A huge reason for the low power usage is the iPhone.
Apple spent years incrementally improving efficiency and performance of their chips for phones. Intel and AMD were more desktop based so power efficiency wasnt the goal. When Apple's chips got so good they could transition into laptops, x86 wasn't in the same ballpark.
Also the iPhone is the most lucrative product of all time (I think) and Apple poured a tonne of that money into R&D and taking the top engineers from Intel, AMD, and ARM, building one of the best silicon teams.
Apple purchased Palo Alto Semi which made the biggest difference. One of their best acquisitions ever in my opinion… not that they make all that many of those anyway.
Equally (arguably) importantly, Johny Srouji joined Apple the same year as PA Semi's acquisition - '08 – and led Apple A4. (He previously worked at IBM on POWER7, which is a fascinating switch in market segment.)
I vaguely remember Intel tried to get into the low power / smartphone / table space at the time with their Atom line [0] in the late 00's, but due to core architecture issues they could never reach the efficiency of ARM based chips.
Intel and Nokia partnered around 2007 .. 09 to introduce x86 phone SoCs and the required software stack. Remember MeeGo? Nokia engineers were horrified by the power consumption and were convinced it wouldn't work. But Nokia management wanted to go to a dual supplier model instead of just relying on TI at all cost.
MeeGo proceeded far too slowly and Elop chose his former employers' Windows instead in 2011. Nokia's decline only increased and Intel hired many Nokia engineers.
Soon Nokia made no phone anymore and Intel did not even manage to make their first mass-selling product.
ARM-based SoCs were 10 years ahead in power saving. The ARM ecosystem did not make any fatal mistakes, Intel never caught up.
I don't think it was core architecture issues. My impression is that over the years their efforts to get into low-power devices never got the full force of their engineering prowess.
I worked for an IP vendor that was in some Atom SoCs (over a decade ago now though) - from what I remember the perf/w was actually pretty competitive for contemporary ARM devices when we supplied the IP, but then took so long to actually end up in products it ended up behind others - other customers were already on the next generation by that point, even if the initial projects started at about the same time. And the atoms were buggy as hell, never had more problems with dumb cache/fabric/memory controller issues.
To me the Atom team always felt like a dead-end inside intel - everyone seemed to be trying to get in to a different higher-status team ASAP - our engineering contacts often changed monthly, if we even knew who our "contacts" were meant to be at any time. I think any product developed like that would struggle.
When they bought PA Semi the company worked on IBM Power architecture chips. It was very much the team Apple was after, not any one particular technology.
Besides Apple's SoCs they also have made dedicated silicon for secure enclaves, wifi, bluetooth, ultra-wideband, and cellular radios, and motion coprocessors.
Apple bought PA Semi a long time ago. They have a significant silicon development group. Their architecture license (they were an early investor in ARM) for ARM means they get to basically do whatever they want using the ARM ISA. The SoCs in pretty much all their devices are designed in-house.
Were they ARM investors at the time they needed CPU for Newton? Was that before or after e.g. iPaq PDA-s? And latter - was it that it looked that Apple maybe in danger of going under, and then they sold their ARM stake and got a cash injection that way?
I remember iPaq PDA fondly. Wrote a demo to select a song from a playlist with few thousand author-album-song with voice query. The WiFi add-on was a big plastic "sleeve", that the iPaq slid into, not the other way around. Could run the ASR engine for about whole 10 mins before it drained the battery flat, haha. :-)
IIRC Apple originally invested in ARM during the development of the Newton. The original Newtons used ARM 610 CPUs. I don't know exactly when they sold their ARM stake but they kept their architecture license.
The Newton was long before the iPaq, the MessagePad was released in 1993.
Q> And latter - was it that it looked that Apple maybe in danger of going under, and then they sold their ARM stake and got a cash injection that way?
A> And yes. In the late-1990s turnaround, Apple sold down its ARM stake in multiple tranches after ARM’s 1998 IPO, realizing hundreds of millions of dollars that helped shore up finances (alongside the well-known $150 million Microsoft deal in Aug 1997).
