Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

leman

macrumors Core
Oct 14, 2008
19,523
19,680
Really? How many times have you seen "up to …" in their device descriptions?

I really have no idea what you are getting at. Is there any precedent for them to ship an in-house designed chip that was obviously severely underclocked? The do clock the iPad/Mac chips slightly higher than the iPhone chips but Apple Silicon has a much lower dynamic frequency range (around 10%) than anything in the x86 land (where 80% is not uncommon).
 

deconstruct60

macrumors G5
Mar 10, 2009
12,493
4,053
Did you guys see these allegedly leaked slides from Luke Miani? What do we think?

View attachment 1792483
View attachment 1792484
If production is Q2, I don’t see how we can get a summer launch. However this is in contradiction with other rumours saying production had started in April. Wait & see I guess.

PS: EFI refers to the number of low power cores out of the core total (CPU). Don’t addition CPU and EFI.

If 'upcoming' means going into test/production in some future quarter don't want to talk about yet. Also that this slide dates from late 2020 or Q1 2021 ( Janurary - Feburary).

It is pretty much aligned with what was in the Bloomberg report a while ago.
The 8-9-4 and 8-10-4 are the successor to the M1 which is probably called M2. TSMC 5nm+ so not much of a process shrink. Apple could add 2 more GPU cores and hit the same power level.

10-16/32-2 also lines up. upper MacBooPro 13 and upper Mac Mini along with MBP 16" . Coin flip whether that is 5nm ( filling gap where M1 volume may be slowing down slightly ) or some 5nm+ . [ I suspect M1"some adjective" , 5nm . ] .

20-64-8 is a bit different ( higher number of E cores. 8 instead of 4 ). Probably iMac 27" (large screen)

40-128-8 is haven't thinned out the iMac completely could put this in a iMac 27" also. ( as well as possible half-sized Mac Pro ).


If there are chiplets, somewhat indicative that are possibly split into CPU ( base ) and GPU core focused camps.
[ for example chiplet 1: 8E-NPU-Thunderbolt-core I/O (USB etc) - one or two memory controllers
chiplet 2: 6 P cores + 32 GPU cores + memory controllers
So to 'crank up' the P-GPU count you keep adding chiplet 2's. ]

If not chiplets then they can tune to whatever combination they want.

The 128 CPU core option. That looks like a "science experiment" that is not heading to any product. Apple could have fabricated several wafers to see how it works, but probably practicing for 4nm or 5nm . If the others aren't chiplets then this one is. [ e.g. building block with 30 P - 64 GPU cores ] . If they want to "test" that as non product they could put just a chiplet into a 'test jig' and get feed back as opposed to a whole product. So it is somewhat of a "path finding" device.

That is why full blown high volume production testing isn't schedule. It isn't meant for high volume. [ And it is disconnected from Darwin kernel (macOS) core limits. ]

[ If Apple goes one last iteration on the Xeon W-3300 then takes the pressure off of that option also. ]
 

deconstruct60

macrumors G5
Mar 10, 2009
12,493
4,053
When does FY-Q2 start? December or January?

Either way, the graphic looks pretty ridiculous. Even 40 cores is way too high to be realistic for any package.

Apple could be taking this "system on a chip" relatively seriously. :). If you go to a modern larger screen laptop and add up the board space area of the following :

CPU's BGA 'socket'
RAM soldered on or so-DIMM slots
discrete GPU
soldered on VRAM for GPU.
T2 chip ( on Mac laptops )

Add all of that up into one grand total mm^2 capacity. That is what Apple is going to sit on top of. It will be a relatively ( to a mobile CPU) larger package , but Apple has also subsumed significant other stuff off the main logic board also. So it is a swap of space. Less discrete packages for bigger package.

Probably explains their JadeC-Chop option which might be a smaller package for the MBP 14" that didn't have a dGPU and VRAM to swap for a larger package. [ but the T2 + RAM + CPU area may be enough. ]


IMHO, Apple is out to migrate a larger number of entry-midrange desktop users to laptops. Apple mainly sells laptops and probably won't mind to get more into that form factor. Probably can't do 40 cores in a laptop on this fab process but 2-3 fab process shrinks down the road then maybe. For now, they can get more folks into iMacs.
Apple already pursued that with the iMac Pro. Going from 18 -> 40 cores would draw even more in.
 

deconstruct60

macrumors G5
Mar 10, 2009
12,493
4,053
Never, because those are obviously desktop chips. A 20/64/8 is around 120 watts combined TDP, way too hot for a 16” chassis. Maybe on 2nm when the power consumption goes down by 50%…

Errr, What do you think is in the current MBP 16" chassis?

i9-9980HK 45-90 W
RX-5700M 150W ( 5500M 85W )

So upper range is 195-240W now. ( 130-175W). At 120W, that would be 50% less than the current , "old", top mark. And right in-line with the entry-mid GPU option.

