Arm's partners develop 2nm AI processor using Arm's Neoverse — Arm's Total Design ecosystem doubles in size on first anniversary

Arm logo
(Image credit: Arm)

When Arm introduced its Total Design initiative in October 2023, its success was not guaranteed. According to a blog post over at Arm, it looks like Total Design is indeed a success. Arm's Total Design ecosystem has rapidly expanded, doubling its size within a year. The ecosystem now involves over 30 companies including Samsung Foundry, ADTechnology and Rebellions who have teamed up for an AI solution to be made on Samsung's 2nm-class process technology.

One of the most notable collaborations within the ecosystem involves Arm, Samsung Foundry, ADTechnology, and Rebellions, which developed an AI CPU chiplet platform based on the Neoverse CSS V3. This chiplet platform is designed for AI/ML training, cloud computing, and high-performance computing workloads. Samsung Foundry will make it on its 2nm-class process technology. 

"AI and HPC designs require technology solutions that deliver maximum performance, high transistor density, and energy efficiency," said Taejoong Song, Vice President and Head of Foundry Business Development at Samsung Electronics. "Samsung Foundry's 2nm GAA process is designed precisely to satisfy the most stringent HPC and AI design requirements, and we’re excited to leverage the flexibility of Arm CSS and the power of the Arm Total Design ecosystem to deliver an AI CPU chiplet platform, which will further accelerate adoption of our leading-edge technology and design solutions for hyperscalers and cloud service providers."

Arm Total Design is a global initiative launched by Arm to accelerate development of custom silicon solutions for AI, HPC, and datacenter applications. Arm Total Design is an ecosystem of partners, including design houses, and software companies, to collaboratively build advanced chiplets and system-on-chips. The ecosystem supports development process from design to manufacturing and uses Arm's designed Compute Subsystems (CSS) to simplify silicon development by providing standard building blocks and ability to integrate third-party components. 

Arm Total Design partners, such as Alphawave, Cadence, and proteanTecs, are validating their third-party IP products with Compute Subsystems (CSS) on cutting-edge nodes to meet Arm's specifications and standards. This enables partners to develop custom silicon based on CSS using the latest technology, ensuring a smooth and ready-to-use software experience. With input from 30 industry partners, Arm Total Design can enable building Arm-based multi-chiplet designs quickly. 

TOPICS
Anton Shilov
Contributing Writer

Anton Shilov is a contributing writer at Tom’s Hardware. Over the past couple of decades, he has covered everything from CPUs and GPUs to supercomputers and from modern process technologies and latest fab tools to high-tech industry trends.

  • Pierce2623
    Am I the only person that finds it strange that a chip made up of general purpose CPU cores is being called an “AI” chip? It seems more like just a cloud based CPU to control a system containing real AI focused chips.
    Reply
  • subspruce
    speaking of the 2nm, it doesn't exist yet
    Reply
  • Pierce2623
    subspruce said:
    speaking of the 2nm, it doesn't exist yet
    Yes it does. It’s not in mass production but any foundry not capable of making a 2nm chip is WAY behind.
    Reply
  • subspruce
    Pierce2623 said:
    Yes it does. It’s not in mass production but any foundry not capable of making a 2nm chip is WAY behind.
    cough cough SMIC, UMC and GloFo
    Reply
  • Pierce2623
    subspruce said:
    cough cough SMIC, UMC and GloFo
    Yeah those are all extremely far behind bleeding edge fabs. SMIC has fake 7nm, GloFo has real 12nm with similar performance and UMC basically only does 28nm and up right? Intel, Samsung and TSMC all have 2nm fabs capable of generating working chips, even if not necessarily at good yields yet at Intel or Samsung.
    Reply
  • subspruce
    Pierce2623 said:
    Yeah those are all extremely far behind bleeding edge fabs. SMIC has fake 7nm, GloFo has real 12nm with similar performance and UMC basically only does 28nm and up right? Intel, Samsung and TSMC all have 2nm fabs capable of generating working chips, even if not necessarily at good yields yet at Intel or Samsung.
    Yeah, TSMC's basically risk producing N2 at this point
    Reply