{"id":21066,"date":"2026-01-23T15:16:22","date_gmt":"2026-01-23T06:16:22","guid":{"rendered":"https:\/\/aireviewirush.com\/?p=21066"},"modified":"2026-01-23T15:16:22","modified_gmt":"2026-01-23T06:16:22","slug":"accelerating-ethernet-native-ai-clusters-with-intel-gaudi-3-ai-accelerators-and-cisco-nexus-9000","status":"publish","type":"post","link":"https:\/\/aireviewirush.com\/?p=21066","title":{"rendered":"Accelerating Ethernet-Native AI Clusters with Intel\u00ae Gaudi\u00ae 3 AI Accelerators and Cisco Nexus 9000"},"content":{"rendered":"<p> <br \/>\n<\/p>\n<div>\n<p>Fashionable enterprises face important infrastructure challenges as giant language fashions (LLMs) require processing and transferring large volumes of information for each coaching and inference. With even essentially the most superior processors restricted by the capabilities of their supporting infrastructure, the necessity for sturdy, high-bandwidth networking has develop into crucial. For organizations aiming to make the most of high-performance AI workloads effectively, a scalable, low-latency community spine is essential to maximizing accelerator utilization and minimizing pricey, idle assets.<\/p>\n<div id=\"ez-toc-container\" class=\"ez-toc-v2_0_53 counter-hierarchy ez-toc-counter ez-toc-grey ez-toc-container-direction\">\n<div class=\"ez-toc-title-container\">\n<p class=\"ez-toc-title \" >Table of Contents<\/p>\n<span class=\"ez-toc-title-toggle\"><a href=\"#\" class=\"ez-toc-pull-right ez-toc-btn ez-toc-btn-xs ez-toc-btn-default ez-toc-toggle\" aria-label=\"Toggle Table of Content\" role=\"button\"><label for=\"item-69e80e260e6f5\" ><span class=\"\"><span style=\"display:none;\">Toggle<\/span><span class=\"ez-toc-icon-toggle-span\"><svg style=\"fill: #999;color:#999\" xmlns=\"http:\/\/www.w3.org\/2000\/svg\" class=\"list-377408\" width=\"20px\" height=\"20px\" viewBox=\"0 0 24 24\" fill=\"none\"><path d=\"M6 6H4v2h2V6zm14 0H8v2h12V6zM4 11h2v2H4v-2zm16 0H8v2h12v-2zM4 16h2v2H4v-2zm16 0H8v2h12v-2z\" fill=\"currentColor\"><\/path><\/svg><svg style=\"fill: #999;color:#999\" class=\"arrow-unsorted-368013\" xmlns=\"http:\/\/www.w3.org\/2000\/svg\" width=\"10px\" height=\"10px\" viewBox=\"0 0 24 24\" version=\"1.2\" baseProfile=\"tiny\"><path d=\"M18.2 9.3l-6.2-6.3-6.2 6.3c-.2.2-.3.4-.3.7s.1.5.3.7c.2.2.4.3.7.3h11c.3 0 .5-.1.7-.3.2-.2.3-.5.3-.7s-.1-.5-.3-.7zM5.8 14.7l6.2 6.3 6.2-6.3c.2-.2.3-.5.3-.7s-.1-.5-.3-.7c-.2-.2-.4-.3-.7-.3h-11c-.3 0-.5.1-.7.3-.2.2-.3.5-.3.7s.1.5.3.7z\"\/><\/svg><\/span><\/span><\/label><input aria-label=\"Toggle\" aria-label=\"item-69e80e260e6f5\"  type=\"checkbox\" id=\"item-69e80e260e6f5\"><\/a><\/span><\/div>\n<nav><ul class='ez-toc-list ez-toc-list-level-1 ' ><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-1\" href=\"https:\/\/aireviewirush.com\/?p=21066\/#Cisco_Nexus_9000_Sequence_Switches_for_AIML_workloads\" title=\"Cisco Nexus 9000 Sequence Switches for AI\/ML workloads\">Cisco Nexus 9000 Sequence Switches for AI\/ML workloads<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-2\" href=\"https:\/\/aireviewirush.com\/?p=21066\/#An_built-in_answer_with_excessive_efficiency_scalability_and_openness\" title=\"An built-in answer with excessive efficiency, scalability, and openness\">An built-in answer with excessive efficiency, scalability, and openness<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-3\" href=\"https:\/\/aireviewirush.com\/?p=21066\/#Cisco_Clever_Packet_Circulation_to_optimize_AI_site_visitors\" title=\"Cisco Clever Packet Circulation to optimize AI site visitors\">Cisco Clever Packet Circulation to optimize AI site visitors<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-4\" href=\"https:\/\/aireviewirush.com\/?p=21066\/#Unified_operations_with_Nexus_Dashboard\" title=\"Unified operations with Nexus Dashboard\">Unified operations with Nexus Dashboard<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-5\" href=\"https:\/\/aireviewirush.com\/?p=21066\/#A_scalable_compliant_future-ready_infrastructure\" title=\"A scalable, compliant, future-ready infrastructure\">A scalable, compliant, future-ready infrastructure<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-6\" href=\"https:\/\/aireviewirush.com\/?p=21066\/#For_extra_info_or_to_guage_how_this_reference_structure_will_be_tailor-made_to_your_group%E2%80%99s_wants_see_specs_for_Cisco_Nexus_9300_Sequence_Switches_and_Intel_Gaudi_3_AI_accelerators\" title=\"For extra info or to guage how this reference structure will be tailor-made to your group\u2019s wants, see specs for Cisco Nexus 9300 Sequence Switches and Intel Gaudi 3 AI accelerators.