It would be hard for them to do "whisper quiet" MBP 16", but if willing to tolerate the same noise levels out of the laptop; that is what was designed for. ( for better or worse. ). Can also put the current MBP 16" chassis into the "bad" catagory where the current usage isn't acceptable ( just that Apple had to ship it anyway.) But in that case the whole chassis fails; not the TDP limit.

Moderate chance the problem is the package size if can't get rid of the two fans. There is lots of board space to swap for though ( T2 , VRAM , GPU , CPU , 1/2 the RAM ) . The heat is also dense packed so would need a better heat pipe and cooling. ( not exactly an Apple strong suit. ). if the 20 and 40 core options come in the same size package ( so that can share common logic board ), then the package is likely too big for a laptop. ( package is sized to fit on a iMac 27" and/or Mac Pro logicboard class of area. )

Larger chance though they don't want to give up the battery life gains for the approximately same size form factor.

However, if Apple did a MBP 16" package size that could take 10 or 20 cores in same package they could have a logicboard which took both.
 

leman

macrumors Core
Oct 14, 2008
19,523
19,680
Errr, What do you think is in the current MBP 16" chassis?

i9-9980HK 45-90 W
RX-5700M 150W ( 5500M 85W )

So upper range is 195-240W now. ( 130-175W). At 120W, that would be 50% less than the current , "old", top mark. And right in-line with the entry-mid GPU option.


The 16” chassis is limited to around 85W TDP. This is very clear from empirical tests and is also indirectly confirmed by wattage of the supplied power adapter. The 16” limits the CPU to 60W PL1 TDP and the GPU to roughly 50W TDP, with combined power of 85W (if both the CPU and the GPU are used the CPU tends to be deprioritized a bit). I did a lot of tests on my 16” when I received it, plus, there are third party reviews where these things are reported (e.g. notebookcheck).

The Pro 5500M/5600M used by Apple are very different from the regular RX GPUs. They are better binned, use much lower clocks and slower VRAM (or, in case of Pro 5600M HBM2 that’s faster but uses less power). They are about 20-30% slower than their gaming counterparts but have superior power efficiency.

P.S. The TDP of Apple laptops was pretty much unchanged in the last 10 years (although the 16” does increase it by 10 watts or so). The high-end 13” was always limited to 30W and the 15”/16” was in range of 65-80W. I am quite confident these are also the power limits we will see with the new hardware. Which incidentally maps very well to the leaked configuration as a 8+2/32 SoC would use exactly 40(CPU) + 40(GPU) watts. Apple is very precise about their numbers :)
 
Last edited:

Falhófnir

macrumors 603
Aug 19, 2017
6,146
7,001
Errr, What do you think is in the current MBP 16" chassis?

i9-9980HK 45-90 W
RX-5700M 150W ( 5500M 85W )

So upper range is 195-240W now. ( 130-175W). At 120W, that would be 50% less than the current , "old", top mark. And right in-line with the entry-mid GPU option.

It would be hard for them to do "whisper quiet" MBP 16", but if willing to tolerate the same noise levels out of the laptop; that is what was designed for. ( for better or worse. ). Can also put the current MBP 16" chassis into the "bad" catagory where the current usage isn't acceptable ( just that Apple had to ship it anyway.) But in that case the whole chassis fails; not the TDP limit.

Moderate chance the problem is the package size if can't get rid of the two fans. There is lots of board space to swap for though ( T2 , VRAM , GPU , CPU , 1/2 the RAM ) . The heat is also dense packed so would need a better heat pipe and cooling. ( not exactly an Apple strong suit. ). if the 20 and 40 core options come in the same size package ( so that can share common logic board ), then the package is likely too big for a laptop. ( package is sized to fit on a iMac 27" and/or Mac Pro logicboard class of area. )

Larger chance though they don't want to give up the battery life gains for the approximately same size form factor.

