Solved! ARM Apple High-End CPU - Intel replacement

[Richie Rich]

There is a first rumor about Intel replacement in Apple products:

• ARM based high-end CPU
• 8 cores, no SMT
• IPC +30% over Cortex A77
• desktop performance (Core i7/Ryzen R7) with much lower power consumption
• introduction with new gen MacBook Air in mid 2020 (considering also MacBook PRO and iMac)
• massive AI accelerator

Source Coreteks:

 

[awesomedeluxe]

View attachment 26709

HBM2/E is still the overall lower power option.

There is also multiple speed bin options available; Speed bins: 0.5, 1.0, 1.5, 1.6, 1.8, 2.0, 2.4, 3.0, 3.2 Gbps

LPDDR5 is 6.4 Gb/s => (128 * 6.4 ) / 8 => 102.4 GB/s
HBM2E to do the above => ( 102.4 * 8 ) / 1024 => 0.8 Gb/s
Samsung's 110mm2 HBM2E 16GB archieves 5 Gbps @ 1.1V and SK Hynix's 110mm2 HBM2E 16GB achieves 4 Gbps @ 1.14V.

If Samsung/SK Hynix ever come out with a Low-Power 512-bit it only needs 1.6 Gb/s to beat LPDDR5. With the Low-Power & High-Bandwidth Memory probably being even lower power than HBM2/HBM2E is at it's higher speed.

Thanks for sharing this. I don't take all of JEDEC's info about their own memory at face value, but understand that lower speeds could be used for less power. Still, I recommend you take a look at this study, especially Section 5.3. In broad strokes, you'll see HBM2 using much more power than LPDDR4. As HBM2E uses more power than HBM2, and LPDDR5 uses less power than LPDDR4, this gap will widen in 2021. Let me know if you still think this can be surmounted with lower speeds.

I think it depends on the form factor. Design considerations for the MacBook Air are VERY different than the MacBook Pro.

It’s still amazing to me that the iMac Pro essentially has a 1U server running quiet hanging off the back of a display.

I expect there will be some form of deviation from everything being SOC when we look at say the Mac Pro. They could theoretically develop their own version of AMDs Infinity Fabric Link to maximize the ability for separate SOCs to communicate. The MPX Connector could lead the way to a separate card that has multiple purpose built chips like large blocks of Apple designed GPU cores. Granted they could have their own memory dedicated to them.

I think the Infinity Link style solution is likely if not simply necessary at some point. I doubt Apple wants to print lots of 200+ mm2 dies. Granted, they could use some more sophisticated smart interposer, but Infinity Link is simple and within immediate reach.

Moreover, Apple has always found creative ways to reuse parts. I don't think they want to design a zillion APUs when they could conceivably get away with 4 parts on different chiplet packages.

A14: Goes in the iPhone and Air.

A14X: The Bloomberg APU, 8 perf cores and 8-16 GPU cores. Goes in the iPad Pro, MBP13, MBP16, Mac Mini, iMac.

GPU: Die with only GPU cores. One goes on package with with the A14X in the MBP16 and upgraded iMacs. Two get slapped on a package together for (i)Mac Pros.

CPU: Die with only CPU perf cores, ~16. (i)Mac Pro part; can slap two on a package for high end configs.

 

[Eug]

The developers are saying the A12Z dev box (Mac mini) runs very, very cool.

They are also saying their ports to Arm are going really fast. They are also saying performance is excellent, and even Rosetta 2 performance is decent. No benchmarks though, since it violates the NDA.

 

[Doug S]

The developers are saying the A12Z dev box (Mac mini) runs very, very cool.

They are also saying their ports to Arm are going really fast. They are also saying performance is excellent, and even Rosetta 2 performance is decent. No benchmarks though, since it violates the NDA.

Considering that SoC was designed for a form factor that lacks active cooling, if it ran anything but cool there would be a lot of pissed off iPad Pro owners.

