Why Cloud Run with GPUs is a Game-Changer for Developers on Google

With Cloud Run and GPUs, it's possible to deploy and scale open-source models like Gemma 2, Llama3, and Mistral 7B extremely quickly and easily, enabling real-time predictions and insights in a few clicks or commands. This combination provides the necessary speed and performance for even the most demanding AI-driven projects.

Suppose you work with fine-tuned large language models. In that case, Cloud Run also simplifies the process of hosting and serving these models, making it easier to create AI experiences tailored to specific needs.

However, it’s not just for AI workloads, other capabilities include image recognition, video transcoding, real-time streaming, and high-performance data processing, making it suitable for a wide range of compute-intensive tasks.

Why use GPUs with Cloud Run?

Speed and Agility:

In today's fast-paced web app industry, minimising latency is crucial. Cloud Run ensures that your models and applications run smoothly by reducing infrastructure delays. With the ability to scale GPU instances in seconds, it easily manages traffic spikes and fluctuating workloads, maintaining optimal performance without interruption.

Cost Efficiency:

Keeping costs under control is essential for any project. Cloud Run’s pay-as-you-go model means you're only charged for the resources you use. The platform's automatic scaling further enhances cost efficiency by reducing instances to zero when not in use, eliminating unnecessary expenses. Running just a container or a few, with little other infrastructure or set-up, drastically reduces TCO.

Developer-Friendly:

Cloud Run's serverless architecture streamlines the development process, allowing you to deploy, test, and iterate on applications quickly, without dealing with complex infrastructure. This results in faster releases, more frequent updates, and a more agile development workflow.

Security and Compliance:

Beyond performance and cost advantages, Cloud Run on GPU prioritises security and compliance. Google Cloud’s comprehensive security features—such as data encryption, identity and access management, and support for regulatory compliance—ensure that your applications and data remain secure.

How to Get Started

Sign Up:
Excited to get hands-on with Cloud Run on GPUs? Express your interest and secure your spot in the public preview by signing up at g.co/cloudrun/gpu. This is your chance to be among the first to explore this powerful new feature.

Explore:
Want to dive deeper into what Cloud Run on GPU can do? Explore the official documentation, which offers comprehensive guides and resources to help you make the most of this technology: GPU Support for Cloud Run Documentation.

Experiment:
Ready to see Cloud Run with GPUs in action? Follow a detailed tutorial to deploy models like Gemma 2 with GPUs on Cloud Run using Ollama: Running Gemma 2 with GPUs on Cloud Run using Ollama. This hands-on guide will walk you through the process step by step, making it easy to get started.

OVERALL…

The introduction of GPUs to Cloud Run was more than just a technical upgrade; it was a paradigm shift in how developers approached AI and high-performance computing. 

Whether startups aim to scale their AI capabilities or enterprises seek to streamline compute-heavy operations, Cloud Run with GPUs provides the tools needed to innovate without the overhead of traditional infrastructure.

Read our success story featuring Protraders Ltd to witness how Mesoform harnessed the power of Cloud Run and see how we can unleash its potential for your business!

If you would like to discuss any of these topics in more detail, please feel free to get in touch