Mastodon

Gabriele SveltoNvidia has been doing a lot of useless stuff lately, but this is actually a big deal. I wonder what the latency looks like on these switches. Traditionally direct-attach copper has always been the preferred choice for low-latency applications, with optics used for longer connections where latency matters less. I'm curious if this is going to change that.<a href="https://www.techpowerup.com/334337/nvidia-commercializes-silicon-photonics-with-infiniband-and-ethernet-switches" rel="nofollow noopener" translate="no" target="_blank">https://www.techpowerup.com/334337/nvidia-commercializes-silicon-photonics-with-infiniband-and-ethernet-switches</a><a href="https://fosstodon.org/tags/Infiniband" class="mention hashtag" rel="nofollow noopener" target="_blank">#Infiniband</a> <a href="https://fosstodon.org/tags/Ethernet" class="mention hashtag" rel="nofollow noopener" target="_blank">#Ethernet</a>

Peter J. WelcherNew blog “How AI Ate My Blog on RoCEv2”. <a href="https://techfieldday.net/tags/PeterWelcher" class="mention hashtag" rel="nofollow noopener" target="_blank">#PeterWelcher</a> <a href="https://techfieldday.net/tags/CCIE1773" class="mention hashtag" rel="nofollow noopener" target="_blank">#CCIE1773</a> <a href="https://techfieldday.net/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#AI</a> <a href="https://techfieldday.net/tags/ECN" class="mention hashtag" rel="nofollow noopener" target="_blank">#ECN</a> <a href="https://techfieldday.net/tags/PFC" class="mention hashtag" rel="nofollow noopener" target="_blank">#PFC</a> <a href="https://techfieldday.net/tags/RoCEV2" class="mention hashtag" rel="nofollow noopener" target="_blank">#RoCEV2</a> <a href="https://techfieldday.net/tags/Infiniband" class="mention hashtag" rel="nofollow noopener" target="_blank">#Infiniband</a>. URL: <a href="https://www.linkedin.com/pulse/ai-ate-my-blog-rocev2-peter-welcher-uu9ue/" rel="nofollow noopener" translate="no" target="_blank">https://www.linkedin.com/pulse/ai-ate-my-blog-rocev2-peter-welcher-uu9ue/</a>.

Asterfusion🚀 RoCE vs. InfiniBand: The Game-Changing Data Center Switch Test Results Revealed! ⚡In AI and HPC networks, RoCE (RDMA over Converged Ethernet) and InfiniBand (IB) are often the go-to choices. Both offer low-latency, lossless transmission, but they come with key differences.🔍 InfiniBand: A mature, low-latency protocol with specialized hardware, but higher TCO due to single-vendor reliance.🔍 RoCEv2: More cost-effective, interoperable, and ideal for large-scale deployments like xAI’s AI cluster in Memphis!Which one fits your needs? See the full comparison! 🔥<a href="https://techhub.social/tags/RoCE" class="mention hashtag" rel="nofollow noopener" target="_blank">#RoCE</a> <a href="https://techhub.social/tags/InfiniBand" class="mention hashtag" rel="nofollow noopener" target="_blank">#InfiniBand</a> <a href="https://techhub.social/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#AI</a> <a href="https://techhub.social/tags/HPC" class="mention hashtag" rel="nofollow noopener" target="_blank">#HPC</a> <a href="https://techhub.social/tags/DataCenter" class="mention hashtag" rel="nofollow noopener" target="_blank">#DataCenter</a> <a href="https://techhub.social/tags/Networking" class="mention hashtag" rel="nofollow noopener" target="_blank">#Networking</a> <a href="https://techhub.social/tags/TechComparison" class="mention hashtag" rel="nofollow noopener" target="_blank">#TechComparison</a> <a href="https://techhub.social/tags/Ethernet" class="mention hashtag" rel="nofollow noopener" target="_blank">#Ethernet</a> <a href="https://techhub.social/tags/NetworkOptimization" class="mention hashtag" rel="nofollow noopener" target="_blank">#NetworkOptimization</a> <a href="https://cloudswit.ch/blogs/how-will-deepseek-shake-up-white-box-networking/" rel="nofollow noopener" translate="no" target="_blank">https://cloudswit.ch/blogs/how-will-deepseek-shake-up-white-box-networking/</a>

