis part is strictly for configuring Dell servers. If you have landed on this page and you're looking for a specific part number. Please send us a \"Request a Quote\" with the part number you need. \u003c\/p\u003e\n\u003c\/div\u003e","brand":"Dell CTO","offers":[{"title":"Default Title","offer_id":52401177985336,"sku":null,"price":0.0,"currency_code":"USD","in_stock":false}]},{"product_id":"nvidia-bluefield-3-1x-400-gbe-b3140h","title":"NVIDIA BlueField-3 1x 400 GbE B3140H","description":"\u003cdiv class=\"eight medium-down--one-whole-NG \"\u003e\n\u003cb\u003eCTO Item\u003c\/b\u003e\u003cp\u003eThis part is strictly for configuring Dell servers. If you have landed on this page and you're looking for a specific part number. Please send us a \"Request a Quote\" with the part number you need. \u003c\/p\u003e\n\u003c\/div\u003e","brand":"Dell CTO","offers":[{"title":"Default Title","offer_id":52405643706680,"sku":null,"price":0.0,"currency_code":"USD","in_stock":false}]}],"url":"https:\/\/expresscomputersystems.com\/collections\/dpu.oembed","provider":"Express Computer Systems","version":"1.0","type":"link"}