\">For extra info or to guage how this reference structure will be tailor-made to your group\u2019s wants, see specs for Cisco Nexus 9300 Sequence Switches and Intel Gaudi 3 AI accelerators.<\/a><\/li><\/ul><\/nav><\/div>\n<h2><span class=\"ez-toc-section\" id=\"Cisco_Nexus_9000_Sequence_Switches_for_AIML_workloads\"><\/span>Cisco Nexus 9000 Sequence Switches for AI\/ML workloads<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>Cisco Nexus 9000 Sequence Switches ship the high-radix, low-latency switching cloth that AI\/ML workloads demand. For Intel\u00ae Gaudi\u00ae 3 AI accelerator<sup>1<\/sup> deployments, Cisco has validated particular Nexus 9000 switches and configurations to make sure optimum efficiency.<\/p>\n<p>The Nexus 9364E-SG2 (Determine 1), for instance, is the premier AI networking change from Cisco, powered by the Silicon One G200 ASIC. In a compact 2RU type issue, it delivers:<\/p>\n<ul>\n<li><strong>64 dense ports of 800 GbE<\/strong> (or 128 x 400 GbE \/ 256 x 200 GbE \/ 512 x 100 GbE by way of breakouts)<\/li>\n<li><strong>51.2 Tbps mixture bandwidth<\/strong> for non-blocking leaf-spine materials<\/li>\n<li><strong>256 MB shared on-die packet buffer<\/strong>, which is important for absorbing the synchronized site visitors bursts attribute of collective operations in distributed coaching<\/li>\n<li><strong>512 high-radix structure<\/strong> that reduces the variety of switching tiers required, decreasing latency and simplifying cloth design<\/li>\n<li><strong>Extremely Ethernet prepared:<\/strong> Cisco is a founding member of the Extremely Ethernet Consortium (UEC) and Nexus 9000 switches are forward-compatible with rising UEC specs<\/li>\n<\/ul>\n<figure id=\"attachment_484392\" aria-describedby=\"caption-attachment-484392\" style=\"width: 600px\" class=\"wp-caption aligncenter\"><img fetchpriority=\"high\" decoding=\"async\" class=\"lazy lazy-hidden wp-image-484392\" data-lazy-type=\"image\" src=\"https:\/\/blogs.cisco.com\/gcs\/ciscoblogs\/1\/2026\/01\/nexus-9000-blog-image-768x615.jpg\" alt=\"\" width=\"600\" height=\"480\"><noscript><img fetchpriority=\"high\" decoding=\"async\" class=\"wp-image-484392\" src=\"https:\/\/blogs.cisco.com\/gcs\/ciscoblogs\/1\/2026\/01\/nexus-9000-blog-image-768x615.jpg\" alt=\"\" width=\"600\" height=\"480\"><\/noscript><figcaption id=\"caption-attachment-484392\" class=\"wp-caption-text\">Determine 1. Cisco Nexus 9364E-SG2: Optimized for scalability and open connectivity, supporting Intel\u00ae\ufe0f Gaudi\u00ae\ufe0f 3 AI accelerator deployments<\/figcaption><\/figure>\n<p>The Intel Gaudi 3 AI accelerator addresses the necessity for scalable, open AI methods. It was designed to supply state-of-the-art knowledge heart efficiency for AI workloads, together with generative purposes like LLMs, diffusion fashions, and multimodal fashions. The Intel Gaudi 3 accelerator demonstrates important enhancements over earlier generations, delivering as much as 4x AI compute efficiency for Mind Floating Level 16-bit (BF16) workloads and a 1.5x enhance in reminiscence bandwidth in comparison with the Intel Gaudi 2 processor.<\/p>\n<p>A key differentiator is its networking infrastructure: every Intel Gaudi 3 AI accelerator integrates 24 x 200 GbE Ethernet ports, supporting large-scale system enlargement with normal Ethernet protocols. This method eliminates a reliance on proprietary networking applied sciences and offers 2x the networking bandwidth in comparison with the Intel Gaudi 2 accelerator, enabling organizations to construct clusters from a number of nodes to a number of thousand seamlessly.<\/p>\n<h2><span class=\"ez-toc-section\" id=\"An_built-in_answer_with_excessive_efficiency_scalability_and_openness\"><\/span>An built-in answer with excessive efficiency, scalability, and openness<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>Cisco Nexus 9364E-SG2 switches and OSFP-800G-DR8 transceivers are licensed to help Intel Gaudi 3 AI accelerators in scale-out configurations for LLM coaching, inference, and generative AI workloads.<\/p>\n<p>Key technical highlights of the validated structure embody:<\/p>\n<ul>\n<li><strong>Excessive-speed and non-blocking connectivity:<\/strong> 256 x 200 Gbps interfaces on Cisco Nexus 9364E-SG2 switches permit high-speed and non-blocking community design for interconnecting Intel Gaudi 3 accelerators<\/li>\n<li><strong>Lossless cloth:<\/strong> Full help for RDMA over Converged Ethernet model 2 (RoCEv2) with Precedence Circulation Management (PFC) prevents packet loss as a result of congestion, thereby bettering the completion instances of distributed jobs<\/li>\n<li><strong>Simplified operations:<\/strong> Nexus Dashboard permits configuring Intel Gaudi 3 AI accelerators for scale-out networks utilizing the built-in AI cloth kind. It additionally affords templates for additional customizations and a single operations platform for all networks accessing an AI cluster.<\/li>\n<\/ul>\n<h2><span class=\"ez-toc-section\" id=\"Cisco_Clever_Packet_Circulation_to_optimize_AI_site_visitors\"><\/span>Cisco Clever Packet Circulation to optimize AI site visitors<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>AI workloads generate site visitors patterns not like conventional enterprise purposes\u2014large, synchronized bursts, \u201celephant flows,\u201d and steady GPU-to-GPU communication that may overwhelm standard networking approaches. Cisco addresses these challenges with <a href=\"https:\/\/www.cisco.com\/c\/en\/us\/products\/collateral\/ios-nx-os-software\/nx-os-software\/intelligent-packet-flow-solution-overview.html\" target=\"_blank\" rel=\"noopener\">Cisco Clever Packet Circulation<\/a>, a sophisticated site visitors administration framework constructed into NX-OS.<\/p>\n<p>Clever Packet Circulation incorporates a number of load balancing methods designed for AI materials:<\/p>\n<ul>\n<li><strong>Dynamic load balancing (flowlet-based):<\/strong> Actual-time site visitors distribution primarily based on hyperlink utilization telemetry<\/li>\n<li><strong>Per-packet load balancing:<\/strong> Packet spraying throughout a number of paths for optimum throughput effectivity<\/li>\n<li><strong>Weighted Value Multipath (WCMP):<\/strong> Clever path weighting mixed with Dynamic Load Balancing (DLB) for uneven topologies<\/li>\n<li><strong>Coverage-based load balancing:<\/strong> Assigns particular traffic-handling methods to combined workloads primarily based on ACLs, DHCP markings, or RoCEv2 headers, creating custom-fit effectivity for numerous wants<\/li>\n<\/ul>\n<p>These capabilities work collectively to attenuate <em>job completion time<\/em>\u2014the important metric that determines how rapidly your AI fashions prepare and the way effectively your inference pipelines reply.<\/p>\n<h2><span class=\"ez-toc-section\" id=\"Unified_operations_with_Nexus_Dashboard\"><\/span>Unified operations with Nexus Dashboard<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>Deploying and working AI infrastructure at scale requires visibility and different options that go far past conventional community monitoring. Cisco Nexus Dashboard serves because the centralized administration platform for AI materials, offering end-to-end RoCEv2 visibility and built-in templates for AI cloth provisioning.<\/p>\n<p>Key Cisco Nexus Dashboard operational capabilities embody:<\/p>\n<ul>\n<li><strong>Congestion analytics:<\/strong> Actual-time congestion scoring, Precedence Circulation Management and Specific Congestion Notification (PFC\/ECN) statistics, and microburst detection<\/li>\n<li><strong>Anomaly detection:<\/strong> Proactive identification of efficiency bottlenecks with instructed remediation<\/li>\n<li><strong>AI job observability:<\/strong> Finish-to-end visibility into AI workloads from community to GPUs<\/li>\n<li><strong>Sustainability insights:<\/strong> Vitality consumption monitoring and optimization suggestions<\/li>\n<\/ul>\n<p><em><em>\u201cAI at scale calls for each compute effectivity and high-performance AI networking cloth. Intel\u00ae Gaudi\u00ae 3 AI accelerator mixed with Cisco Nexus 9000 switching delivers an optimized, open answer that lets prospects construct at scale LLM inference clusters with uncompromising cost-efficient efficiency.\u201d<br \/><\/em><\/em>\u2014Anil Nanduri, VP, AI Get-to-Market &amp; Product Administration, Intel<\/p>\n<h2><span class=\"ez-toc-section\" id=\"A_scalable_compliant_future-ready_infrastructure\"><\/span>A scalable, compliant, future-ready infrastructure<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>Cisco Nexus 9000 switches paired with Intel Gaudi 3 AI accelerators present enterprises with a safe, open, and future-ready community and compute setting. This mixture of applied sciences allows organizations to deploy scalable, high-performance AI clusters that meet each present and rising workload necessities.<\/p>\n<p>\u00a0<\/p>\n<blockquote>\n<h2 style=\"text-align: center;\"><span class=\"ez-toc-section\" id=\"For_extra_info_or_to_guage_how_this_reference_structure_will_be_tailor-made_to_your_group%E2%80%99s_wants_see_specs_for_Cisco_Nexus_9300_Sequence_Switches_and_Intel_Gaudi_3_AI_accelerators\"><\/span>For extra info or to guage how this reference structure will be tailor-made to your group\u2019s wants, see specs for <a href=\"https:\/\/www.cisco.com\/site\/us\/en\/products\/networking\/cloud-networking-switches\/nexus-9300-series-switches\/index.html\" target=\"_blank\" rel=\"noopener\">Cisco Nexus 9300 Sequence Switches<\/a> and <a href=\"https:\/\/www.intel.com\/content\/www\/us\/en\/products\/details\/processors\/ai-accelerators\/gaudi.html\" target=\"_blank\" rel=\"noopener\">Intel Gaudi 3 AI accelerators<\/a>.<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<\/blockquote>\n<p>Extra assets:<\/p>\n<p><sup>1<\/sup><span style=\"font-size: 10px;\">\u00a0Intel, the Intel emblem, and Gaudi are logos of Intel Company or its subsidiaries.<\/span><\/p>\n<\/p><\/div>\n\n","protected":false},"excerpt":{"rendered":"<p>Fashionable enterprises face important infrastructure challenges as giant language fashions (LLMs) require processing and transferring large volumes of information for each coaching and inference. With even essentially the most superior processors restricted by the capabilities of their supporting infrastructure, the necessity for sturdy, high-bandwidth networking has develop into crucial. For organizations aiming to make the [&hellip;]<\/p>\n","protected":false},"author":1,"featured_media":21068,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[22],"tags":[],"class_list":{"0":"post-21066","1":"post","2":"type-post","3":"status-publish","4":"format-standard","5":"has-post-thumbnail","7":"category-iot"},"_links":{"self":[{"href":"https:\/\/aireviewirush.com\/index.php?rest_route=\/wp\/v2\/posts\/21066","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/aireviewirush.com\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/aireviewirush.com\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/aireviewirush.com\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/aireviewirush.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=21066"}],"version-history":[{"count":1,"href":"https:\/\/aireviewirush.com\/index.php?rest_route=\/wp\/v2\/posts\/21066\/revisions"}],"predecessor-version":[{"id":21067,"href":"https:\/\/aireviewirush.com\/index.php?rest_route=\/wp\/v2\/posts\/21066\/revisions\/21067"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/aireviewirush.com\/index.php?rest_route=\/wp\/v2\/media\/21068"}],"wp:attachment":[{"href":"https:\/\/aireviewirush.com\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=21066"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/aireviewirush.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=21066"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/aireviewirush.com\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=21066"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}