HabrНевероятная мощь NVIDIA GB200 NVL72: Внутри гиганта ИИ-вычисленийПривет, Хабр! Если вас всегда интересовало, как устроены по-настоящему производительные системы , вы попали по адресу. В сегодняшней статье мы расскажем, как Nvidia объединила сразу 72 ускорителя B200 в единый CUDA процессор GB200 NVL72 . Узнаем, как для создания эффективного интерконнекта используются технологии NVLink , Ethernet и Infiniband . Предметный разговор об аппаратной части уже ждет вас под кнопкой «Читать далее».<a href="https://habr.com/ru/companies/serverflow/articles/864314/" rel="nofollow noopener" translate="no" target="_blank">https://habr.com/ru/companies/serverflow/articles/864314/</a><a href="https://zhub.link/tags/%D1%81%D0%B5%D1%80%D0%B2%D0%B5%D1%80_%D1%84%D0%BB%D0%BE%D1%83" class="mention hashtag" rel="nofollow noopener" target="_blank">#сервер_флоу</a> <a href="https://zhub.link/tags/GB200_NVL72" class="mention hashtag" rel="nofollow noopener" target="_blank">#GB200_NVL72</a> <a href="https://zhub.link/tags/nvlink" class="mention hashtag" rel="nofollow noopener" target="_blank">#nvlink</a> <a href="https://zhub.link/tags/blackwell" class="mention hashtag" rel="nofollow noopener" target="_blank">#blackwell</a> <a href="https://zhub.link/tags/nvidia_grace" class="mention hashtag" rel="nofollow noopener" target="_blank">#nvidia_grace</a> <a href="https://zhub.link/tags/Nvidia_Superchip" class="mention hashtag" rel="nofollow noopener" target="_blank">#Nvidia_Superchip</a> <a href="https://zhub.link/tags/NVLink_Spine" class="mention hashtag" rel="nofollow noopener" target="_blank">#NVLink_Spine</a> <a href="https://zhub.link/tags/infiniband" class="mention hashtag" rel="nofollow noopener" target="_blank">#infiniband</a> <a href="https://zhub.link/tags/llm" class="mention hashtag" rel="nofollow noopener" target="_blank">#llm</a> <a href="https://zhub.link/tags/SeverFlow" class="mention hashtag" rel="nofollow noopener" target="_blank">#SeverFlow</a>

David GraylessVMEBus Switched Serial, commonly known as VXS, is an <a href="https://mastodon.social/tags/ANSI" class="mention hashtag" rel="tag">#ANSI</a> standard (ANSI/VITA 41) that improves the performance of standard parallel <a href="https://mastodon.social/tags/VMEbus" class="mention hashtag" rel="tag">#VMEbus</a> by enhancing it to support newer switched serial fabrics. The base specification (ANSI 41) defines all common elements of the standard, while "dot"-specifications (ANSI 41.n) define extensions which use specific serial fabrics (such as <a href="https://mastodon.social/tags/PCIExpress" class="mention hashtag" rel="tag">#PCIExpress</a>, <a href="https://mastodon.social/tags/RapidIO" class="mention hashtag" rel="tag">#RapidIO</a>, StarFabric from <a href="https://mastodon.social/tags/DolphinInterconnectSolutions" class="mention hashtag" rel="tag">#DolphinInterconnectSolutions</a> and <a href="https://mastodon.social/tags/InfiniBand" class="mention hashtag" rel="tag">#InfiniBand</a>) or additional functionality.

