The new cloud service, called Azure ND A100 v4 Cloud GPU, is powered by Nvidia’s A100 Tensor Core GPU and targets users that demand high performance and heavy workloads, and allows customers to Microsoft’s most demanding customers scale services without delay.
Microsoft says that the ND A100 v4 is built to leverage industry-standard high-performance computing (HPC) and Artificial Intelligence (AI) tools and libraries, and can be used without need any special software or framework.
According to Microsoft’s internal benchmarks, 164 ND A100 v4 VMs yielded a high-performance Linpack (HPL) result of 16.59 petaflops. With this HPL result, the cloud service entered the Top 20 in the list of Top 500 fastest supercomputers in the world as of November 2020.
Microsoft says that the ND A100 v4 starts with a single virtual machine and eight A100 Tensor Core GPUs that depend on the Nvidia Ampere architecture and can scale up to 1000 GPUs in a single cluster.
Anyone can bring AI and HPC-intensive workloads to the cloud via ND A100 v4 with minimal fuss, but for customers who prefer an Azure-native approach, Azure Machine Learning offers a tuned virtual machine (pre-installed with necessary drivers and libraries) and optimized container-based environment for the ND A100 v4 series.
Currently, the ND A100 v4 service is available in four Azure regions: US East, US West, Western Europe, and South Central America.
#Microsoft #Azure #unveils #fastest #public #cloud #supercomputer #powered #Nvidia #A100