Blog

Maecenas sed diam eget risus varius blandit sit amet non magna. Sed posuere consectetur est at lobortis. Donec ullamcorper nulla non metus auctor fringilla. Curabitur blandit tempus porttitor. Sagittis lacus vel augue laoreet rutrum faucibus dolor auctor.

Educated Guesses Re: A7 Transistor Density and Foundry

A forum post at MacRumors contains pretty interesting guesswork about the A7. If it is right, Apple will be moving fab from Samsung to TMSC for the A7. This is an important and expected move that may be here sooner than most observers thought it would be.

First, if the die sizes are roughly the same, how are they getting twice the amount of transistors in there? We do know a few things about Apple A-series SoC designs. They've been getting progressively more custom from the A4 to the A6, with the A6 having fully hand designed CPU cores. This is opposed to the "place and route" approach where companies allow a CAD tool to automatically floorplan their device based on their functional description of the processor and the use of standard library blocks. Going full custom allows you to get denser because you're manually designing, but it also takes a lot more time, which is why almost no one does it.

Taking the time to make the inside as beautiful and perfect as the outside is classic Apple. It also helps to differentiate their products when "almost no one does it".

The post continues:

So, how do we get the rest? TSMC. TSMC is known for having denser processes at the same feature size. This can easily be seen by comparing standard ARM cores and their die sizes across processes. TSMC is noted for having a 20% or better density efficiency. So, if we compound the 20% density improvements, we get to about 1.5x. This is about as best as we can do with simple heuristics. We don't know how much custom circuitry apple will do to further improve density. It would be overly laborious and likely fruitless to try and weigh die share versus circuitry density (CPU and GPU) to get an overall idea. In either event, it seems obvious that the move to 28nm and TSMC are both necessary to get the claimed 2x transistor density.

But Apple won't use TSMC until A8 you say! Well, the A7 leak had a new chip letter identifier that suggested a different fab . When macrumors consulted chipworks about this change, Chipworks suggested that it meant the chip was TSMC. So, that seems enough of a smoking gun to me.

Daniel Eran Dilger has written in detail about the likely transition to TSMC.

The move to a 20nm process is a vastly expensive and incredibly complex new technological frontier, one that caused NVIDIA's CEO Jen-Hsun Huang to publicly question whether the move even made financial sense just last year.

One problem Apple doesn't have is a lack of resources, particularly overseas where it has over $90 billion flying in a holding pattern. Few companies have TSMC's chip fab technology, and few companies have Apple's resources to fund such technological endeavors.

If it's possible, Apple and TSMC would likely be the ones to pull it off, heavily motivated by Apple's desire to pull its funding from Samsung, and equally motivated by TSMC's (and Taiwan's) desire to establish itself as one of the world's few major chip foundries.

His article includes a pretty convincing observation:

Perhaps the most convincing evidence of a TSMC-built A7 is that the news of Apple's new 64-bit chip was also apparently news to Samsung. In fact, it was largely news to everyone.

Either that is true or Samsung is engaging in some pretty wild theatrics.