ARM based MacBook Pro discussion [merged]

[raknor]

Then please also show me definitive video proof that Geekbench scores show actual performance difference in all cases.

If you cannot, then please stop trying to ask me to provide evidence.

How about I just show you how full of crap your assertion is instead?

This is a comparison of the Surface Pro X vs Surface Pro 7. The Intel chip here is the Icelake core i5 which should out perform the comet lake you are BSing about.

The reviewer is super biased against the Pro X even then you can see not only does the Surface Pro X not show any obvious performance issues in comparison, the Intel based surface pro 7 in fact shows random pauses and glitches instead.

EDIT: adding more evidence.

Windows 10 runs like dog turd even on Intel Core i7 based tablet. You can see this reviewer bitch about it.

 

[bill-p]

How about I just show you how full of crap your assertion is instead?

This is a comparison of the Surface Pro X vs Surface Pro 7. The Intel chip here is the Icelake core i5 which should out perform the comet lake you are BSing about.

The reviewer is super biased against the Pro X even then you can see not only does the Surface Pro X not show any obvious performance issues in comparison, the Intel based surface pro 7 in fact shows random pauses and glitches instead.

I have been polite, but okay, let's play that game. You're saying the Surface Pro X has comparable performance to the Surface Pro 7?

This review says otherwise:

Surface Pro X vs. Pro 7: ARM needs some legs

A deep dive into the Surface Pro X experience.

www.theverge.com

Elsewhere, I’ve felt like the Pro X performance has been a little erratic. Occasionally, I’d resume from standby and switch between apps, and things would take a solid minute to settle and not feel laggy. Discord isn’t exactly the most highly optimized app for Windows, but it struggles at times on the Pro X. I’ve never experienced erratic performance on the Pro 7 like this. Likewise, Spotify can be painful to use initially until it settles down and stops pegging the CPU on the Pro X.

Here's another one:

Surface Pro X Review: The Surface I really tried to love, but ended up hating

In this article Toggle TL;DR Very good design USB-C is great (depending on how you look at it) The Microsoft SQ1 Processor powering the Surface Pro X isn’t what Microsoft says it is The battery life is just alright This isn’t a Surface for gaming The Type Cover is decent, at the right angle The […]

www.onmsft.com

I’ll start the technical side of things by first by mentioning some of the Geekbench testing. I tested the Surface Pro X against my Surface Pro 2017, which is powered by the Seventh Gen Intel Core i5-7300U processor. That’s one full generation behind what Microsoft has tested, but I’ve also linked to other reviewers’ Surface Pro 6 and Surface Pro 7 benchmarks too. Typically, I also try to test my laptops with 3D Mark to test the GPU performance, but that benchmark failed to run on the Surface Pro X.

The best case of all this is how I used my Surface Pro X in my daily life. I typically use both the Chromium-Edge browser (I call it Credge) as well as the 32-bit versions of Slack, and Microsoft Teams. Again, I’m not a graphic designer or engineer. I’m using some very basic lightweight programs.

So, much to my disappointment, opening a few tabs on the 32-bit version of Credge, and running the emulated version of the 32-bit Teams app, the Surface Pro X felt a bit slow. Slack was also slow to update when I clicked to it. Much to my sadness, Teams was constantly reloading when I clicked it to the foreground in front of Credge.

Typing in WordPress was also slow at times, although closing out tabs sometimes speeded it up. There even was a situation where the Surface Pro X froze on me in the middle of typing an article. But, if you have a look at the task manager image below, you can see the victims here. All these 32-bit apps were eating up 34% of the SQ1 processor and nearly 70% of the RAM.

Please feel free to write to those people and ask them to provide video proof of their assertion as well. Actually, I'll stop claiming things as you suggested. I don't need to claim anything. Others have already said enough on that topic.

 

[raknor]

I have been polite, but okay, let's play that game. You're saying the Surface Pro X has comparable performance to the Surface Pro 7?

Please feel free to write to those people and ask them to provide video proof of their assertion as well. Actually, I'll stop claiming things as you suggested. I don't need to claim anything. Others have already said enough on that topic.

I already posted video proof handily disproving your claims. Windows 10 is an abysmal performer on a tablet whether it runs ARM or The latest x86. No one sane can draw conclusions about CPU performance in this case, yet you seem to be and that’s dubious at best.

even your own article proves you wrong.

