ISC 2023: Expertise the ability of HPC within the cloud with Microsoft Azure | Azure Weblog

The ISC Excessive-Efficiency Occasion (ISC), previously often known as Worldwide Supercomputing Convention, is among the most necessary occasions within the high-performance computing (HPC) business, bringing collectively a number of the world’s most outstanding researchers, builders, distributors, and customers to debate the most recent developments in HPC know-how. Microsoft Azure has had a serious presence at ISC for a number of years now, and this yr’s convention will probably be no exception.

Wanting again at ISC 2022

At ISC 2022, we showcased the most recent HPC options for Azure HPC, together with the overall availability of a number of new merchandise. These options are designed to assist organizations of all sizes reap the benefits of the ability and adaptability of Azure for his or her HPC workloads.

Key themes for us prior to now have been round showcasing our newest developments in HPC and AI infrastructure, and the way our options can assist companies speed up their analysis, streamline their operations, and obtain their enterprise objectives. As well as, we demonstrated the worth of our partnerships with varied business leaders and the way we collectively allow clients to leverage the most recent applied sciences in HPC and AI. We additionally highlighted our dedication to offering safe and scalable cloud options to satisfy the calls for of essentially the most complicated workloads.

Key themes to assist companies at ISC 2023

As we look ahead to the upcoming occasion, these key themes are much more necessary to debate with clients and discover methods for them to do extra with much less on this surroundings of macroeconomic uncertainty. Our consultants mixed with our key companions on this house are really unimaginable at serving to companies optimize their cloud spend and leverage the perfect know-how for the suitable ROI to get the quickest outcomes potential.

AI can be a key theme in our present panorama, with huge breakthroughs taking place just lately on this house, and can move via into our presence on the occasion. Whether or not you’re engaged on picture and speech recognition, pure language processing, or predictive modeling, a strong infrastructure is required to energy and speed up these workloads.

This yr we’re a silver sponsor for the ISC 2023 occasion. We’ll have each a sales space on-site (C318) and a digital presence. We’ll have a number of displays per day round key subjects within the HPC and AI areas, given by Microsoft consultants and companions. We’ll even have some displays from clients, the place you possibly can study extra about how different firms are accelerating their workloads on the cloud and ask questions.

Vendor Showdown

Antigoni Chrysostomou, Director Specialist Administration, Microsoft.

Come hear in regards to the following subjects on this yr’s Vendor Showdown:

• Microsoft’s technique to proceed investing within the latest CPUs and graphic processing items (GPUs) with AMD, Intel, and NVIDIA—constructing real HPC and AI infrastructure at scale.
• How we’re working Linux on our HPC methods and we will run message passing interface (MPI) jobs on Azure.
• How Azure is attaining world-class efficiency optimized usually with particular variations of CPUs not obtainable on prem or with different cloud distributors in the marketplace.
• Microsoft’s technique round AI infrastructure for coaching AI fashions.
• Quantum Computing and the flexibility to run quantum use instances on right now’s HPC infrastructure.
• Examples of how we assist clients obtain extra.

HPC Options Discussion board

Corridor H, Sales space K1001—Floor Ground on Tuesday, Could 23, 2023, from 1:40 PM to 2:00 PM CEST.

HPC infrastructure for AI outcomes: Studying at massive scale—Monday, Could 22, 2023. Time to be introduced.

Gabrielle DavelaarSenior AI Lead World Black Belt, Microsoft.	Karl PodestaSenior Specialist, Microsoft Azure HPC.

Microsoft Azure claims 5 entries within the prime 50 supercomputers on this planet (November 2022). These entries are every smaller components of bigger methods, at separate websites. As part of public cloud infrastructure, they’re utilized by clients all around the world for working each HPC and enormous AI fashions. They’re additionally utilized by Microsoft itself. Our partnerships (akin to with NVIDIA and OpenAI) assist us to run a number of the largest AI fashions on this planet and convey this intelligence straight into your entire suite of Microsoft merchandise, from Home windows to Workplace to Bing (and Azure itself, too). Come and study a few of what now we have realized from coaching, deploying, and working the world’s largest AI fashions, on a number of the world’s largest HPC infrastructure—and our ideas on how this pertains to sustainability and powering the longer term.

Wish to regulate what’s taking place? Try our ISC 2023 web site, which we’ll proceed to replace as plans are solidified.