Mike PApparently if you push a wookie, you can expect to get a cookie in response.I'm not sure I'll be trying this one myself.<a href="https://mastodon.social/tags/rdma" class="mention hashtag" rel="tag">#rdma</a> <a href="https://mastodon.social/tags/infiniband" class="mention hashtag" rel="tag">#infiniband</a>

Benjamin Carr, Ph.D. 👨🏻‍💻🧬What If <a href="https://hachyderm.io/tags/OmniPath" class="mention hashtag" rel="nofollow noopener" target="_blank">#OmniPath</a> Morphs Into The Best <a href="https://hachyderm.io/tags/UltraEthernet" class="mention hashtag" rel="nofollow noopener" target="_blank">#UltraEthernet</a>? Many <a href="https://hachyderm.io/tags/HPC" class="mention hashtag" rel="nofollow noopener" target="_blank">#HPC</a> centers in <a href="https://hachyderm.io/tags/US" class="mention hashtag" rel="nofollow noopener" target="_blank">#US</a> – importantly <a href="https://hachyderm.io/tags/Sandia" class="mention hashtag" rel="nofollow noopener" target="_blank">#Sandia</a> and <a href="https://hachyderm.io/tags/LawrenceLivermore" class="mention hashtag" rel="nofollow noopener" target="_blank">#LawrenceLivermore</a> as well as the Texas Advanced Computing Center (<a href="https://hachyderm.io/tags/TACC" class="mention hashtag" rel="nofollow noopener" target="_blank">#TACC</a>) – wanted an alternative to <a href="https://hachyderm.io/tags/InfiniBand" class="mention hashtag" rel="nofollow noopener" target="_blank">#InfiniBand</a> or proprietary interconnects like <a href="https://hachyderm.io/tags/HPE" class="mention hashtag" rel="nofollow noopener" target="_blank">#HPE</a>/#Cray’s Slingshot, and they have been funding the redevelopment of Omni-Path. And now, <a href="https://hachyderm.io/tags/CornelisNetworks" class="mention hashtag" rel="nofollow noopener" target="_blank">#CornelisNetworks</a> is going to be intersecting its roadmap with Omni-Path switches and adapters with the <a href="https://hachyderm.io/tags/UEC" class="mention hashtag" rel="nofollow noopener" target="_blank">#UEC</a> roadmap. <a href="https://www.nextplatform.com/2024/06/26/what-if-omni-path-morphs-into-the-best-ultra-ethernet/" rel="nofollow noopener" translate="no" target="_blank">https://www.nextplatform.com/2024/06/26/what-if-omni-path-morphs-into-the-best-ultra-ethernet/</a>

aijobs.netHIRING: Principal GPU Capacity and Resource Management Engineer / US, CA, Santa Clara 💰 USD 272K+👉 <a href="https://ai-jobs.net/J326446/" rel="nofollow noopener" translate="no" target="_blank">https://ai-jobs.net/J326446/</a><a href="https://mstdn.social/tags/Ansible" class="mention hashtag" rel="nofollow noopener" target="_blank">#Ansible</a> <a href="https://mstdn.social/tags/ComputerScience" class="mention hashtag" rel="nofollow noopener" target="_blank">#ComputerScience</a> <a href="https://mstdn.social/tags/CUDA" class="mention hashtag" rel="nofollow noopener" target="_blank">#CUDA</a> <a href="https://mstdn.social/tags/DeepLearning" class="mention hashtag" rel="nofollow noopener" target="_blank">#DeepLearning</a> <a href="https://mstdn.social/tags/Engineering" class="mention hashtag" rel="nofollow noopener" target="_blank">#Engineering</a> <a href="https://mstdn.social/tags/GPU" class="mention hashtag" rel="nofollow noopener" target="_blank">#GPU</a> <a href="https://mstdn.social/tags/HPC" class="mention hashtag" rel="nofollow noopener" target="_blank">#HPC</a> <a href="https://mstdn.social/tags/InfiniBand" class="mention hashtag" rel="nofollow noopener" target="_blank">#InfiniBand</a> <a href="https://mstdn.social/tags/Kubernetes" class="mention hashtag" rel="nofollow noopener" target="_blank">#Kubernetes</a> <a href="https://mstdn.social/tags/Linux" class="mention hashtag" rel="nofollow noopener" target="_blank">#Linux</a>

