Arm today announced the launch of the Arm®v9 architecture to meet the global demand for increasingly powerful security, artificial intelligence (AI) and ubiquitous dedicated processing. Armv9 is based on the successful foundation of Armv8 and is the latest Arm architecture in the past ten years. Armv8 is driving the best performance per watt in areas that require computing today.
Arm’s CEO Simon Segars said, “As we look to the future defined by AI, we must lay a solid foundation for advanced computing to meet the unique challenges of the future. Armv9 is the answer we gave. Based on design freedom and accessibility, the market needs universal, dedicated, safe and powerful processing capabilities, which will drive the development of the next 300 billion chips based on the Arm architecture, and Armv9 is the technological pioneer of these chips."
Shipments of chips based on the Arm architecture continue to accelerate, and the shipments of devices based on the Arm architecture have exceeded 100 billion in the past five years. At the current speed, whether it is on the terminal, data network or cloud, 100% of the world's shared data will soon be processed through Arm technology. Such a wide range of applications allows Arm to shoulder more responsibilities and missions. For this reason, Arm provides more security and performance in Armv9, conforms to the strong development of AI, Internet of Things and 5G on a global scale, and accelerates each industrial application. From general calculations to special calculations.
Security: the biggest challenge in computing
In order to solve today's biggest technical challenge-protecting global data security, the Armv9 architecture roadmap introduces the Arm Confidential Compute Architecture (CCA). Confidential computing performs calculations by creating a hardware-based secure operating environment to protect part of the code and data from being accessed or modified, or even affected by privileged software.
Arm CCA will introduce the concept of dynamically creating realms, which are for all applications and run in an environment independent of the secure or non-secure environment to achieve the purpose of protecting data security. For example, in commercial applications, the confidential domain can protect commercial confidential data and codes in the system, regardless of whether they are being used, idle, or in transit. In fact, in a recent Pulse survey of corporate executives, more than 90% of respondents believe that confidential computing can help reduce the cost of companies’ security investments, so that they can invest heavily in engineering Innovation.
Henry Sanders, corporate vice president and chief technology officer of Microsoft Azure Edge and Platform Division, said, "As use cases from the edge to the cloud become more and more complex, we can no longer use a one-size-fits-all solution to solve all use cases. Therefore. , Heterogeneous computing is becoming more and more common, which requires enhanced collaboration between hardware and software developers. The Armv9 confidential computing function developed in close cooperation between Arm and Microsoft is a good example of collaboration between hardware and software. Arm In a unique position, it can accelerate heterogeneous computing at the core of the ecosystem and promote open innovation in the computing architecture that drives billions of devices."
Ubiquitous AI: Calling for dedicated, scalable solutions
The universality and versatility of AI workloads require more diverse and specialized solutions. It is estimated that by the mid-2020s, there will be more than 8 billion devices equipped with AI voice assistance worldwide. And 90% or more of the applications on devices will contain AI elements and AI-based interfaces such as vision or voice.
To meet this demand, Arm and Fujitsu have developed a Scalable Vector Extension (SVE) technology, which drives the world's fastest supercomputer "Fuyue". On this basis, Arm developed SVE2 for Armv9 in order to achieve enhanced machine learning and digital signal processing capabilities in a wider range of applications.
SVE2 enhances the processing capabilities of 5G systems, virtual and augmented reality, and ML workloads running locally on the CPU, such as image processing and smart home applications. In the next few years, Arm will further expand the AI ??capabilities of its technology. In addition to continuous AI innovation in its Mali™ GPU and Ethos™ NPU, it will also greatly enhance the matrix multiplication in the CPU.
Maximize performance through system design
In the past five years, Arm technology has improved CPU performance at a speed faster than the industry every year. Arm will maintain this speed on the next-generation architecture Armv9, and it is expected that the performance of the next two generations of mobile and infrastructure CPUs will increase by more than 30%.
However, as the industry develops from general-purpose computing to general dedicated processing, the annual double-digit CPU performance improvement is not enough. In addition to enhancing dedicated processing capabilities, Arm’s Total Compute design method will accelerate overall computing performance through centralized system-level hardware and software optimization and use case performance improvement.
By applying the design principles of comprehensive computing to the entire IP portfolio including automotive, client, infrastructure and IoT solutions, Armv9 system-level technology will pervade the entire IP solution and improve individual IP. In addition, Arm is also developing a number of technologies to increase frequency, bandwidth, cache, and reduce memory latency, thereby maximizing the performance of Armv9-based CPUs.
Vision for the next decade of computing
Richard Grisenthwaite, senior vice president, chief architect and technical academician of Arm, said, "More complex AI-based workload requirements are driving the development of more secure and dedicated processing, which will be the key to opening new markets and seizing new opportunities. Armv9 will empower developers to build and program future trusted computing platforms by bridging the key gap between software and hardware, while achieving standardization, helping our partners to strike a balance between faster time to market and cost control, and at the same time Be able to create your own unique solution."
Tags：