Mobile Technology
Choosing the right architecture
D
esigning embedded and mobile products is difficult. Products are getting more complex, competition is driving down margins, and for portable devices there’s a constant pressure to preserve and increase battery life. On top of all of this, time to market is decreasing – with shorter design cycles making it harder to innovate and deliver great products. Of course it’s not just about design. Companies need to create products that they can support, that can scale upwards or downwards if new versions are required, and that are going to be reliable in the field. And they need to be able to recruit engineers with the right skills to develop within their chosen hardware and software environments.
What engineers say
Thousands of design engineers have recently attended technical seminars run by ARM and Avnet in the US, UK, Northern Europe and Asia. These seminars helped attendees understand how to incorporate ARM technologies in their embedded and mobile designs. ARM and Avnet took the opportunity to survey the 2500 engineers who attended and find out which hardware platforms were most popular. The results are summarised in the graph below, which shows both current technologies used, and the engineers’ plans for their next generation of designs that they expect to start in the next one to two years.
sessions, which were packed every time. Why is the ARM architecture proving
increasingly popular? There are several answers. Firstly, the ARM architecture is highly efficient. It allows engineers to develop in C or other high level languages, yet generate compact, efficient code – enabling them to write good software more quickly whilst keeping memory requirements to a minimum. In contrast, some competing architectures require code to be written in assembler for the best efficiency, which slows down development and reduces maintainability. Most applications are not simple
anymore – we demand more out of our devices, with increased functionality, more complicated and richer user interfaces and longer battery life. This means that power efficiency is vital, and the ARM architecture provides features to help. For example, it has an efficient interrupt structure, and good mechanisms for waking a processor up quickly from sleep modes. This allows ARM to address a lot of applications with 32-bit devices that would traditionally require an 8-bit design to achieve low power consumption.
Power consumption has two main aspects: the power used while a device is awake and performing a task, and that consumed while it is asleep. Typically, what makes a design power efficient is spending as much time as possible in a sleep mode. This means that it’s not always possible to work out which device will have the lowest overall power consumption by comparing figures of mW per MHz. A higher performance processor may be able to complete a task much more quickly and return to sleep faster than an 8 bit processor, so the overall power used is less – particularly for processing-intensive tasks.
The most significant trend that can be seen in the graph is the current popularity of the ARM Cortex cores, and how a growing number of engineers are expecting to adopt them for new designs in the near future.
The seminar attendees were also asked about their preferred operating system. Here, there was a tremendous interest in Linux and Android. This was reinforced by the attendance at the Linux-related
28 February 2012
Scalability is another important factor, as most companies have multiple products or different versions of a product. In the past, this has often meant they would need different development tools for the various products, with different code, which added to time to market and made maintaining code more difficult. With the ARM architecture, a company can invest in software tools and develop code on the platform all the way down to a
Components in Electronics
low-end, affordable Cortex-M0 core, and up to a powerful Cortex-A8 device. This means a smaller engineering team can create and reuse code faster, cutting time to market. They can also be confident that they are creating software that will be supported by future generations of the ARM architecture.
Software availability The success of the ARM architecture has created a virtuous circle: more hardware devices means more customers, which means more software is available, which makes the architecture even more attractive. With the wide range of ARM cores now available, many third parties have taken the opportunity to develop software including operating systems, middleware and communications stacks. Engineers know if they pick ARM, it’s not an architecture that’s going to go away, and software support will be there for them, while the software companies know if they develop for ARM they will have a broad customer base to market to.
The availability of good software has become increasingly important, not least to help companies meet time to market pressures. To help engineers find the right solutions, Avnet and ARM recently launched the Embedded Software Store (
www.embeddedsoftwarestore.com), an online marketplace for software IP. This helps engineers to improve their efficiency, with ready-to-go software for the ARM architecture that’s easy to find. Using pre- written software reduces development costs and cuts time to market, as well as freeing up time for companies to focus on the differentiators that will make their product stand out.
Developing software platforms All of the topics highlighted here are contributing to the overall industry trend of developing software on a project by project basis and also investing in reusable software platforms. In the past this has not been possible because of the many proprietary architectures, and tools, incompatible peripheral sets and the need to often program at a very low level.
In contrast, many OEMs have identified they can develop a reusable platform, making use of an internal or off the shelf RTOS, using the silicon vendors’ rich peripheral libraries and middleware support and purchasing software components such as those available through the new Embedded Software Store. This is on top of
Embedded and mobile designers are facing increasing pressures to create complex, power-efficient products. Richard York and Jim Carver look at why chips based on ARM’s architecture are becoming an increasingly popular solution
internal teams creating software modules that are more portable and sharing those from project to project.
Hardware innovation The ARM architecture is embraced by many large semiconductor companies, with almost all the key hardware players offering ARM processor-based solutions. Customers appreciate this because they can use a power efficient Cortex-M4 device in one project, which might replace DSPs in a processing-intensive application, then in their next application they might choose an ultra low power core such as the Cortex- M0. The huge choice of silicon gives them many options.
All the suppliers use the same core intellectual property (IP), but they add their own innovations around the core – which means customers can re-use software and utilise common development tools. While a proprietary architecture might offer a couple of dozen devices, there are over 1000 ARM processor-based devices to choose from. This is true scalability, with everything from extremely power efficient 50c parts, up to dual-core Cortex-A15 processors running at more than a Gigahertz.
Although ARM processor-based devices have been available for many years, the pace of innovation in hardware based around ARM cores is accelerating. With more companies developing chips based on the architecture, there are new devices being launched every month that help improve performance and battery life for embedded and mobile applications. To give just a few examples, NXP has developed dual-core microcontrollers with Cortex-M0 and Cortex-M4 cores on the same die, while Energy Micro has energy-saving microcontrollers and transceivers, and Atmel’s ARM processor-based devices include differentiated security features. For the past decade or so, ARM has
increasingly become the standard architecture for embedded and mobile applications. Switching architectures is not a trivial decision, but engineers are being persuaded by the performance, power efficiency and scalability of ARM devices, and the enormous range of software that is available.
ARM |
www.arm.com Avnet |
www.avnet.com Richard York is Director of Embedded Processor Products at ARM and Jim Carver is Technical Director, Advanced Architectures at Avnet
www.cieonline.co.uk
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48