iPad Pro Why do people believe the M1 iPad Pro is going to be 65%, 70% of even 4 times faster than the current A12Z model?

[isoft7]

Why are ridiculous numbers like this tossed around here and on other forums? Even on youtube... insane numbers being projected.

The M1 is going to push big benchmarks and all, but when it comes to "rendering" times (3D work, photo filters, video compression, etc.) it's going to be around 15% faster than the A12Z. Maybe 20%.

Those alone are impressive performance increases, so why do so many seem to think we're in for some revolutionary new hardware philosophy or something here?

The M1 is ARM, the A12Z was ARM... it's all ARM. The M1 isn't quantum or something.

Have to imagine some people will be disappointed.

I bought my M1 iPad Pro yesterday because I always buy the new one hah, but lets be real about what we're getting here.

 

[UBS28]

Because an iPhone 12 Pro Max is 50% faster than a 2020 iPad Pro in single core test according to benchmarks. So give the A14 the same number of cores and there you have your 50% overall faster CPU.

The M1 is the A14X so it is basically 2 generations newer than the A12Z.

 

[Falhófnir]

Via geekbench, the A12Z has about a 1,115 single core score, around 4,650 multi core, the M1 has a single core score of 1,700 in the fanless MacBook Air, and a multi core score of 7375. So depending on what you're looking at, it's about a 1.5x or 1.6x (i.e. 50-60%) improvement.

 

[tann]

Via geekbench, the A12Z has about a 1,115 single core score, around 4,650 multi core, the M1 has a single core score of 1,700 in the fanless MacBook Air, and a multi core score of 7375. So depending on what you're looking at, it's about a 1.5x or 1.6x (i.e. 50-60%) improvement.

The iPad will likely come in lower than the Mac though due to power/thermal limitations. I reckon about 1500 / 6500.

But either way a big jump from the existing iPad, but again, don’t think we can just assume it’ll be the same scores as the Macs.

 

[doboy]

The iPad will likely come in lower than the Mac though due to power/thermal limitations. I reckon about 1500 / 6500.

But either way a big jump from the existing iPad, but again, don’t think we can just assume it’ll be the same scores as the Macs.

You can reckon all you want, but we won't know for sure until it lands.

 

[AutomaticApple]

You can reckon all you want, but we won't know for sure until it lands.

May 21 is coming! Time to see if the OP is right or not...

 

[rui no onna]

The iPad will likely come in lower than the Mac though due to power/thermal limitations. I reckon about 1500 / 6500.

But either way a big jump from the existing iPad, but again, don’t think we can just assume it’ll be the same scores as the Macs.

The Air 4 with A14 is closer to 1600 than 1500.

https://browser.geekbench.com/v5/cpu/search?page=1&q=ipad+air+4th&utf8=✓

The iPad Pro with M1 won't drop below that. The 12.9 might even bench better due to potentially better thermals with the bigger chassis.

 

[isoft7]

Via geekbench, the A12Z has about a 1,115 single core score, around 4,650 multi core, the M1 has a single core score of 1,700 in the fanless MacBook Air, and a multi core score of 7375. So depending on what you're looking at, it's about a 1.5x or 1.6x (i.e. 50-60%) improvement.

Geekbench is a synthetic benchmark and not in any way indicative or fully representative of real world performance.

Thats my point in my post, sure, MaxTech on youtube can go bonkers all he wants over geekbench scores, but when it comes to rendering your 4K video? 15%, maybe 20%.

We will all see soon enough.

 

[Gandek]

Geekbench is a synthetic benchmark and not in any way indicative or fully representative of real world performance.

Thats my point in my post, sure, MaxTech on youtube can go bonkers all he wants over geekbench scores, but when it comes to rendering your 4K video? 15%, maybe 20%.

We will all see soon enough.

Well any single threaded basic tasks like opening apps are faster on the A14 vs A12Z while the 12 pulls ahead In tasks like video rendering. That’s the closest real world usage you get on a mobile device

 

[isoft7]

Well any single threaded basic tasks like opening apps are faster on the A14 vs A12Z while the 12 pulls ahead In tasks like video rendering. That’s the closest real world usage you get on a mobile device

Right, single thread is improved between the various models, no doubt and that contributes some to the performance people experience, even when involving just the launching of applications.