• Surface Pro X (Microsoft SQ1): 3,524 Single Core/ 11,440 Multi-Core Score
• Surface Pro 7 (Core i5-1035G4): 5,342 Single Core/ /16,859 Multi-Core Score
• Surface Pro 6 (Core i5-8250U): 4,074 Single Core/ 13,567 Multi-Core Score
• Surface Pro 5 (Core i5-7300U): 4,242 Single Core/ 8,298 Multi-Core Score

If you are claiming GeekBench doesn’t imply better performance. Yet you highlighted all the CPUs that have higher single core scores than the Surface Pro X as better performers according to the reviews. So GeekBench seems to be quite indicative of real world performance , the slowest system on Geekbench performed worse. Again handily disproving your claim. It is no surprise the i5-7300u performed better, it scores higher in GeekBench single core than the SQ1..

 

[bill-p]

I already posted video proof handily disproving your claims. Windows 10 is an abysmal performer on a tablet whether it runs ARM or The latest x86. No one sane can draw conclusions about CPU performance in this case, yet you seem to be and that’s dubious at best.

even your own article proves you wrong.

• Surface Pro X (Microsoft SQ1): 3,524 Single Core/ 11,440 Multi-Core Score
• Surface Pro 7 (Core i5-1035G4): 5,342 Single Core/ /16,859 Multi-Core Score
• Surface Pro 6 (Core i5-8250U): 4,074 Single Core/ 13,567 Multi-Core Score
• Surface Pro 5 (Core i5-7300U): 4,242 Single Core/ 8,298 Multi-Core Score

If you are claiming GeekBench doesn’t imply better performance. Yet you highlighted all the CPUs that have higher single core scores than the Surface Pro X as better performers according to the reviews. So GeekBench seems to be quite indicative of real world performance , the slowest system on Geekbench performed worse. Again handily disproving your claim. It is no surprise the i5-7300u performed better, it scores higher in GeekBench single core than the SQ1..

i5 7300U Geekbench 5 scores:

Intel Core i5-7300U Benchmarks - Geekbench

browser.geekbench.com

Microsoft SQ1 Geekbench 5 scores:

Microsoft Corporation Surface Pro X - Geekbench

Benchmark results for a Microsoft Corporation Surface Pro X with a Microsoft SQ1 processor.

browser.geekbench.com

I rest my case.

 

[raknor]

i5 7300U Geekbench 5 scores:

Intel Core i5-7300U Benchmarks - Geekbench

browser.geekbench.com

Microsoft SQ1 Geekbench 5 scores:

Microsoft Corporation Surface Pro X - Geekbench

Benchmark results for a Microsoft Corporation Surface Pro X with a Microsoft SQ1 processor.

browser.geekbench.com

I rest my case.

That’s premature and ill adviced. Same Geekbench version 5.2.5.

i5-7300u
922 Single Core

Microsoft Corporation Surface Book 2 - Geekbench

Benchmark results for a Microsoft Corporation Surface Book 2 with an Intel Core i5-7300U processor.

browser.geekbench.com

One more from a Surface Pro showing 921 single core.

Microsoft SQ1 Geekbench 5 scores:
754 Single Core

 

[bill-p]

That’s premature and ill adviced. Same Geekbench version 5.2.5.

i5-7300u
922 Single Core

Microsoft Corporation Surface Book 2 - Geekbench

Benchmark results for a Microsoft Corporation Surface Book 2 with an Intel Core i5-7300U processor.

browser.geekbench.com

One more from a Surface Pro showing 921 single core.

Microsoft SQ1 Geekbench 5 scores:
754 Single Core

SQ1 multi-core:
2960

i5 7300U multi-core:
2072

And even if I were to take your premise that Geekbench does indeed show real world performance, then here are the scores of the i5 10210U:
981 Single Core
3067 Multi Core

Intel Core i5-10210U Benchmarks - Geekbench

browser.geekbench.com

So either Geekbench does not show real world performance, or my original claim is substantiated by Geekbench scores. We can go at this all day.

 

[raknor]

SQ1 multi-core:
2960

i5 7300U multi-core:
2072

And even if I were to take your premise that Geekbench does indeed show real world performance, then here are the scores of the i5 10210U:
981 Single Core
3067 Multi Core

Intel Core i5-10210U Benchmarks - Geekbench

browser.geekbench.com

So either Geekbench does not show real world performance, or my original claim is substantiated by Geekbench scores. We can go at this all day.

Not unless you live in some alternate universe.

The Surface Pro X scores very similar to Intel i5 10210U but the latter runs circles around Surface Pro X in Windows 10, at literally everything. In fact, the Surface Pro X is a classic example of how synthetic benchmarks cannot be trusted to represent real world usage.

SQ1 single-core: 754
SQ1 multi-core: 2960


i5 10210U:
981 Single Core
3067 Multi Core

You are just about wrong on everything so far. You claim the i5 is faster (without any evidence to determine workoads) even though it has similar numbers on GeekBench, any one with half a brain can see a big difference in single core performance between the SQ1 and i5. Either you don’t know what your are talking about or deliberately trying to mislead by just quoting the multi core numbers. I’ll go with the former since you wouldn’t be deliberately trying to mislead, would you?

