Users of the Google Cloud Platform that are looking to use powerful Nvidia GPUs on Google Compute Engine will now be able to.
Support is being added for Nvidia P100 GPUs in Beta, and Nvidia K80 GPUs are now generally available.
Basically, Google is looking to give businesses GPUs in the cloud that can be paid for in a flexible manner, the more powerful the GPU the greater the capacity for carrying out complex machine learning workloads.
To sweeten the deal, Google is also adding a sustained pricing model which means that GPU’s that are running for a sustained period of time will be given an up to 30% discount, although this depends on usage – details on that can be found here.
Google reckons that bare-metal performance can be provided: “Cloud GPUs are offered in passthrough mode to provide bare-metal performance. Attach up to 4 P100 or 8 K80 per VM (we offer up to 4 K80 boards, that come with 2 GPUs per board). For those looking for higher disk performance, optionally attach up to 3TB of Local SSD to any GPU VM,” the company said in a blog post.
In further good news for users of Google’s cloud, there’s also the introduction of Google Cloud Dataprep in a public beta.
Content from our partners
Is cloud computing a victim of its own success?
Google’s been revealing more details about its cloud strategy
Why the UK is ground zero for the Cloud Wars
GCD is basically a serverless data preparation tool that’s run as a fully-managed cloud service that was developed in collaboration with Trifacta.
The service is designed to give analysts and data scientists the ability to visually explore and prepare data for analysis within GCP.
One of the most valuable elements of the technology is its speed, it takes away the need for manual cleaning of data by detecting schemas, types, joins, and things such as missing values all automatically.
The service will be natively integrated with GCP services such as BigQuery and others.
Google’s move follows in the footsteps of market pace setter Amazon Web Services that launched AWS Glue last year.
