What makes ARM superior?

[Yebubbleman]

I think we're still some way away from producing ARM-based machines that are "twice as powerful" as comparable Intel/AMD x86 machines, but they are certainly getting closer:

Amazon's Arm-based Graviton2 Against AMD and Intel: Comparing Cloud Compute

www.anandtech.com

)

100+ Benchmarks Of Amazon's Graviton2 64-Core CPU Against AMD's EPYC 7742 - Phoronix

www.phoronix.com

No one is saying "twice as powerful". I certainly wasn't. But the A12X and A12Z both outperform all eighth gen Intel Macs that are i7 or less. They likely outperform the 10th Gen Y series chips in the 2020 MacBook Air and I'm sure they give the 10th Gen U series chips in the 2020 Intel 4-port 13" MacBook Pro a run for its money if not totally outperforming those as well. What we'll get in an A14 based Mac specific SoC will only be better. Twice as powerful is unlikely at first. But it's likely that the second gen Mac specific SoCs will be at least twice as fast as contemporary Intel processors at native workflows. Though, Apple will, at that point, return to the Apples and Oranges of comparison between its Macs and the PCs produced by the rest of the PC industry that was commonplace during the PowerPC era.

The above is for cloud servers, so is not necessarily very relevant for most consumers, but we can draw the following conclusions:

1) An Apple Silicon SoC will probably be able to match or slightly exceed Intel/AMD CPUs in the same market segment, whilst using significantly less power.

They won't "slightly" exceed. The first wave of Apple Silicon Mac SoCs will clearly exceed the processors employed by the predecessor machines. You're not going to see a first generation Apple Silicon 13" MacBook Pro handily outperform an Intel 16" MacBook Pro, but you will see it handily outperform the last Intel 13" MacBook Pro.

2) Apple will probably balance performance vs battery life in ASi laptops - we may see something that is "about the same speed as Intel Tiger Lake" but with 15-20 hour battery life.

Again, Apple Silicon destined for the MacBook Air and 13" MacBook Pro will handily outperform Intel's chips. Tiger Lake isn't THAT MUCH faster than Ice Lake to change that as far as those Macs are concerned. Where Apple may have work they need to do is in competing with the H series CPUs. But the Apple Silicon 16" MacBook Pro has AT LEAST eight months more in the oven if current rumors are to be believed.

3) Desktops SoCs will probably try to increase the number of CPU cores and custom SoC features to beat the current high-core-count Intel & AMD CPUs.

Can't argue there. That sounds both correct and sensible.

 

[theorist9]

No one is saying "twice as powerful". I certainly wasn't. But the A12X and A12Z both outperform all eighth gen Intel Macs that are i7 or less.

Unless the following was in some other context I didn't understand....

Performance per watt is huge. They'll be able to produce a computer that's twice as powerful while consuming a fraction of the power of your average Intel processor.

 

[jdb8167]

That's a very good point. But I'm not sure there is a huge gap between Apple Silicon and Intel Tiger Lake, particularly the 28W version (https://www.tomshardware.com/features/intel-11th-gen-tiger-lake-superfin-10nm-benchmarks). I actually think the first ASi Mac may perform slightly worse than this in Geekbench scores, but will do better is some tasks where custom silicon makes a difference, and will offer better life.

It will exciting to see what the reality turns out to be!

The A12Z running at 2.5 GHz vs. a 3 GHz 1185G7 (top of the line Tiger Lake) in Geekbench is: 1120/4650 vs 1560/6080. Adjusting for the additional 500 MHz clock, the the A12Z would be about: 1350/5580. So with a 2 year old microarchitecture the A12X/Z is about 10% slower than a new design at similar clocks. Given the new 5 nm design and probably more cores, I would expect the high end mobile ASi SoC to have significantly better performance than the top of the line Tiger Lake.

A more interesting point will be the integrated graphics performance. Geekbench Compute shows about 19600 for the G7 Xe graphics. My iPad Pro with the A12Z shows about 1/2 of that performance at 10000. That is a pretty large difference to overcome.

Edit: Never mind on the graphics. The A12Z was Metal and the Tiger Lake is OpenCL. They can't be directly compared.

Then there is the new Microsoft/Qualcomm SQ2 to add into the mix. I'm interested to see if that CPU can come close to either Apple's or Intel's performance.

 

[johngwheeler]

That's a very good point. But I'm not sure there is a huge gap between Apple Silicon and Intel Tiger Lake, particularly the 28W version (https://www.tomshardware.com/features/intel-11th-gen-tiger-lake-superfin-10nm-benchmarks). I actually think the first ASi Mac may perform slightly worse than this in Geekbench scores, but will do better is some tasks where custom silicon makes a difference, and will offer better life.

It will exciting to see what the reality turns out to be!

The A12Z running at 2.5 GHz vs. a 3 GHz 1185G7 (top of the line Tiger Lake) in Geekbench is: 1120/4650 vs 1560/6080. Adjusting for the additional 500 MHz clock, the the A12Z would be about: 1350/5580. So with a 2 year old microarchitecture the A12X/Z is about 10% slower than a new design at similar clocks. Given the new 5 nm design and probably more cores, I would expect the high end mobile ASi SoC to have significantly better performance than the top of the line Tiger Lake.

A more interesting point will be the integrated graphics performance. Geekbench Compute shows about 19600 for the G7 Xe graphics. My iPad Pro with the A12Z shows about 1/2 of that performance at 10000. That is a pretty large difference to overcome.

Edit: Never mind on the graphics. The A12Z was Metal and the Tiger Lake is OpenCL. They can't be directly compared.

Then there is the new Microsoft/Qualcomm SQ2 to add into the mix. I'm interested to see if that CPU can come close to either Apple's or Intel's performance.

It will be interesting to how it shapes up!

Of course, Geekbench is just a baseline...that may or may not be indicative of real-world performance with specific apps. We won't know for sure until production models are available.

 

[jdb8167]

It will be interesting to how it shapes up!

Of course, Geekbench is just a baseline...that may or may not be indicative of real-world performance with specific apps. We won't know for sure until production models are available.

Geekbench is mostly just very convenient. It would be nice if there was a more comprehensive cross-platform benchmark but Geekbench is the best I've found. For graphics we'll be able to use Cinebench which is well regarded soon after ASi Macs are released.

 

[thingstoponder]

That's a very good point. But I'm not sure there is a huge gap between Apple Silicon and Intel Tiger Lake, particularly the 28W version (https://www.tomshardware.com/features/intel-11th-gen-tiger-lake-superfin-10nm-benchmarks). I actually think the first ASi Mac may perform slightly worse than this in Geekbench scores, but will do better is some tasks where custom silicon makes a difference, and will offer better life.

It will exciting to see what the reality turns out to be!

They won’t be worse even if the Macs ran at the same frequency as the a14. ~1550 single core in geekbench is rough estimate for the a14 if we go by the numbers Apple provided which is a12 plus 40%. The leaked score from a few months ago was 1658 but i don’t know if it’s real, the single core looks plausible but the multi core seemed high at 4612 for it considering it’s still 2+4. Although l0vetodream did tweet it out, so it may be real.

The A12Z running at 2.5 GHz vs. a 3 GHz 1185G7 (top of the line Tiger Lake) in Geekbench is: 1120/4650 vs 1560/6080. Adjusting for the additional 500 MHz clock, the the A12Z would be about: 1350/5580. So with a 2 year old microarchitecture the A12X/Z is about 10% slower than a new design at similar clocks. Given the new 5 nm design and probably more cores, I would expect the high end mobile ASi SoC to have significantly better performance than the top of the line Tiger Lake.

It’s not a 500 MHz difference because the Tiger Lake boosts up to 4.8GHz. If the Tiger Lake was capped at 3GHz it would lose to even the a12.

A12 has 70% better IPC than Skylake. I’m not sure how this compares to Tiger Lake but Tiger Lake is actually a slight regression in IPC from Ice Lake and Ice Lake was disappointing as is from an IPC perspective, Intel exaggerated. And keep in mind the a12 is 2 gens old for Apple.

https://twitter.com/i/web/status/1307645405883183104

 

[Yebubbleman]

Unless the following was in some other context I didn't understand....

I didn't say that they WOULD produce a machine that's twice as powerful. Only that they could. And while it may not be in the initial round, certainly, at the rate of growth that they have, they'll be more than twice as powerful as Intel equivalent machines. Do you realize how much more powerful the A14 is than the A9? Do you realize that Intel is still providing only minor improvements over the chips they put out circa the A9's advent? That's the point I'm trying to make here.

 

[theorist9]

https://twitter.com/i/web/status/1307645405883183104

Mais non!

 

[Andropov]

That's a very good point. But I'm not sure there is a huge gap between Apple Silicon and Intel Tiger Lake, particularly the 28W version (https://www.tomshardware.com/features/intel-11th-gen-tiger-lake-superfin-10nm-benchmarks). I actually think the first ASi Mac may perform slightly worse than this in Geekbench scores, but will do better is some tasks where custom silicon makes a difference, and will offer better life.

It will exciting to see what the reality turns out to be!

It would make more sense to compare the A12Z to the 15W version of Tiger-Lake, since they have the same TDP. The Geekbench score of the A12Z is ~4% slower, and that CPU was released more than 2 years ago. In comparison, the A13 Bionic has improved the performance of the A12 by ~20% in just one year. If Apple is able to maintain the performance improvements per year they've been making for the last few years, the A14/A14X processors would be hugely faster than TDP-equivalent Intel CPUs.

We'll see soon enough, the iPad Air with the A14 Bionic is set to be released this month.

 

[johngwheeler]

It would make more sense to compare the A12Z to the 15W version of Tiger-Lake, since they have the same TDP. The Geekbench score of the A12Z is ~4% slower, and that CPU was released more than 2 years ago. In comparison, the A13 Bionic has improved the performance of the A12 by ~20% in just one year. If Apple is able to maintain the performance improvements per year they've been making for the last few years, the A14/A14X processors would be hugely faster than TDP-equivalent Intel CPUs.

We'll see soon enough, the iPad Air with the A14 Bionic is set to be released this month.

That's a good point; it was an "unfair" comparison. I'm pretty sure the first ASi Mac laptop will be competing against the 15W Intel products, such as the Dell XPS 13. I'm confident that it will perform better than these - but my instinct tells me that Apple will balance performance against battery life - i.e. 30-50% faster, but double the battery life of the competition.

 

[Carrotcruncher]

When you build your own chip, to work only on your hardware, designed to work only on your software there are going to be gains in performance, no doubt about that. That comes simply from the control you have over it.

The problem with claims about an A12Z matching a 3.5GHz Skylake is that people are comparing tablet performance, nobody actually knows yet how any Apple chip will perform in a desktop. Because right now, the actual desktop chip is being kept under wraps.

Until there is an ARM device running macOS any comparisons are going to be flawed.

Isn't hindsight a wonderful thing !

 

[LeeW]

Isn't hindsight a wonderful thing !

Well, you know, I hate being right