#nvlink — Public Fediverse posts

RT @QuixiAI: Wer ist Vibe-Hardware? Die Welt braucht NVLink-Brücken (flexibel, bitte, damit sie von 2 bis 4 Steckplätzen funktioniert). Auf eBay kosten sie bereits über 1.000 US-Dollar – da kann man richtig Geld verdienen, Eric Hartford (@QuixiAI). Es gibt 2-Steckplatz-Modelle auf eBay, aber nur wenn man 2-Steckplätze (chinesische Modifikationen) oder wassergekühlte RTX 3090-Grafikkarten hat. Die 3-Steckplatz- und 4-Steckplatz-Modelle (für Retail-GPUs) scheinen inzwischen unerschwinglich geworden zu sein. (Vielleicht ist es Zeit für Vibe-Hardware) — https://nitter.net/QuixiAI/status/2073474471529160905#m

mehr auf Arint.info

#Grafikkarten #HardwareModding #NVLink #RTX3090 #TechNews #VibeHardware #arint_info

https://x.com/QuixiAI/status/2073529939333144939#m

RT @QuixiAI: Wer ist Vibe-Hardware? Die Welt braucht NVLink-Brücken (bitte flexibel, damit sie von 2 bis 4 Steckplätzen funktionieren). Auf eBay kosten sie bereits über 1.000 Dollar, da kann man richtig Geld verdienen. Eric Hartford (@QuixiAI) sagt, es gibt 2-Steckplatz-Modelle auf eBay, aber das gilt nur, wenn man 2-Steckplatz-Grafikkarten (chinesische Modifikationen) oder wassergekühlte RTX 3090er hat. Die 3- und 4-Steckplatz-Modelle (für Retail-GPUs) scheinen inzwischen zu teuer geworden zu sein. (Vielleicht ist es Zeit für Vibe-Hardware.) — https://nitter.net/QuixiAI/status/2073474471529160905#m

mehr auf Arint.info

#Grafikkarten #Hardware #NVLink #PCBuilding #TechNews #VibeHardware #arint_info

https://x.com/QuixiAI/status/2073529939333144939#m

Нейро сети для самых маленьких. Часть первая (которая после нулевой). Удобство в прокрустовом ложе оптимизации

Это первая (после нулевой) статья из серии Нейро сети для самых маленьких , в которой мы разбираем инфраструктуру для запуска нейронных сетей. Для обучения и инференса нейросетей и для любых видов High Performance Computing используются специализированные технологии: GPU/TPU, RDMA, Kernel bypass, NVLink, InfiniBand, RoCE и другие. Про некоторые из них большинство только что-то слышали, но сталкиваться с ними не приходилось. Нельзя просто взять ванильный стек Linux, воткнуть в него 400Gb Ethernet+IP и получить рабочее решение. Почему? Потому что общее решение на масштабе в большинстве случаев проигрывает специализированным как в скорости, так и в стоимости. Как бы странно последнее ни звучало.

https://habr.com/ru/companies/yandex/articles/1047072/

#rdma #gpudirect_rdma #infiniband #tcp #ethernet #zero_copy #roce #nvlink #nvidia #gpu

【NVIDIA Vera Rubin Pod発表】AIファクトリーを再定義、電力効率と計算能力で競合を圧倒 — BigGo ファイナンス https://www.yayafa.com/2829292/ #AgenticAi #AI #ArtificialGeneralIntelligence #ArtificialIntelligence #dsx #MGX #NVFP4 #NVIDIA #NVLink #RubinGPU #VeraCPU #VeraRubinPod #エージェント #エージェント型AI #ジェンスン・フアン #人工知能 #汎用人工知能

NVIDIA and Marvell alliance scales AI-RAN infrastructure https://www.telecomstechnews.com/news/nvidia-and-marvell-alliance-ai-ran-infrastructure/ #nvidia #marvell #airan #telecoms #nvlink #5g #6g #ai #technology

NVIDIA and Marvell alliance scales AI-RAN infrastructure https://www.telecomstechnews.com/news/nvidia-and-marvell-alliance-ai-ran-infrastructure/ #nvidia #marvell #airan #telecoms #nvlink #5g #6g #ai #technology

NVIDIA’s new Vera Rubin platform brings together specialized chips (Vera CPUs, Rubin GPUs, Groq LPUs, and BlueField-4 DPUs) into coordinated, rack-scale systems designed for real-time AI.

