Rumored pro chip

[Zdigital2015]

No they are not and I am sure one will be able to spend many thousands on BTO options (32 CPU core option, 64GB of RAM, 8TB of storage, Nanotexture Glass).

Frankly I think $2499 is fine for the entry level model because it would offer enough screen size, cores, RAM and storage over the 24" model to justify that $1000. Especially with the 24" iMac being so much better than the 21.5" in every way I think it can stand on its own vis-a-vis the larger model than the old 21.5" could against the 27".

The first Retina 5K 27” iMac came out at $2499 and no one blinked an eye as the 5K display was a breakthrough at the time. A 32” 6K display with some form of the XDR’s brightness levels is going to be a breakthrough at $2499, contrary to any forum poster’s comment to the contrary. I would expect it to be 16GB/1TB/16 GPU cores as the base model. Hopefully, we’ll see soon enough.

 

[Fomalhaut]

Apple wouldn't have to downclock as much. But those are the parts they are working with now (and those are the mobile TDP numbers for the GPUs. Those are down from desktop. ) . The RAM is slower anyway. The size of that gap is where could loose more on down clocking.

Apple's power supply is somewhat capped on USB-C power standards (~100W) . Going to MagSafe means don't have to keep those. Going to 100-110W wouldn't be hard it keep the noise levels with small tweaks to the chassis.

[/QUOTE]

I don't think that Apple will increase the power requirements for the MBP16 because this would mean worse battery life. The battery capacity will be capped at 100Wh to allow the computer to be taken on commercial aircraft without special approval.

The battery life on the MBP16 is already pretty uninspiring, especially when using the dGPU. I get 4-5 hours of typical use (including a couple of short video conferences), so I would be hoping that the new Apple Silicon models are aiming for the 50-70W range for total TDP.

 

[richinaus]

Apple wouldn't have to downclock as much. But those are the parts they are working with now (and those are the mobile TDP numbers for the GPUs. Those are down from desktop. ) . The RAM is slower anyway. The size of that gap is where could loose more on down clocking.

Apple's power supply is somewhat capped on USB-C power standards (~100W) . Going to MagSafe means don't have to keep those. Going to 100-110W wouldn't be hard it keep the noise levels with small tweaks to the chassis.

I don't think that Apple will increase the power requirements for the MBP16 because this would mean worse battery life. The battery capacity will be capped at 100Wh to allow the computer to be taken on commercial aircraft without special approval.

The battery life on the MBP16 is already pretty uninspiring, especially when using the dGPU. I get 4-5 hours of typical use (including a couple of short video conferences), so I would be hoping that the new Apple Silicon models are aiming for the 50-70W range for total TDP.
[/QUOTE]

I can see this 16" blowing the competition out the water in terms of speed and longevity. No doubt Apple will be going for a good balance [target the 11th gen i9 and a 3060 GPU] and make it last all day.

 

[leman]

I don't think that Apple will increase the power requirements for the MBP16 because this would mean worse battery life. The battery capacity will be capped at 100Wh to allow the computer to be taken on commercial aircraft without special approval.

The battery life on the MBP16 is already pretty uninspiring, especially when using the dGPU. I get 4-5 hours of typical use (including a couple of short video conferences), so I would be hoping that the new Apple Silicon models are aiming for the 50-70W range for total TDP.

Peak TDP has little benefit for battery life. 150W vs. 50W on a 100Wh battery is a difference between 40 minutes and 2 hours under load, both very bad. Good battery life comes from excellent idle power consumption, which these Macs will undoubtedly have.

But I agree that we will likely see the 16” being capped to 60-70 watt, that’s the sweet spot for the chassis abs more than enough to offer excellent performance.

 

[leman]

It is easily possible to bring that chip well under 80W. We know from the macbook air the power usage of the cores when it is throttled. Going down to ~2.6Ghz reduces power usage by almost 3X compared to 3.2Ghz.

That's what high core chips have done for a long time. For example a threadripper running with 32 cores uses almost the same as when all 64 cores are loaded. Why? They decrease the frequency when running with 64 cores.

We need to stay realistic though. The big chip will never reach its full performance potential in the 16” chassis, there is an obvious issue with the memory interface (the big chip will have more on-package RAM and space+power in a laptop is limited), and there is obviously a price consideration as well. While many things are possible in theory there is very little practical chance that Apple decides to put a big desktop chip into a laptop.

 

[Appletoni]

I see no TDP problem.

1. Because it’s Apple
2. The competitors have much much higher TDP
3. Apple can easily build a new air cooler with double speed and still be 50% quieter than the cooler it has now
4. The 16-inch case will be obviously air flow improved
5. The air cooler/case will have noise canceling so it could theoretically run at 5000 rpm
6. The 16-inch case will come with two air cooler

= The chip can use 300 watts and still be better and safe.

 

[leman]

= The chip can use 300 watts and still be better and safe.

Why 300? 1000! Will still suck for chess though, they should probably just use a Threadripper in the 16”

 

[thunng8]

We need to stay realistic though. The big chip will never reach its full performance potential in the 16” chassis, there is an obvious issue with the memory interface (the big chip will have more on-package RAM and space+power in a laptop is limited), and there is obviously a price consideration as well. While many things are possible in theory there is very little practical chance that Apple decides to put a big desktop chip into a laptop.

We would it need to reach its full potential? going from 3.2Ghz to 2.6Ghz (with all cores loaded) would put it under what the current 16" macbook pro is at and provide much more performance and is only 20% less than its "full potential". I don't see any issues if that is what Apple decides to do. They would still need to robust cooling solution - but nothing more than what the current 16" Intel has.

If they wants to make the Apple Silicon 16" much thinner than the current 16" Intel - then it would be an issue.

 

[One2Grift]

Nope, you're not wrong, power consumption doesn't really mean much to me. It's pure processing power I care about the most.

I don’t blame you. I normally don’t look at power consumption as a first second or third step. Over a 5 year lifespan it could be…um…at most 50$ in power cost difference(just a guess)? It’s not enough $$$ to accept a performance step down (devices named dryer or hot water heater or central air conditioner are a different power consumption discussion ?)

With that said, power consumption can definitely have a direct relationship to heat generation. While laptop Thermal management has come along nicely, those x86 fans can still start rockin and rollin when the heat ups. IMHO, Intel absolutely should be working on increasing clock/benchmark while keeping a strong eye on power consumption (the same oower or much more ideally even less power). My guess is increasing speed by increasing power consumption is not going to go over well. Do they want to let Apple dunk on them regarding the per watt discussion?

Ball is in Intel’s court. Let’s see what they can do…

 

[AgentMcGeek]

Razer managed to fit a Ryzen 5900HX and a mobile RTX3080 into a 14’ chassis. The CPU alone has a 45W TDP. It’s not as thin as a MBP, but I’m sure Apple can fit a 50W overall TDP in the next MBP 14’.

 

[Appletoni]

Razer managed to fit a Ryzen 5900HX and a mobile RTX3080 into a 14’ chassis. The CPU alone has a 45W TDP. It’s not as thin as a MBP, but I’m sure Apple can fit a 50W overall TDP in the next MBP 14’.

 

[Appletoni]

The Ryzen 5900HX has 54W. The mobile RTX 3080 has 150W. 54+150=204W.
The watt might be even higher.
The other things inside will use also some watt.
And that’s only a 14-inch case.

Of course Apple can fit 50W overall TDP in the next MBP 14-inch. They could probably fit 150W in 14-inch.
Think what Apple can do with his 16-inch case and 2 air cooler.

 

[One2Grift]

Razer managed to fit a Ryzen 5900HX and a mobile RTX3080 into a 14’ chassis. The CPU alone has a 45W TDP. It’s not as thin as a MBP, but I’m sure Apple can fit a 50W overall TDP in the next MBP 14’.

This is a gaming laptop with a base msrp, I believe, in the ballpark of 2000. Looks like it may be top notch. I’d be curious how well they handle the thermals(they’re squeezing out the juice here).
Apple’s MBP m2 or m1x isn’t going to be gaming geared. Given the Nvidia on board Razer is sweet, I’d be curious if the MBP can actually compete in those areas with soc/m2.

 

[AgentMcGeek]

Apparently Razer uses a huge vapor chamber. You can see it in the video.

 

[leman]

We would it need to reach its full potential? going from 3.2Ghz to 2.6Ghz (with all cores loaded) would put it under what the current 16" macbook pro is at and provide much more performance and is only 20% less than its "full potential". I don't see any issues if that is what Apple decides to do. They would still need to robust cooling solution - but nothing more than what the current 16" Intel has.

If they wants to make the Apple Silicon 16" much thinner than the current 16" Intel - then it would be an issue.

Because a throttled down 20/64 part would be barely faster than a 10/32 part, while bringing significant engineering and power-management challenges, not to mention expenses. It just won't happens. Apple never operated like this and they won't start now.

 

[thunng8]

Because a throttled down 20/64 part would be barely faster than a 10/32 part, while bringing significant engineering and power-management challenges, not to mention expenses. It just won't happens. Apple never operated like this and they won't start now.

I don’t think you can do your maths properly.

for multithreaded performance 2.6x20 >> 3.2 x 10 and actually use significantly less power for the cpu portion. I don’t know about the gpu portion though. They could a similar technique to get gpu power under control

 

[AgentMcGeek]

The Blade 14 stands at 16.8mm vs 15.6mm for the M1 MBP 13".
If we can fit 150W+ in the Blade we can definitely do a lot in the new MBP. The possibility of having 32 GPU cores in the 14' seems more a marketing decision than a thermal constraint.

The Ultimate 14-Inch Laptop for Gamers and Creators: Razer Blade 14💻 | Razer United Kingdom

Slim and Lightweight ✅ Travel Friendly ✅ Fast Performance ✅ Razer Blade 14, the compact laptop perfect for gaming and streaming on the go.

www.razer.com

 

[leman]

I don’t think you can do your maths properly.

for multithreaded performance 2.6x20 >> 3.2 x 10 and actually use significantly less power for the cpu portion. I don’t know about the gpu portion though. They could a similar technique to get gpu power under control

As I said before, that's not how Apple does stuff. Yes, you might get 30-50% more CPU performance from taking a bigger cluster and severely underclocking it. But it's a laptop, not a headless server. Just because one aspect makes sense to you and it technically possible, it does not mean that the entire product would make sense. There is memory, the GPU, the power management issue (since you are asking this chip to run a much wider dynamic clock than usually), battery life concerns...

The Blade 14 stands at 16.8mm vs 15.6mm for the M1 MBP 13".
If we can fit 150W+ in the Blade we can definitely do a lot in the new MBP. The possibility of having 32 GPU cores in the 14' seems more a marketing decision than a thermal constraint.

Who cares what Razer Blade does? They (and others) have been shipping comparable laptops for years and we all know what the tradeoffs are: you cut down the battery to an absolute minimum, fill 50% of the laptop with the cooling system, cut open huge holes in your laptop to circulate the hot air, and scrape everything else you can. This is literally the opposite design to a MacBook Pro.

Apple has used constant TDP factors for their laptops for years. They even retained them as CPUs and GPUs were getting hotter and hotter, which has earned them a lot of criticism. Now they have their own hardware that is much more power efficient, and you are talking that they might increase the TDP on their laptops? Come on, let's be serious. If anything, the 16" will have it's TDP reduced.

Here is my prediction:

- 14" will have the TDP of 30W, give or take
- 16" will have the TDP of 60W, give or take
- The new 30" (or whatever) iMac will use the same chip as the 16", possibly with the 20/64 option
- Mac Pro is something else entirely

 

[thunng8]

As I said before, that's not how Apple does stuff. Yes, you might get 30-50% more CPU performance from taking a bigger cluster and severely underclocking it. But it's a laptop, not a headless server. Just because one aspect makes sense to you and it technically possible, it does not mean that the entire product would make sense. There is memory, the GPU, the power management issue (since you are asking this chip to run a much wider dynamic clock than usually), battery life concerns...