HabrInfiniBand в Windows — это простоК написанию этой небольшой инструкции меня привела статья на Хабре - Быстрая сеть в домашней лаборатории или как я связался с InfiniBand . Я был очень заинтригован данным вопросом, но каково было моё удивление, когда я не мог найти почти никакой информации по InfiniBand на Windows в домашних условиях, например, в домашней лаборатории или в небольшом офисе. Информация была, конечно. Было описано, как использовали InfiniBand, какое оборудование использовали и о полученной производительности сети. Всё точно как у товарища из упомянутой статьи выше. Но не было информации о том, как поднять домашнюю сеть на IB, как настроить её и вообще с чего начать. Проведя время в интернете, я пришёл к выводу, что большинство пользователей, даже тех, кто знаком с сетями и их настройкой, тупо боятся слова InfiniBand. Для них это что-то сложное, используемое мегакорпорациями для создания суперсетей для суперкомпьютеров. А сочетание слов "InfiniBand дома" приводит их в ужас. А если ещё и коммутатор неуправляемый... Ну, вы поняли. Из той немногой информации, что мне удалось найти, я вычленил и написал простую инструкцию для новичков по InfiniBand в домашних условиях: какое оборудование нужно, как установить драйвера, как настроить сеть IB между несколькими ПК и неуправляемым коммутатором. Итак, давайте начнем!<a href="https://habr.com/ru/articles/797759/" rel="nofollow noopener" translate="no" target="_blank">https://habr.com/ru/articles/797759/</a><a href="https://zhub.link/tags/infiniband" class="mention hashtag" rel="nofollow noopener" target="_blank">#infiniband</a> <a href="https://zhub.link/tags/40g" class="mention hashtag" rel="nofollow noopener" target="_blank">#40g</a> <a href="https://zhub.link/tags/mellanox" class="mention hashtag" rel="nofollow noopener" target="_blank">#mellanox</a> <a href="https://zhub.link/tags/%D0%BA%D0%BE%D0%BC%D0%BC%D1%83%D1%82%D0%B0%D1%82%D0%BE%D1%80" class="mention hashtag" rel="nofollow noopener" target="_blank">#коммутатор</a> <a href="https://zhub.link/tags/%D1%81%D0%B5%D1%82%D0%B5%D0%B2%D0%B0%D1%8F_%D0%BA%D0%B0%D1%80%D1%82%D0%B0" class="mention hashtag" rel="nofollow noopener" target="_blank">#сетевая_карта</a> <a href="https://zhub.link/tags/%D0%B2%D1%8B%D1%81%D0%BE%D0%BA%D0%B0%D1%8F_%D0%BF%D1%80%D0%BE%D0%B8%D0%B7%D0%B2%D0%BE%D0%B4%D0%B8%D1%82%D0%B5%D0%BB%D1%8C%D0%BD%D0%BE%D1%81%D1%82%D1%8C" class="mention hashtag" rel="nofollow noopener" target="_blank">#высокая_производительность</a>

Kevin Karhan :verified:<a href="https://aus.social/@melissabeartrix" class="u-url mention" rel="nofollow noopener" target="_blank">@melissabeartrix</a> then consider <a href="https://infosec.space/tags/iSCSI" class="mention hashtag" rel="nofollow noopener" target="_blank">#iSCSI</a> or <a href="https://infosec.space/tags/FCoE" class="mention hashtag" rel="nofollow noopener" target="_blank">#FCoE</a> (<a href="https://infosec.space/tags/FibreChannel" class="mention hashtag" rel="nofollow noopener" target="_blank">#FibreChannel</a> over <a href="https://infosec.space/tags/Ethernet" class="mention hashtag" rel="nofollow noopener" target="_blank">#Ethernet</a>) over <a href="https://infosec.space/tags/OM5" class="mention hashtag" rel="nofollow noopener" target="_blank">#OM5</a>-fiber - based <a href="https://infosec.space/tags/100GBASE" class="mention hashtag" rel="nofollow noopener" target="_blank">#100GBASE</a>-SR1.2 as per 802.3bm-2015 Ethernet.Just make shure your devices have <a href="https://infosec.space/tags/QSFP28" class="mention hashtag" rel="nofollow noopener" target="_blank">#QSFP28</a> ports to chug' in the LC-Duplex connectors of the fibers amd support 64k Jumbo Frames...Cuz unlike <a href="https://infosec.space/tags/InfiniBand" class="mention hashtag" rel="nofollow noopener" target="_blank">#InfiniBand</a> that stuff is at least long-term useful and salvageable...<a href="https://en.wikipedia.org/wiki/100_Gigabit_Ethernet" rel="nofollow noopener" translate="no" target="_blank">https://en.wikipedia.org/wiki/100_Gigabit_Ethernet</a> <a href="https://en.wikipedia.org/wiki/ISCSI" rel="nofollow noopener" translate="no" target="_blank">https://en.wikipedia.org/wiki/ISCSI</a> <a href="https://en.wikipedia.org/wiki/Fibre_Channel" rel="nofollow noopener" translate="no" target="_blank">https://en.wikipedia.org/wiki/Fibre_Channel</a> <a href="https://en.wikipedia.org/wiki/Fibre_Channel_over_Ethernet" rel="nofollow noopener" translate="no" target="_blank">https://en.wikipedia.org/wiki/Fibre_Channel_over_Ethernet</a> <a href="https://en.wikipedia.org/wiki/Fibre_Channel_over_IP" rel="nofollow noopener" translate="no" target="_blank">https://en.wikipedia.org/wiki/Fibre_Channel_over_IP</a> <a href="https://en.wikipedia.org/wiki/ISCSI_Extensions_for_RDMA" rel="nofollow noopener" translate="no" target="_blank">https://en.wikipedia.org/wiki/ISCSI_Extensions_for_RDMA</a> <a href="https://en.wikipedia.org/wiki/InfiniBand" rel="nofollow noopener" translate="no" target="_blank">https://en.wikipedia.org/wiki/InfiniBand</a>

