Google, Microsoft, and Intel establish UALink to develop open standard for AI accelerator chip interconnect

Google, Microsoft, and Intel establish UALink to develop open standard for AI accelerator chip interconnect

INTEL, Google, Microsoft, Meta, AMD, Hewlett Packard Enterprise, Broadcom, and Cisco have formed the Ultra Accelerator Link (UALink) Promoter Group to develop open-source components that connect AI accelerator chips in data centers.

This group aims to standardize a protocol used by the industry to connect AI accelerators in the form of CPUs or GPUs needed in data centers to train AI models.

The standard is necessary to prevent innovation from being monopolized by just one company.

With the involvement of major technology companies in using UALink, each member can contribute their expertise for mutual benefit.

Therefore, the UALink Consortium will be established in Q3 2024 to develop UALink specifications in the future.

With the formation of this group, the UALink 1.0 specification has been announced, and it can connect 1024 AI accelerators within a single compute pod in a data center.

UALink 1.0 utilizes open standards such as AMD Infinity Fabric. The first UALink 1.0 products will be launched in the coming years.

NVIDIA did not join the UALink Promoter Group because, at this time, they are using their own AI accelerator connections.

At present, NVIDIA holds 90% of the data center GPU market, helping them achieve a 262% revenue increase in the last quarter compared to the same quarter last year.

The UALink Promoter Group is another industry effort to reduce NVIDIA’s power in the AI world.

Last March, Intel, Google, Arm, Qualcomm, Samsung, and several other technology companies established the Unified Acceleration Foundation (UXL).

Through UXL, AI development will no longer be tied to the use of a specific programming language, code base, and other tools that bind them to proprietary architectures, such as Nvidia’s CUDA platform.

Artificial intelligence (AI) should be open source for several reasons:

Transparency and Accountability: Open source AI allows researchers, developers, and users to inspect the code, understand how algorithms work, and identify any biases or errors. This transparency promotes accountability and helps mitigate the potential negative impacts of AI systems.

Accessibility: Open source AI democratizes access to advanced technology by making it freely available to anyone. This encourages innovation and allows a broader range of individuals and organizations, regardless of financial resources, to contribute to and benefit from AI development.

Collaboration and Innovation: Open source fosters collaboration among diverse communities of developers, researchers, and practitioners. By sharing code, knowledge, and best practices, contributors can collectively solve complex problems, accelerate innovation, and build upon each other’s work more effectively.

Customization and Adaptation: Open source AI provides flexibility for customization and adaptation to specific use cases, industries, or user needs. Organizations can modify and extend existing AI models and algorithms to better suit their requirements, improving performance, efficiency, and relevance.

Interoperability: Open source standards and protocols promote interoperability and compatibility between different AI systems and platforms. This enables seamless integration and data exchange between disparate systems, fostering ecosystem development and preventing vendor lock-in.

Ethical Considerations: Open source AI encourages ethical practices and responsible AI development. Communities can collaborate to establish guidelines, standards, and governance frameworks that prioritize fairness, accountability, transparency, and privacy in AI systems.

Overall, open source AI promotes transparency, collaboration, innovation, and ethical AI practices, making AI more accessible, inclusive, and beneficial for society as a whole.

CATEGORIES
Share This

COMMENTS

Wordpress (0)