The pressure to deliver longer battery life in mobile handsets and tablets is forcing chipmakers to put spare processor cores onto their devices with the intention of using their signature, power-hungry cores only on an occasional basis.
Market leaders ARM and NVIDIA have independently chosen to use the same approach for forthcoming processor designs in which high-speed processors are only called upon to run code if more efficient ‘shadow’ processors run out of steam. The graphics chipmaker unveiled the Kal-El architecture it plans for the Tegra-3 system-on-chip (SoC) a couple of weeks ago. Now ARM has followed with what it calls the ‘big-little processing model’.
“The concept is being driven by the insatiable consumer demand for more performance at the same time as a demand for more power efficiency,” says Tom Cronk, executive vice president of ARM’s processor division. “Moving down in process geometry helps a bit and new processor architectures help a bit but the consumer is demanding more than the combination of those things.”
Avner Goren, director of strategic marketing at Texas Instruments, said a problem for typical mobile devices is that they “are always running some tasks in the background. We tried to use more dedicated hardware units to perform some of those tasks but even then we still had the fundamental issue of what to do with the main CPU”.
The answer that ARM and NVIDIA chose is to move processing onto a dedicated low-energy processor whenever possible, but keep one or more fast cores in reserve for situations where the user needs to play graphics-intensive games or watch high-definition video. Both companies decided, for the moment, to hide this migration from the operating system to make it easier for handset makers to use the technique.
The NVIDIA Kal-El processor uses five ARM Cortex-A9 processors in a symmetric multiprocessing arrangement; but one of them is tuned for lower overall power consumption, and so runs more slowly than the others. This is the core that is used for basic background tasks. As soon as activity increases beyond the capability of that slower processor, one or more of the others start up.
ARM decided to develop new processor core rather than retrofit its existing designs, launching the Cortex-A7 last week. The A7 is designed to act as the shadow for the faster A15. The A7's simpler instruction pipeline coupled with some optimisations learned from experience with the older A8 and A9 processors means it uses less energy per instruction than the larger A15 core.
Jae Cheol Son, vice president of the SoC platform development team at Samsung, told E&T: “The beauty of the A7 is that it is the default processor. The A7 will be the processor that handles most of the software tasks.”
Asked if the die space allocated to the larger A15 core is going to waste, Son replied: “If we don't have the A15, we can't answer the demand for an increase in performance. It’s better to have both.
According to ARM, the combination of A7 and 15 can offer power savings of up to 70 per cent compared with the processors in smartphones available today, although some of that improvement comes from an improvement in process technology. Today's devices use 40nm or 45nm processes whereas the combination that companies such as Samsung Electronics and Texas Instruments have now decided to licence is aimed at the 28nm high-k, metal-gate processes now moving into volume production.