Compilando PodcastInfiniBand <a href="https://www.eduardocollado.com/2024/01/02/infiniband/" target="_blank" rel="nofollow noopener" translate="no">https://www.eduardocollado.com/2024/01/02/infiniband/</a> <a href="https://mastodon.social/tags/Hpc" class="mention hashtag" rel="tag">#Hpc</a>, <a href="https://mastodon.social/tags/IA" class="mention hashtag" rel="tag">#IA</a>, <a href="https://mastodon.social/tags/Infiniband" class="mention hashtag" rel="tag">#Infiniband</a>, <a href="https://mastodon.social/tags/InteligenciaArtificial" class="mention hashtag" rel="tag">#InteligenciaArtificial</a>, <a href="https://mastodon.social/tags/Supercomputaci%C3%B3n" class="mention hashtag" rel="tag">#Supercomputación</a>

HPC GuruNvidia data center roadmap: <a href="https://mastodon.social/tags/Hopper" class="mention hashtag" rel="tag">#Hopper</a> refresh in 2024, transition to <a href="https://mastodon.social/tags/Blackwell" class="mention hashtag" rel="tag">#Blackwell</a> later in 2024 with another architecture in 2025Nvidia now breaks out its Arm-based products and its x86-based products, with <a href="https://mastodon.social/tags/ARM" class="mention hashtag" rel="tag">#ARM</a> on topOn the <a href="https://mastodon.social/tags/networking" class="mention hashtag" rel="tag">#networking</a> side, both <a href="https://mastodon.social/tags/Infiniband" class="mention hashtag" rel="tag">#Infiniband</a> and <a href="https://mastodon.social/tags/Ethernet" class="mention hashtag" rel="tag">#Ethernet</a> are going to progress from 400Gbps to 800Gbps in 2024 and then to 1.6Tbps in 2025 Something missing from the roadmap is the NVSwitch/ NVLink roadmap<a href="https://servethehome.com/nvidia-data-center-roadmap-with-gx200nvl-gx200-x100-and-x40-ai-chips-in-2025/" target="_blank" rel="nofollow noopener" translate="no">https://servethehome.com/nvidia-data-center-roadmap-with-gx200nvl-gx200-x100-and-x40-ai-chips-in-2025/</a><a href="https://mastodon.social/tags/AI" class="mention hashtag" rel="tag">#AI</a> <a href="https://mastodon.social/tags/HPC" class="mention hashtag" rel="tag">#HPC</a> <a href="https://mastodon.social/tags/GPU" class="mention hashtag" rel="tag">#GPU</a>

