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

Tyler O'Bannon

macrumors 6502a
Original poster
Nov 23, 2019
893
1,504
Just wanted to share some thoughts about Apple Silicon that I haven't really seen:

Apple Silicon for Mac (M series):

It would be very Apple for the M1 to be the only 1st Generation Apple Silicon. They clearly designed the low end of the spectrum (but brought the baseline power for "low end" way up). It's a chip for colorful portables and desktops at the lower end. Take away the fan and you've got an ultra portable machine. Add a fan and you've got a low end MacBook Pro and a Low End Desktop, both with a huge leap in power. But I could see the M1 being the only 1st Gen Apple Silicon for Mac.

Round 2 pushes up to the Pros.

This Fall, M2 comes in the same timeframe as A15. Redesigned MacBook or MacBook Air that is colorful like iMac 4.5K.

M2 has 3 variants. M2 for all machines running M1. M2X for 14" and 16" MacBook Pro that are now redesigned around Apple Silicon (like 4.5K iMac was). And M2Z for the larger iMac (presumably now called the iMac Pro so we don't have to say 4K and 5K or 4.5K and 5.5K, but just iMac and iMac Pro), (I won't speculate about hardware beyond this: 27" minimum, 32" maximum, 30" likely, 5.5K? who knows).

Mac Pro will just be it's own thing, just Apple Silicon with crazy specs, no fancy name? Makes sense in my brain.

When 14" and 16" drop, they will be in the fall right after M2 generation, and get M2X. And next iMac at some point after that with M2Z. All before June-September of 2022 within the promised 2 year transition window.


Apple Silicon for iPhone and iPad (A Series):

I have not seen this mentioned anywhere. I don't think we've seen the last of "X" series A chips. I think the stage is set for the iPhone Pro and Pro Max to have X series chips. It will further separate the devices, and it wouldn't surprise me at all to see this happen. AS on Mac spreads the non-pro and pro between non-X and X + Z, and now iPhones as well.

The much larger battery capacity in the coming 13 (presumably) series of iPhones may be for this purpose. It at least makes sense, even if it's 100% wrong lol. And that's also true for everything in this thread.

Low end iPad may always just have a 1-2 generation older chip, and iPad Air the current non X/Z. But the nerd in me would love the low end iPad to have current chip, and iPad Air to have X/Z, and then Pro are now on M. It separates the products. I can see Apple getting to where all AS production has all products on latest chip every year. No more products behind a couple of generations. That would be a huge reason for them to make this transition in the first place. Once production is in place. Maybe this won't happen in the M2 and A15 generation, but maybe this will all happen over the next few years as Apple produces all of their own chips.

Lastly, regarding Intel, Apple may offer Intel Mac pro a couple of years longer than the 2 year transitions for pros that simply need Intel. However, they would also offer the full realization of their Trash Can Mac Pro dream with AS. And, eventually, even Intel Mac Pro goes away.

There will be either 0 or 1 Mac that still offers intel after 2022, and if it is 1, that 1 will be the Mac Pro. Some professionals would just require it for their specific work.

Looking forward to comments.
 
My biggest issue is the Pro market is very GPU heavy. ML, Metal, OpenCL, Video Rendering, Color Grading, 3D rendering, ML/AI all require the latest and greatest in GPUs and will always want more power.

The latest MacPro with Vega Pro Duo level GPUs was the first MacPro to compete on the higher end in quite a while.

I have a brand new M1 MacBook Pro and I love it, I’m about to buy some M1 iMacs as well. The cpu and wattage and overall MacOS experience is phenomenal, but for GPU intensive tasks I still have to use my eGPUs on Intel for rendering. I did some tests and it’s 1/5 to 1/3 my Vega Frontier in and eGPU. My 20min GPU render is now over an hour.

Apple needs to figure out a way to compete in high end cutting edge graphics aiming for 3090ti++ level or it can’t really create a Modern Pro workstation.

iMacs, MacBook Pros, MacMini will be more than adequate, but high Pro level workstations will need high end graphics.