The big shift: AI isn’t just about training models anymore — it’s about orchestrating entire systems to power intelligent, autonomous agents in real time.
https://www.buysellram.com/blog/the-agentic-ai-era-how-nvidia-rubin-vera-cpu-groq-3-lpus-bluefield-4-redefine-the-inference-factory/
#NVIDIAGTC #AgenticAI #VeraRubin #DataCenter #GPU #InferenceFactory #AIInfrastructure #Groq #NVIDIA #NVLink #AIHardware #technology

NVIDIA’s new Vera Rubin platform brings together specialized chips (Vera CPUs, Rubin GPUs, Groq LPUs, and BlueField-4 DPUs) into coordinated, rack-scale systems designed for real-time AI.

The big shift: AI isn’t just about training models anymore — it’s about orchestrating entire systems to power intelligent, autonomous agents in real time.
https://www.buysellram.com/blog/the-agentic-ai-era-how-nvidia-rubin-vera-cpu-groq-3-lpus-bluefield-4-redefine-the-inference-factory/
#NVIDIAGTC #AgenticAI #VeraRubin #DataCenter #GPU #InferenceFactory #AIInfrastructure #Groq #NVIDIA #NVLink #AIHardware #technology

#AWS announced #Trainium3, a new #AItrainingchip with significant performance and energy efficiency improvements. #Trainium4, already in development, will offer even better performance and support #Nvidia’s #NVLink Fusion technology, potentially attracting more AI applications to AWS. https://techcrunch.com/2025/12/02/amazon-releases-an-impressive-new-ai-chip-and-teases-a-nvidia-friendly-roadmap/?eicker.news #tech #media #news

#AWS announced #Trainium3, a new #AItrainingchip with significant performance and energy efficiency improvements. #Trainium4, already in development, will offer even better performance and support #Nvidia’s #NVLink Fusion technology, potentially attracting more AI applications to AWS. https://techcrunch.com/2025/12/02/amazon-releases-an-impressive-new-ai-chip-and-teases-a-nvidia-friendly-roadmap/?eicker.news #tech #media #news

#Arm and #Nvidia are #partnering to integrate Arm-based #Neoverse #CPUs with Nvidia’s #GPUs using Nvidia’s #NVLink Fusion technology. This #collaboration will benefit customers, particularly #hyperscalers, who prefer custom infrastructure setups. The partnership highlights Nvidia’s strategy of collaborating with major tech companies to expand its influence in the #AIindustry. https://www.cnbc.com/2025/11/17/arm-nvidia-nvlink-partnership-custom-cpu.html?eicker.news #tech #media #news

#Arm and #Nvidia are #partnering to integrate Arm-based #Neoverse #CPUs with Nvidia’s #GPUs using Nvidia’s #NVLink Fusion technology. This #collaboration will benefit customers, particularly #hyperscalers, who prefer custom infrastructure setups. The partnership highlights Nvidia’s strategy of collaborating with major tech companies to expand its influence in the #AIindustry. https://www.cnbc.com/2025/11/17/arm-nvidia-nvlink-partnership-custom-cpu.html?eicker.news #tech #media #news

8x AMD Instinct #MI355X (288GB @8TB/s) take back the lead over 8x Nvidia #B200 (180GB @8TB/s) in #FluidX3D #CFD, achieving 362k MLUPs/s (vs. 219k MLUPs/s). Thanks to Jon Stevens from Hot Aisle to run the benchmarks! 🖖😊

In single-GPU, both perform about the same, but in 8x #GPU config, MI355X is 65% faster. The difference comes from PCIe bandwidth - MI355X does 55GB/s, B200 only 14GB/s. #Nvidia leaves a lot of perf on the table by not exposing #NVLink P2P to #OpenCL.

https://github.com/ProjectPhysX/FluidX3D

8x AMD Instinct #MI355X (288GB @8TB/s) take back the lead over 8x Nvidia #B200 (180GB @8TB/s) in #FluidX3D #CFD, achieving 362k MLUPs/s (vs. 219k MLUPs/s). Thanks to Jon Stevens from Hot Aisle to run the benchmarks! 🖖😊

In single-GPU, both perform about the same, but in 8x #GPU config, MI355X is 65% faster. The difference comes from PCIe bandwidth - MI355X does 55GB/s, B200 only 14GB/s. #Nvidia leaves a lot of perf on the table by not exposing #NVLink P2P to #OpenCL.

https://github.com/ProjectPhysX/FluidX3D

NVIDIA DGX Spark: o mini supercomputador de IA com chips Blackwell já tem preço e data de lançamento
🔗 https://tugatech.com.pt/t72827-nvidia-dgx-spark-o-mini-supercomputador-de-ia-com-chips-blackwell-ja-tem-preco-e-data-de-lancamento

