iPad Pro A realistic glimpse into M1 iPad Pro vs A12Z iPad Pro performance...

[isoft7]

Wasting time clicking through youtube on performance data relating to the Apple M1 processor, I found an interesting video from several month ago comparing the performance data of the M1 in the Mac Air to the A12Z in the iPad Pro and it had me thinking... this is probably very close to what we can expect in the M1 iPad Pro, since it's more or less just an Air without a keyboard attached.

It's entirely possible the iPad Pro M1 might be clocked down some, but probably not by much.

Anyway, here is the relevant screenshot:

From of this video:

 

[Surne]

Wasting time clicking through youtube on performance data relating to the Apple M1 processor, I found an interesting video from several month ago comparing the performance data of the M1 in the Mac Air to the A12Z in the iPad Pro and it had me thinking... this is probably very close to what we can expect in the M1 iPad Pro, since it's more or less just an Air without a keyboard attached.

It's entirely possible the iPad Pro M1 might be clocked down some, but probably not by much.

Anyway, here is the relevant screenshot:

View attachment 1766575

From of this video:

The iPad Pro M1 probably won't be downclocked, but it will probably also thermal throttle quicker than a MBA in those instances of use where thermal throttling will occur.

 

[Tsepz]

That is an interesting one.
Also keep in mind the iPad Pro 12.9 has more pixels to push versus the MBA.

 

[one more]

M1 will certainly outpace A12Z on the spreadsheet. The question is - so what? I can’t think of any real life scenario where I would be sitting in front of an iPad Pro 2020 tweaking my fingers and saying: “I wish that snail could just move faster!”. As Apple executives recently mentioned, they put the M1 into an iPad to prove to everyone they could. It would be great if there are some notable iPadOS 15 advances to harvest that power, but I would not be betting on it. We will know soon enough.

 

[isoft7]

M1 will certainly outpace A12Z on the spreadsheet. The question is - so what? I can’t think of any real life scenario where I would be sitting in front of an iPad Pro 2020 tweaking my fingers and saying: “I wish that snail could just move faster!”. As Apple executives recently mentioned, they put the M1 into an iPad to prove to everyone they could. It would be great if there are some notable iPadOS 15 advances to harvest that power, but I would not be being on it. We will know soon enough.

Right, we will know soon enough.

It's been difficult to find solid output data from the various Apple processors beyond "Geekbench" which is a synthetic benchmark and not reflective of anything important, just like 3D Mark on a desktop computer, it's fun to see the numbers but it doesn't mean much.

What this comparison suggests is best case the M1 iPad Pro would render this video workflow roughly 14% faster than the previous generation iPad Pro.

If there is thermal throttling or reduced speed for battery concerns, then perhaps it'll be 10%?

It's an interesting argument, 10% is a healthy increase in performance for a new model, but the internet seems ripe with idea that we're going to see 50% to 70% increases in performance, and I'm just not to sure.

But like you said, we will know for sure soon enough.

 

[Ludatyk]

It's an interesting argument, 10% is a healthy increase in performance for a new model, but the internet seems ripe with idea that we're going to see 50% to 70% increases in performance, and I'm just not to sure.

I sure as you know what… don’t buy into we will see an incredible leap in performance. I think 10 to 20% increase in performance over previous generation seems like a reasonable guess. Even by Apple standards, they know that the performance gap isn’t as huge judging by their iPad Pro presentation.

On the slide show… they were comparing their statistics to the 1st generation iPad not the previous generation.

 

[rui no onna]

For me, it's not even the M1 itself but the other stuff (RAM, Thunderbolt and Qualcomm/5G). I've always thought Apple was way too stingy on RAM. Pretty much putting just enough for acceptable operation. 8GB is quite good for use now (with multitasking) while 16GB actually offers some headroom. It's not unusual for me to have 100+ Safari tabs open so I can probably benefit from 16GB right away.

Of course, I'm still using a 2017 model so for me, the jump is quite considerable. The 1TB 2018 Pro was $1749/1899 at launch which is a measly $50-100 less than the 1TB 2021 Pro at $1799/1999. I'm getting more bang for the upgrade buck now with the 2021 than I would've gotten with the 2018 Pro. ?‍♀️

 

[Ludatyk]

It's not unusual for me to have 100+ Safari tabs open so I can probably benefit from 16GB right away.