Latest developments since ISC 2022

Azure HBv4-series and HX-series digital machines, now in preview

In November 2022, the most recent developments in purpose-built digital machines focused for HPC workloads introduced the HBv4-series and HX-series, powered by lengthy awaited, latest 4th Gen EPYCTM AMD processors, codenamed ‘Genoa.’ The HBv4-series is the following technology of our lengthy working, flagship HB-series, and brings vital enhancements to key workloads akin to computational fluid dynamics (CFD), finite factor evaluation, frontend and backend digital design automation (EDA), rendering, molecular dynamics, computational geoscience, climate simulation, AI inference, and monetary threat evaluation. The model new HX-series is a big step in the direction of delivering the perfect platform for silicon design, with larger reminiscence capability and is designed to help ever rising fashions which might be turning into extra commonplace amongst chip designers.

Azure Managed Lustre, now in preview

In February 2023 we launched Azure Managed Lustre in preview, a brand new storage providing for Azure HPC. Lustre is an open-source parallel file system famend for HPC and is adept at large-scale cluster computing. Azure Managed Lustre (preview) offers high-performance storage of Lustre with the management and consistency of Azure. In consequence, clients can deal with their enterprise objectives, whether or not that’s constructing a fraud detection system based mostly on statistical evaluation system (SAS) analytics or decoding the human genome to create the following breakthrough in drugs.

Azure NVads A10 v5 digital machines, now usually obtainable

The brand new Azure NVads A10 v5 digital machines turned usually obtainable in June 2022, that includes NVIDIA A10 Tensor Core GPUs and enhanced efficiency capabilities for graphics-intensive processing workloads. These digital machines are perfect for graphics-intensive functions for CAD, structure engineering and building (AEC), gaming, and digital desktop infrastructure (VDI) and supply clients with versatile, scalable options to satisfy their computing wants. With GPU partitioning (1/8 to 2 full GPUs obtainable), these new digital machines decrease the barrier to entry for firms eager to leverage GPUs, the place a full GPU is just not essential. We encourage clients to discover the brand new capabilities of Azure NVads A10 v5 digital machines and expertise the advantages of superior graphics processing for his or her enterprise.

See what our clients have been as much as

		Moody’s Analytics Moody’s Analytics collaborated with Microsoft Azure to develop a scalable cloud HPC answer to help its insurance coverage monetary intelligence platform. By leveraging Azure digital machine scale units and Azure HPC SKUs, Moody’s was capable of effectively scale its infrastructure to help dynamic workloads, enhance software efficiency, and scale back operational prices. Because of this partnership, Moody’s Analytics was capable of improve its buyer expertise and supply extra correct, dependable, and well timed monetary evaluation to the insurance coverage business.
		Climavision Climavision partnered with Microsoft Azure to develop a scalable and dependable cloud answer to help its superior climate forecasting know-how. With Azure’s AI and analytics capabilities, Climavision was capable of enhance its climate prediction accuracy and supply real-time, hyperlocal climate insights to its clients. By leveraging Azure’s safe and versatile cloud platform, Climavision was capable of improve its operational effectivity and obtain vital enterprise development.
		UD Vehicles UD Vehicles collaborated with Microsoft to develop a complicated HPC and AI infrastructure to enhance the effectivity and accuracy of its manufacturing processes. By leveraging Azure’s HPC capabilities and machine studying algorithms, UD Vehicles was capable of streamline its provide chain operations and improve its car design course of. With Azure’s safe and scalable cloud platform, UD Vehicles was capable of obtain vital value financial savings and place itself for future development within the automotive business.
		Encina Chemical compounds Encina Chemical compounds partnered with Microsoft Azure to develop a scalable, safe, and versatile cloud answer to optimize its manufacturing processes, scale back time to answer, and enhance the opportunity of working bigger and extra refined area fashions that incorporate extra detailed physics and extra complicated chemical reactions. Azure’s built-in AI and analytics capabilities enabled Encina Chemical compounds to automate high quality management, optimize manufacturing, and scale back waste—resulting in vital value financial savings, elevated effectivity, and higher useful resource sustainability. Utilizing Azure HPC allowed taking leaps in sustainability initiatives as a substitute of slowly creeping ahead with preexisting know-how. The simulations generated exact, actionable information used within the design of a cutting-edge recycling facility permitting Encina Chemical compounds and different sustainability startups to make use of sooner, cheaper modeling in Azure to assist them deploy and scale modeling revolutionary concepts with confidence.

Be taught extra about Azure HPC and ISC 2023

Listed here are some key locations so that you can study extra about Azure HPC and our presence on the ISC Excessive-Efficiency Occasion 2023: