performance2

Give every idea the opportunity to grow

The Power of Two

Ensuring the right capabilities and the raw processing power are always at hand where and when you need them, is the ultimate purpose behind the NVIDIA and Google Cloud partnership.

The NVIDIA T4, V100, P4 and P100 GPUs are all generally available on Google Cloud Platform to accelerate multiple flavors of deep learning, analytics, physical simulation, video transcoding, and molecular modeling.

NVIDIA V100 GPUs

The world's most powerful GPU

NVIDIA Tesla V100 gpus

Best performance for dense workloads like ML training and high performance computing (HPC).

With the power to accelerate the most demanding AI, high performance computing, data science, and graphics workloads, the NVIDIA V100 sits at the very forefront of data center GPU performance.

power2

acceleration2

NVIDIA T4 GPUs

Unrivaled technology and performance

NVIDIA T4 GPUs

Universal GPU, with best performance for distributed workloads like ML inference and model development, and next-generation graphics visualization.

NVIDIA T4 GPUs deliver revolutionary, multi-precision performance to accelerate a wide range of cloud workloads including analytics, virtualized graphics, video transcoding, and 3D rendering.

NVIDIA P100 GPUs

Advanced specifications and capabilities

NVIDIA P100 GPUs

Entry-level performance for dense workloads like ML training and HPC.

Crafted to deliver the world’s fastest compute node, the NVIDIA P100 powers superior strong-scaling performance for HPC and hyper scale applications.

optimization2

intelligence2

NVIDIA P4 GPUs

Exceptional agility and inspired design

NVIDIA P4 GPUs

Entry-level performance for distributed workloads like ML inference and model development, and video transcode.

Purpose-built to run deep learning workloads and enabling smart, responsive AI-based services, the NVIDIA P4 is the GPU for ultra-efficient scale out servers.

Accessing Processors on Google Cloud

Beyond directly selecting the processor you want via the Google Cloud console, you can select these GPUs on Google Cloud platforms and services, designed to provide you complete solutions for your workloads.
The top Google Cloud services are captured below.

Google Cloud services

Service

Description

GPUs
integrated

Artificial Intelligence

Jupyter Notebooks for BERT training and inference | UNet inference

Yes

NVIDIA GPU Images: Tensorflow, Pytorch, MXNET, Caffe1, CNTEK, XGBoost, RAPIDS (NV ML library for Data Analytics) and generic CUDA  

Yes

Managed JupyterLab environments with TensorFlow/PyTorch/R and generic GPU flavors | Working to integrate TRTIS  (TensorRT Inference Server)

Yes

AI Platform CMLE Prediction

NVIDIA and Google Cloud joint development on Kubeflow

Yes

AI Platform Training

Yes

AutoML Video Edge

Seamless integration of AutoML Edge Video with NVIDIA edge devices like Jetson Xavier, T4

Private
Alpha

NVIDIA GPUs

NVIDIA
V100

NVIDIA
T4

NVIDIA
P4

NVIDIA
P100

Artificial Intelligence

 
AI Platform CMLE Prediction
AI Platform CMLE Prediction

AutoML Video Edge

Google Cloud services

Service

Description

GPUs
integrated

High performance computing

Compute EngineAI Hub

Scalable, high performance VMs 

Yes

Managed Kubernetes

Yes

Batch on GKE
Job submission tool for HPC and HTC

Yes

Anthos – GKE on Prem
In certification phase

From
2Q-2020

NVIDIA GPUs

NVIDIA
V100

NVIDIA
T4

NVIDIA
P4

NVIDIA
P100

High performance computing

 

Compute EngineAI Hub

Batch on GKE

Anthos – GKE on Prem

Google Cloud services

Service

Description

GPUs
integrated

Graphics rendering

Zync

Batch rendering service

Yes

NVIDIA GPUs

NVIDIA
V100

NVIDIA
T4

NVIDIA
P4

NVIDIA
P100

Graphics rendering

 

Zync

Google Cloud services

Service

Description

GPUs
integrated

Data analytics

Cloud DataFlow

Real-time batch and stream data processing.

From
2Q-2020

Cloud DataProc

Managed SPARK and Hadoop service

Yes

Big Query ML

Yes

NVIDIA GPUs

NVIDIA
V100

NVIDIA
T4

NVIDIA
P4

NVIDIA
P100

Data analytics

 

Cloud DataFlow

Cloud DataProc

Big Query ML

Google Cloud services

Service

Description

GPUs
integrated

Management tools
Stackdriver
Can support via plugins

NVIDIA GPUs

NVIDIA
V100

NVIDIA
T4

NVIDIA
P4

NVIDIA
P100

Management tools
 

Zync

The
Power of Two

and you

Book your discovery session now.

Sunpower

A faster route to solar power

SunPower harnesses the power of NVIDIA V100 GPU-enabled virtual machines on Google Cloud to help homeowners design their ideal solar solution in seconds, and bring clean, sustainable energy to homes across the US and beyond.

Jake wachman

Using NVIDIA GPUs through Google Cloud has helped us focus on core design problems, achieve our goals faster, and create new designs quickly.

Jake Wachman, Vice President, Digital, SunPower Corporation

Unleash the
Power of Two

Technology
showcase

Artificial intelligence from NVIDIA and Google Cloud:
A technology showcase

Technology
showcase

Transform your business with artificial intelligence from NVIDIA and Google Cloud:
A technology showcase

Executive
guide

Graphics visualization from NVIDIA and Google Cloud:
The business benefits

How to guide

Streamline your Machine Learning projects with our AI Platform:
NVIDIA and Google Cloud

Scroll to Top

Want to stay up to date?