However, if Apple did a MBP 16" package size that could take 10 or 20 cores in same package they could have a logicboard which took both.
Yeah the chips can run up way over their nominal TDPs, but they are limited somewhat by the 96W power supply. I haven't heard routine reports of the 16" draining its battery while plugged in (as some previous MBPs were known to do) so I assume it can't be going too mad with that?
 

thedocbwarren

macrumors 6502
Nov 10, 2017
430
378
San Francisco, CA
Yeah the chips can run up way over their nominal TDPs, but they are limited somewhat by the 96W power supply. I haven't heard routine reports of the 16" draining its battery while plugged in (as some previous MBPs were known to do) so I assume it can't be going too mad with that?
I have and seen it happen. Mostly nasty chrome stuff destroying the performance. Luckily not my issue.
 

deconstruct60

macrumors G5
Mar 10, 2009
12,493
4,053
The 16” chassis is limited to around 85W TDP. This is very clear from empirical tests and is also indirectly confirmed by wattage of the supplied power adapter. The 16” limits the CPU to 60W PL1 TDP and the GPU to roughly 50W TDP, with combined power of 85W (if both the CPU and the GPU are used the CPU tends to be deprioritized a bit). I did a lot of tests on my 16” when I received it, plus, there are third party reviews where these things are reported (e.g. notebookcheck).
[/QUOTE]

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.



P.S. The TDP of Apple laptops was pretty much unchanged in the last 10 years (although the 16” does increase it by 10 watts or so). The high-end 13” was always limited to 30W and the 15”/16” was in range of 65-80W.

World has changed in 10 years. There is a bigger , broader class of "workstation" / "high end" laptops now. More people are doing hybrid/host desk commutes where "desktop replacement" is acceptable ( move laptop from one desktop to another with little to no use inbetween). USB just cranked their limits to 240W on the next standard upgrade.


Performance wise the MBP 13-14" sit above the 16" sits now. Why should Apple hold the bigger MBP back?

It goes against the whole "anorexic are us" laptop theme Apple generally has going on. Capping at 10 cores will be less disruptive. But it is also less disruptive ( competitors aren't trying to hold the 90W line on power. ) Apple has the more pretty laptop and others have the ones that get more work done.
 

leman

macrumors Core
Oct 14, 2008
19,523
19,680
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.

Be it as it may, you are still looking at the wrong GPUs. Here are the one's that Apple uses: https://www.amd.com/en/graphics/radeon-apple-5000m-series

TDP of 16" MBP components as well as the chassis TDP are known quantities. There is no point in speculating about them. They have been measured and discussed in detail.

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.

This argument keeps repeated over and over, despite it clearly being nonsensical. First, Apple can implement whatever power standards they want, they are not beholden to the USB consortium. They cap their charger to 90W because that's how they want to build their laptops. Second, USB power delivery has been just updated to 240W. Third, Apple now builds their own CPUs and GPUs that are 2-3 times more efficient than any other company can deliver. If you think that they will use this new power-efficient technology to increase the TDP of their computers, you might get quite disappointed.

World has changed in 10 years. There is a bigger , broader class of "workstation" / "high end" laptops now. More people are doing hybrid/host desk commutes where "desktop replacement" is acceptable ( move laptop from one desktop to another with little to no use inbetween).

It's not that the laptops got hotter, just the technology got better to the point where a) the difference between desktop and mobile is not as dramatic as it used to be and b) everything is fast enough for almost any use. The amount of bulky, large laptops actually went down dramatically. Almost nobody makes heavy 17" desktop replacement anymore. It's all about thin, light, compact stuff. Power usage of high end gaming and workstation laptops has been sitting on ~ 120-150W for last 10 years, without any changes.
 

Sydde

macrumors 68030
Aug 17, 2009
2,563
7,061
IOKWARDI
Wikipedia says that the mini draws 39W peak, 7W idle. For a MB, add a generous 10W for the display and you have over 50W before you get to the A/C limit. Given what Apple has done with energy efficiency, I suspect that gives them plenty of headroom to double performance without closing the gap.
 

aeronatis

macrumors regular
Sep 9, 2015
198
152
If you think that they will use this new power-efficient technology to increase the TDP of their computers, you might get quite disappointed.

Do you think 16" and the rumoured 14" will have the same chip? Given that the 14" will not have a considerable bigger chassis than the current 13", if not even thinner, I think 14" Pro could simply replace the current 13" Pro at the end of the year, only with more ports. As Macbook Air M1 is more popular than MacBook Pro M1 now, this would make the Pro model more appealing compared to the Air, even with the same chip.

