After releasing the flagship Windows PC processor Snapdragon X Elite at the Snapdragon Technology Summit last year, Qualcomm recently launched a new product for the high-end PC market – the Snapdragon X Plus processor.
According to information released by Qualcomm, Snapdragon X Plus is a high-end ARM PC processor product.Its CPU part adopts a 10-core design, with a maximum main frequency of 3.4GHz, and a leading 4nm process technology.
Advertisement
The Qualcomm Adreno GPU delivers 3.8TFLOPS, and the Qualcomm Hexagon NPU brings 45TOPS, which is the fastest NPU among known laptop processors.
It is worth noting that in multiple public benchmark tests, the overall performance of Snapdragon X Plus has surpassed Intel’s current mainstay in mobile PCs, the Core Ultra 7 155H.
In contrast, under the same power consumption, the CPU core of Snapdragon X Plus provides higher single-thread and multi-thread processing capabilities, and the GPU graphics computing performance is also stronger. Computers can bring longer battery life, higher processing efficiency, and a smarter user experience. But unfortunately, Qualcomm did not show more details of practical applications like it did when introducing Snapdragon X Elite.
Advertisement
The launch of Snapdragon X Plus can be said to be another round of strong challenges launched by ARM to the traditional x86 processor + WIndows camp after X Elite.
Although Qualcomm has not revealed its long-term plan in the PC processor field in detail, through the release of two consecutive products, it can be clearly felt that Qualcomm is accelerating its entry into the PC processor field and is actively deploying in the ARM Windows ecosystem. The rise of artificial intelligence technology is also accelerating the reshaping of traditional PCs. AI PCs have brought historic development opportunities for Qualcomm to enter the PC field.
From chip architecture to market structure, AI is reshaping the PC industry
In the traditional PC era, Intel x86 architecture and Microsoft Windows operating system formed the dual core of the PC ecosystem. As the two major engines that drive the development of the PC industry, their alliance has determined that over the past few decades, the main line of development of PC processors has always centered around improving CPU performance.
Whether it is the reduction of process technology, the increase in the number of cores, or the optimization of instruction set technology, the CPU has always been a key device that determines PC performance. Even though GPUs later began to help handle graphics and parallel computing loads, and heterogeneous computing architectures emerged, the CPU is still the most important indicator of overall machine performance.
This “CPU is PC” pattern is largely derived from the software ecological advantages of the x86 architecture. Most applications and games are developed for x86, and the improvement of CPU performance naturally becomes the decisive factor in the application experience. Therefore, from the original 4004 to the current Core, Intel has always regarded the CPU as the core value of x86 chip design.
However, the advent of the AI era is overturning this traditional pattern. AI computing has extremely high demands for large-scale data parallel processing capabilities, making it unsustainable for the CPU to work alone. Specialized AI accelerators such as GPU, TPU, and NPU are incorporated into chip heterogeneous designs to jointly make up for the CPU's lack of AI computing power.
Among them, NPU, as a chip dedicated to AI training and reasoning, plays a pivotal role in the AI era. CPU is no longer the only indicator to measure chip performance, AI computing power is the new trump card. This change has created opportunities for the ARM camp in the PC field.
The ARM architecture itself has relatively mature support for AI accelerators, and ARM chip manufacturers such as Qualcomm have been working in mobile AI, edge AI and other fields for a long time. Especially Qualcomm, as early as 2015, the Snapdragon 820 mobile phone chip was equipped with the first generation AI Engine.
By 2018, the Hexagon NPU in Snapdragon 855 added the Hexagon tensor accelerator to further improve AI performance. In 2020, Qualcomm Hexagon NPU ushered in a revolutionary architectural update, integrating vector, tensor, and scalar accelerators to achieve better performance and power consumption.
These accumulations have allowed Qualcomm to accumulate rich experience in AI chip design on the mobile phone side, and also laid the foundation for its subsequent expansion of AI acceleration chips into broader fields such as PCs.
Therefore, when AI reshaped the chip architecture paradigm and AI accelerators became the new trump card, ARM also gained a key opportunity to take a share of the traditional x86 Windows PC field.
Apple lays out a prototype for ARM-based PCs
As a pioneer in personal computers, Apple has been conducting independent design and innovation in the field of processors since the 1980s.
Early Apple II and Macintosh computers used Motorola 68000 series processors. Although the performance was average, the 68000 architecture was ahead of the Intel x86 chips at the time in terms of simplicity in design and functional integration.
In the 1990s, Apple teamed up with IBM and Motorola to create PowerPC RISC architecture processors, which were used in the PowerMac series. PowerPC once dominated Apple's computer product line with its high computing power.
However, in the late 21st century, the development of PowerPC was restricted. In order to gain support from the huge x86 software ecosystem, Apple announced in 2005 that it would switch to Intel Core processors, ending the 20-year PowerPC era.
With the support of Intel chips, Apple has launched landmark products such as MacBook, MacBook Pro, and iMac. However, Apple's long-term strategy of relying on Intel chips has limited Apple to a certain extent by the “ceiling” of architecture and performance development. In order to further improve product performance and power consumption experience, in 2020, Apple announced its migration from x86 processors to the self-designed ARM instruction set chip M1.
Apple's M1 processor relies on TSMC's 5-nanometer process technology, 8-core CPU, 7-core GPU, 16-core neural network engine and other powerful configurations to enable MacBook Air/Pro notebooks to achieve a qualitative leap in performance, battery life, etc., leading the way in PC processing A new trend in device performance.
Today, Apple's ARM architecture processor has developed to the third generation M3 chip, which demonstrates ARM's ability to compete with x86 chips in the PC field.
The problem is in Microsoft's hands
As the old leader of personal computer operating systems, Microsoft has actually seen the potential of ARM processors in the field of mobile devices quite early and began to lay out ARM versions of Windows systems.
In 2012, Microsoft launched the Windows RT system, which was the first Windows version specifically designed for the ARM chip architecture. Although Windows RT also has some traditional Windows programs built-in, due to limited ecological support and hardware defects such as battery life, it ended up being tepid in the market.
In 2015, Microsoft launched the Windows 10 ARM version to better support the application of ARM chips in notebook computers and other scenarios. The system allows running 32-bit ARM native applications and most traditional x86 applications. However, there is still a big gap in software compatibility and performance experience.
In order to solve the performance bottleneck of ARM Windows, Microsoft announced the launch of a new ARM64 mobile phone instruction set architecture in 2017, and launched the Always Connected PC plan in 2018 to integrate the advantages of mobile chips to achieve sustainable connections and ultra-long battery life.
However, the market response to Always Connected PC has been mediocre due to factors such as a lack of hardware ecology and unsatisfactory performance. In 2021, Microsoft will further launch the ARM64 EC plan to improve the performance of the ARM64 version of Windows App.
Although Microsoft has been promoting the development of the ARM Windows ecosystem, ARM Windows has not been widely recognized by the market for many years due to constraints such as software ecosystem support and insufficient chip performance.
However, the official benchmark test results of Qualcomm’s ARM PC processors Snapdragon X Elite and Snapdragon Preparation for head-on competition.
Next, whether the ARM Windows ecosystem can truly achieve a breakthrough in the PC field has been handed over to Microsoft. Whether Microsoft can finally solve pain points such as ARM Windows software compatibility and developer support will determine the future development prospects of ARM processors in the PC market.
Only by paying equal attention to both hardware and software and taking into account each other can the ARM architecture truly gain a foothold in the field of PC processors and get a share of the pie.