One of the complex issues between those types of benchmarks demonstration on youtube videos with A12 vs A14 type products and how "quickly" they launch applications is the variances that aren't being accounted for.

These differences include the type of storage, be it Toshiba KICM, or Kioxia TC58, etc. Not to mention the amount of data being stored on the devices any the time of benchmarking. This even gets down to the process of installation of tested applications, the order in which they are installed, or even how many times they're accessed throughout the life of the product and the time of the benchmark. It's so complicated and so misinterpreted.

 

[Digitalguy]

there is absolutely no reason why the M1 should be only 15-20% faster in rendering while in geekbench is 50%, it has exactly the same TDP as the A12X/Z and the same number of cores.... so if the A12X did not throttle on sustained load this won't either... so if this is 50-60% faster in single core it will be in multi-core too... as simple as that

 

[Falhófnir]

Geekbench is a synthetic benchmark and not in any way indicative or fully representative of real world performance.

Thats my point in my post, sure, MaxTech on youtube can go bonkers all he wants over geekbench scores, but when it comes to rendering your 4K video? 15%, maybe 20%.

We will all see soon enough.

It's the simplest way of comparing relative performance. It tells us roughly speaking, for every 100 units of performance the A12Z has, the M1 has about 150. You're right that doesn't necessarily mean a video that will render in 60 minutes on an A12Z will render in 40 minutes on an M1, as there are other factors that come into play, but it is still useful as a guide to how chips are progressing in power from generation to generation.

Particularly in single core performance, Apple keeping up a 20-25% generation to generation improvement means that over the span of 5 years the horizon of what's possible on a tablet or small laptop is still shifting dramatically, even if its no longer true from year to year.

This reminds me a bit of what people were saying before Apple Silicon - you can't compare iPad and Mac scores, it works differently, it's because it's only running iOS. Lo and behold, as it turns out, the benchmarks did what they're designed to do and gave an indication of relative performance between Intel and Apple's chips, even using different platforms.

 

[gnomeisland]

Apple themselves said 50% Cpu Increase. They’ve been pretty conservative so far with M1 claims so given that the math checks out (comparing the A12 to M1) I’m confident we’ll see ~50% increase regardless of thermal throttling.

the increased RAM and faster SSD should also have an outsized influence on real-world usage that might not show up in benchmarks.

 

[UltimateSyn]

Apple touts 50% CPU speed increase and 40% GPU speed increase over the A12Z. I'm sure that's fairly accurate.

 

[TheRealAlex]

Why are ridiculous numbers like this tossed around here and on other forums? Even on youtube... insane numbers being projected.

The M1 is going to push big benchmarks and all, but when it comes to "rendering" times (3D work, photo filters, video compression, etc.) it's going to be around 15% faster than the A12Z. Maybe 20%.

Those alone are impressive performance increases, so why do so many seem to think we're in for some revolutionary new hardware philosophy or something here?

The M1 is ARM, the A12Z was ARM... it's all ARM. The M1 isn't quantum or something.

Have to imagine some people will be disappointed.

I bought my M1 iPad Pro yesterday because I always buy the new one hah, but lets be real about what we're getting here.

It’s gonna be an iPad Air like performance period. But the Air has a 7Core GPU the iPad Pro XDR will have the 8 Core GPU and that’s proof to me of binning. Where in CPU manufacturing some wafers will not come out as good as others so instead of ruining a whole wafer of CPUs they offer a lower spec product like the 8 Core 7Core GPU MacBook Air. Now for the XDR iPad Pro, Apple is in my guess using the absolute BEST binned M1 CPUs but the star of the show isn’t its performance. It’s the XDR Display.

 

[doboy]

the increased RAM and faster SSD should also have an outsized influence on real-world usage that might not show up in benchmarks.

Is the SSD different for 2021 models?

