Is It Me, Or Is The New iPad Pro More Powerful Than The New Base Model MacBook Air?

[TiresomeToe933]

Look, I'm no hardware engineer or software optimisation engineer.... However, there is something that I have come questioning of.

 

Is it me, or is the new iPad Pro more powerful than the base model MacBook Air?

 

Lets think about this.

 

1. The iPad Pro has hella more cores, 8 CPU cores VS 2 on the Air. Now sure, you could make the possible statement that the Air's cores are higher clocked or has more cache, so yeah, it will win in single-threaded performances.... But what about multi-core...... Surely the iPad Pro should tower over that little 2 core CPU
2. The iPad Pro has a dedicated GPU VS the Air which has "integrated graphics". Now I don't wanna make any comparisons on how the interred graphic is better.... because 1. I don't have the time to write up how Integrated Graphics works and 2. Even at the end of the day... a dedicated GPU is better, even if its on a SOC in this case.
3. The iPad Pro has a Neural Network Engine Chip, this is different from the T2 chip. The NNEC (as I'll call it now) can do up to Quote: "1 Trillion Calculations" and whilst the NNEC is for more specific purposes, that chip can help in certain compute tasks
4. The iPad Pro is possibly/I think is more optimised. Think about it, the iPad Pro SOC is designed by Apple engineers so they can have a better optimisation with the OS. MacBook Air's on the other hand use Intel chips, meaning they cannot optimise the chip as much.

Like I said, I'm no hardware engineer or software optimisation engineer, but I can definitely tell based of my current knowledge of CPU Clock Speed/Cores ratio that something is just off...

 

I don't know, what do you guys think.

[Sauron]

We can't know for sure since there are no reliable benchmarks and you can't just guess it from the specs alone. I don't think the performance would be too far off though. Unfortunately, this isn't because the iPad is fast - it's because the mba is trash in terms of performance, at least by the standards of a 1200$+ ultrabook.

[Guest]

Dang it! I'm still on an iPad 2! I definitely need some more power. Really though, apple does make the best mobile processors. If you check out the Verge video, they say that you should be comparing the new iPad Pro to laptops, not other tablets. It's way to far ahead for that.

[Mira Yurizaki]

2 hours ago, TiresomeToe933 said:

The iPad Pro has hella more cores, 8 CPU cores VS 2 on the Air. Now sure, you could make the possible statement that the Air's cores are higher clocked or has more cache, so yeah, it will win in single-threaded performances.... But what about multi-core...... Surely the iPad Pro should tower over that little 2 core CPU

Core count doesn't really mean anything if the single core performance is lacking enough. Take for instance the PS4 and Xbox One. They both had 8-core CPUs. Even if you took say two away for the OS (which I'm not really certain of considering how the homebrew market hacked the PS4), you could still have dual-core PCs of the time keep up with it in gaming performance.

2 hours ago, TiresomeToe933 said:

The iPad Pro has a dedicated GPU VS the Air which has "integrated graphics". Now I don't wanna make any comparisons on how the interred graphic is better.... because 1. I don't have the time to write up how Integrated Graphics works and 2. Even at the end of the day... a dedicated GPU is better, even if its on a SOC in this case.

It doesn't matter if the GPU is dedicated or not. All that matters is how it actually performs. Besides that, the GPU in the A12X is technically "integrated" as much as the Intel HD is on the MBA.

2 hours ago, TiresomeToe933 said:

The iPad Pro has a Neural Network Engine Chip, this is different from the T2 chip. The NNEC (as I'll call it now) can do up to Quote: "1 Trillion Calculations" and whilst the NNEC is for more specific purposes, that chip can help in certain compute tasks

It could help, but it'll only be useful in certain situations. Also the 1 trillion calculations (I'm presuming per second) number only means that. It's likely doing matrix math on tiny numbers in the same way NVIDIA can claim Turing has 114 TFLOPS of compute power... on FP16 values in 4x4 matrices meaning that per tensor core operation they're really doing 16 operations at once, at least.

 

2 hours ago, TiresomeToe933 said:

The iPad Pro is possibly/I think is more optimised. Think about it, the iPad Pro SOC is designed by Apple engineers so they can have a better optimisation with the OS. MacBook Air's on the other hand use Intel chips, meaning they cannot optimise the chip as much.

That's about the only advantage they have making their own SoC, they know how their own hardware works. I'm sure Intel's holding back some secret sauces from Apple, but at the same time, there's only a handful of MBA configurations, so it's not like they can't highly tune the hardware and software.

 

2 hours ago, TiresomeToe933 said:

Like I said, I'm no hardware engineer or software optimisation engineer, but I can definitely tell based of my current knowledge of CPU Clock Speed/Cores ratio that something is just off...

