The last four years have seen Arm on a relentless pursuit of the very best CPU performance and efficiency. It started with the Arm Cortex-X1, which was our first high-performance Cortex-X CPU. Since then, we have successfully delivered double digit IPC performance gains with the Arm Cortex-X2 and Cortex-X3 CPUs that were built on the Armv9 architecture. All of the Cortex-X CPUs are part of our Cortex-X Custom program. This creates performance first designs, provides customization capabilities for different market needs, and extends the capabilities of the Arm ecosystem.
Alongside the second and third-generation Cortex-X CPUs, we introduced two generations of Armv9 CPU clusters. These combine the high-performance Cortex-X CPU cores with the efficient premium performance of the Cortex-A700 CPU series and highly efficient Cortex-A500 CPU series. The clusters deliver industry-leading flexibility, with the highest levels of performance and efficiency possible. The different clusters are already enriching the Arm ecosystem, enabling new technologies and experiences that are transforming our digital lives.
Now, we are delighted to launch the latest Cortex-X CPU, the Arm Cortex-X4, alongside brand-new CPU clusters that combine the Cortex-X4 with our new Cortex-A720 and A520 CPUs. We are not only delivering the highest performing Arm Cortex CPU through the Cortex-X4, but also introducing the most powerful Cortex CPU compute cluster ever built. The backbone of the cluster is a brand new DynamIQ Shared Unit, the DSU-120, which builds on the DSU-110 with enhancements across scalability and PPA, as well as brand-new features. The foundation of these new IP and clusters is the latest Armv9.2 architecture, continuing our CPU architecture leadership.
There is an insatiable drive for more performance across the consumer computing landscape. Current and future use cases across premium and flagship laptop and smartphone devices are driving this demand.
For flagship smartphones, the drive to continuously boost battery life means users can enjoy their favorite digital experiences for longer. There are also a growing number of applications that require more performance and advanced computing capabilities, such as new AI and machine learning (ML) based applications and more immersive AAA gaming.
For premium laptop devices, which include Windows on Arm (WoA) laptops that have come to market since 2018, multi-day battery life and ‘always on’ capabilities remain key considerations. We are focused on improving multi-threaded workloads in our CPU designs, which are especially important for the slim, lightweight form factors of WoA laptop devices. Additionally, we are making sure that all our premium CPU designs and clusters support video and audio rich entertainment and gaming applications on laptops. And we also continue support for the most commonly used productivity applications on laptop devices.
As with every Cortex-X CPU, the Cortex-X4 is designed for ultimate performance and represents yet another generation of double-digit IPC growth, with 15 percent performance improvements compared to last year’s Android flagship¹. The double-digit IPC growth provides peak performance for single-threaded ‘bursty’ workloads, like launching applications, web browsing, page loading and rendering, and camera functions on smartphones. This best-in-class performance is possible due to a step increase in the instruction bandwidth in the front-end microarchitecture. Alongside performance, Cortex-X4 is the most efficient Cortex-X CPU core ever built, with 40 percent better power efficiency². These performance and efficiency benefits sit alongside features supporting greater scalability, with up to 2MB per core L2 cache to target a broad range of consumer devices.
The new DSU-120 is based on the DSU-110 that was launched two years ago and offers enhancements across all aspects of the DynamIQ Shared Unit. It enables scalability from single core all the way to 14 cores for a broad range of CPU clusters for consumer devices, from entry-level smartphones and wearables to flagship smartphones and premium laptops. The DSU-120 provides advanced features like new intelligent power-saving modes to improve efficiency across the CPU cluster. These new efficiency levels increase days of use for whatever consumer device a particular cluster is targeting. DSU-120 also adds larger L3 cache configuration options with 24MB and 32MB. This expands the cluster capability with higher performance and reduced system power.
With up to 14 cores and larger L3 cache, the new DSU-120 is the catalyst for our most powerful, highest performing CPU clusters ever. From a 10x Cortex-X4 and 4x Cortex-A720 CPU cluster for high-performance laptops to the 1x Cortex-X4, 4x Cortex-A720 and 4x Cortex-A520 cluster for flagship smartphones, a highly scalable DSU is at the cornerstone of Arm’s new Total Compute Solutions (TCS23). We are also providing an 8 core CPU cluster consisting of 1x Cortex-X4, 5x Cortex-A720 and 2x Cortex-A520. Alongside these CPU clusters, partners have the complete freedom to innovate their own for the next generation of consumer devices. These can be developed and utilized based on a variety of use cases and their own computing requirements.
Arm’s new CPU products are all built on the foundation of the new Armv9.2 architecture. This represents the fastest ever pace that we have brought brand new architecture features to market. Alongside Memory Tagging Extension (MTE) and SVE2, the Armv9.2 architecture adds a new QARMA3 algorithm for Pointer Authentication (PAC) for improved performance when activating the security feature on our Armv9 CPU designs.
This year’s latest Armv9 CPUs, DSU, clusters, and new Armv9.2 architecture represent Arm’s relentless engineering innovation. This is delivering the very best performance and efficiency for the next generation of consumer devices. We continue to provide ultimate performance through the Cortex-X4. Meanwhile, our new CPU clusters combine the latest Cortex-A720 and Cortex-A520 cores to balance efficiency and performance across a broad range of consumer devices. To top it all off, the new DSU-120 delivers enhanced scalability of up to 14 cores for the highest performing CPU cluster ever. The foundation of these CPU designs is the new Armv9.2 architecture that brings advanced security features and computing capabilities.
Our commitment to push the frontiers of CPU performance and efficiency means we continue to deliver the advanced experiences that users know and love on their favorite consumer devices. These are enriching our ecosystem and creating technologies that change lives and redefine the consumer computing experience. Through high performance and advanced Armv9 features and computing capabilities, we remain at the leading edge of the innovative consumer technology landscape.
¹ Performance claims are for SPECRate®2017_int_base.
² Indicates Cluster Power savings at iso-performance for SPECRate®2017_int_base. Comparing Peak SPECRate®2017_int_base performance for Cortex-X3-based Android flagship device shipping as of March 2023 vs Cortex-X4: 2MB L2, 8MB L3, 3.4GHz, 100ns latency.