Source: Arm Author: Arm
    Arm Holdings, Inc. (NASDAQ: ARM, "Arm") announced the launch of the first public specification for its Core System Architecture (CSA) to further standardize core technology and reduce industry fragmentation. At present, more than 60 industry leaders, such as ADTechnology, Alphawave Semi, AMI, Kayden Electronics, Yunbao Intelligence, Kalray, Rebellions, Siemens and Sunsys Technologies, are actively involved in the work of CSA. Help develop and adhere to unified standards for chip strategies in different areas.
    Eddie Ramirez, vice president of Infrastructure at Arm, said: "Artificial intelligence (AI) has enormous potential to lead the next industrial revolution, and its market penetration has reached unprecedented levels of depth and breadth. To achieve this ambitious goal, we must be able to handle a wide and diverse range of AI workloads in different markets. This comes with a wide range of computing requirements, which means that we need to offer far more than one computing solution, each optimized for the needs of a specific market. "With the increasing demand for custom chips, coupled with the increasing cost and complexity of chip production, Chiplet is becoming a widely adopted solution."

Promote core ecosystem development through standardization and ecological collaboration

    By reusing dedicated cores to develop a variety of customized system-on-a-chip (SoC), higher performance, lower power consumption and lower overall design cost can be achieved compared to traditional monolithic chips. However, in the absence of an industry-consistent standard and framework, differences between chipsets can lead to compatibility issues that hinder the pace of innovation. To address this fragmentation, Arm launched CSA last year. The CSA provides a set of standards for system segmentation and core interconnection developed in conjunction with the ecosystem to enable the industry to agree on the base choices for building cores. With CSA, new core designs can be seamlessly adapted and reused in standards-compliant systems, which not only accelerates core based system innovation, but also effectively reduces the risk of fragmentation.

CSA's first public specification further accelerates the evolution of chip technology and effectively reduces the risk of fragmentation.
    The extensive involvement of these innovative technology companies in CSA forms the cornerstone of the core pellet ecosystem based on Arm technology. The ecosystem is committed to revolutionizing system design to make SOCs more flexible, accessible, and cost-effective, while significantly reducing the risk of fragmentation. With the publication of the public specification, designers are able to agree on how to define and connect cores to build composable SOCs with a high degree of flexibility to meet the diverse needs of AI workloads and ensure that the final chip product is precisely tailored to the needs of a specific market.
    Several partners participating in the CSA are also actively building solutions in the Arm Total Design eco-program. Arm Total Design is a commitment to seamless delivery by Arm? Neoverse? An ecosystem of custom chips powered by computing subsystems (CSS). To date, Arm Total Design has achieved significant success in deploying core based CSS solutions that support strategic implementation for specific markets, including:
    Customizing AI workloads for diverse markets: Alphawave Semi's customers have a pressing need for high-performance chips for AI workloads in areas such as networking, edge computing, storage and security. In response to these needs, Alphawave Semi combines Arm Neoverse CSs-based cores with proprietary I/O grains (dies) that utilize AMBA? CHI C2C technology connects accelerators tailored to different market needs. These market-specific, customized chips are based on a standardized foundation that effectively spreads the cost of calculating grains while maintaining the flexibility to build multiple systems.
    - Revolutionizing large-scale AI training and reasoning workloads: ADTechnology, Samsung Foundries, Rebellions and Arm have teamed up to create an AI CPU core platform for training and reasoning for large-scale AI workloads in the data center. It is expected to provide a 2-3 times energy efficiency advantage for generative AI workloads (Llama3.1 405B parameter LLMs). The multi-vendor core platform integrates Rebellions' REBEL AI accelerator, a consistent NPU using AMBA CHI C2C interconnect technology, and ADTechnology's Neoverse CSS V3 based compute core, It is manufactured using Samsung Foundry's 2nm full surround gate (GAA) process. This result is due to the progress of CSA's standardization work.

Cores play a key role in the custom chips needed for growing AI workloads
    CSA has a proven track record of delivering efficient solutions for diverse AI-powered workloads across multiple market segments, including infrastructure, automotive and consumer electronics. With the flexibility of the Arm computing platform, the seamless communication technology enabled by standards such as AMBA CHI C2C, and the integrated innovation trend led by CSA, the ARM-based core ecosystem is well positioned to respond to the growing AI needs in various markets. As the CSA ecosystem continues to grow and the industry continues to deepen standardization and collaboration, Arm will work with partners to significantly reduce fragmentation and accelerate the development and deployment of custom chip solutions.

About Arm
    As the industry's most powerful and energy-efficient computing platform, Arm covers 100% of the world's connected population with unparalleled scale. Arm delivers advanced solutions to meet the never-ending demand for computing, enabling the world's leading technology companies to unlock unprecedented AI experiences and performance. Arm is working with the world's most extensive computing ecosystem and 20 million software developers to build the future of AI on the Arm platform.