Erdal Cosgun-co author of this blog- is the Lead Data Scientist in the Microsoft Genomics Team and working on the Interactive Analysis component of Terra on Microsoft Azure.
Terra on Microsoft Azure cloud computing continues to evolve to meet the demand for high-performance virtual machines (VMs) capable of handling intensive workloads. We’re thrilled to announce the availability of NCSv3 (NC6s_v3 and NC12s_v3) series GPU VMs, which bring unprecedented computational power to Terra, making it ideal for a wide range of applications such as Artificial Intelligence (AI), Machine Learning (ML), and Deep Learning (DL) for biomedical analysis.
Harnassing the power of NCSv3-series GPU VMs
NCSv3-series GPU VMs provide an exceptional level of performance by leveraging NVIDIA’s advanced GPU technology. You can take advantage of these updated GPUs powered by NVIDIA Tesla V100 GPUs on Terra for traditional High Performance Computing (HPC) workloads such as reservoir modeling, DNA sequencing, protein analysis, Monte Carlo simulations, and others. Working in Terra lets you easily access several datasets in the cloud such as Microsoft Genomics Data Lake without having to compete for fixed HPC resources. In addition to GPUs, the NCv3-series VMs are also powered by Intel Xeon E5-2690 v4 (Broadwell) CPUs. [Ref]
Accelerate model training and inferencing
Whether training neural networks or deploying real-time inferencing solutions, these virtual machines accelerate model training and inferencing tasks on Terra. With the NCSv3-series GPUs, you can bring your ML/DL models to real-life problems faster, empowering breakthroughs in biomedical research.
A new era of High-Performance Computing (in the cloud)
The availability of NCSv3-series GPU VMs on Terra on Microsoft Azure brings a new era of high-performance computing to the biomedical computing landscape. With their exceptional GPU capabilities and extensive memory, these virtual machines let you tackle the most demanding workloads – from AI and ML to deep learning – all within the ecosystem of Terra on Microsoft Azure.
Try it out in a Microsoft Azure Public Workspace
To get started, try our featured workspace, which includes a sample notebook for training the deep learning models. As of July 2023, running this notebook on the NC6s_v3 GPU VM in your own Terra on Microsoft Azure workspace costs about $3.06 per hour with the included data.Please use Microsoft Azure Pricing calculator for the latest price information.
Important Note: Be sure to check GPU quotas from the Azure portal before deploying the GPU VMs on Terra on Microsoft Azure. Please check the Terra Support Article for step-by-step instructions as well as caveats and troubleshooting about quota increase requests.
View quotas – Azure Quotas | Microsoft Learn