 

[CluelessOne]

Yes, unfortunately. My former Uncle, an Engineer at Nokia and Alcatel was insistent that performance = GHz. The arguments I had with him growing up was cute.

It got tedious when I had much deeper understanding of chip architecture, and his line was very clear “performance is based on clock speed. Period”




Sent from my iPad using Tapatalk Pro

Off topic, but I am curious. How does one become a former uncle?

 

[DrMrLordX]

Off topic, but I am curious. How does one become a former uncle?

Death? Divorce? Or maybe it's an Ethereum thing.

 

[lobz]

Off topic, but I am curious. How does one become a former uncle?

What a catch! Thanks for the laughs, guys 😂

 

[Eug]

Considering that SoC was designed for a form factor that lacks active cooling, if it ran anything but cool there would be a lot of pissed off iPad Pro owners.

And I think Apple is going to leverage this. Developers are already very satisfied at this level of performance... from a chip from 2018... but are marveling at the low wattage.

Like I said before, I think Apple will lower the power utilization of their mobile Macs as opposed to stuffing as much horsepower in them as they can with existing TDPs. I won't completely rule it out, but I don't think there is any need for something like an A14X with 4 performance cores in a MacBook Air. That would be used in a lower end MacBook Pro. The chips with 6 to 8 performance cores would be used in mid to higher end MacBook Pros.

To put this in perspective: I mentioned in another thread that I just got a 2006 MacPro with 2 x dual-core 2.66 GHz Xeon, and that does really well for general business and home computing in terms of speed. If my workplace issued to me a machine with this level of performance for my daily work activities, I'd have no major complaints, as long as it had SSD and sufficient RAM. My prediction is that A14 will be roughly THREE TIMES as fast as that machine.

The only way an X/Z chip makes sense to me in a MacBook Air is if they use an older one like A12X. But even then, I'd think they'd just be better off just shipping it with A12 instead of A12Z, since even A12 is quite fast for this type of use.

Putting the same chip in the MacBook Air and the iPad Pro doesn't make much sense to me either from a marketing perspective or a performance perspective.

 

[Thala]

And I think Apple is going to leverage this. Developers are already very satisfied at this level of performance... from a chip from 2018... but are marveling at the low wattage.

Like I said before, I think Apple will lower the power utilization of their mobile Macs as opposed to stuffing as much horsepower in them as they can with existing TDPs. I won't completely rule it out, but I don't think there is any need for something like an A14X with 4 performance cores in a MacBook Air. That would be used in a lower end MacBook Pro. The chips with 6 to 8 performance cores would be used in mid to higher end MacBook Pros.

I do believe they are going to beat the performance of the Intel derivatives in each tier by a significant amount - and still have headroom left to lower the power. Should be a non issue putting 4 A14 performance cores into MacBook Air and still stay well below 10W TDP.

 

[ThatBuzzkiller]

$800 for their new supposed ARM Macbook is still too expensive to pay for this garbage that developers and everyone else will have to pay more in pain for over time. The Apple silicon itself is worth exactly $0 so we shouldn't have to pay more than $600 for the entire system (ideally it should be $500) when there will likely be cheaper laptops that will work better than their more expensive solution ...

Some in the market might've been willing to overlook the potential flaws but there's absolutely no excuse for Apple's historically poor software quality now that they're comparatively priced against mainstream segment. Customers don't expect to pay a pretty penny for just the hardware, they expect to see good software support as well which will be highly lacking ...

If consumers are going to have to pay as much as they did for a new platform that has yet to establish itself compared to the old platform then ARM Macs are already a dead end ...

 

[NTMBK]

$800 for their new supposed ARM Macbook is still too expensive to pay for this garbage that developers and everyone else will have to pay more in pain for over time. The Apple silicon itself is worth exactly $0 so we shouldn't have to pay more than $600 for the entire system (ideally it should be $500) when there will likely be cheaper laptops that will work better than their more expensive solution ...