Apple might have something to compete, like Metal specific render pucks or tb3 render modules, who knows, but there is nothing projected at the moment.
 
My biggest issue is the Pro market is very GPU heavy. ML, Metal, OpenCL, Video Rendering, Color Grading, 3D rendering, ML/AI all require the latest and greatest in GPUs and will always want more power.

The latest MacPro with Vega Pro Duo level GPUs was the first MacPro to compete on the higher end in quite a while.

I have a brand new M1 MacBook Pro and I love it, I’m about to buy some M1 iMacs as well. The cpu and wattage and overall MacOS experience is phenomenal, but for GPU intensive tasks I still have to use my eGPUs on Intel for rendering. I did some tests and it’s 1/5 to 1/3 my Vega Frontier in and eGPU. My 20min GPU render is now over an hour.

Apple needs to figure out a way to compete in high end cutting edge graphics 3090ti++ or it can’t really create a Modern Pro workstation.

iMacs, MacBook Pros, MacMini will be more than adequate, but high Pro level workstations will need high end graphics.

Apple might have something to compete, like Metal specific render pucks or tb3 render modules, who knows, but there is nothing at the moment.

Apple has the most energy-efficient GPU IP on the market at the moment. M1 GPU is a 10 watt parts, your Vega Frontier edition is 300W. An RTX 3090 is around 10-15 times faster than M1, but consumes 35 times more power.

So in the end, it depends on how far Apple wants to go. There is no technical reason why they wouldn't be able to deliver 3090-like performance at half the power consumption. All it takes to get there from where they stand now is money, engineering resources and a business decision. First two Apple has more than plenty, the last part you never know with those guys.
 
My biggest issue is the Pro market is very GPU heavy. ML, Metal, OpenCL, Video Rendering, Color Grading, 3D rendering, ML/AI all require the latest and greatest in GPUs and will always want more power.

The latest MacPro with Vega Pro Duo level GPUs was the first MacPro to compete on the higher end in quite a while.

I have a brand new M1 MacBook Pro and I love it, I’m about to buy some M1 iMacs as well. The cpu and wattage and overall MacOS experience is phenomenal, but for GPU intensive tasks I still have to use my eGPUs on Intel for rendering. I did some tests and it’s 1/5 to 1/3 my Vega Frontier in and eGPU. My 20min GPU render is now over an hour.

Apple needs to figure out a way to compete in high end cutting edge graphics aiming for 3090ti++ level or it can’t really create a Modern Pro workstation.

iMacs, MacBook Pros, MacMini will be more than adequate, but high Pro level workstations will need high end graphics.

Apple might have something to compete, like Metal specific render pucks or tb3 render modules, who knows, but there is nothing projected at the moment.

I am not a pro, and I understand from the outside looking in that this is by far the biggest issue. I would love to know a frame of reference for it. If Apple could make an AS Mac Pro with 64 and 128 core GPU, is that still less than you can pull in the current Mac Pro with maxed out Graphics? How do they compare?
 
I thought all of this was obvious. People complain about the performance of the current M1 chip but don’t understand this is replacing base model computers like the old MacBook Air where simple tasks like video calls would cause the fans to ramp up. Everything is speculation on what will come out next but the performance leap of the models Apple has replaced with M1 is significant
 
Apple has the most energy-efficient GPU IP on the market at the moment. M1 GPU is a 10 watt parts, your Vega Frontier edition is 300W. An RTX 3090 is around 10-15 times faster than M1, but consumes 35 times more power.

So in the end, it depends on how far Apple wants to go. There is no technical reason why they wouldn't be able to deliver 3090-like performance at half the power consumption. All it takes to get there from where they stand now is money, engineering resources and a business decision. First two Apple has more than plenty, the last part you never know with those guys.

How do the mathematics work with more AS GPU cores? By straight math (which may not translate in real world usage), if Apple could take just what exists now in the first generation, give 128 cores, that would be 16 times the power of an 8 core (which i know that doesn't account for any inefficiency). Would that directly compete with the 10-15x more power of the RTX 3090?

