Microsoft powers transformation at NVIDIA’s GTC Digital Convention

The world of supercomputing is evolving. Work as soon as restricted to high-performance computing (HPC) on-premises clusters and conventional HPC situations, is now being carried out on the edge, on-premises, within the cloud, and in all places in between. Whether or not it’s a producer working superior simulations, an power firm optimizing drilling by means of real-time properly monitoring, an structure agency offering skilled digital graphics workstations to staff who have to work remotely, or a monetary companies firm utilizing AI to navigate market threat, Microsoft’s collaboration with NVIDIA makes entry to NVIDIA graphics processing items (GPU) platforms simpler than ever.

These fashionable wants require superior options that have been historically restricted to some organizations as a result of they have been laborious to scale and took a very long time to ship. At this time, Microsoft Azure delivers HPC capabilities, a complete AI platform, and the Azure Stack household of hybrid and edge choices that instantly deal with these challenges.

This yr throughout GTC Digital, we’re spotlighting a few of the most transformational purposes powered by NVIDIA GPU acceleration that spotlight our dedication to edge, on-prem, and cloud computing. Registration is free, so signal as much as learn the way Microsoft is powering transformation.

Visualization and GPU workstations

Azure allows a variety of visualization workloads, that are essential for desktop virtualization in addition to skilled graphics reminiscent of computer-aided design, content material creation, and interactive rendering. Visualization workloads on Azure are powered by NVIDIA’s world-class GPUs and Quadro know-how, the world’s preeminent visible computing platform. With entry to graphics workstations on Azure cloud, artists, designers, and technical professionals can work remotely, from wherever, and from any related system. See our NV-Collection digital machines (VMs) for Home windows and Linux.

Synthetic intelligence

We’re sharing the discharge of the up to date execution supplier in ONNX Runtime with integration for NVIDIA TensorRT 7. With this replace, ONNX Runtime can execute open Open Neural Community Alternate (ONNX) fashions on NVIDIA GPUs on Azure cloud and on the edge utilizing the Azure Stack Edge, benefiting from the brand new options in TensorRT 7 like dynamic form, blended precision optimizations, and INT8 execution.

Dynamic form assist allows customers to run variable batch measurement, which is utilized by ONNX Runtime to course of recurrent neural community (RNN) and Bidirectional Encoder Representations from Transformers (BERT) fashions. Blended precision and INT8 execution are used to hurry up execution on the GPU, which allows ONNX Runtime to raised steadiness the efficiency throughout CPU and GPU. Initially launched in March 2019, TensorRT with ONNX Runtime delivers higher inferencing efficiency on the identical {hardware} when in comparison with generic GPU acceleration.

Moreover, the Azure Machine Studying service now helps RAPIDS, a high-performance GPU execution accelerator for information science framework utilizing the NVIDIA CUDA platform. Azure builders can use RAPIDS in the identical approach they at present use different machine studying frameworks, and together with Pandas, Scikit-learn, PyTorch, and TensorFlow. These two developments characterize main milestones in the direction of a really open and interoperable ecosystem for AI. We’re working to make sure these platform additions will simplify and enrich these developer experiences.


Microsoft offers numerous options within the Clever Edge portfolio to empower prospects to ensure that machine studying not solely occurs within the cloud but additionally on the edge. The options embrace Azure Stack Hub, Azure Stack Edge, and IoT Edge.

Whether or not you’re capturing sensor information and inferencing on the Edge or performing end-to-end processing with mannequin coaching in Azure and leveraging the educated fashions on the edge for enhanced inferencing operations Microsoft can assist your wants nevertheless and wherever you might want to.

Supercomputing scale

Time-to-decision is extremely essential with a world economic system that’s continuously on the transfer. With the accelerated tempo of change, firms are in search of new methods to collect huge quantities of knowledge, practice fashions, and carry out real-time inferencing within the cloud and on the edge. The Azure HPC portfolio consists of purpose-built computing, networking, storage, and software companies that can assist you seamlessly join your information and processing wants with infrastructure choices optimized for numerous workload traits.

Azure Stack Hub introduced preview

Microsoft, in collaboration with NVIDIA, is asserting that Azure Stack Hub with Azure NC-Collection Digital Machine (VM) assist is now in preview. Azure NC-Collection VMs are GPU-enabled Azure Digital Machines out there on the sting. GPU assist in Azure Stack Hub unlocks quite a lot of new answer alternatives. With our Azure Stack Hub {hardware} companions, prospects can select the suitable GPU for his or her workloads to allow Synthetic Intelligence, coaching, inference, and visualization situations.

Azure Stack Hub brings collectively the complete capabilities of the cloud to successfully deploy and handle workloads that in any other case will not be potential to convey right into a single answer. We’re providing two NVIDIA enabled GPU fashions through the preview interval. They’re out there in each NVIDIA V100 Tensor Core and NVIDIA T4 Tensor Core GPUs. These bodily GPUs align with the next Azure N-Collection VM varieties as follows:

  • NCv3 (NVIDIA V100 Tensor Core GPU): These allow studying, inference and visualization situations. See Standard_NC6s_v3 for the same configuration.
  • TBD (NVIDIA T4 Tensor Core GPU): This new VM measurement (out there solely on Azure Stack Hub) allows mild studying, inference, and visualization situations.

Hewlett Packard Enterprise is supporting the Microsoft GPU preview program as a part of its HPE ProLiant for Microsoft Azure Stack Hub answer.“The HPE ProLiant for Microsoft Azure Stack Hub answer with the HPE ProLiant DL380 server nodes are GPU-enabled to assist the utmost CPU, RAM, and all-flash storage configurations for GPU workloads,” stated Mark Evans, WW product supervisor, HPE ProLiant for Microsoft Azure Stack Hub, at HPE. “We sit up for this collaboration that can assist prospects discover new workload choices enabled by GPU capabilities.” 

Because the main cloud infrastructure supplier1, Dell Applied sciences helps organizations take away cloud complexity and lengthen a constant working mannequin throughout clouds. Working carefully with Microsoft, the Dell EMC Built-in System for Azure Stack Hub will assist further GPU configurations, which embrace NVIDIA V100 Tensor Core GPUs, in a 2U type issue. This may present prospects elevated efficiency density and workload flexibility for the rising predictive analytics and AI/ML markets. These new configurations additionally include automated lifecycle administration capabilities and distinctive assist.

To take part within the Azure Stack Hub GPU preview, please ship us an e mail at this time. 

Azure Stack Edge preview

Azure Stack Edge Hi-Res

We additionally introduced the enlargement of our Microsoft Azure Stack Edge preview with the NVIDIA T4 Tensor Core GPU. Azure Stack Edge is a cloud managed equipment that gives processing for quick native evaluation and insights to the information. With the addition of an NVIDIA GPU, you’re capable of construct within the cloud then run on the edge. For extra details about this thrilling launch please see the detailed weblog.

GTC Digital

Microsoft session recordings will likely be out there on the GTC Digital web site beginning March 26. You could find an inventory of the Microsoft digital classes together with corresponding hyperlinks within the Microsoft Tech Neighborhood weblog right here.

1 IDC WW Quarterly Cloud IT Infrastructure Tracker, Q3 2019, January 2020, Vendor Income

Show More

Related Articles

Leave a Reply

Your email address will not be published. Required fields are marked *

Back to top button