Breaking

Tuesday, May 1, 2018

Google Cloud expands GPU portfolio with Nvidia Tesla V100

Nvidia Tesla V100 GPUs are now publicly available in beta on Google Compute Engine and Kubernetes Engine.


Google Cloud on Monday announced the expansion of its GPU portfolio. Nvidia Tesla V100 GPUs are now publicly available in beta on Google Compute Engine and Kubernetes Engine, while Nvidia Tesla P100 GPUs are now generally available.

The Tesla V100 GPU equates to 100 CPUs, giving customers more power to handle computationally demanding applications, like machine learning, analytics, and video processing. As an example, Google points to one of its customers, the marketing automation provider LeadStage, that's using V100 and P100 GPUs on Google Compute Engine to make poor-quality data from handwritten documents and survey drawings into machine-readable data.

As GPUs become increasingly important for running AI workloads, public cloud providers like Google, IBM, and Oracle have been competing to offer the latest from Nvidia on their cloud infrastructure.

As for Google's new offering, customers can now get as many as eight Tesla V100 GPUs, 96 vCPU, and 624GB of system memory in a single VM. Meanwhile, NVLINK interconnects deliver up to 300GB per second of GPU-to-GPU bandwidth, boosting performance on deep learning and HPC workloads by up to 40 percent, Google says. The V100s are now available in the us-west1, us-central1 and Europe-west4 regions.

In Google's Kubernetes Engine, V100s can be used to turbocharge containers, and they can be scaled down with Google's Cluster Autoscaler when not in use.

As for the Tesla P100, customers can select up to four P100 GPUs, 96 vCPUs and 624GB of memory per VM. The P100 is also now available in Europe-west4 in addition to us-west1, us-central1, us-east1, Europe-west1, and Asia-east1.


1 comment: