Apple’s Sept. 7 Event

[exoticspice1]

For the first time ever, the GPU increase is <1%. Really sad. They only impoved the RAM. Qualcomm will beat them in GPU with Gen 2 Snapdragon.

 

[Cmaier]

If this Geekbench result is to be believed, the speed improvements look minimal. But it's obviously unconfirmed yet. Also weird that single core perf went up by ~10% while multicore stayed the same.

Yeah that looks doubtful to me. A 10 percent increase I could believe. I would expect performance plus power reduction to add up to around 30 percent. But multi core should also increase (unless apple did something unprecedented for them like reduce the top clock speed when multiple cores are churning).

Anyway, if A16 performance cores increase performance by 10 percent while reducing power by 20 percent, and taking into account that maybe 5-10 percent of that is from process improvements, you’d probably end up with M3 getting a decent performance bump at the same power consumption as M2. A really nice speed bump if M3 is on a better node than N4 (which seems likely).

In other words, 10 percent performance plus 20 percent power savings is pretty good.

Of course, numbers still look suspicious and we need more data and confirmed data to speculate more intelligently.

 

[Cmaier]

For the first time ever, the GPU increase is <1%. Really sad. They only impoved the RAM. Qualcomm will beat them in GPU with Gen 2 Snapdragon.

Could be. I guess the question is whether it makes any difference on a phone.

 

[DT]

Qualcomm will beat them in GPU with Gen 2 Snapdragon.

Oh my! And then what happens?

 

[Andropov]

For the first time ever, the GPU increase is <1%. Really sad. They only impoved the RAM. Qualcomm will beat them in GPU with Gen 2 Snapdragon.

I don't think it's possible that they improved GPU memory bandwidth by 50% without it affecting performance significantly. Maybe in some cases where there's not a lot of geometry and memory bandwidth usage is low the improvement is not noticeable. But it has to be significant in memory-bound workloads.

Yeah that looks doubtful to me. A 10 percent increase I could believe. I would expect performance plus power reduction to add up to around 30 percent. But multi core should also increase (unless apple did something unprecedented for them like reduce the top clock speed when multiple cores are churning).

Anyway, if A16 performance cores increase performance by 10 percent while reducing power by 20 percent, and taking into account that maybe 5-10 percent of that is from process improvements, you’d probably end up with M3 getting a decent performance bump at the same power consumption as M2. A really nice speed bump if M3 is on a better node than N4 (which seems likely).

In other words, 10 percent performance plus 20 percent power savings is pretty good.

Of course, numbers still look suspicious and we need more data and confirmed data to speculate more intelligently.

Where's the 20% reduction in power usage number come from? Did they say so in the keynote? If that's true then the improvement it's not minimal at all.

 

[Andropov]

More weirdness in the benchmark result. If we believe it, Apple improved multicore AES-XTS encryption by a huge margin: 12.5GB/s on iPhone 14 Pro vs 9.15GB/s on iPhone 13 Pro, a +36% improvement. But, at the same time, single core AES-XTS is only improved by 3.7% (4.83 to 5.01GB/s). Other than that, the multicore score is worse (or equal, at best) in almost every other subset of the multicore Geekbench tests vs the A15 Bionic. The overall multicore score is only about the same as the iPhone 13 Pro because of the unusually high AES-XTS subtest score, take that test out and it would actually be worse than the iPhone 13 Pro.

 

[Joe]

I was tempted by the Watch Ultra but the more I look at it the more it just looks way too big (for me) to also use as an everyday watch alongside the main focus of it. I don’t consider myself some big athlete but I do run, hike and stuff.

I’ll keep my S4 for now. It’s still in great condition and does what I need. I’ll see what they have next year.

 

[Cmaier]

I don't think it's possible that they improved GPU memory bandwidth by 50% without it affecting performance significantly. Maybe in some cases where there's not a lot of geometry and memory bandwidth usage is low the improvement is not noticeable. But it has to be significant in memory-bound workloads.


Where's the 20% reduction in power usage number come from? Did they say so in the keynote? If that's true then the improvement it's not minimal at all.

Yeah, they said that in the keynote.