100+ Safari tabs open?? Really? I think at most… I use 8.

 

[LogicalApex]

Wasting time clicking through youtube on performance data relating to the Apple M1 processor, I found an interesting video from several month ago comparing the performance data of the M1 in the Mac Air to the A12Z in the iPad Pro and it had me thinking... this is probably very close to what we can expect in the M1 iPad Pro, since it's more or less just an Air without a keyboard attached.

It's entirely possible the iPad Pro M1 might be clocked down some, but probably not by much.

Anyway, here is the relevant screenshot:

View attachment 1766575

That has to be about the worst chart I've ever seen. Visually it looks like the iPhone 12 is the fastest device with the shortest line although it is the slowest. Then the bars don't align at all with the time it took the devices to run and instead measures maybe the timing difference between the devices? Really an all around mess.

I'd agree though we should see similar to MacBook Air performance depending on how they do the SOC.

 

[AutomaticApple]

100+ Safari tabs open?? Really? I think at most… I use 8.

Haha, same!

 

[isoft7]

For me, it's not even the M1 itself but the other stuff (RAM, Thunderbolt and Qualcomm/5G). I've always thought Apple was way too stingy on RAM. Pretty much putting just enough for acceptable operation. 8GB is quite good for use now (with multitasking) while 16GB actually offers some headroom. It's not unusual for me to have 100+ Safari tabs open so I can probably benefit from 16GB right away.

Of course, I'm still using a 2017 model so for me, the jump is quite considerable. The 1TB 2018 Pro was $1749/1899 at launch which is a measly $50-100 less than the 1TB 2021 Pro at $1799/1999. I'm getting more bang for the upgrade buck now with the 2021 than I would've gotten with the 2018 Pro. ?‍♀️

Yeah, the extra RAM is going to be nice, even if 90% of the apps available would never use it, it'll still be nice to have.

Your Safari tab usage is wild to me though lol, it's odd if I have even 4 or 5 tabs open, usually it's just 1. It's nice to know you can do that if you want though.

 

[rui no onna]

Haha, same!

Well, looks like I'm under 100 at the moment. ?

I tend to open threads and notifications in new tabs in one go and just close the tabs as I read them. I can have 20 tabs open for MacRumors forums alone. Same goes when I do a search (Google, shopping, etc). I just go through all the search results first and open everything that seems relevant in a new tab, then check each tab and close when I'm done.

 

[rui no onna]

It's an interesting argument, 10% is a healthy increase in performance for a new model, but the internet seems ripe with idea that we're going to see 50% to 70% increases in performance, and I'm just not to sure.

50-70% doesn't make sense. A12 to A14 is around 40% best case, I think. Realistically, the gains from A12X/Z to M1 would be around the same or lower.

 

[jmgregory1]

If I had to guess, I would say that the M1 powered iPad Pros are going to perform almost exactly like the keyboard-less M1 MacBook Airs that they are, albeit with a MUCH better screen. The added ram is certainly going to help with Safari, but it’s not going to be the same huge jump that we saw M1 provide over Intel versions of the Macs. The A12X/Z were already performing at levels above many Intel chipped computers, so the M1 iPP is going to be iterative in improvements (other than the screen which should show a noticeable improvement).

 

[Ludatyk]

I tend to open threads and notifications in new tabs in one go and just close the tabs as I read them. I can have 20 tabs open for MacRumors forums alone. Same goes when I do a search (Google, shopping, etc). I just go through all the search results first and open everything that seems relevant in a new tab, then check each tab and close when I'm done.

I use Tapatalk along side the website… just depends on where I’m at during the moment. But just by looking at your screenshot, not trying to critic your method… but why not have these toggles on? it appears you might have it off (maybe I’m wrong).

 

[rui no onna]

I use Tapatalk along side the website… just depends on where I’m at during the moment. But just by looking at your screenshot, not trying to critic your method… but why not have these toggles on? it appears you might have it off (maybe I’m wrong).