I also believe that the 27" or 30" iMac, whichever it will be, will have the same chip as the 16" MacBook Pro and the higher end Mac Mini will. During WWDC 2020, they emphasised that they wanted to cover the area between performance and power consumption, which makes me believe that they will not just make more power hungry chips for the desktop (I am excluding Mac Pro).
 

leman

macrumors Core
Oct 14, 2008
19,523
19,680
Do you think 16" and the rumoured 14" will have the same chip? Given that the 14" will not have a considerable bigger chassis than the current 13", if not even thinner, I think 14" Pro could simply replace the current 13" Pro at the end of the year, only with more ports. As Macbook Air M1 is more popular than MacBook Pro M1 now, this would make the Pro model more appealing compared to the Air, even with the same chip.

I also believe that the 27" or 30" iMac, whichever it will be, will have the same chip as the 16" MacBook Pro and the higher end Mac Mini will. During WWDC 2020, they emphasised that they wanted to cover the area between performance and power consumption, which makes me believe that they will not just make more power hungry chips for the desktop (I am excluding Mac Pro).

Apple still sells the higher-TDP Intel 13” model which tells us that they have plans for a more powerful compact laptop in their lineup. I am fairly confident that the upcoming 14” and 16” will use the same chip, but the 16” will have an option for a bigger GPU and will offer significantly better sustained performance. The rumored 8+2/16 chip will require about 40-50 watts to reach its full performance potential, which probably won’t be possible within a 14” chassis.
 

thenewperson

macrumors 6502a
Mar 27, 2011
992
912
The rumored 8+2/16 chip will require about 40-50 watts to reach its full performance potential, which probably won’t be possible within a 14” chassis.
So how would they handle that, considering it’s the minimum spec of the rumoured chips?
 

Sydde

macrumors 68030
Aug 17, 2009
2,563
7,061
IOKWARDI
The rumored 8+2/16 chip will require about 40-50 watts to reach its full performance potential, which probably won’t be possible within a 14” chassis.
As others have hinted at, the 14" could well be 10/12 bins of the 10/16.
 

leman

macrumors Core
Oct 14, 2008
19,523
19,680
So how would they handle that, considering it’s the minimum spec of the rumoured chips?

It will run thermally constrained, just like the M1 in the MBA. Will still outperform any Zen 3 or Tiger Lake laptop ;)

But of course, this is all just speculation. It is not certain that power/performance characteristics of the upcoming chip is going to be comparable to M1. What I wrote is just a projection based on M1’s power consumption that we know.
 

aeronatis

macrumors regular
Sep 9, 2015
198
152
Apple still sells the higher-TDP Intel 13” model which tells us that they have plans for a more powerful compact laptop in their lineup.

Still couldn't imagine 8+2/16 chip inside 14" chassis. I can imagine being thermally limited version, though. Could be like the difference between M1 Air vs M1 Pro, only with a little bit more gap maybe. Not sure about iMac and Mac Mini ?
 

AgentMcGeek

macrumors 6502
Jan 18, 2016
374
305
London, UK
Doesn’t the 14’ need to be a tad thicker than the current 13’ in order to accommodate a full HDMI port?

The HDMI female port is 4.55mm thick, whereas USB-C is 2.56mm.

If that allows a 16 core GPU option I wouldn’t complain.
 

deconstruct60

macrumors G5
Mar 10, 2009
12,493
4,053
Doesn’t the 14’ need to be a tad thicker than the current 13’ in order to accommodate a full HDMI port?

The HDMI female port is 4.55mm thick, whereas USB-C is 2.56mm.

If that allows a 16 core GPU option I wouldn’t complain.

Current MBP 13” is 15mm thick . Half of that is 7.5mm ( back of envelope split between top and bottom ) . There is room in terms of thickness. Even more if square off the edges a bit.

Going even thinner . Yeah, that’s a problem .
 

deconstruct60

macrumors G5
Mar 10, 2009
12,493
4,053
As others have hinted at, the 14" could well be 10/12 bins of the 10/16.

how does 4 GPU cores buy 10W of savings when the 8 cores in the M1 aren’t that high ? The 16 is already itself a “chop” . Binning down on what is already chopped …
 

thunng8

macrumors 65816
Feb 8, 2006
1,032
417
Never, because those are obviously desktop chips. A 20/64/8 is around 120 watts combined TDP, way too hot for a 16” chassis. Maybe on 2nm when the power consumption goes down by 50%…
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.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.