New Azure HPC and AI Collaboration Center Program

New Azure HPC & AI Collaboration Centers Will Share Best Practices for Unlocking Customer Innovations & Productivity with HPC & AI

The Azure HPC & AI Collaboration Centers Program

Microsoft Azure is proud to have recently launched the Azure HPC & AI Collaboration Centers program, which funds collaborations with well-known institutions to develop and share these best practices with the HPC & AI communities. The Collaboration Centers program is delivered in partnership with our friends at AMD, Intel, and NVIDIA.

The first five Azure Collaboration Centers are already underway with Duke University, Purdue University, Red Oak Consulting, Numerical Algorithms Group (NAG), and RedLine Performance. We plan to announce further exciting collaboration centers over the coming months.

Full Spectrum of HPC & AI in the Cloud

Microsoft Azure delivers differentiated cloud capabilities that are purpose-built for the full spectrum of HPC & AI workloads, including the most demanding HPC & AI use cases.

At the core of the infrastructure are HPC & AI clusters using HPC-class InfiniBand interconnect, with genuine low latency (~1us) and high bandwidth (typically 200Gbps), using the latest processor technologies from AMD, Intel and NVIDIA. These performance advantages have enabled Azure to show scaling of supercomputing workloads to 12x more CPU cores than other clouds, and to deliver cost-to-solution leadership over other solutions. Azure has also built genuine HPC & AI expertise across the spectrum from engineering teams, to customer engagements, to through-life customer support. We are also committed to nurturing a thriving ecosystem of partners who can help our customers get the maximum impact using Azure.

A strong enabler for the success of our customers will be demonstrating and sharing best practices around how Azure HPC & AI capabilities can be used to unlock customer innovations and productivity.

Nidhi Chappell, Microsoft Azure’s Head of Product for Azure HPC & AI, has described how the program will benefit customers: “It is exciting to see experts from the HPC & AI communities collaborating with Azure and our partners at AMD, Intel, and NVIDIA to showcase best practices for how customers can unlock innovations and productivity using Azure HPC & AI products”.

Kumaran Siva, Corporate Vice President, EPYC Cloud Business, AMD, said: “AMD has an extensive relationship with Azure and we are excited to continue our work together to push the growth of HPC in the cloud with the Microsoft’s Azure HPC & AI Collaboration Centers Program. This program will further reiterate how customers can use the leadership performance capabilities that the AMD EPYC processors provide the HB-series VMs from Azure, to enable HPC workloads and applications that deliver impact and value to their business.”

Trish Damkroger, Vice President, Intel Data Center Group, said: “Intel is proud to be a leading partner in Microsoft’s Azure HPC & AI Collaboration Centers Program. This program will demonstrate how customers can address challenging workloads in these areas through Azure’s instance types ranging from general purpose to high performance computing based on Intel’s Xeon Scalable processors utilizing market leading features like Intel Turbo Boost Technology, Intel AVX-512, Intel DL Boost, and Intel SGX.”

“Extending NVIDIA’s partnership with Azure in support of Microsoft’s Azure HPC & AI Collaboration Centers Program opens interesting possibilities for the future of accelerated computing.” said Paresh Kharya, Senior Director of Product Management for AI Systems at NVIDIA. "This program will enable customers to push the boundaries of their AI, ML and HPC applications by using the highest-performance Azure instances powered by NVIDIA Ampere architecture GPUs and NVIDIA InfiniBand networking technology.”

The Five Inaugural Centers

• The Duke University Azure HPC Collaboration Center, in partnership with NVIDIA, will use Azure GPU capabilities to develop personalized hemodynamics simulation capabilities in the cloud with an initial target of improving our understanding and treatment of coronary bifurcation lesions.
• The Purdue University Azure HPC & AI Collaboration Center, in partnership with AMD, will demonstrate best practices for using the flexibility and scalability of Azure HPC cloud to augment on-premises HPC systems, such as Purdue’s AMD-based Anvil supercomputer.
• The Red Oak Consulting Azure HPC+AI Collaboration Center, in partnership with Intel, will develop and showcase best practices for successfully building HPC clusters on Azure using Intel technology, along with key educational materials.
• The NAG Azure HPC+AI Collaboration Center, in partnership with NVIDIA, will be focused on developing best practices for the deployment of scalable Machine Learning.
• The RedLine Performance Azure HPC+AI Collaboration Center, in partnership with Intel, will implement and demonstrate the National Weather Service (NWS) Finite Volume Cubed Sphere Global Forecast System (FV3GFS) workflow on Azure, from the Global Data Assimilation System (GDAS) through post-processing.

A series of short videos introducing each Collaboration Center and their planned work is available here: Duke University, Purdue University, Red Oak Consulting, NAG, Redline Performance  

Next Steps

We look forward to sharing the best practices and success stories from the inaugural Collaboration Centers with the wider HPC & AI communities, and to announcing further exciting Collaboration Centers in the coming months.