Apple has bought startups with various technologies like Anobit, that developed advanced flash memory controllers, and have funded development efforts by partners. For example Apple worked hand in glove with Sharp to develop the tech for their 5K display panels. They also now have their own cellular chip designs in some models, in their quest for independence from Qualcomm. That’s all from memory, I’m sure there are many more examples.
Outsourced to who? The only companies with the engineers you’d need are the other CPU makers like Intel, AMD, Qualcomm, and Nvidia. And none of them make a CPU as efficient as Apple does.
They design much more in house than any other smartphone brand, except maybe Samsung.
CPU, GPU, neural processor, image signal processor, U1 chip for device tracking, Secure Enclave for biometrics, a 5G modem (only used in the 16e so far)…
They don’t manufacture the chips in house of course. They contract that out to TSMC and other companies.
Apple took the ARM base design (they licensed it), and then they modified and tweaked it.
You get the ARM ISA, and compilers that work for ARM will compile to Apple Silicon. It's just that the actual hardware you get, is better than the base design, and therefore beats other ARM processors in benchmarks.
It's more than that. They have an unlimited license to arm designs, and can change them as they see fit, since they were an early investors (or something along those lines). Other manufacturers can't get these terms, or if they can, it will be prohibtly expensive
The thing about Apple having a “special license” due to being a partial founder of Arm is an urban legend. They have an architectural license, just like several other companies making custom Arm CPUs do.
It is very unlikely Apple uses anything from ARM’s core designs, since that would require paying an additional license fee and Apple was able to design superior cores using its architectural license.
I don't think it is so much efficiency of their chips for their hardware (phones) so much as efficiency of their OS for their chips and hardware design (like unified memory).
It is likely the hardware effiency of their chips. Apple SoCs running industry-standard benchmarks still run very cool, yet still show dominant performance. The OS efficiency helps, but even under extreme stress tests like SPEC, the Apple SoCs dominate in perf & power.
> It is likely the hardware effiency of their chips. Apple SoCs running industry-standard benchmarks still run very cool, yet still show dominant performance
I have heard that Apple Silicon chips are designed around the retain-release cycle that goes back to NeXT and is still here today (hidden by ARC compilation), but I don't think that's the whole story. Back when the M1's came out, many benchmarks showed virtualized Windows blowing the doors off of market-equivalent x86 CPUs.
Also, there's the obvious benefits of being TSMC's best customer. And when you design a chip for low power consumption, that means you've got a higher ceiling when you introduce cooling.
The SoC benefits are being ignored by some people here. Apple doesn't control every piece of software as some here posit, however, OS optimizations and utilization of extra-efficiency cores (though still requiring SoC design they do also need specific OS code support) are part of the performance.
All instructions across x86 and Arm are being decoded to micro-operations, which are implementation specific. You could have an implementation which prioritizes performance, or an implementation that prioritizes power consumption, regardless of the ISA.
Decoding instructions, particularly on a modern die, doesn’t consume a significant amount of area or power, even for complicated variable length instructions.
Because it’s a take thst sounds like someone who has been reading comp.sys.mac.advocacy from 1995 when the PPC vs x86 wars were going on (and when PPC chips were already behind in performance) up through 2005 when Apple gave up and went to Intel.
You are wrong. The Snapdragon X Elite is actually a great example, unlike M1 it's performance isn't particularly great and it eats 50W under load. That makes its CPU cores a fair bit less efficient that AMDs even on the same production node. If Apple Silicon didn't exist then you might instead argue that x86-64 is more efficient than ARM.
If all that's true then why does Snapdragon have better battery life? As I said in my comment the great battery life comes from when the CPU isn't being used. It's everything else around it. That's where AMD is still significantly behind.
Apple spent years incrementally improving efficiency and performance of their chips for phones. Intel and AMD were more desktop based so power efficiency wasnt the goal. When Apple's chips got so good they could transition into laptops, x86 wasn't in the same ballpark.
Also the iPhone is the most lucrative product of all time (I think) and Apple poured a tonne of that money into R&D and taking the top engineers from Intel, AMD, and ARM, building one of the best silicon teams.