Google Cloud has revealed what it describes as “the world’s largest publicly available megacluster for machine learning” that will deliver nine exaflops of computing power to Google Cloud Platform users. Artificial intelligence workloads are becoming increasingly important to many cloud users, and Google Cloud will hopefully find its new cluster attractive to customers.

Google Cloud has launched what it sees as the world’s most powerful machine learning cluster. (Photo by Sean Gallup / Getty Images)

Unveiled as part of the company’s Google I / O developer conference, the machine learning cluster will be powered by the latest version v4 of Google’s internal Tensor Processing Units, which is designed to perform its cloud services as well as other platforms such as YouTube.

The Google Cloud Machine Learning Cluster

The Google Cloud ML Cluster allows researchers and developers to break into the forefront of AI, enabling them to train increasingly sophisticated workload power models such as large-scale natural language processing (NLP), referral systems, and computer vision algorithms, ”Sachin Gupta, vice president and general manager of infrastructure, and Max Sapozhnikov, product manager for cloud TPU, said in a joint statement. “With nine exaflops peak aggregate performance, we believe our cluster of Cloud TPU v4 Pods is the largest publicly available ML hub in the world in terms of cumulative computing power.

Based at one of the company’s data centers in Oklahoma, the ML cluster is powered by a series of Cloud TPU v4 pods, each consisting of 4096 chips connected in an ultra-fast interconnected network. Google says each capsule has an “industry-leading” bandwidth of six terabits per second, allowing it to quickly absorb information and train large AI models. The Oklahoma Data Center runs on 90% renewable energy.

The machine learning cluster is available for visualization today.

Why AI and machine learning are important to cloud providers

An increasing number of companies are turning to AI and ML to help improve efficiency and digitize their operations. According to McKinsey study, “The state of AI in 2021”published last year, 56% of more than 1,800 organizations surveyed worldwide said they had embraced AI in at least one function of their business, up from 50% in 2020.

Content from our partners
How businesses can best prepare for the digitization of finance

How AI can enable Middle East energy operators to deliver Oil & Gas 4.0

How should businesses get out of their data centers?

The role of cloud computing in the successful implementation of AI was also highlighted in a McKinsey study, which identified companies it described as “highly productive on AI” – businesses that attribute at least 20% of their revenue to AI implementation. This high-performance group performs an average of 64% of its AI workloads in the cloud, compared to 44% for other respondents, suggesting that cloud-based AI may offer better returns than on-premises systems.

The high-performance team “also has access to a wider range of AI capabilities and techniques in the public cloud,” according to a McKinsey study. “For example, they are twice as likely as others to say they use the cloud to understand natural language and face recognition capabilities.”

AR, VR and database innovations in Google I / O

Google Cloud has made several other I / O announcements, including the launch of AlloyDB, a new PostgreSQL-compliant database service for extremely demanding workloads. He claims that this offers twice the processing of the comparable transaction load service offered by AWS to Amazon, the world’s leading cloud platform.

Also available to Google Cloud Platform users today is Immersive Stream, a new service that makes immersive 3D and augmented reality experiences and allows them to stream to mobile devices.

For security, Google Cloud has Network Analyzer, a new module in the Platform Network Intelligence Center that allows developers to detect network failures and prevent outages by identifying potential issues such as accidental incorrect configurations and overuse of services.

Read more: Google Cloud is creating a web3 unit to stay ahead of AWS and Azure

Google Cloud claims ‘most powerful’ publicly available machine learning cluster

Previous articleTop 10 time management applications for 2022
Next articleFertility: Rat testicular cells produce sperm after freezing for 23 years