The major problems I have with comparing the two are:

• One's running a full desktop OS, the other is running a tablet OS. Even if the differences in resource usage are minimal as possible, I still can't find this to be a fair comparison
• The only benchmark I've seen that's comparing the two is GeekBench. However I'm inclined to believe GeekBench is heavily weighted towards the iOS devices since there are benchmarks where a 2.3GHz Apple SoC is able to keep up with 3.1GHz Intel chips in single core tests. So unless Intel's IPC is really that bad, I can't really take this with any more than a grain of salt. And there hasn't been any other cross platform benchmarks that I'm aware of.

[jonrosalia]

the iPad Pro needs to replace the AIR at this point with a hard keyboard dock like the surface book and a switcher that allowed Mojave when docked and ran IOS when it was in tablet move would sink surface and be one of the best options in the market. the new AIR I would like to see it stacked up to a 2012 13" MBP with a ssd and a few others newer models see if its worthy of existing at all most likely the answer is no.

[Mira Yurizaki]

1 minute ago, jonrosalia said:

the iPad Pro needs to replace the AIR at this point with a hard keyboard dock like the surface book and a switcher that allowed Mojave when docked and ran IOS when it was in tablet move would sink surface and be one of the best options in the market. the new AIR I would like to see it stacked up to a 2012 13" MBP with a ssd and a few others newer models see if its worthy of existing at all most likely the answer is no.

Having two separate OSes wouldn't really make it useful. The transition from tablet to laptop mode needs to be seamless down to basically "absolutely nothing changed other than what inputs I'm using now."

[TiresomeToe933]

8 hours ago, M.Yurizaki said:

Core count doesn't really mean anything if the single core performance is lacking enough. Take for instance the PS4 and Xbox One. They both had 8-core CPUs. Even if you took say two away for the OS (which I'm not really certain of considering how the homebrew market hacked the PS4), you could still have dual-core PCs of the time keep up with it in gaming performance.

It doesn't matter if the GPU is dedicated or not. All that matters is how it actually performs. Besides that, the GPU in the A12X is technically "integrated" as much as the Intel HD is on the MBA.

It could help, but it'll only be useful in certain situations. Also the 1 trillion calculations (I'm presuming per second) number only means that. It's likely doing matrix math on tiny numbers in the same way NVIDIA can claim Turing has 114 TFLOPS of compute power... on FP16 values in 4x4 matrices meaning that per tensor core operation they're really doing 16 operations at once, at least.

 

That's about the only advantage they have making their own SoC, they know how their own hardware works. I'm sure Intel's holding back some secret sauces from Apple, but at the same time, there's only a handful of MBA configurations, so it's not like they can't highly tune the hardware and software.

 

The major problems I have with comparing the two are:

• One's running a full desktop OS, the other is running a tablet OS. Even if the differences in resource usage are minimal as possible, I still can't find this to be a fair comparison
• The only benchmark I've seen that's comparing the two is GeekBench. However I'm inclined to believe GeekBench is heavily weighted towards the iOS devices since there are benchmarks where a 2.3GHz Apple SoC is able to keep up with 3.1GHz Intel chips in single core tests. So unless Intel's IPC is really that bad, I can't really take this with any more than a grain of salt. And there hasn't been any other cross platform benchmarks that I'm aware of.

 

10 hours ago, PacketMan said:

Now compare the AMD FX 8350 (8 cores @ 4GHz) with the i5 8600K (6 cores @ 3.6GHz)

The numbers mean nothing, and those are different architectures

If the Air finally comes with the Vega graphics, it will be a beast, take a look at the AMD 2200G, a top seller, and it has integrated graphics

A dedicated GPU means nothing, compare the AMD 2200G integrated graphics to the GT 710

I call this bullshit, Apple is doing marketing with the Neural Network, same with Huawei with their AI CPU (I own a Mate 10, and if there is an AI, it's stupid)

Apple is the king of optimisation, I got a MacBook Pro 15 Mid '14 (iirc it had an I7-4980HQ, which is a beast) and the power delivered when needed, and the battery life was even a bigger beast (it lasted me almost two days with a correct use of it) so I don't think Apple will improve performance on the Intel CPU but to improve battery life

It's a lot more than just numbers, at first I didn't know neither and I was like let's go for a FX 8350 instead of the overpriced i7

 

10 hours ago, Sauron said:

We can't know for sure since there are no reliable benchmarks and you can't just guess it from the specs alone. I don't think the performance would be too far off though. Unfortunately, this isn't because the iPad is fast - it's because the mba is trash in terms of performance, at least by the standards of a 1200$+ ultrabook.

Now thinking about it, you guys are totally right and I forgot about architecture! Thanks guys, sorry if this question was stupid, I'm really bad at formulating questions... So yeah I guess that the MacBook Air is better. Thanks guys!

[jonrosalia]

49 minutes ago, M.Yurizaki said:

Having two separate OSes wouldn't really make it useful. The transition from tablet to laptop mode needs to be seamless down to basically "absolutely nothing changed other than what inputs I'm using now."

I guess im weird then I would like it to switch and basically the files stay in the same place but the way you access them  would change basically the optimization the two platforms offer rolled together.