The new series is designed for training and interface. Microsoft says the ND-series is powered by NVIDIA’s Tesla P40 GPU and is built on the Pascal Architecture. This allows customers to get double the performance compared to older generations. Improvements in AI workloads are also available with the new series. Microsoft is expanding the range because its NVIDIA Tesla-based GPUs have been a huge success. The company first launched these GPUs last year. The power and performance of the GPUs allowed Azure customers to expand their deep learning and artificial intelligence workloads. Among the applications that have leveraged the Tesla-powered technology are speech training, autonomous vehicle technology, and image recognition. The ND-series builds on these capabilities and brings customers more power.
Improving Performance in Azure
Because of the larger memory size (24GB), customers can accommodate bigger neural net models. There is also RDMA and InfiniBand connectivity, which is a staple on the NC-series. HPC workloads also get performance boosts from the new GPU series. This benefits modelling such as DNA sequencing, protein analysis, rendering, and more. Microsoft says the new series shows how the cloud is transforming data and how people work: “One of the promises of the cloud has always been agility. As your computational needs change and expand/shrink and as the models improve/mature, you want to leverage the latest and greatest hardware for computation without waiting for existing hardware to age. With Azure, this will now become possible.”