Some in the market might've been willing to overlook the potential flaws but there's absolutely no excuse for Apple's historically poor software quality now that they're comparatively priced against mainstream segment. Customers don't expect to pay a pretty penny for just the hardware, they expect to see good software support as well which will be highly lacking ...

If consumers are going to have to pay as much as they did for a new platform that has yet to establish itself compared to the old platform then ARM Macs are already a dead end ...

Higher IPC than Intel, world's best power efficiency... How the hell do you figure that is worth $0?

 

[Eug]

I do believe they are going to beat the performance of the Intel derivatives in each tier by a significant amount - and still have headroom left to lower the power. Should be a non issue putting 4 A14 performance cores into MacBook Air and still stay well below 10W TDP.

How much do you think they need to beat the previous model by?

A14 with 2 performance cores should probably be roughly in the same ballpark as I7-1060NG7, which is in the current top tier MacBook Air. However, it is loud with lots of fan noise.

A14 could go into the lower end MacBook Air fanless. But what about the top tier model? How about a higher clock? I'm thinking a part with 3 performance cores doesn't quite make sense since the volume would be too low, unless it's a 4-perf-core model with one (defective) core deactivated.

That would be faster than the current Intel offering but would still differentiate itself from A14X with 4 performance cores in a low end MacBook Pro.

I hope you turn out to be right, but I'm just not that as optimistic as some of you here.

Higher IPC than Intel, world's best power efficiency... How the hell do you figure that is worth $0?

Don't take the bait my friend.

 

[ThatBuzzkiller]

Higher IPC than Intel, world's best power efficiency... How the hell do you figure that is worth $0?

How about incompatible software or bugs ? I'm not sure if anybody is willing to deal with $800 worth of likely headaches even if it has higher performance or power efficiency ...

I reiterate that Apple silicon is worth $0 until proven otherwise that it can run anything worth a damn. Even Intel does a better job at supporting AVX-512 in professional software while the number of software partners that could count on the same level support from Apple can be counted with your fingers. Regular Mac systems sucks plenty as it is despite featuring the same CPU/GPU architecture compared to Windows systems but now that they're changing both the CPU/GPU architecture I suppose that they're worth effectively nothing now ...

That is why ARM Macbooks should be priced at no more than $600 because otherwise it's going to face competition against other alternatives in higher price segments that just straight up work better and are more stable ...

 

[Thala]

How much do you think they need to beat the previous model by?
A14 with 2 performance cores should probably be roughly in the same ballpark as I7-1060NG7, which is in the current top tier MacBook Air. However, it is loud with lots of fan noise.

You have to keep in mind that an iPhone 11 with A13 already beats the 2020 MacBook Air with i7-1060NG7 in both ST and MT by roughly 10%.

 

[Doug S]

I do believe they are going to beat the performance of the Intel derivatives in each tier by a significant amount - and still have headroom left to lower the power. Should be a non issue putting 4 A14 performance cores into MacBook Air and still stay well below 10W TDP.

While it would not be a problem to do that, if they want to make the Macbook Air the low end, the Macbook the midrange, and the Macbook Pro the high end, they have to leave some room between each tier.

Having 4 cores in the low end, 6 in the midrange, and 8 in the high end (as an example) isn't really much separation if the clock rates are basically the same, since most people aren't running software that would allow them to tell much difference between 4 and 8 cores.

I don't know exactly how Apple will segment things, but they are lucky that they have a lot of options how to do so thanks to the low bar set by the Intel CPU in the Air.

 

[Eug]

You have to keep in mind that an iPhone 11 with A13 already beats the 2020 MacBook Air with i7-1060NG7 in both ST and MT by roughly 10%.

What performance benchmark are you comparing?

A13 gets around 1340/3575 in Geekbench 5.
i7-1060NG7 on a good day can hit about 1225/4000.

I am thinking ballpark A14 might be around 1550/4125. That would be a good starting point for the entry level MacBook Air, and a faster version (faster clocked, or else triple-perf-core?) could be reasonable for the top end MacBook Air.

