Google cloud platform

Google Cloud adds support for more powerful Nvidia GPUs

Google Cloud Platform reported help for some capable Nvidia GPUs on Google Compute Engine today.

First of all, the organization is making Nvidia K80 GPUs for the most part accessible. In the meantime, it’s starting help for Nvidia P100 GPUs in Beta alongside another managed evaluating model.


For organizations working with machine learning workloads, approaching GPUs in the cloud furnishes them with adaptability to pay for what they use with incrementally evaluating, and the maintained estimating model means in the event that they wind up running the GPUs for a supported timeframe, they get up to a 30 percent rebate, contingent upon the utilization. At the end of the day, they don’t get whacked with a tremendous bill since they like the administration.

Furthermore, Google is guaranteeing that this approach could give exposed metal sort of execution. “Cloud GPUs are offered in passthrough mode to give exposed metal execution. Connect up to 4 P100 or 8 K80 per VM (we present to 4 K80 sheets, that accompany 2 GPUs per board),” Google wrote in the blog entry declaring the GPU bolster.

Likewise, Google is putting forth the adaptability to run GPU workloads in virtual machines or compartments and they are conveying the administration in four worldwide areas.

Google is seeing clients utilize this administration for an assortment of figure escalated undertakings, for example, genomics, computational back and preparing and surmising on machine learning models. The two unique chips give groups the adaptability to pick the chips they have to ideally run their workloads while adjusting pace and estimating. Among their initial GPU cloud clients is Shazam, which utilizes GPU to help fuel its music ID benefit.

Leave a Reply

Your email address will not be published. Required fields are marked *

%d bloggers like this: