Thanks to its superfast servers in their mega data centers equipped with GPUs Facebook, Amazon and Google can now recognize voices and images. Nevertheless, many companies turn to cloud, as they cannot afford that level of resources for deep learning. In this case, the servers in remote data centers take care of the heavy tasks. Microsoft is one of the few companies offering remote servers with GPUs with its cloud service Azure which excel in machine learning tasks. Microsoft is getting competition from Nimbix which is also using GPUs in remote servers but unlike Microsoft’s older versions of Nvidia GPUs, the Nimbix is using Nvidia’s latest Pascal architecture GPUs. Customers can get a virtual machine with an admittance to bare metal server hardware after leasing time in the cloud. Among the Nvidia’s fastest GPUs the P100s in IBM Power S822LC servers is been used by Nimbix to offer cloud service for its customers.
There are other benefits to the cloud service as well such as on the server side, a high-speed NVLink interconnect links the CPU to the GPU along with other components at speeds 2.5 times faster than PCI-Express 3.0. High-performance computing application is the target of Nimbix cloud service.