 

[Cmaier]

More weirdness in the benchmark result. If we believe it, Apple improved multicore AES-XTS encryption by a huge margin: 12.5GB/s on iPhone 14 Pro vs 9.15GB/s on iPhone 13 Pro, a +36% improvement. But, at the same time, single core AES-XTS is only improved by 3.7% (4.83 to 5.01GB/s). Other than that, the multicore score is worse (or equal, at best) in almost every other subset of the multicore Geekbench tests vs the A15 Bionic. The overall multicore score is only about the same as the iPhone 13 Pro because of the unusually high AES-XTS subtest score, take that test out and it would actually be worse than the iPhone 13 Pro.

Right, that makes no sense at all.

 

[Runs For Fun]

But did you order them in all available colors?

Haha no but I think we know someone who did.

 

[Runs For Fun]

So the look of the Ultra has really grown on me. I love the look of it. So 100% CAD renders are just pure ugly.

 

[Pumbaa]

So the look of the Ultra has really grown on me. I love the look of it. So 100% CAD renders are just pure ugly.

The look works. I’m mostly worried about the size. Maybe there’s an Ultra in my future if I get used to the 45mm 8 first after the tinier 3.

 

[Nycturne]

The look works. I’m mostly worried about the size. Maybe there’s an Ultra in my future if I get used to the 45mm 8 first after the tinier 3.

It is the size I’m worried about as well. But for me, it’s the bump for the sensors that’s the biggest fit issue. Hoping things have improved since the S3 in that regard.

 

[Citysnaps]

I think Watch Ultra looks terrific and has an outstanding set of features. It's very tempting just from an exploring the tech perspective.

But ultimately think it would look weird on my wrist. If I was into running, cross country hiking, diving, etc that would be different - it would then become an essential tool.

Maybe I'll just by another Knipex hand tool, instead. I have my eye on this one. Knipex makes the absolute best hand tools for a wide variety of applications. All are made in Germany.

 

[Yoused]

Could be. I guess the question is whether it makes any difference on a phone.

Or, quite frankly, a computer. SC/MC performance has run up against a wall-of-irrelevance, where the heavy workloads run off-core and CPU cores are already well beyond good enough for what needs to be done. Who really cares about breaking 2000 when the difference it makes will be imperceptible to, honestly, anyone.

 

[Herdfan]

What? You saying that just cuz I gots a big ass truck, I can’t have a lovely dainty white iPhone?

I have a white 12 Pro and a big ass truck, so sure you can.

I do have it in a black case so the only way you know is by the camera opening though.

 

[tomO2013]

Ugh…. pre-orders strart at 5am PT.
My wife jokes that it’s the only time I’m able to haul ass out of bed so early, when there is a big apple gadget release that I want!

 

[Chew Toy McCoy]

After thinking about the iPhone 14 Pro Max and its upgraded cameras for awhile, and deciding I'll likely purchase one, thought I should make a mockup of it from a scrap piece of wood to compare with my 12 Pro that I use now. It fits comfortably in the pocket of the t-shirts I like to wear, so it looks like it's an official go.

View attachment 17405

At one point I had whatever the largest iPhone with a home button was, about the size of a checkbook. I really liked it. I skipped the first version that nixed the home button. When I finally ordered a version that didn’t have a home button I assumed that since it didn’t have that and they decreased the bezel that the 2nd to biggest would have at least the same screen size as my old checkbook sized phone. It did not and I was mad.

I know a lot depends on need and preference but for me Apple has been the grandmaster of making me pay a premium slight upgrade on top of their premium price to get something that is at least a little better than casual use. It's like they are going "What is another $100 on top of $1,000?" I'll tell you what it is. It's another $100 for what I feel I shouldn't have to pay another $100 for.

 

[Cmaier]

Ugh…. pre-orders strart at 5am PT.
My wife jokes that it’s the only time I’m able to haul ass out of bed so early, when there is a big apple gadget release that I want!

Used to be worse when it was midnight but the system didnt actually start working until 1am

 

[DT]

I'm kind of thinking about changing over to Air Pods [...]

I did not even realize I couldn't order these yet, hahahaha ...