Google announced the availability of Nvidia H100 GPUs powered A3 GPU Supercomputer VMs in the private preview. In an official blog post Company announced the launch of A3 GPU Supercomputer service, which is designed for running most demanding AI models such as today’s generative AI and large language models.
On the official Google Post, Roy Kim, director of product management, and Chris Kleban, group product manager at Google Cloud, announced the launch of A3 GPU Supercomputer for the private preview. The A3 GPU Supercomputer VMs are paired with the Nvidia H100 GPUs, which can be used for large scale AI Model training. A3 GPU Supercomputer by Nvidia H100 GPUs is the next generation supercomputing offering from Google Cloud. Currently this service is available for the private preview.
The A3 GPU Supercomputer is developed to provide a solution to the enterprise customers for training today’s most demanding AI models. These models are based on today's generative AI which uses large scale processing for training and deployment. With the launch of this service Google Cloud brings very high processing capabilities to the enterprise customers.
Google Cloud's new service will allow companies to use it for building and training the most demanding AI models, such as generative AI and big language models. In the last few years we have seen tremendous growth in the AI and ML field. Now enterprises will be able to build similar types of models due to availability of A3 GPU Supercomputer service.
Implementation of artificial intelligence (AI) and machine learning (ML) models requires lots of GPU processing power both for model training and model deployment. So, companies planning to develop such models need huge computing power and now A3 GPU Supercomputers will be very common. Companies will be able to use the high computing power of Good Cloud for training next generation AI Models.
Google's new A3 GPU Supercomputer can be used to solve large number use cases such as very language models (LLMs), generative AI and many other complex models. This solution from Google Cloud is the first company to offer new NVIDIA L4 Tensor Core GPUs, which can be used to train and server large generative AI models/workloads. Currently Google's new A3 GPU Supercomputer is available for private preview, later Google might release this service or modified service to all the users on the Google Cloud.
Google Compute Engine A3 supercomputers are so powerful that they can be used for training and serving large Generative AI models to meet today’s business requirements. Google Cloud offers A3 VMs which combine NVIDIA H100 Tensor Core GPUs and Google’s leading networking advancements to help all its customers to run AI workloads.
According to the Google Cloud official blog post Google Compute Engine A3 is the first GPU instance which uses custom-designed 200 Gbps IPUs. The 200 Gbps IPUs further uses GPU-to-GPU data transfers without using CPU running on the host machine, which is awesome and provides a powerful mechanism to train huge models. It uses separate interfaces from other VM networks for the fast data transfer. All these makes network transfer as 10x network bandwidth is available.
NVIDIA official further stated "Google Cloud's A3 VMs, powered by next-generation NVIDIA H100 GPUs, will accelerate training and serving of generative AI applications,"
"On the heels of Google Cloud’s recently launched G2 instances, we're proud to continue our work with Google Cloud to help transform enterprises around the world with purpose-built AI infrastructure."
The A3 supercomputer is very powerful which scales up to 26 exaFlops of AI performance. Google Cloud further improves the time and costs for training large ML models. This solution is expected to change the way people are using Cloud computing for AI/ML solutions.
Companies around the world use this platform for A3 GPU Supercomputer AI Application Development. Companies are already working on the AI/ML technologies for various AI solution development. A3 GPU Supercomputer will further help organisations to use these solutions in their business environment.