I also understand another angle of this is ARM vs x86. And I'm not a Pro, so AS is just fine for me. I know many will need x86 for more than the next couple of years, and am curious how Apple's solutions will work for them.
 
I thought all of this was obvious. People complain about the performance of the current M1 chip but don’t understand this is replacing base model computers like the old MacBook Air where simple tasks like video calls would cause the fans to ramp up. Everything is speculation on what will come out next but the performance leap of the models Apple has replaced with M1 is significant
It is VERY significant. Raising the bottom bar is just as vitally significant as raising the top bar. While we are all excited to see what the top end will look like, it's important to realize that Apple just tripled the bottom line of CPU and GPU in an insanely thin notebook with no fan. And they are just getting started. M2 and redesign of MacBook or MacBook Air will no doubt be something that could not have existed less than 2 years before its debut without AS. Moving up the bottom line 3x is HUGE.
 
I thought all of this was obvious. People complain about the performance of the current M1 chip but don’t understand this is replacing base model computers like the old MacBook Air where simple tasks like video calls would cause the fans to ramp up. Everything is speculation on what will come out next but the performance leap of the models Apple has replaced with M1 is significant
Right but that's also on Apple for designing something that can't properly disperse heat or do proper GPU driver optimization. The Intel m3 CPU in a Surface Go can decode video streams just fine without sounding like a wind tunnel.
 
Right but that's also on Apple for designing something that can't properly disperse heat or do proper GPU driver optimization. The Intel m3 CPU in a Surface Go can decode video streams just fine without sounding like a wind tunnel.
I'm guessing Apple knew they were headed back toward no fan with AS (like the 12" MacBook). The M1 and no fan solves that now for sure. And will only continue to moving forward. But you are right, it does leave the people who have current MBA design with Intel with that issue though.
 
How do the mathematics work with more AS GPU cores? By straight math (which may not translate in real world usage), if Apple could take just what exists now in the first generation, give 128 cores, that would be 16 times the power of an 8 core (which i know that doesn't account for any inefficiency). Would that directly compete with the 10-15x more power of the RTX 3090?

GPU are parallel processing device and by definition they specialize on highly parallelizable tasks. This means that Gpu performance scales linearly with the available computational resources. Keep adding cores and your performance will go up accordingly. The only condition is that you scale up your other resource (memory bandwidth, caches, interconnect performance) as well, and of course that your tasks are large enough to saturate the larger GPU. This is true for any GPU out there. You can do the math for any AMD or Nvidia GPU and you will see that their real-world performance scales with the number of cores (don't forget to account for different clocks and memory specifications though!).

Since Apple Silicon is very power efficient, scaling it up should not provide significant engineering challenges. These are solved problems. So yes, 16x as many cores (128 cores, 16384 compute units) would put the hypothetical Apple GPU on roughly the same level as the 3090.

I also understand another angle of this is ARM vs x86. And I'm not a Pro, so AS is just fine for me. I know many will need x86 for more than the next couple of years, and am curious how Apple's solutions will work for them.

Some might (especially people relying on legacy plugins or software), many will not.
 
Apple has the most energy-efficient GPU IP on the market at the moment. M1 GPU is a 10 watt parts, your Vega Frontier edition is 300W. An RTX 3090 is around 10-15 times faster than M1, but consumes 35 times more power.

So in the end, it depends on how far Apple wants to go. There is no technical reason why they wouldn't be able to deliver 3090-like performance at half the power consumption. All it takes to get there from where they stand now is money, engineering resources and a business decision. First two Apple has more than plenty, the last part you never know with those guys.
I read somewhere than GPUs can scale easier than CPU, so if Apple has an efficient GPU then could just add more GPU cores until it is at
 
I read somewhere than GPUs can scale easier than CPU, so if Apple has an efficient GPU then could just add more GPU cores until it is at

I think the GPU could easily scale to compete with a 3090, but Apple is going to have to decide if it makes sense for them to create a High End MacPro with State of art GPU performance, or come to the realization that the market is just too small.