#Acer #asus #CES #ChatGPT #computador #desktop #docker #escalabilidade #google #hardware #ia #JetBrains #Lenovo #Meta #microsoft #MSI #mundo #nvidia #NVLink #OpenAI #software #spark #startup #startups #supercomputadores #tecnologia 

NVIDIA DGX Spark: o mini supercomputador de IA com chips Blackwell já tem preço e data de lançamento
🔗 https://tugatech.com.pt/t72827-nvidia-dgx-spark-o-mini-supercomputador-de-ia-com-chips-blackwell-ja-tem-preco-e-data-de-lancamento

#Acer #asus #CES #ChatGPT #computador #desktop #docker #escalabilidade #google #hardware #ia #JetBrains #Lenovo #Meta #microsoft #MSI #mundo #nvidia #NVLink #OpenAI #software #spark #startup #startups #supercomputadores #tecnologia 

Nvidia e Intel unem-se para dominar os PCs com IA: a memória unificada é a arma secreta
🔗 https://tugatech.com.pt/t72561-nvidia-e-intel-unem-se-para-dominar-os-pcs-com-ia-a-memoria-unificada-e-a-arma-secreta

#AMD #apple #computador #cpu #CUDA #encapsulamento #gaming #gpu #hardware #ia #Intel #llms #motherboard #mundo #nvidia #NVLink #processador #processadores #qualcomm #ram #snapdragon #software #tecnologia #x86 

Nvidia e Intel unem-se para dominar os PCs com IA: a memória unificada é a arma secreta
🔗 https://tugatech.com.pt/t72561-nvidia-e-intel-unem-se-para-dominar-os-pcs-com-ia-a-memoria-unificada-e-a-arma-secreta

#AMD #apple #computador #cpu #CUDA #encapsulamento #gaming #gpu #hardware #ia #Intel #llms #motherboard #mundo #nvidia #NVLink #processador #processadores #qualcomm #ram #snapdragon #software #tecnologia #x86 

https://blog.gslin.org/archives/2025/09/25/12636/intel-%e8%b7%9f-nvidia-%e5%90%88%e4%bd%9c%ef%bc%8c%e6%89%80%e4%bb%a5-intel-%e8%87%aa%e5%b7%b1%e7%9a%84%e9%a1%af%e5%8d%a1%e9%82%84%e8%a6%81%e6%90%9e%e5%97%8e/

Intel 跟 NVIDIA 合作，所以 Intel 自己的顯卡還要搞嗎...

#ai #cuda #gpu #intel #nvidia #nvlink #soc

https://blog.gslin.org/archives/2025/09/25/12636/intel-%e8%b7%9f-nvidia-%e5%90%88%e4%bd%9c%ef%bc%8c%e6%89%80%e4%bb%a5-intel-%e8%87%aa%e5%b7%b1%e7%9a%84%e9%a1%af%e5%8d%a1%e9%82%84%e8%a6%81%e6%90%9e%e5%97%8e/

Intel 跟 NVIDIA 合作，所以 Intel 自己的顯卡還要搞嗎...

#ai #cuda #gpu #intel #nvidia #nvlink #soc

NVIDIA e Intel unem forças: investimento de 4,6 mil milhões de euros para revolucionar a IA e os PCs
🔗 https://tugatech.com.pt/t71953-nvidia-e-intel-unem-forcas-investimento-de-46-mil-milhoes-de-euros-para-revolucionar-a-ia-e-os-pcs

#ia #Intel #LiDAR #nvidia #NVLink #processadores #semicondutores #servidores #soc #tecnologia #x86 

NVIDIA e Intel unem forças: investimento de 4,6 mil milhões de euros para revolucionar a IA e os PCs
🔗 https://tugatech.com.pt/t71953-nvidia-e-intel-unem-forcas-investimento-de-46-mil-milhoes-de-euros-para-revolucionar-a-ia-e-os-pcs

#ia #Intel #LiDAR #nvidia #NVLink #processadores #semicondutores #servidores #soc #tecnologia #x86 

Gaming-Fokus 🎮 Die neuen „#Intel #x86RTX SoCs“ kombinieren CPU und GPU in einem Chip und zielen auf #Gaming-Laptops und kompakte PCs – direkte Konkurrenz für AMDs APUs.

Datenzentren & AI 🌐 Intel fertigt künftig auch maßgeschneiderte #x86Serverchips für #Nvidia, die per #NVLink besonders schnell mit GPUs kommunizieren.