Benjamin Carr, Ph.D. 👨🏻‍💻🧬<a href="https://hachyderm.io/tags/Meta" class="mention hashtag" rel="nofollow noopener" target="_blank">#Meta</a> Platforms Is Determined To Make <a href="https://hachyderm.io/tags/Ethernet" class="mention hashtag" rel="nofollow noopener" target="_blank">#Ethernet</a> Work For <a href="https://hachyderm.io/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#AI</a> <a href="https://hachyderm.io/tags/MetaPlatforms" class="mention hashtag" rel="nofollow noopener" target="_blank">#MetaPlatforms</a> is one of the founding companies behind the <a href="https://hachyderm.io/tags/UltraEthernet" class="mention hashtag" rel="nofollow noopener" target="_blank">#UltraEthernet</a> Consortium. What unites these companies – <a href="https://hachyderm.io/tags/Broadcom" class="mention hashtag" rel="nofollow noopener" target="_blank">#Broadcom</a>, <a href="https://hachyderm.io/tags/Cisco" class="mention hashtag" rel="nofollow noopener" target="_blank">#Cisco</a>, and <a href="https://hachyderm.io/tags/HewlettPackardEnterprise" class="mention hashtag" rel="nofollow noopener" target="_blank">#HewlettPackardEnterprise</a> for switch <a href="https://hachyderm.io/tags/ASIC" class="mention hashtag" rel="nofollow noopener" target="_blank">#ASIC</a> (and soon <a href="https://hachyderm.io/tags/Marvell" class="mention hashtag" rel="nofollow noopener" target="_blank">#Marvell</a> we think), <a href="https://hachyderm.io/tags/Microsoft" class="mention hashtag" rel="nofollow noopener" target="_blank">#Microsoft</a> and Meta among the titans, and Cisco, <a href="https://hachyderm.io/tags/HPE" class="mention hashtag" rel="nofollow noopener" target="_blank">#HPE</a>, and <a href="https://hachyderm.io/tags/Arista" class="mention hashtag" rel="nofollow noopener" target="_blank">#Arista</a> Networks among the switch makers – is common enemy: <a href="https://hachyderm.io/tags/InfiniBand" class="mention hashtag" rel="nofollow noopener" target="_blank">#InfiniBand</a>. Figuring out way make Ethernet as good as IB for AI & <a href="https://hachyderm.io/tags/HPC" class="mention hashtag" rel="nofollow noopener" target="_blank">#HPC</a> <a href="https://hachyderm.io/tags/networking" class="mention hashtag" rel="nofollow noopener" target="_blank">#networking</a> <a href="https://www.nextplatform.com/2023/09/26/meta-platforms-is-determined-to-make-ethernet-work-for-ai/" rel="nofollow noopener" translate="no" target="_blank">https://www.nextplatform.com/2023/09/26/meta-platforms-is-determined-to-make-ethernet-work-for-ai/</a>

flatplanetNice to see that they are working to avoid <a href="https://mastodon.ie/tags/infiniband" class="mention hashtag" rel="nofollow noopener" target="_blank">#infiniband</a> for <a href="https://mastodon.ie/tags/ai" class="mention hashtag" rel="nofollow noopener" target="_blank">#ai</a> <a href="https://mastodon.ie/tags/networks" class="mention hashtag" rel="nofollow noopener" target="_blank">#networks</a> <a href="https://mastodon.ie/tags/ultraethernet" class="mention hashtag" rel="nofollow noopener" target="_blank">#ultraethernet</a> <a href="https://ultraethernet.org/" rel="nofollow noopener" target="_blank">https://ultraethernet.org/</a>

HPC GuruUltra Ethernet Consortium<a href="https://mastodon.social/tags/InfiniBand" class="mention hashtag" rel="tag">#InfiniBand</a> is controlled by one vendor, and the hyperscalers and cloud builders hate that, and it is not Ethernet, and they hate that, tooThey want one protocol to scale to 1 million endpoints in a single networkThey also want a new implementation of RDMA that is more efficient and more scalable than either InfiniBand or Ethernet with RoCE<a href="https://www.nextplatform.com/2023/07/20/ethernet-consortium-shoots-for-1-million-node-clusters-that-beat-infiniband/" target="_blank" rel="nofollow noopener" translate="no">https://www.nextplatform.com/2023/07/20/ethernet-consortium-shoots-for-1-million-node-clusters-that-beat-infiniband/</a><a href="https://mastodon.social/tags/HPC" class="mention hashtag" rel="tag">#HPC</a> <a href="https://mastodon.social/tags/AI" class="mention hashtag" rel="tag">#AI</a> <a href="https://mastodon.social/tags/network" class="mention hashtag" rel="tag">#network</a>

