Oliver Funk

It is now easier than ever to build inference pipelines for #generativeAI applications. 📣 Introducing: NVIDIA NIM Operator which gives you the ability to automate the deployment, configuration and management of our NIM microservices.

1

I’m thrilled to share that Pure Storage recently joined the Ultra Ethernet Consortium (UEC), an initiative aimed at building an open, Ethernet-based architecture to power AI and HPC applications. As AI workloads grow, so do the challenges of managing and scaling network solutions. UEC is tackling these issues head-on by advancing Ethernet technology to make it more scalable, efficient, and futureproof. By contributing to UEC, Pure is committed to helping shape the standards that will drive faster innovation, more cost-effective operations, and greater flexibility for enterprises - and that includes ensuring our data storage platform supports Ultra Ethernet standards long-term. Whether it’s speeding up AI model development or making sure our customers have the best networking options, we’re committed to delivering high-performance storage to power the most advanced large-scale AI initiatives. Exciting times ahead as we work to redefine what’s possible with Ethernet! https://lnkd.in/duN9Y2ub #AI #HPC #Innovation #UltraEthernet

140

3 Kommentare

🚀 Power your #AI projects with #PowerScale! Robust storage is key to feeding your AI models the data they need, and with the new F210 & F710 all-flash systems you've got scalable, secure, high-performance storage tailored for AI. Learn more! #Iwork4Dell 👇 https://dell.to/4bDJxtK #powerscalerocks!

26

Explore the trade-offs between throughput and latency in #deeplearning models in this new self-paced course from #NVIDIA: Sizing LLM Inference Systems. Learn how to measure and optimize performance, deploy models using NVIDIA NIM, and analyze cost benefits of on-prem vs. cloud. 🌟 Take the new course for FREE (until Nov 30th). Enroll now: https://nvda.ws/4fomyod #NVIDIATraining #DLI #LLM

8

1 Kommentar

Great keynote from Ankur Gupta, sharing how EDA is using Ai to fuel the todays VLSI design.

51

Experience Mixtral 8x22B and Mistral Large API endpoints, powered by NVIDIA NIM microservice, from our NVIDIA API catalog ➡️ https://lnkd.in/edx97Up6 These performance-optimized #generativeAI models by Mistral AI, excel in complex multilingual reasoning tasks, and more.

2

In this technical overview, learn how our NVIDIA NIM microservices and the Outerbounds platform together enable efficient, secure management of #LLMs and systems built around them.

1

Interesting move and appears a win/win for both companies. That said, may take awhile to implement. I've toured many HPC & AI and enterprise computing datacenters, as well as a number of cryptomining facilities, and the infrastructure is starkly different. HPC & AI datacenters are typically air conditioned, raised floor rooms with direct liquid cooling to chip and high speed (e.g. InfiniBand) interconnect. This is very, very different from the shoeboxes on shelves model, often in former warehouse buildings, with a simple Ethernet connection to the outside world. https://lnkd.in/g-GFNCZ4

12

2 Kommentare

In addition to our keynote we will also show our customers tips and tricks on simplifying the use of several of our VSE products. #GSE

33

On Market View today: Broadcom’s better than expected Q4 profit; Amazon donates US$1M to Trump’s inaugural fund; China’s Central Economic Work Conference; US wholesale inflation accelerates unexpectedly; Keppel DC Reit’s preferential offering 159.9% oversubscribed; South Korean equities ahead of second impeachment vote Singapore stocks slipped into the negative territory this morning, after China deepened its stimulus drive in a bid to brace itself for trade tensions with the United States. In early trade, the Straits Times Index (STI) fell 0.2 per cent to 3,803.55 points after 21.4 million securities changed hands in the broader market. In terms of counters to watch, we have Keppel DC Real Estate Investment Trust, after its manager announced yesterday that the Reit’s non-renounceable preferential offering was about 159.9 per cent subscribed. Elsewhere, from a quick look at developments out of China’s annual Central Economic Work Conference, to how Chip supplier Broadcom posted better-than-expected profit in the fourth quarter on the back of growing AI computing demand, more corporate and international headlines remain in focus. Also on deck – what to watch in the US after the country’s wholesale inflation unexpectedly accelerated in November. Plus – volatilities in South Korea as investors’ focus there turns to a second impeachment vote of President Yoon Suk Yeol planned for tomorrow. Glad to unpack the key movements of the day with Juncheong Too (涂骏祥), Assistant Dealing Manager from Moomoo SG. Catch the podcast episode on Awedio, Apple podcast and Spotify at the links below. Podcast links: https://lnkd.in/gF_5NMSt https://lnkd.in/d2cbuvcM https://lnkd.in/duppfFyN Gavin Chia #singapore #sgx #sti #stockmarket #us #inflation #tightening #cooling #federalreserve #fed #interestrates #ratecut #ppi #wholesaleprice #acceleration #producerpriceindex #broadcom #earnings #artificialintelligence #ai #amazon #donation #trump #china #centraleconomicworkconference #keppeldcreit #preferentialoffering #oversubscribed #southkorea #second #impeachment #vote

21

3 Kommentare

AMD has unveiled an exciting collaboration with Oracle Cloud Infrastructure (OCI) by integrating AMD Instinct MI300X accelerators with ROCm open software into the latest OCI Compute Supercluster instance, BM.GPU.MI300X.8. This partnership is set to elevate AI capabilities significantly. Find out more: [Link to the article] #AMD #Oracle #AI #CloudComputing

18

1 Kommentar

Google Cloud NetApp Volumes offer an ideal solution for your evolving cloud needs, whether you're scaling up your usage with new workloads or initiating your cloud journey and seeking a unified storage platform for various tasks. This solution facilitates smooth migration of all your enterprise workloads, irrespective of their size or performance requirements, with no need for any code restructuring. #NetApp #GoogleCloud #Collaboration

12

Check out how the world’s first #generativeAI models for #OpenUSD development will be available as NVIDIA NIM microservices.

76

Hewlett Packard Enterprise Builds Next-Generation 100% Direct Liquid Cooled Supercomputer at the Leibniz Supercomputing Center. The new supercomputer “Blue Lion” is part of the German national HPC infrastructure of the Gauss Center for Supercomputing and will serve a wide range of research projects, combining classic simulations with artificial intelligence. #AI #HPC #DLC #Sustainability https://lnkd.in/dnK4Ptfi

4