iPhone 5 A6 chip uses a custom ARMv7 core, faster than A5X chip on the new iPad: AnandTech

Ever since the iPhone 5 has been launched there has been a lot of confusion about the A6 chip Apple uses in it. Some thought it was a quad-core ARM Cortex A9, and then some thought it was an ARM Cortex A15. Benchmark experts, AnandTech have uncovered an even more puzzling revelation.
According to them, the A6 chip inside the iPhone 5 is a completely custom designed Apple chip using an ARMv7 core. Additionally, some GeekBench benchmarks have surfaced courtesy MacRumors that suggest this new chip is significantly faster than the A5X chip on the new iPad and even faster than the Exynos chip found on the Galaxy S III.
The Geekbench results also point towards a unique SoC design as it points towards a dual-core design with a clock speed of 1.02GHz in comparison to the iPhone 4S’ 800MHz chip. In fact, it looks like Apple could be ramping the instructions executed per clock (IPC) and it could be using some technology similar to Intel’s Turbo Boost, to ramp up speeds when needed. In the GeekBench results one can clearly decipher that the A6 chip is significantly faster than the A5X chip seen in the new iPad, but it is also a hair faster than the Nvidia Tegra 3 on the Asus Nexus 7 tablet and the Exynos chip on the Samsung Galaxy S III.
This is important because while most of the current generation chips use an ARMv7 core, they are basically offshoots of ARM’s Cortex A9 and Cortex A15 designs as seen in the Nvidia Tegra 3 and the Qualcomm Snapdragon S4. Apple’s solution is a completely custom, and it also manages to integrate LTE radios from Qualcomm. It was previously suspected that Apple might be using licensed technology from Qualcomm for the A6 chip, but it looks like Apple has really stepped up the game with a completely custom solution.
The report notes that there are hints for the same inside Xcode 4.5, the software development platform used by iOS developers. In fact, support for a custom ARMv7s architecture has been found which contains some key differences from the standard ARMv7 core.
Additionally while Apple has licenses for both ARM cores and the ARM instruction set, it could have gone ahead and designed its own solution. AnandTech further emphasizes that because the A15 architecture was designed for servers Apple could have used an A9 core and ramped up the CPU clock speeds, effecting the 2x performance boost over the A5 chipset used in the iPhone 4S. In fact this is not an unusual way of gaining generational performance bumps with out a generational change in architecture as it is more power efficient. Consumers have already noticed this in the early 2000s when Intel’s Pentium 4 processor attained massive performance gains by ramped up clock speeds.
Read