So yes Geekbench does show real world performance, the chip you claim shows better performance in real world tasks also scores higher on GeekBench.

 

[Maximara]

I have been polite, but okay, let's play that game. You're saying the Surface Pro X has comparable performance to the Surface Pro 7?

These examples have the same issue - Microsoft's 32-bit translator (Microsoft would have be total idiots to use an emulator) blows goats. Given Apple killed 32-bit code just before the ARM Mac makes me suspect 32-bit translation by its very nature is more of a PITA then it is worth.

I will point out that any reviewer that uses "emulator" rather then "translator" clearly doesn't know the important difference raising the question of what else they are doing wrong. QEMU is an emulator while both versions of Rosetta and whatever name Microsoft assigns to that part of Windows 10 are translators.

 

[nquinn]

A14x geekbench leak:

A14X Bionic allegedly benchmarked days before Apple Silicon Mac event | AppleInsider

The chip expected to be at the core of the first Apple Silicon Mac — the "A14X" — may have been benchmarked just days before the next Apple event.

appleinsider.com

Looks similar to an i9 macbook pro 16"!

2019 16" MBP i9 (8-core, 2.4ghz) geekbench scores:
- 1096 single / 6870 multi

A14x leak scores:
- 1634 single (+49%) / 7220 multi (+5%)

If they throw that A14x in a 13" model it will be faster than last year's 16" ! Would be nice to see more multi-core performance, but not a huge deal for most users. Obviously the other wildcard is what they can do on the iGPU side - that's the biggest risk for a new 16" model it seems. John Siracusa from ATP felt that Apple would be capable of just ramping up their existing graphics, but who knows.

 

[yanksrock100]

Are we expecting there to be multiple variants of the A14x? I don't want the Mac to be like the iPad line - all iPad Pro models have the same processor and only vary in memory size.

 

[Maximara]

Are we expecting there to be multiple variants of the A14x? I don't want the Mac to be like the iPad line - all iPad Pro models have the same processor and only vary in memory size.

It would be strange for Apple to do that on the laptop/desktop line. Logic indicates variations in the Ghz just like we see with the Intel chips.

 

[Unregistered 4U]

Are we expecting there to be multiple variants of the A14x? I don't want the Mac to be like the iPad line - all iPad Pro models have the same processor and only vary in memory size.

Possibly. I’m still guessing that it’s the same chip for both the MacBook and the MacBook Pro, like the iPad. The difference becomes what features you want on top of that. (Perhaps the Pro is OLED?)

Think of it like this. If they’ve produced only one variant of the 14x at a speed in a MacBook Pro 13 that would challenge an i9, would you want that same processor in the MacBook 12 or something weaker?

 

[ascender]

Are we expecting there to be multiple variants of the A14x? I don't want the Mac to be like the iPad line - all iPad Pro models have the same processor and only vary in memory size.

Definitely.... Tuesday should give some clarity around their plans, but I expect there will be different chips or versions of chips for the different lines - presumably one for the MacBook Air and lower Pro, then one with more cores for the higher performance Pros and a desktop one etc etc.

A14x, y, z etc?

 

[PharmlyDoc]

I am going to be utterly disappointed if the next MacBook Pro 16-inch is ARM. And if Apple is going to stoop so low as to compare their ARM chips to an Intel Core i9 9th generation, then everyone needs to broadcast that out to consumers and publicly shame the company. Apple is doing themselves and their customers a disservice by not producing a MBP 16-inch with an Intel Core i9 11th generation (Tiger-Lake) H-series chip, WiFi-6, and AMD Radeon Pro 5600M with 8GB of HBM2 memory - I don't care if I have to wait until early 2021 for Tiger Lake-H to come out.

 

[PharmlyDoc]

The iPhone already matches the fastest of Tiger Lake and Zen 3 in single core. Add two more high performance cores and the A14 will be running circles around Tiger Lake - even at the same 3.0ghz frequency the phones use.

Fastest of Tiger Lake? The fastest of Tiger Lake is an Intel Core i9 Tiger Lake H-series, and has yet to be released.

 

[leman]

Fastest of Tiger Lake? The fastest of Tiger Lake is an Intel Core i9 Tiger Lake H-series, and has yet to be released.

I am only talking about existing CPUs. What will maximal clocks be on Tiger Lake H anyway? 5.1ghz at 40-50 watts power consumption? It won’t make any difference.

 

[PharmlyDoc]

I am only talking about existing CPUs. What will maximal clocks be on Tiger Lake H anyway? 5.1ghz at 40-50 watts power consumption? It won’t make any difference.