View attachment 1766694

I actually hate that I can't have multiple tabs open in Tapatalk so I never use it.

I intentionally disabled the tab bar for the screenshot for privacy. I normally have those options enabled and I've got Close Tabs set to 1 week.

 

[PBG4 Dude]

That has to be about the worst chart I've ever seen. Visually it looks like the iPhone 12 is the fastest device with the shortest line although it is the slowest. Then the bars don't align at all with the time it took the devices to run and instead measures maybe the timing difference between the devices? Really an all around mess.

I'd agree though we should see similar to MacBook Air performance depending on how they do the SOC.

Yeah, that graphic is all kinds of messed up. The timelines shown on the graph do not mesh with the times reported at all. The iPhone 12 line stops around the 4:00 mark yet no exports took this little amount of time. It makes no sense unless the 4:00 is supposed to represent something other than export time.

 

[Ludatyk]

I actually hate that I can't have multiple tabs open in Tapatalk so I never use it.

I intentionally disabled the tab bar for the screenshot for privacy. I normally have those options enabled and I've got Close Tabs set to 1 week.

Fair enough… I’m starting to think you revel in the chance to have 100+ Safari tabs open. Apple made the 16GB RAM iPad specifically for you.

 

[rui no onna]

Fair enough… I’m starting to think you revel in the chance to have 100+ Safari tabs open. Apple made the 16GB RAM iPad specifically for you.

I don't. They just seem to multiply like mushrooms.

 

[t0pher]

I don't. They just seem to multiply like mushrooms.

I do exactly the same, open a page like macrumors, and open in background a bunch of headlines, iOS starts to moan when you approach 500 tabs, every so often I used to go through closing individual tabs, I now just ditch the lot.

on the MacBook I noticed the fan would kicking frequently etc. I installed a tab counter & I had ~ 1400 tabs open, mainly for amazon as I'm looking for a few particular items.

I have 14 pinned tabs for IT stuff I run at home & every new window adds 15 to the count, I have 16 windows with the 14 pinned tabs plus others so in reality have 196 other tabs.

the biggest victim is WindowServer that shows 25-35% CPU (need to divide by 8 threads to get actual) when more than 300 windows are reported open

 

[Homme]

That has to be about the worst chart I've ever seen. Visually it looks like the iPhone 12 is the fastest device with the shortest line although it is the slowest. Then the bars don't align at all with the time it took the devices to run and instead measures maybe the timing difference between the devices? Really an all around mess.

I'd agree though we should see similar to MacBook Air performance depending on how they do the SOC.

but you forget that A12Z gives better multicore scores, 8 core CPU and also has 2GB of RAM more than 12 Mini. So no it’s not the worst chart I ( and others) have seen?

 

[rui no onna]

but you forget that A12Z gives better multicore scores, 8 core CPU and also has 2GB of RAM more than 12 Mini. So no it’s not the worst chart I ( and others) have seen?

The criticism is for the way the data is presented, not the data itself. The values on the X-axis don't match the posted results.

 

[isoft7]

Haha, yeah, the chart depictions are rather odd... but the numbers themselves are at least easy enough to understand. Not the best presentation.

 

[TheRealAlex]

MacBook Air is missing 1 GPU Core.

 

[isoft7]

MacBook Air is missing 1 GPU Core.

It depends on the configuration, base models are 7 core GPU, anything above that is 8.

However it doesn't appear LumaFusion utilizes GPU cores for video file export or compression, only for playback... one reason we can be confident with that opinion is render times when benchmarked on the 2018 iPad Pro (7 GPU cores) vs the 2020 iPad Pro (8 GPU cores) show absolutely no differences. So that extra core doesn't appear to make any difference.

That said, I imagine Luma could update and leverage the GPU cores for additional export performance. At the cost of heat and battery though.

I think what is shown in this videos is pretty decent... we will see Geekbench scores through the roof on the M1 compared to the A12Z, but we won't see those drastic increases when comparing functional productivity applications. It's no different on desktop computers... often times benchmarks will show unbelievable increases in performance that just don't fully realize on workloads. It's normal.