To say that Apple gains nothing by licensing and developing their own ARM chips is disingenuous. They make some of the best, if not the best, ARM processors for mobile/general purpose systems. Consistently great battery life and performance packages, compared to competing ARM packages. After all, iOS is specifically tuned for A*-chips, and vice versa.

By continuing to develop their own chips based on existing ARM specs, they can take advantage of the rising tide while maintaining the freedom to make changes as they want, without waiting on the industry to deliver. I doubt that Apple has that sort of influence, if any at all, over the architectural minutiae of Intel chips. It seems more likely they just take the latest of whatever Intel dishes out, like virtually everyone else.

ARM may never fully compete with Intel processors, but if it ever does, Apple will be in a great position to own its entire processor lineup. That amount of top-to-bottom control costs money, but it gives them a greater position in the market. It doesn't sound crazy at all to me to imagine Apple continuing to blur the lines between their ARM and Intel systems, to the point at which maintaining separate architectures no longer makes sense.

Well this is Engadget. But there's a lot to be said about the articles claiming how Apple is close to surpassing Intel, based Geekbench scores. Geekbench payloads are different for mobile and desktop, ao there is no comparison.

That was on old versions of Geekbench 3, on recent versions the difference between mobile and desktop is minor according to Primate. JPoole provided an example of Geekbench of the A9 running mobile and desktop workload sizes [1].

[1] http://forums.anandtech.com/showpost.php?p=37728909&postcoun...

"ARM and x86 are simply instruction sets (RISC and CISC, respectively). There's nothing about either set that makes one or the other more efficient."

This senior editor should go back to reporting on vaporizers and video games.

Newer Intel CPUs are even using RISC cores with wrappers for CISC functions because it is more efficient.

The opposite is increasingly true, too - micro-op and macro-op fusion joins multiple instructions into one.

There's no clear-cut line no more.

The implementations may have converged but there's still hard differences in the ISA. RISC (including ARM) has fixed-width instructions, x86 has variable width. RISC (including ARM again) is load-store, x86 allows memory to be sources or destinations. So ARM can have simpler decoder logic, if nothing else.

Sure, but it's not true that CISC machines are just proxies to RISC ones.

> no more ?

I can't remember there being a clean line for decades now.

> Newer Intel CPUs are even using RISC cores with wrappers for CISC functions because it is more efficient.

Not even newer chips. The Pentium Pro was released in what, 1996?

> Newer Intel CPUs are even using RISC cores with wrappers for CISC functions because it is more efficient.

this has been going on for quite a while now, iirc approx 20 years or thereabouts...

And x86 started ~20 years before that? Thanks for the info though.

Not only that—when people speak of ARM vs. x86, they aren't really talking about the instruction sets at all—they're talking about the processor lines that have been developed and evolved over period of many years.

What impresses me most about this years upgrade is that the graphics performance of the iPad Pro is better than my 13" Macbook Pro in some benchmarks. This is why they can say "desktop class performance".

Anyway, the title is misleading; the article talks only about the CPU whereas chip implies the whole SoC. But that doesn't matter. The whole article is written rather one sided by someone who just copied what a clueless analyst told him.

I can't wait for Anandtech's full review on the A9X SoC. They already have a short preview on their site: http://www.anandtech.com/show/9780/taking-notes-with-ipad-pr...

In all fairness, in a normal office use, the CPU is very rarely max-ed out, and an ipad pro may be able to do the task as well as a laptop.

It matters more to users doing graphics and videos, which I think is also probably more the target population of the ipad pro.

How much did Intel pay engadget for this article?

I mean how much does Apple pay to all tech blogs to pretend that iPad is still relevant.

Cynicism is easy.

It's specially not a big deal considering they shipped with iOS 9 who stutters in UI operations. Hardware is nothing without software and it seems Apple is forgetting that more and more every year in search of revenue.

This. When you're making enormous margins on the hardware it's hard to prioritize the software platform which indirectly drives the majority of long term revenue.