Published On: Thu, Sep 21st, 2017

Google Cloud adds support for some-more absolute Nvidia GPUs

Google Cloud Platform announced support for some absolute Nvidia GPUs on Google Compute Engine today.

For starters, a association is creation Nvidia K80 GPUs generally available. At a same time, it’s rising support for Nvidia P100 GPUs in Beta along with a new postulated pricing model.

For companies operative with appurtenance training workloads, carrying entrance to GPUs in a cloud provides them with coherence to compensate for what they use with by a notation pricing, and a postulated pricing indication means if they finish adult using a GPUs for a postulated duration of time, they get adult to a 30 percent discount, depending on a usage. In other words, they don’t get whacked with a outrageous check since they like a service.

And Google is claiming that this proceed could yield unclothed steel kind of performance. “Cloud GPUs are charity in passthrough mode to yield bare-metal performance. Attach adult to 4 P100 or 8 K80 per VM (we offer adult to 4 K80 boards, that come with 2 GPUs per board),” Google wrote in a blog post announcing a GPU support.

What’s more, Google is charity a coherence to run GPU workloads in practical machines or containers and they are delivering a use in 4 tellurian locations.

Screenshot 2017-09-21 12.28.20.png

Map: Google


Google is saying business use this use for a accumulation of compute-intensive tasks such as genomics, computational financial and training and deduction on appurtenance training models. The dual opposite chips give teams a coherence to select a chips they need to optimally run their workloads while balancing speed and pricing. Among their early GPU cloud business is Shazam, that uses GPU to assistance fuel the song marker service.

Featured Image: Bloomberg/Getty Images

About the Author

Leave a comment

XHTML: You can use these html tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>