BrandedLogoDesigns Reviews
: Customers from Google's Cloud Platform can also rely on additional computing power from Tesla GPUs in the future. The Group has started GPU support for the first regions as part of a beta phase.Google now also offers GPU support for its Cloud Platform customers. This is especially for companies who want to perform computer-intensive tasks in the areas of machine learning or AI in a virtualized cloud environment. The new offer is initially available as part of a beta phase in the regions us-east1, asia-east1 and europe-west1.
The use of up to eight Nvidia GPUs will dramatically shorten the development time of machine learning solutions in the Google Cloud.
As a hardware, Google is using Nvidia's Tesla K80 GPUs. They each have 2,496 stream processors and 12 GB of GDDR5 memory. On the software side, the Nvidia cards support common AI frameworks such as TensorFlow, Theano, Torch, MXNet and Caffe as well as Nvidia's Cuda software for creating GPU-driven applications. Through the use of the powerful Hadware, Google promises dramatically shortened development times in conjunction with its in-house Cloud-Machine-Learning service (Cloud ML).
As required, customers can connect one, two, four or eight GPUs for additional computing power. Billing is carried out in the minute minute with a minimum time of 10 minutes. For each booked GPU, Google is charging US $ 0.77 in the Europe and Asia regions, while the US is priced at US $ 0.7 per hour.New software and hardware for powerful machine learning
It was only in the middle of the month that the IT company presented the new version of its AI library with TensorFlow 1.0.0. The open-source project is to enable even more powerful solutions and to be deployed more flexibly at the same time. Google uses the technology in practice, for example, for its intelligent image search in Google Photos or its translator software Google Translate.
Nvidia had already introduced new models of its Quadro series with Pascal architecture at the beginning of February. The more powerful versions are to be used in workstations for the calculation of complex simulations and VR- or AI-workflows.
BrandedLogoDesigns Reviews
No comments:
Post a Comment