Arm responds to AI attacks on PCs and smartphones with CSS for Client

The growing presence of AI-enabled desktops and smartphones shows that the AI-driven demand for more powerful computing and greater energy efficiency is not a trend limited to data centers. As AI experiences increasingly become a part of people’s daily lives, AI support must extend to all devices that end users rely on for these experiences.

This makes client devices a suitable target for Arm’s newest compute subsystems architecture, Arm Compute Subsystems (CSS) for Client, which aims to combine the AI ​​performance, security and other benefits of the Armv9 processor architecture with “validated and verified implementations ready for production”. new Arm CPUs and GPUs on 3nm process nodes,” the company said in a statement, adding: “CSS for Client provides core compute elements for flagship SoCs and includes the latest Armv9.2 CPUs and Immortalis GPUs, as well as physical-ready manufacturing solutions 3nm CPU and GPU implementations and the latest Corelink System Interconnect modules and system memory management units (SMMU).”

Specifically, the processor cluster that underpins CSS for Client is based on Armv9.2 and includes Cortex X925, Cortex A725, and Cortex A520 processor cores, as well as the new Arm Immortalis G925 GPU. As for how this all translates into the end-user experience, an Arm representative said during the briefing that for Android devices, CSS for Client delivers “a more than 30% increase in compute and graphics performance and 59% faster AI inference for broader AI/ML and computer vision (CV) processing.

The use of 3nm process node technology is important because it provides greater performance and reduced form factor and power consumption for AI.

Additionally, as AI software becomes increasingly important in extending the performance benefits of AI-optimized processors across the entire user experience, Arm also announced the KleidiAI software library, which supports popular AI platforms to make it easier for developers to use the software. Combined with CSS for Client, KleidiAI helps increase application performance by leveraging Arm acceleration technologies such as the NEON instruction set architecture and extensions such as SVE2 and SME2.

“We’re seeing significant innovation from phones to laptops, which are now referred to as both AI smartphones and AI PCs,” said Chris Bergey, senior vice president and general manager of Arm’s customer division. In particular, he noted that AI capabilities in new phone models such as the Samsung S24 are “redefining smartphone ‘smartness’.”

He added: “With greater integration of AI in everything from productivity to entertainment applications, what is possible is changing at our fingertips. We are now seeing new use cases and enablers of interaction between people across language barriers, which suggests that it is not here yet. At the Samsung S24 launch earlier this year, we showed off real-time multilingual translation for better communication (as well as new demonstrations of how) smartphone cameras collect and process light for photos and videos. These features were powered by Samsung’s artificial intelligence model, Galaxy AI, which runs on the Google Gemini platform.”

Bergey also noted that Google Gemini is not the only AI assistant emerging, and that AI’s impact will be felt not only in desktop computers and smartphones, but also in the televisions in every consumer home and in an increasingly wide range of wearable devices.