👉 https://eicker.TV ▹ #Technik #Medien #Politik #Wirtschaft ⏻ https://eicker.BE/ratung von Gerrit Eicker aus Münster (2/2)

NVIDIA 開放自家互連架構 NVLink Fusion 登場，GPU 直連高通, 富士通 CPU
NVIDIA 最近推出全新 NVLink Fusion 技術，首次開放其高速互連技術予合作夥伴使用，讓第三方處理器及加速器能與 NVIDIA 晶片共同運作。今次也是 NVIDIA 生態系統重大轉變，企業可根據需求選擇搭配不同廠商晶片，提升系統靈活性。
The post NVIDIA 開放自家互連架構 NVLink Fusion 登場，GPU 直連高通, 富士通 CPU appeared first on 香港 unwire.hk 玩生活．樂科技.
#科技新聞 #NVIDIA #NVLink Fusion
https://unwire.hk/2025/05/20/nvidia-announces-nvlink-fusion/fun-tech/?utm_source=rss&utm_medium=rss&utm_campaign=nvidia-announces-nvlink-fusion

When PCIe is just too slow, #NVLink is now available to third parties. CPU manufacturers can build chips that tightly couple with their GPUs and participate in the cluster-wide coherent memory mesh. #nvidia https://social.heise.de/@heiseonlineenglish/114533282638521156

When PCIe is just too slow, #NVLink is now available to third parties. CPU manufacturers can build chips that tightly couple with their GPUs and participate in the cluster-wide coherent memory mesh. #nvidia https://social.heise.de/@heiseonlineenglish/114533282638521156

#UALink debuts its first #AI interconnect spec – usable in just 18 short months
Consortium (UAC) was formed in May 2024 by vendors – AMD, AWS, Broadcom, Cisco, Google, HPE, Intel, Meta, Microsoft, and Astera Labs – who think world needs an open alternative to #Nvidia's #NVLink tech that allows creation of networked GPU clusters needed to run AI workloads at scale.
The Ultra Accelerator Link Consortium has delivered first GPU interconnect specification: UALink 200G 1.0.
https://www.theregister.com/2025/04/08/ualink_200g_version_1/

#UALink debuts its first #AI interconnect spec – usable in just 18 short months
Consortium (UAC) was formed in May 2024 by vendors – AMD, AWS, Broadcom, Cisco, Google, HPE, Intel, Meta, Microsoft, and Astera Labs – who think world needs an open alternative to #Nvidia's #NVLink tech that allows creation of networked GPU clusters needed to run AI workloads at scale.
The Ultra Accelerator Link Consortium has delivered first GPU interconnect specification: UALink 200G 1.0.
https://www.theregister.com/2025/04/08/ualink_200g_version_1/

[Перевод] Подробное руководство по облачной инфраструктуре для ИИ-проектов

Как вычислительные ресурсы, системы хранения, сетевые технологии и AI-фреймворки влияют на производительность AI-проектов. Почему понимание этой инфраструктуры даёт конкурентное преимущество

https://habr.com/ru/articles/896194/

#облачная_инфраструктура #искусственный_интеллект #gpu #tpu #dpu #хранение_данных #сетевые_технологии #nvlink #rdma #ai_фреймворки

Невероятная мощь NVIDIA GB200 NVL72: Внутри гиганта ИИ-вычислений

Привет, 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».

https://habr.com/ru/companies/serverflow/articles/864314/

#сервер_флоу #GB200_NVL72 #nvlink #blackwell #nvidia_grace #Nvidia_Superchip #NVLink_Spine #infiniband #llm #SeverFlow

Мы потратили 320 тысяч рублей ради Nvlink для нейросетей. Но был ли в этом смысл?

Привет Хабр, с вами снова ServerFlow. И да, мы потратили 320 тысяч рублей, чтобы проверить, есть ли какой-то толк от NVLink для развертывания нейросетей? Тесты мы провели на двух видеокартах Nvidia A5000, сначала объединённые NVLink , а после без него. Для наглядности теста нами была выбрана последняя языковая модель LLaMA 3.2 на 90 миллиардов параметров. Что у нас в итоге вышло – сможете узнать ниже в посте.

https://habr.com/ru/companies/serverflow/articles/860082/

#ServerFlow #Машинное_обучение #Искуственный_интеллект #ИИ #Серверное_администрирование #nvidia #A5000 #видеокарты #llama #nvlink

