All about smartphone chipsets
Gone are the days when a mobile phone was judged by its design. In the past, no one really cared about the speed of the interface or the number of pixels in the display, a phone was popular because of the brand and the design. Skip to the present and the mobile phone industry (tablets included) is actually competing with the PC market. Companies have always talked about how the mobile phone is like a computer in your pocket and looking at the latest trend of smartphones, that line is starting to blur. Naturally, just like any computer, the CPU is the first to go through major changes since at the end of the day, your camera, screen resolution, apps and OS can only improve if the CPU gets more powerful.
Smartphones have gone from single-core to dual-core and now quad-core and it’s only going to keep increasing. With companies throwing around terms like Snapdragon, Tegra 3, Mali-400, Cortex-A9, etc., it’s a daunting task to keep up with the different CPUs and chipsets in the market, even for us! You’ve heard the terms SoC being used very often in spec sheets and reviews, so what is it?
What is a SoC?
SoC is short for System on a Chip, an integrated circuit that combines all the primary components of a computer into a single chip. Think of your CPU, graphics card, memory controller and other components all rolled into a single chip, that’s essentially an SoC. This way, handset manufactures can simply drop this chip in their device and reap the benefits of the chipset straight away, rather than having to implement the CPU, GPU, etc. manually. Below, we have two popular SoCs, the first being an Nvidia Tegra 3 and the second, a Qualcomm Snapdragon S4 SoC.
What goes on inside an SoC
Notice that Nvidia’s SoC only features the GPU, memory controller and video out streams whereas Qualcomm has managed to package all of that as well as the wireless radios like GPS, Wi-Fi, LTE, etc as well in that single chip. This is because the Qualcomm uses a smaller fabrication process (28nm) for the S4 chip allowing them to add more components into a single piece of silicon. Nvidia on the other hand is still on the 40nm fabrication for Tegra 3 so that’s why they can’t cram in more than that without increasing the size of the chip.
ARM is where the heart is
Today, a majority of mobile phones, tablets, wireless routers, digital media players, hand-held gaming consoles and many computer peripherals are powered by an ARM microprocessor. Just like Intel and AMD, ARM is a 32-bit microprocessor originally created by Acorn Computers back in 1987. Since then, ARM has been the preferred microchip by any and all companies looking for a cheap and more importantly, low-powered chip for portable devices. ARM uses a completely different architecture as compared to Intel and AMD, who’ve stuck with the x86 architecture and hence, all three companies have happily co-existed. With Windows 8 all set to support ARM devices as well, that equation may soon change.
One chip to rule them all
Since the ARM architecture is licensable, companies can either obtain an architectural licence to design their own, customized CPU or simply choose from one of ARM’s ready-to-ship core designs. Popular core design from ARM include ARMv7, ARM9, ARM11, Coretex-A8, Cortex-A9 and the most recent, Cortex-A15. Companies like Nvidia, Texas Instruments, Samsung and Apple (to name a few) simply choose one of these designs and packages them along with a GPU and other components, into their own branded SoC. However, SoC manufacturers like Qualcomm prefer to design their own custom CPUs rather than go with the crowd. ‘Krait’ is the latest CPU, designed by Qualcomm and is used in their new Snapdragon S4 SoC.