This post has been republished via RSS; it originally appeared at: Azure Compute Blog articles.
Strategic investments bolster versatility and relevance of Azure HPC for Microsoft
As I look at SC20 quickly approaching on my calendar, I cannot help but reflect on what has happened since the last one. And there is a lot to think about. So many changes to our lifestyle, our environment, and our safety. But, despite everything that has happened over the last year, we continue to be amazed per our customers and people who continually put their best foot forward to drive and solve complex problems across all industries.
You’ll hear us refer to Azure HPC as purpose-built. What we mean is that, instead of figuring out how to apply commodity hardware toward complex HPC & AI workloads, Azure’s formula for HPC/AI customer success starts with infrastructure that is purpose-built for scalable HPC & AI workloads. For example, our solutions can deliver a 10x performance or higher advantage compared to products elsewhere on the public cloud, which gives us a really strong foundation of performance and scale leadership. We ensure this level of performance by aligning with partners like Intel, so our HPC instances provide the most powerful architecture available to accelerate both HPC and AI workloads. Our team of HPC experts then add on a broad set of Azure technologies to leverage that infrastructure in an agile and secure manner. That’s what leads to big impact for customers. These are the pillars of HPC & AI on Azure. Some great examples are customers who are modelling complex problems like drug discovery, weather simulation, crash test analysis, and state of the art AI training.
To better understand Microsoft’s vision and strategy that led to their HPC investments, check out the video links below:
Azure HPC Overview, featuring Nidhi Chappell, Head of Product and Engineering at Azure HPC, summarizes Microsoft’s recent investments into building their comprehensive HPC cloud platform.
Azure HPC Vision, featuring Andrew Jones, Lead, Future HPC & AI Capabilities at Microsoft, outlines how we see the future of HPC and its convergence with AI in the cloud.
Many people will adapt their behaviors to capitalize on new capabilities offered by technology advancements. Often times, the technologies themselves will establish new lifestyle patterns and practices. But in HPC, it is the other way around. Technologies must clearly identify, align, and support the traditional behaviors of HPC engineers and service managers in order to be deemed relevant or useful. This is one reason the ramp to HPC in the cloud has been slower than some expected…cloud platforms have taken a long time to become engineered in line and in step with how veteran HPC organizations expect to be able to work.
These videos illustrate a much deeper understanding and alignment of Azure HPC relevance to specific application workloads within known industry verticals:
Link to video: How to bring your risk workloads to Azure
Presented by: Stephen Richardson, EMEA HPC & AI Technology Specialist at Microsoft, and Greg Ulepic, FSI Lead for Risk Workloads at Microsoft.
Summary: The pandemic has not just hit our health, but also our pockets. Financial institutions need to re-invent how they help their clients mitigate and manage risk across their portfolios. To do that, many institutions are examining both a movement of their workload to the cloud as well as setting up native cloud-based architectures to operate moving forward. This video illustrates how the financial services marketplace is evolving towards cloud-driven solutions, and how Azure is well-furnished to support that model.
Autonomous Vehicle Development
Link to video: Accelerating Autonomous Vehicle Development
Presented by: Kurt Niebuhr, Principal Program Manager for Azure HPC
Summary: There’s few other areas in modern digital transformation that have more buzz than self-driving cars. Given the miniscule room for error when it comes to driving safely, the very notion of self-driving cars can easily raise concerns even in the most liberal mind. In this video, Kurt Niebuhr discusses how Microsoft Azure provides an end-to-end autonomous driving development and engineering solution with Azure HPC, giving carmakers an excellent understanding of how they need to plan for the next phases of automobile development and manufacturing.
Presented by: Tim Carroll, Director of HPC and AI for Research for Microsoft Azure
Summary: The global community of researchers has come into greater spotlight recently with the rise in large scale health and environmental issues. When data scientists and researchers are incumbered by IT limitations, they are being denied their potential to innovate and discover new remedies. This video illustrates how Microsoft is better enabling universities and research institutions around the world to massively speed up their processing tasks without building out their hardware data centers, thereby eliminating barriers of limitation for scientists and researchers everywhere.
Link to video: Empowering exploration and production with Azure
Presented by: Hussein Shel, Principal Program Manager for Azure Global Energy
Summary: One of the biggest areas of digital transformation in energy is optimizing how companies explore and discover new untapped reservoirs of fossil fuels. Additionally, the energy industry as a whole has been put under a giant microscope in recent decades with more voices calling for a phase out of fossil fuel production and transitioning to clean & renewable energy. These are all initiatives that can be enabled by Azure HPC. In this video, Hussein Shel discusses some of Azure HPC’s offerings in exploration and production for the oil and gas industry.
Presented by: Prashant Varshney, Sr. Director, Product Management for Azure Engineering
Summary: The semiconductor industry is the best example of technology dependence on itself. The pace at which micro-processing architectures increase in density and capability has direct impact on the viability of entire product and service ecosystems. As such, silicon engineering & manufacturing companies must be highly capable, yet highly nimble, with their electronic design automation (EDA) to remain competitive. In this video, Prashant Varshney outlines how Azure HPC has been built with some of these considerations in mind, touching on how semiconductor companies can build the best chips by using Azure for their EDA processes.
Link to video: Microsoft Azure HPC discrete Manufacturing
Presented by: Karl Podesta, HPC Technical Specialist, AzureCAT
Summary: Manufacturers are being asked to visualize real-time products and product performance from anywhere. We also need to understand what's happening more with physical products and use things like digital twins to really understand what could happen in those products and to really optimize the products. In this session we will discuss how customers are leveraging Azure HPC to solve these types of challenges in the Manufacturing industry.
Guides & Demos
The Azure HPC stack is comprehensive and full featured: purpose-built infrastructure, high performing storage options, fast & secure networking, workload orchestration services, and data science tools integrated across cloud, hybrid, and the edge make Azure a force with which to be reckoned. It is important to see where your opportunity is in this stack, to understand what Azure can do for your HPC needs.
The following videos are excellent deep dives into some of the infrastructure and services well-aligned for HPC use case scenarios:
Presented by: Rob Futrick, Principal Program Manager, Azure HPC
Summary: This is an excellent summary of all the HPC oriented services residing on Azure, and how you can take advantage of them for your needs.
Presented by: Mike Kiernan, Sr. Program Manager for Azure HPC
Summary: Learn how you can use the Azure Batch Service with Containers to better orchestrate your workload in the cloud
Presented by: Cormac Garvey, Sr. Program Manager for Azure HPC
Summary: This video outlines the basics for you to be able to stand up an end-to-end HPC environment using Azure CycleCloud.
Presented by: Scott Jeschonek, Cloud Storage Specialist at Microsoft Azure
Summary: This video provides an overview of a variety of different storage options to run HPC workloads on including Blob Storage, Azure NetApp Files, Azure HPC Cache, and others.
Things are shaping up to be a very productive 2021, and we are tremendously excited and honored to participate!