flatplanet<a href="https://www.nextplatform.com/2023/06/22/cisco-guns-for-infiniband-with-silicon-one-g200/" rel="nofollow noopener" target="_blank">https://www.nextplatform.com/2023/06/22/cisco-guns-for-infiniband-with-silicon-one-g200/</a> <a href="https://mastodon.ie/tags/networking" class="mention hashtag" rel="nofollow noopener" target="_blank">#networking</a> <a href="https://mastodon.ie/tags/ethernet" class="mention hashtag" rel="nofollow noopener" target="_blank">#ethernet</a> <a href="https://mastodon.ie/tags/ai" class="mention hashtag" rel="nofollow noopener" target="_blank">#ai</a> <a href="https://mastodon.ie/tags/infiniband" class="mention hashtag" rel="nofollow noopener" target="_blank">#infiniband</a> <a href="https://mastodon.ie/tags/cisco" class="mention hashtag" rel="nofollow noopener" target="_blank">#cisco</a>

MinaI'm looking for recommendations of a cheap / used PC that can run <a href="https://cathode.church/tags/FreeBSD" class="mention hashtag" rel="nofollow noopener" target="_blank">#FreeBSD</a> and has at least two PCIe slots for <a href="https://cathode.church/tags/InfiniBand" class="mention hashtag" rel="nofollow noopener" target="_blank">#InfiniBand</a> cards(also looking for hints for cheap FreeBSD compatible Infiniband cards that aren't Mellanox ;)

Gabriele SveltoEleven years ago I volunteered to add native <a href="https://fosstodon.org/tags/Infiniband" class="mention hashtag" rel="nofollow noopener" target="_blank">#Infiniband</a> / <a href="https://fosstodon.org/tags/RDMA" class="mention hashtag" rel="nofollow noopener" target="_blank">#RDMA</a> support to <a href="https://fosstodon.org/tags/ZeroMQ" class="mention hashtag" rel="nofollow noopener" target="_blank">#ZeroMQ</a>. At the time I was working on high-performance networking and I thought it was a nice challenge... but shortly afterwards I landed my job at <a href="https://mastodon.social/@mozilla" class="u-url mention" rel="nofollow noopener" target="_blank">@mozilla</a> and never finished it.Since then I've been contacted multiple times by people who wished to finish my work but none succeeded. Last time was yesterday. Maybe I should give it a spin again: <a href="https://zeromq-dev.zeromq.narkive.com/a3hbU2H1/contributing-native-infiniband-rdma-support-to-0mq" rel="nofollow noopener" target="_blank">https://zeromq-dev.zeromq.narkive.com/a3hbU2H1/contributing-native-infiniband-rdma-support-to-0mq</a>

Glenn K. LockwoodThe dates for the 2023 MVAPICH User Group (<a href="https://mast.hpc.social/tags/MUG23" class="mention hashtag" rel="nofollow noopener" target="_blank">#MUG23</a>) have been set as August 21-23 and, as usual, it will be near OSU in Columbus. Despite its name, MUG a great snapshot of the state of the practice in <a href="https://mast.hpc.social/tags/InfiniBand" class="mention hashtag" rel="nofollow noopener" target="_blank">#InfiniBand</a> in <a href="https://mast.hpc.social/tags/HPC" class="mention hashtag" rel="nofollow noopener" target="_blank">#HPC</a> and <a href="https://mast.hpc.social/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#AI</a> overall.See <a href="https://mug.mvapich.cse.ohio-state.edu/" rel="nofollow noopener" target="_blank">https://mug.mvapich.cse.ohio-state.edu/</a>.

MinaFreeBSD vs cloud-init

Recent searches

Search options

Administered by:

Server stats:

#infiniband