 

[Thala]

What performance benchmark are you comparing?

A13 gets around 1340/3575 in Geekbench 5.
i7-1060NG7 on a good day can hit about 1225/4000.

I am thinking ballpark A14 might be around 1550/4125. That would be a good starting point for the entry level MacBook Air, and a faster version (faster clocked, or else triple-perf-core?) could be reasonable for the top end MacBook Air.

I believe they are putting more cores into the Air than into an iPhone 12 - we are talking 2x power budget here. iPad Pro and MacBook Air i can imagine using the same SoC.

 

[Eug]

I believe they are putting more cores into the Air than into an iPhone 12 - we are talking 2x power budget here. iPad Pro and MacBook Air i can imagine using the same SoC.

I would be a little annoyed if my brand new 2021 iPad Pro wasn't faster than the MacBook Air, considering a similarly sized (12.9") model with Magic Keyboard is much, much more expensive than the MacBook Air.

 

[Thala]

I would be a little annoyed if my brand new 2021 iPad Pro wasn't faster than the MacBook Air, considering a similarly sized (12.9") model with Magic Keyboard is much, much more expensive than the MacBook Air.

I dont think that Apple sees the iPad as competitor to the MacBooks. In practice this means they could go with whatever the thermal headroom allows - and thermal boundaries should be very similar between MacBook Air and iPad Pro i assume.
Of course in theory they could artificially cripple the Air.

 

[Eug]

I dont think that Apple sees the iPad as competitor to the MacBooks. In practice this means they could go with whatever the thermal headroom allows - and thermal boundaries should be very similar between MacBook Air and iPad Pro i assume.
Of course in theory they could artificially cripple the Air.

I guess I should point out that Tim Cook has already said that his vision is that the iPad Pro will replace desktops and notebooks, including Apple's own notebooks. This is a direct quote:

I think if you’re looking at a PC, why would you buy a PC anymore? No really, why would you buy one? Yes, the iPad Pro is a replacement for a notebook or a desktop for many, many people. They will start using it and conclude they no longer need to use anything else, other than their phones. I think it clearly creates some cannibalization—which we knew would occur—but we don’t really spend any time worrying about that, because as long as we cannibalize ourselves, it’s fine.

Tim Cook on iPad Pro: ‘Why would you buy a PC anymore?’

Apple's CEO has high hopes for the new 12.9-inch tablet, even if it cannibalizes existing iPads and MacBooks.

www.macworld.com

Anyhow, going with a slower SoC in the Air isn't "crippling" the Air. It's about market segmentation and targeting specific groups of customers. In 2020, the point of the Air is to be cheap, not necessarily super fast, and much of the customer base that buys the Air doesn't care that it is not so fast. For example, I just sold a 2019 Air that we won in a raffle to a person just like this. The only thing she was concerned about was the storage space. And you know what? She is very pleased with the responsiveness of her new laptop. It has I5-8210Y, which gets around 810/1700 in Geekbench 5, which BTW is faster than my 2017 MacBook m3-7Y32 at around 740/1640.

 

[Thala]

I guess I should point out that Tim Cook has already said that his vision is that the iPad Pro will replace desktops and notebooks, including Apple's own notebooks. This is a direct quote:

This argument does not resonate with me. I mean if Tim would really think so, one would not need a Mac to develop for iOS. I do really like the form-factor of a tablet but i need a real desktop OS - i guess thats the reason i am mostly using my Surface Pro X when traveling.

 

[Doug S]

I would be a little annoyed if my brand new 2021 iPad Pro wasn't faster than the MacBook Air, considering a similarly sized (12.9") model with Magic Keyboard is much, much more expensive than the MacBook Air.