 

[Tsepz]

Why are ridiculous numbers like this tossed around here and on other forums? Even on youtube... insane numbers being projected.

The M1 is going to push big benchmarks and all, but when it comes to "rendering" times (3D work, photo filters, video compression, etc.) it's going to be around 15% faster than the A12Z. Maybe 20%.

Those alone are impressive performance increases, so why do so many seem to think we're in for some revolutionary new hardware philosophy or something here?

The M1 is ARM, the A12Z was ARM... it's all ARM. The M1 isn't quantum or something.

Have to imagine some people will be disappointed.

I bought my M1 iPad Pro yesterday because I always buy the new one hah, but lets be real about what we're getting here.

The only thing we truly know is that the new iPad Pro much like previous iPad Pros will be more powerful than necessary and will be futureproof AF! That alone is awesome.

 

[Traverse]

Well, the M1 is two generations above the A12Z, but I think most of the day-to-day improvements are just Apple marketing. Based on Apple‘s claims, compared to the A10X I currently use, the M1-based iPads should have:

• 2x the CPU
• 3x the GPU
• More than 2x the storage speed

On paper, it should be night and day, but I doubt I’ll see much difference in casual use. Maybe the SSD speed will help some apps.

 

[rui no onna]

Well, the M1 is two generations above the A12Z, but I think most of the day-to-day improvements are just Apple marketing. Based on Apple‘s claims, compared to the A10X I currently use, the M1-based iPads should have:

• 2x the CPU
• 3x the GPU
• More than 2x the storage speed

On paper, it should be night and day, but I doubt I’ll see much difference in casual use. Maybe the SSD speed will help some apps.

GPU will probably help on pages with tons of animations. I know the Apple product pages often stutter on my A10X.

 

[Digitalguy]

GPU will probably help on pages with tons of animations. I know the Apple product pages often stutter on my A10X.

the GPU in the A10x is actually more powerful than that in the A12 (iPad 8, mini, air 3).
I have to say that in my experience the A12 outperforms the A10X every time in browsing, even complex stuff

 

[rui no onna]

the GPU in the A10x is actually more powerful than that in the A12 (iPad 8, mini, air 3).
I have to say that in my experience the A12 outperforms the A10X every time in browsing, even complex stuff

In GeekBench Metal perhaps.

On 3DMark Wild Life, my Air 3 (6601, 39 ave fps, 25-55 fps) scored higher than the Pro 10.5 (5237, 31 ave fps, 21-42 fps).

 

[Jimbo Limbo]

After getting a 2018 iPad Pro 12.9 1TB LTE for my kid to use while studying at college a couple of years back, I have been wanting one really badly.

I ordered the new M1 12.9 2TB with LTE, and am looking forward to seeing how fast it is in comparison to my kid's 2018 iPad Pro. It will be quite a contrast with my iPhone 7.

 

[rui no onna]

After getting a 2018 iPad Pro 12.9 1TB LTE for my kid to use while studying at college a couple of years back, I have been wanting one really badly.

I ordered the new M1 12.9 2TB with LTE, and am looking forward to seeing how fast it is in comparison to my kid's 2018 iPad Pro. It will be quite a contrast with my iPhone 7.

To be perfectly honest, you probably won't be able to tell the difference for general use.

 

[gnomeisland]

Is the SSD different for 2021 models?

Apple claims twice the speed which seems reasonable if they are keeping the same NAND chips used on my M1 Mac mini. Not only is the M1 SSD fast it hits amazing speeds with low threads and small file sizes which (IME) is a big reason why the M1 machines feel so snappy and need for more swap isn't as noticeable.

EDIT: I went back and checked, "up to 2X" the SSD speed. Still should be a big jump on top of GPU, CPU, and everything else.

 

[Digitalguy]

In GeekBench Metal perhaps.

On 3DMark Wild Life, my Air 3 (6601, 39 ave fps, 25-55 fps) scored higher than the Pro 10.5 (5237, 31 ave fps, 21-42 fps).

yeah geekbench, for some reason...