Cloud GPUs
High-performance GPUs on Google Cloud for machine learning, scientific computing, and 3D visualization.
-
Speed up compute jobs like machine learning and HPC
-
A wide selection of GPUs to match a range of performance and price points
-
Flexible pricing and machine customizations to optimize for your workload
Key features
Key features
A range of GPU types
NVIDIA K80, P100, P4, T4, V100, and A100 GPUs provide a range of compute options to cover your workload for each cost and performance need.
Flexible performance
Optimally balance the processor, memory, high performance disk, and up to 8 GPUs per instance for your individual workload. All with the per-second billing, so you only pay only for what you need while you are using it.
All the benefits of Google Cloud
Run GPU workloads on Google Cloud Platform where you have access to industry-leading storage, networking, and data analytics technologies.
What's new
What's new
Documentation
Documentation
GPUs on Compute Engine
Compute Engine provides GPUs that you can add to your virtual machine instances. Learn what you can do with GPUs and what types of GPU hardware are available.
Adding or removing GPUs on Compute Engine
Learn how to add or remove GPUs from a Compute Engine VM.
Installing GPUs drivers
This guide shows ways to install NVIDIA proprietary drivers after you’ve created an instance with one or more GPUs.
GPUs on Google Kubernetes Engine
Learn how to use GPU hardware accelerators in your Google Kubernetes Engine clusters’ nodes.
Using GPUs for training models in the cloud
Accelerate the training process for many deep learning models, like image classification, video analysis, and natural language processing.
Attaching GPUs to Dataproc clusters
Attach GPUs to the master and worker Compute Engine nodes in a Dataproc cluster to accelerate specific workloads, such as machine learning and data processing.
Pricing
Pricing
For information about GPU pricing for the different GPU
types and regions that are available on Compute Engine,
refer to the
GPU pricing document.