Microsoft and Cray are enabling Azure customers to run certain workloads on Cray supercomputing hardware that's installed in Microsoft datacenters.
The pair announced on October 23 that customers in select Microsoft datacenters will have the option to use Cray XC and Cray CS supercomputers attached to Cray ClusterStor storage systems. The systems will be connected directly to Azure and will integrate with Azure virtual machines, Azure Data Lake storage, as well as with Microsoft's artificial intelligence (AI) and Machine Learning services.
Microsoft and Cray are targeting customers who want to run high-performance computing, artificial intelligence, modeling simulations, and advanced analytics workloads "at supercomputing scale." This new Azure option is for customers who may be unable to purchase and maintain their own on-premises Cray systems.
This isn't the first time Microsoft and Cray have worked together. Last December, Microsoft and Cray teamed on scaling the Microsoft Cognitive Tookit so it could run on Cray XC50 supercomputers. Back in 2005, Cray co-founder Burton Smith joined Microsoft and remains a Technical Fellow with the company, working on hardware, devices, and quantum computing.
Microsoft also is working on giving Azure customers access to field-programmable gate array (FPGA) processing power for data-intensive tasks like deep-neural-networking. Microsoft is bringing its Brainwave AI framework on running on FPGAs to external developers in order to push the boundary of the types of AI-influenced services possible to deploy in the cloud, including computer vision, natural-language processing, and speech. Microsoft is aiming to make this technology commercially available to Azure users in 2018.
PREVIOUS AND RELATED COVERAGE
Microsoft and Cray recently unveiled a collaborative work scaling the Microsoft Cognitive Toolkit to run on a Cray XC50 supercomputer, getting results faster then before.
Today, Microsoft has a lot of projects and technologies focused on parallel and multi-core computing. But this wasn't always the case, as Microsoft Technical Fellow Burton Smith acknowledges in the latest installment of my "Microsoft Big Brains" series.