I don't doubt that the difference will be trivial or that down the line Apple ARM will blast the best Intel Tiger-Lake out of the water. I just want one last Intel-based MBP-16 inch with the best Intel has to offer, I use a lot of different software applications, including Parallels and Windows 10 to run certain software restricted to Windows, and I trust that Intel-based MBP can run these applications.

 

[leman]

I don't doubt that the difference will be trivial or that down the line Apple ARM will blast the best Intel Tiger-Lake out of the water. I just want one last Intel-based MBP-16 inch with the best Intel has to offer, I use a lot of different software applications, including Parallels and Windows 10 to run certain software restricted to Windows, and I trust that Intel-based MBP can run these applications.

H-series 10nm chips are not coming out before spring most likely, and I have my doubts whether Intel will be able to ship enough volume early on to fulfill Apple’s need. And by the end of summer I’d expect 16” ARM MBP to be out. Frankly, I’d be surprised if Tiger Lake ever made it to the 16”.

 

[Unregistered 4U]

H-series 10nm chips are not coming out before spring most likely, and I have my doubts whether Intel will be able to ship enough volume early on to fulfill Apple’s need. And by the end of summer I’d expect 16” ARM MBP to be out. Frankly, I’d be surprised if Tiger Lake ever made it to the 16”.

Yeah, it’s not like we don’t already have a track record of Intel promising to release high performance mobile processors and then missing those expectations.

 

[Maximara]

I am going to be utterly disappointed if the next MacBook Pro 16-inch is ARM. And if Apple is going to stoop so low as to compare their ARM chips to an Intel Core i9 9th generation, then everyone needs to broadcast that out to consumers and publicly shame the company. Apple is doing themselves and their customers a disservice by not producing a MBP 16-inch with an Intel Core i9 11th generation (Tiger-Lake) H-series chip, WiFi-6, and AMD Radeon Pro 5600M with 8GB of HBM2 memory - I don't care if I have to wait until early 2021 for Tiger Lake-H to come out.

Apple has had enough of Intel messing up and there is no promise H-series chip will be out in 2021.

I don't doubt that the difference will be trivial or that down the line Apple ARM will blast the best Intel Tiger-Lake out of the water. I just want one last Intel-based MBP-16 inch with the best Intel has to offer, I use a lot of different software applications, including Parallels and Windows 10 to run certain software restricted to Windows, and I trust that Intel-based MBP can run these applications.

If you are running that many Windows programs why on Earth would you get a Mac?! It's a non sequitur. Never mind it is clear Microsoft wants to go to ARM in the worst way possible. While I doubt it is the case, it would be hysterical if the Bootcamp code seen in the Big Sur wasn't for the x86 version of Windows but the ARM version of Windows.

I strongly suspect Parallels will have a version that translates x86 Windows code to AS code.

 

[ascender]

I am going to be utterly disappointed if the next MacBook Pro 16-inch is ARM. And if Apple is going to stoop so low as to compare their ARM chips to an Intel Core i9 9th generation, then everyone needs to broadcast that out to consumers and publicly shame the company. Apple is doing themselves and their customers a disservice by not producing a MBP 16-inch with an Intel Core i9 11th generation (Tiger-Lake) H-series chip, WiFi-6, and AMD Radeon Pro 5600M with 8GB of HBM2 memory - I don't care if I have to wait until early 2021 for Tiger Lake-H to come out.

How would Apple be doing their customers a disservice by not using Intel if their own chips are faster, more efficient etc?

 

[Maximara]

How would Apple be doing their customers a disservice by not using Intel if their own chips are faster, more efficient etc?

I don't understand the logic ( ) either. From what we saw Rosetta 2 does an excellent job of turning x86 code into something the AS can use. The only thing Intel in a Mac has going for it is the ability to boot into x86 Windows but really unless you have to do that why would you want to? Cloud bases VM seem to be becoming SOP so the whole need to boot doesn't really make sense in the modern world.

 

[waquzy]

I’m expecting Apple Silicon to dominate the industry and be the new benchmark... anyhow, all will be revealed tomorrow.

 

[darkmatter343]

What's the likelihood of Apple having the AS Mac's purchasable right after the announcement? Haven't most Macbook hardware refreshes from Apple been available same day after announcement?

 

[nquinn]

Are we expecting there to be multiple variants of the A14x? I don't want the Mac to be like the iPad line - all iPad Pro models have the same processor and only vary in memory size.

I think we will see maybe 3 variants, but at different clockspeeds:

- A14x low power: clocked lower for a macbook air / fanless, and clocked higher for a macbook pro 13". Apple TV likely will use a cheaper/older chipset like an A12x or A13?
- A14x high power: higher power, bigger igpu for macbook pro 16". Probably also used for iMac and Mac Mini. Probably very similar to the one above.
- Mac Pro: Probably a whole different beast