#AMD & #Intel Team Up For #UALink As Open Alternative To #NVIDIA's #NVLink
UALinks aims to be an open standard for scaling of tens to hundreds of #GPU /accelerators within pods. In addition to Intel and AMD, #Broadcom, #Cisco, #Google, #HPE, #Meta, and #Microsoft have also been involved in the formation of the #UltraAcceleratorLink. #UltraEthernet high speed networking will still have its role for #scaleout purposes.
https://www.phoronix.com/news/Ultra-Accelerator-Link-UALink

#AMD & #Intel Team Up For #UALink As Open Alternative To #NVIDIA's #NVLink
UALinks aims to be an open standard for scaling of tens to hundreds of #GPU /accelerators within pods. In addition to Intel and AMD, #Broadcom, #Cisco, #Google, #HPE, #Meta, and #Microsoft have also been involved in the formation of the #UltraAcceleratorLink. #UltraEthernet high speed networking will still have its role for #scaleout purposes.
https://www.phoronix.com/news/Ultra-Accelerator-Link-UALink

Tech giants form AI group to counter Nvidia with new interconnect standard - Enlarge (credit: Getty Images)

On Thursday, several major tech... - https://arstechnica.com/?p=2027773 #ultraacceleratorlink #machinelearning #hpenterprise #broadcom #chatgpt #chatgtp #biz⁢ #google #nvidia #nvlink #ualink #cisco #intel #meta #amd #ai

Tech giants form AI group to counter Nvidia with new interconnect standard - Enlarge (credit: Getty Images)

On Thursday, several major tech... - https://arstechnica.com/?p=2027773 #ultraacceleratorlink #machinelearning #hpenterprise #broadcom #chatgpt #chatgtp #biz⁢ #google #nvidia #nvlink #ualink #cisco #intel #meta #amd #ai

https://www.hpcwire.com/2024/02/22/a-big-memory-nvidia-gh200-next-to-your-desk-closer-than-you-think/

#gpu #nvlink

From #SLI to #NVLink: The Evolution of #NVidia's Gaming and Multi-#GPU Technologies and Their Influence on #AI
NVLink is crafted for scientific research, data analysis, and, most notably, AI and machine learning applications. The higher bandwidth, lower latency, and unified memory make NVLink a far more flexible and powerful solution for today’s computational challenges. The most recent iterations offer bandwidths up to 900 GB/s, a significant enhancement over SLI.
https://www.storagereview.com/review/from-sli-to-nvlink-the-evolution-of-gaming-and-multi-gpu-technologies-and-their-influence-on-ai

From #SLI to #NVLink: The Evolution of #NVidia's Gaming and Multi-#GPU Technologies and Their Influence on #AI
NVLink is crafted for scientific research, data analysis, and, most notably, AI and machine learning applications. The higher bandwidth, lower latency, and unified memory make NVLink a far more flexible and powerful solution for today’s computational challenges. The most recent iterations offer bandwidths up to 900 GB/s, a significant enhancement over SLI.
https://www.storagereview.com/review/from-sli-to-nvlink-the-evolution-of-gaming-and-multi-gpu-technologies-and-their-influence-on-ai

#Google Launches AI Supercomputer Powered by #Nvidia #H100 #GPUs | #TomsHardware

"Google says the new A3 supercomputers are purpose-built to train and serve the most demanding #AI models that power today's #GenerativeAI and #LargeLanguageModel innovation while delivering 26 exaFlops of AI performance.

Each A3 supercomputer is packed with 4th generation Intel Xeon Scalable processors backed by 2TB of DDR5-4800 memory. But the real brains of the operation come from the eight Nvidia H100 #Hopper GPUs, which have access to 3.6 TBps of bisectional bandwidth by leveraging #NVLink 4.0 and #NVSwitch."

https://www.tomshardware.com/news/google-a3-supercomputer-h100-googleio

#Google Launches AI Supercomputer Powered by #Nvidia #H100 #GPUs | #TomsHardware

"Google says the new A3 supercomputers are purpose-built to train and serve the most demanding #AI models that power today's #GenerativeAI and #LargeLanguageModel innovation while delivering 26 exaFlops of AI performance.

Each A3 supercomputer is packed with 4th generation Intel Xeon Scalable processors backed by 2TB of DDR5-4800 memory. But the real brains of the operation come from the eight Nvidia H100 #Hopper GPUs, which have access to 3.6 TBps of bisectional bandwidth by leveraging #NVLink 4.0 and #NVSwitch."

https://www.tomshardware.com/news/google-a3-supercomputer-h100-googleio

Hidden #Nvlink? #RTX 4090 #Aorus Master #Teardown and #Overclocking to 3GHz

https://www.youtube.com/watch?v=zc-zwQMV8-s