How is that different than Apple selling (non Pro) iPads for half the price of iPhones that have less performance (depending on when the iPad is introduced, if you are buying the n-1 iPhone versus the latest and greatest, etc)

 

[Doug S]

I guess I should point out that Tim Cook has already said that his vision is that the iPad Pro will replace desktops and notebooks, including Apple's own notebooks. This is a direct quote:

I think if you’re looking at a PC, why would you buy a PC anymore? No really, why would you buy one? Yes, the iPad Pro is a replacement for a notebook or a desktop for many, many people. They will start using it and conclude they no longer need to use anything else, other than their phones. I think it clearly creates some cannibalization—which we knew would occur—but we don’t really spend any time worrying about that, because as long as we cannibalize ourselves, it’s fine.

Wow, you sure read a lot into "for many many people".

What would be the point of going ARM on the Mac if Apple was really down with your idiotic idea of replacing desktops and notebooks with the iPad Pro?

 

[Eug]

Wow, you sure read a lot into "for many many people".

What would be the point of going ARM on the Mac if Apple was really down with your idiotic idea of replacing desktops and notebooks with the iPad Pro?

Wasn't my "idiotic idea". It was Tim cook's. If you have a beef with that idea, take that up with him. Meanwhile he's happily moving the iPad-as-primary-computer theme along, with the release of the iPad Pro's Magic Keyboard, and putting the Smart Connector on the entry level iPad. In fact, he's even allowed Logitech to licence that tech and create keyboard mouse combos for the entry level iPad as well as more expensive iPads at a lower price than Apple's Magic Keyboard.

OTOH, Tim is still happy to sell you a Mac in the interim, if that's what you want and won't buy an iPad Pro. And now he can make those Macs for cheaper, with Apple silicon. Make no mistake though. Just because they can thermally stuff a A14X in a MacBook Air doesn't mean they necessarily will. Apple is the king of bean counting at times, with iDevice RAM being a great example.

Sure, I'd be happy to see the MacBook/MacBook Air get A14X at its price point or lower, and stay/become fanless. And yes I agree it's technically feasible. I'm just not convinced they'll do it, for marketing reasons and cost. Furthermore, many, many people simply do not care. A14 in a MacBook/MacBook Air would make for a very nice experience for most general usage. A13 is already fast enough, and A14 will be faster. You probably don't want to be compiling your pro applications on one, but that is not who the MacBook Air is marketed towards.

This argument does not resonate with me. I mean if Tim would really think so, one would not need a Mac to develop for iOS. I do really like the form-factor of a tablet but i need a real desktop OS - i guess thats the reason i am mostly using my Surface Pro X when traveling.

It doesn't resonate fully with me either. I'm just saying that Apple isn't scared of replacing notebooks with iPads. They'll make money either way.

Note however, there are rumours that Xcode mobile is coming. I'm not sure I buy that either, but there ya go. But even if it does come, it might come in true Apple fashion, which is over the course of several years.

How is that different than Apple selling (non Pro) iPads for half the price of iPhones that have less performance (depending on when the iPad is introduced, if you are buying the n-1 iPhone versus the latest and greatest, etc)

? I'm not quite following. iPads with the equivalent chips usually come out a long time after the iPhones, and the entry level ones are usually a couple of generations behind. Remember, Apple's current iPad is A10, and Apple's current iPad Air is A12. Meanwhile, the iPhone is on A13.

Oh and general mainstream performance of A10 is actually quite decent. A10 is already faster than a LOT of Wintel laptops currently for sale.

 

[jpiniero]

There's always the idea of turning the iPad Pro into a 2-in-1 that runs OSX (with the iOS interface still available). Basically a not crappy Surface.

 

[Doug S]

Wasn't my "idiotic idea". It was Tim cook's. If you have a beef with that idea, take that up with him. Meanwhile he's happily moving the iPad-as-primary-computer theme along, with the release of the iPad Pro's Magic Keyboard, and putting the Smart Connector on the entry level iPad. In fact, he's even allowed Logitech to licence that tech and create keyboard mouse combos for the entry level iPad as well as more expensive iPads at a lower price than Apple's Magic Keyboard.

Sorry, I meant to say "your idiotic interpretation of what Tim Cook said".