ARM’s 64-Bit Venture: Why Investors Should Care
Anindya is a member of The Motley Fool Blog Network -- entries represent the personal opinion of the blogger and are not formally edited.
Evercore Partners tech analyst Patrick Wang recently said that ARM Holdings (NASDAQ: ARMH) is expected to benefit from Apple’s introduction of the first 64-bit ARM-based chip this year. Wang added, “This bodes well for ARMH as v8 designs drive richer royalty rates.”
ARM recently introduced its first 64-bit Cortex-A50 series processor designs as the company tries to preserve its dominance in smartphones and tablets, while catching up with Intel (NASDAQ: INTC) in servers. In this post I will dig deeper into ARM’s prospects in the 64-bit game.
64-bit processors were first introduced in 1992, but they’ve been used meaningfully since the past few years. Both Intel and Advanced Micro Devices (NYSE: AMD) have introduced 64-bit chips. 64-bit processors have 64-bit ALUs, 64-bit registers, and 64-bit buses.
One reason why the world wants to embrace 64-bit processors is because of their enlarged address spaces. 32-bit chips are often constrained to a maximum of 2GB or 4GB of RAM access. That may sound like a lot for most home computers. However, a 4GB limit can be a severe problem for server machines and machines running large databases. Even home machines will start bumping up against the 2GB or 4GB limit pretty soon if the current trend continues. A 64-bit chip has none of these constraints, because a 64-bit RAM address space is essentially infinite for the foreseeable future.
ARM’s A57 and A53 SoC processors
Successors to the current generation Cortex-A15, the Cortex-A50 series system-on-chips brings the ARM architecture into a 64-bit capable processor that can be used in a much broader range of devices -- from smartphones to data center workhorses. The Cortex-A57, specifically, is the series’ flagship, and will be a serious entry into the server market.
While ARM processors with 32-bit support have a ceiling of 4GB of memory, current systems with the new ARM processors are being configured to support up to 64GB of memory. Moreover, the hardware-level security features of these processors can be used to authorize mobile transactions without additional tools.
ARM licenses architecture and processor designs to chip companies, which then make the chips that go into tablets, smartphones, and servers. The first Cortex A50-series chips could be available in late 2013, after which companies can start making products. Servers may be the first products to reach the market, and some chip partners are aggressively looking at high-end smartphones and tablets. Ian Forsyth, product manager at ARM, said:
Calxeda and AMD have already announced they would license the 64-bit processor design from ARM and will sell ARM-based server processors in 2014. Samsung, Broadcom, HiSilicon and STMicroelectronics have also licensed Cortex-A50 designs from ARM. Samsung has licensed both the Cortex-A57 and A53 cores.
Licensees will be able to mix and match Cortex-A57 and Cortex-A53 processors. For example, in servers the Cortex-A57 cores could handle a large volume of transactions, while the power-efficient A53 cores could do the quick processing of a transaction when servers are idle.
The Cortex-A57 is a “big core” and is targeted at servers, high-performance tablets, and smartphones. The Cortex-A53 is a “small core” that essentially delivers the same performance as the Cortex-A9 processor used in smartphones and tablets today. However, the core is 40% smaller in design, which could enable more compact and more power-efficient chips.
ARM eyes the server market
Intel dominates the market for personal computers, servers, and workstations with its processors and its x86 architecture (used by Intel and AMD). ARM dominates in smartphones and tablets, but is aiming to make its mark in the server market. There is growing interest in ARM servers as an energy-efficient way to handle large numbers of web requests, such as in search or social networks.
Dell and Hewlett-Packard already offer prototype ARM-based servers for testing to customers looking to deploy ARM servers to cut energy bills. However, Intel is also tweaking its low-power Atom processors to work in cloud servers and will release new Atom S-series chips for microservers later this year.
Intel is trying to adopt its x86 architecture for low-power devices. A major problem for Intel is that ARM-based chips are currently more energy efficient compared to x86-based offerings, and provide a lot of freedom for companies like Apple, who install their own custom-designed system-on-chips into the highly-popular iPhone and iPad.
ARMv8 64-bit architecture is also expected to be more energy-efficient than comparable x86 offerings based on AMD Jaguar or Intel Silvermont. It is noteworthy that AMD recently decided to develop server-class system-on-chips based on ARMv8, just like many other companies, including NVIDIA and Qualcomm. But Intel just does not want to make friends with ARM, complains Robin Saxby, the founding CEO of ARM Holdings.
ARM’s venture into the 64-bit space is a significant move from investors’ point of view since it will result in increased revenue. There will be huge cost benefit to any vendor shipping both mobile devices and traditional desktop and server devices, to having silicon standardized on 64-bit, rather than building or purchasing silicon for 32-bit and 64-bit as distinct parts.
Anindya Batabyal has no position in any stocks mentioned. The Motley Fool recommends Intel. The Motley Fool owns shares of Intel. Try any of our Foolish newsletter services free for 30 days. We Fools may not all hold the same opinions, but we all believe that considering a diverse range of insights makes us better investors. The Motley Fool has a disclosure policy. Is this post wrong? Click here. Think you can do better? Join us and write your own!