Blog Post

Educator Developer Blog
3 MIN READ

Azure Cloud–GPU for DataScience and Academic Activities such as Cloud Rendering

Lee_Stott's avatar
Lee_Stott
Icon for Microsoft rankMicrosoft
Mar 21, 2019
First published on MSDN on Jan 10, 2017

I am really excited in the way some UK Universities are using Azure GPU Cloud services, back in Dec 2016 we announced the general availability of the Azure N-Series. GPU Cloud Compute services.  The availability of Cloud GPU was certainly a interest point for students, academics and researchers and I had a number of UK Universities approach me  around how they could use the Azure N-Series virtual machines to reduce significant costs and time constraints for teaching, learning and research.

As an example one institution wanted to utilise the Azure N series  NVIDIA® GPUs for rendering their design and animation student content. The institution had limited NVIDIA GPU availability and as a a result they faced a issue of hardware availability, costs and time constraints. As a result of this this trialled the N Series as a complete rendering replacement for PCs on Campus with great success from no only the output but also the costs..

One of the key advantages to them was the global access to the variety of sizes of  N-series Hardeare and the availability of  West Europe services. With the availability of Cloud GPU they had the opportunity of running the process and jobs within a European Data Center from the 1st of Dec and being able to test the service for £1 rather than the investment of £1000 for physical hardware of similar specifications. So its great to see that from the thousands of customers who participate in the N-Series preview since we launched it back in August we have a number of UK Universities using a variety of services and applications.

What is key about these services that we are open and want your feedback. The Azure GPU team are working with NVIDIA to ensure that we can offer the enhanced performance and the workload requirements to make this a completely turnkey experience for you.

So what services are available and how can they potentially save you £k on hardware, support and implementation.

Azure NC virtual machines—GPU compute

Azure NC-based instances are powered by NVIDIA Tesla® K80 GPUs and provide the compute power required to accelerate the most demanding high-performance computing (HPC) and AI workloads. Customers can now use these instances to run deep learning training jobs, HPC simulations, rendering, real-time data analytics, DNA sequencing, and many more CUDA® accelerated tasks. Additionally, customers have the option to utilize RDMA (Remote Direct Memory Access) over InfiniBand for scaling jobs across multiple instances. InfiniBand provides close to bare-metal performance even when scaling out to 10s, 100s, or even 1,000s of GPUs across hundreds of machines. This will allow you to submit tightly coupled jobs using frameworks like the Microsoft Cognitive Toolkit (CNTK), Caffe, or TensorFlow, enabling training for natural language processing, image recognition, and object detection.

Azure NV virtual machines—GPU visualization

Azure NV-based instances are powered by NVIDIA Tesla M60 GPUs and provide NVIDIA GRID capabilities on demand. Scientists, designers, and engineers can now utilize these new instances for running hardware-accelerated workstation applications, designing the next concept car, or creating the next blockbuster movie. These instances support applications utilizing both DirectX and OpenGL.

Example of Pricing

For live Pricing see the Azure Price Calculator https://azure.microsoft.com/en-us/pricing/details/virtual-machines/linux/

Learn more:

Resources
Updated Mar 21, 2019
Version 2.0
No CommentsBe the first to comment