NVIDIA Tesla P100 Available on Google Cloud Platform

NVIDIA Tesla P100 and K80 GPUs will be available on Google Cloud Platform, starting early next year.

Delivering the power of our Pascal GPU architecture from the cloud gives businesses another great option for helping to put their data to work and build AI services.

On Google Cloud Platform, Tesla P100 GPUs will be available to Google Compute Engine and Google Cloud Machine Learning users around the world. The Tesla P100 delivers high performance and efficiency to power the most computationally demanding applications — including a 12x increase in neural network training performance compared with a previous-generation offering.

The Tesla K80 GPU accelerator delivers exceptional performance, with increased throughput that allows researchers to advance their scientific discoveries and developers to boost their web services. Using the NVIDIA GPUs on the Google Cloud Platform, customers only pay for what they use on a per-minute basis.

The TensorFlow deep learning framework, originally developed by Google researchers, runs up to 50% faster on NVIDIA Pascal-based GPUs. This can ease the pathway for creating and deploying AI services on the Google Cloud Platform. All other deep learning frameworks are accelerated on NVIDIA GPUs, as well as more than 400 industry-leading applications.

The Power of AI in the Cloud

Cloud computing is transforming industries by making it easier for businesses to analyze and gain insight from their data. NVIDIA is at the center of this trend, offering the most advanced GPUs available, serving as the computational engine for the cloud, and enabling a new era of AI services and applications.

GPU computing lets businesses work with their information in the cloud, avoiding the costs of building local data centers. And it allows them to scale their computing power up and down as business needs dictate.

Learn more about NVIDIA GPU cloud computing and read Google’s announcement.

The post NVIDIA Tesla P100 Available on Google Cloud Platform appeared first on The Official NVIDIA Blog.

by Ian Buck via The Official NVIDIA Blog