For a CPU that hasn't seen the light of day, there's a great deal of debate surrounding Apple's A6—and the suggestion that it may not appear until later in 2012. The A6 is a complex bit of hardware; rumors indicate that the chip is a quad-core Cortex-A9 CPU built on 28nm at TSMC
and utilizing the latter's 3D fabrication technology.
We've previously discussed Intel's own 3D tech
; TSMC's is presumably based on similar concepts.The CPU's competitive performance is surprisingly unclear. While the Cortex-A9 is a proven design, Apple's A6 will be one of the first 28nm chips on the market. That means it'll serve as a test case for TSMC's introduction of both 28nm gate-last technology and 3D chip stacking. Meanwhile this is TSMC's first experience with Apple parts; the A4 and A5 have both historically been manufactured by Samsung.
Image Source: ARM Corp.
The final piece of the puzzle is the Cortex-A15, which is expected to debut at the same time. Qualcomm's MSM8960 will offer dual-core Eagle processors at 1.5-1.7GHz, paired with the company's Adreno 225 GPU and dual-channel LPDDR2 at 500MHz. Given the performance improvements baked into Eagle's silicon, a dual-core Cortex-A15 could prove to be surprisingly robust competition for a slower quad-core Cortex-A9. Power consumption will depend heavily on how well-tuned TSMC's 28nm process technology is at that point.
The Linley group notes that "if Apple launches the iPad 3 in January 2012, following its usual schedule, the tablet will have to use the same A5 processor as the current iPad 2, relying on the rumored new high-resolution “Retina” display to drive the upgrade cycle. Apple could then deliver a midyear iPad upgrade, which would be unusual for the company, or it could plug the A6 into next year’s iPhone first and then into the iPad in early 2013."
We agree with Linley's assessment that a mid-year upgrade would be unusual and suspect that the A6 is more a test vehicle than anything else. Apple's plans for the iPad/iPhone through 2012 are likely designed to give the company flexibility in the event that the A6 doesn't appear on time or deliver as promised. Given TSMC's substantial 40nm difficulties and the inherent challenge of adopting multiple new technologies, treating the A6 as a side bet would be the safest path. Apple's product introductions are typically built around experiences, whether that means a major software upgrade or better battery life courtesy of a lower-power CPU. Given the iPhad's dominant market position, Apple can afford to wait, tune its technology, and release an update when it makes the most sense to do so.