Who cares what Razer Blade does? They (and others) have been shipping comparable laptops for years and we all know what the tradeoffs are: you cut down the battery to an absolute minimum, fill 50% of the laptop with the cooling system, cut open huge holes in your laptop to circulate the hot air, and scrape everything else you can. This is literally the opposite design to a MacBook Pro.

Apple has used constant TDP factors for their laptops for years. They even retained them as CPUs and GPUs were getting hotter and hotter, which has earned them a lot of criticism. Now they have their own hardware that is much more power efficient, and you are talking that they might increase the TDP on their laptops? Come on, let's be serious. If anything, the 16" will have it's TDP reduced.

Here is my prediction:

- 14" will have the TDP of 30W, give or take
- 16" will have the TDP of 60W, give or take
- The new 30" (or whatever) iMac will use the same chip as the 16", possibly with the 20/64 option
- Mac Pro is something else entirely

Still can't do your maths. It is more like 65% not 30-50% and it is not a severe underclock. It is 20% lower when running all cores. When using lightly threaded or single thread, performance will be exactly the same. We'll agree to disagree - Apple could easily do it if they wanted to.

 

[leman]

Still can't do your maths. It is more like 65% not 30-50% and it is not a severe underclock. It is 20% lower when running all cores. When using lightly threaded or single thread, performance will be exactly the same. We'll agree to disagree - Apple could easily do it if they wanted to.

I think you are underestimating the auxiliary power requirement software the larger chip (cache/RAM etc.) while also being very optimistic with other numbers. I don't have my M1 machine anymore unfortunately and can't check, but are you certain that running a Firestorm core at 2.6ghz will only consume 2.5 watts or less?

Anyway, I think we agree that this is technically possible (although we might disagree on how much benefit one would see). My point is just that the resulting product won't make much sense, especially given Apple's (known) design principles.

 

[diamond.g]

@leman is there any reason why Apple couldn't boost one or two cores (like Intel/AMD do) when under the thermal limit? They would have a lower all core clock, but running 1 or 2 cores at higher speeds would retain the single thread dominance.

@thunng8 GPU's, typically, don't clock individual cores (WGP/SM/EU) lower or higher. They will do the whole die, typically, based on load and thermal headroom.

And by die I mean the cores, other components on chip can be ran at different speeds

 

[Homy]

Why 300? 1000! Will still suck for chess though, they should probably just use a Threadripper in the 16”

Threadripper's performance is disappointing. For chess you need dual Fujitsu A64FX.

 

[thunng8]

I think you are underestimating the auxiliary power requirement software the larger chip (cache/RAM etc.) while also being very optimistic with other numbers. I don't have my M1 machine anymore unfortunately and can't check, but are you certain that running a Firestorm core at 2.6ghz will only consume 2.5 watts or less?

Anyway, I think we agree that this is technically possible (although we might disagree on how much benefit one would see). My point is just that the resulting product won't make much sense, especially given Apple's (known) design principles.

Yes. While running at 2.6ghz the cpu uses less than half compared to 3.2ghz. It is actually quite a lot less than half.

powermetrics measures the MacBook Air at 2.5-2.6ghz when thermally throttled running cinebench at 7w. That’s less than 2w per high performance core.

 

[Appletoni]

Here is my prediction:

- 14" will have the TDP of 30W, give or take
- 16" will have the TDP of 60W, give or take
- The new 30" (or whatever) iMac will use the same chip as the 16", possibly with the 20/64 option
- Mac Pro is something else entirely

MacBook Air…
MacBook Pro 14-inch…
MacBook Pro 16-inch could have the 20/64 option. Why not? It doesn’t mean that someone needs to run CPU+GPU in the same time and also both on their limits.
A lot of people will use for their cases only the CPU or GPU. I will use for my most cases only the CPU so it can have even 40 or 60 CPU cores?

 

[leman]

powermetrics measures the MacBook Air at 2.5-2.6ghz when thermally throttled running cinebench at 7w. That’s less than 2w per high performance core.

Thanks for clarifying this. Yes, it sounds like enough thermal headroom to pull a "Threadripper in a laptop". But again. this doesn't sound like a product that Apple is interested in.