A medium article stated the M1 iPad basically destroyed the tablet market by setting the bar super high.. nothing from Samsung, or Google, or Microsoft can even compete with the M1 iPad... When this happens I feel like it is even less of an incentive for Apple to keep making high end Pro Equipment, the market is just so much smaller than the consumer/prosumer market..

Apple has already left the ProMarket and come back a few times, they only create a new MacPro when video editors and CG artists who love macOS complain, the 2019 MacPro for example is probably just for show, I doubt they make much money off of that workstation or that Pro end demographic.

With all its faults Apple, they still make the best computer hardware/software combo on the planet, if my M1 MacBook Pro had high end graphics capabilities, it would just solidify their greatness.
 
Last edited:
It is VERY significant. Raising the bottom bar is just as vitally significant as raising the top bar. While we are all excited to see what the top end will look like, it's important to realize that Apple just tripled the bottom line of CPU and GPU in an insanely thin notebook with no fan. And they are just getting started. M2 and redesign of MacBook or MacBook Air will no doubt be something that could not have existed less than 2 years before its debut without AS. Moving up the bottom line 3x is HUGE.

The bigger takeaway in my opinion is that with the release of the M1 Macs, Apple redefined the entire low-end PC/Mac Market. So many PC sold today are the cheap laptops from HP, Dell, Lenovo, Toshiba, etc. that barely have enough hardware to even run Windows, let alone install any software or do anything productive with. Those systems have massive tradeoffs such as eMMC storage at either 32 or 64GB RAM, 2-4GB RAM, lower resolution displays, and significant keyboard/trackpad flex.
 
The bigger takeaway in my opinion is that with the release of the M1 Macs, Apple redefined the entire low-end PC/Mac Market. So many PC sold today are the cheap laptops from HP, Dell, Lenovo, Toshiba, etc. that barely have enough hardware to even run Windows, let alone install any software or do anything productive with. Those systems have massive tradeoffs such as eMMC storage at either 32 or 64GB RAM, 2-4GB RAM, lower resolution displays, and significant keyboard/trackpad flex.
I agree 100%. It is incredible. When the M2 comes out and we get a colorful MacBook or MacBook Air with no fan, we will have to remind ourselves that this is as powerful as a high end 27" iMac 5K from not even 2 years ago, and now it's the "fun, portable" computer with insane battery life. And nothing comparable will be available on the widows side since they all run Intel. Incredible.
 
I think the GPU could easily scale to compete with a 3090, but Apple is going to have to decide if it makes sense for them to create a High End MacPro with State of art GPU performance, or come to the realization that the market is just too small.

A medium article stated the M1 iPad basically destroyed the tablet market by setting the bar super high.. nothing from Samsung, or Google, or Microsoft can even compete with the M1 iPad... When this happens I feel like it is even less of an incentive for Apple to keep making high end Pro Equipment, the market is just so much smaller than the consumer/prosumer market..

Apple has already left the ProMarket and come back a few times, they only create a new MacPro when video editors and CG artists who love macOS complain, the 2019 MacPro for example is probably just for show, I doubt they make much money off of that workstation or that Pro end demographic.

With all its faults Apple, they still make the best computer hardware/software combo on the planet, if my M1 MacBook Pro had high end graphics capabilities, it would just solidify their greatness.
Truth. And for the next couple of years, they will have pros who want x86 and pros who are all-in on ARM/AS. Which splits that small segment into 2 smaller segments. I am genuinely curious what they WILL do. We know what they CAN do, but will there even be a business model for them to do some of what they CAN do. It's going to be interesting to see what they end up releasing. I'm not a person in this demographic at all, btw. Just love seeing the whole spectrum and what will happen.
 
GPU are parallel processing device and by definition they specialize on highly parallelizable tasks. This means that Gpu performance scales linearly with the available computational resources. Keep adding cores and your performance will go up accordingly. The only condition is that you scale up your other resource (memory bandwidth, caches, interconnect performance) as well, and of course that your tasks are large enough to saturate the larger GPU. This is true for any GPU out there. You can do the math for any AMD or Nvidia GPU and you will see that their real-world performance scales with the number of cores (don't forget to account for different clocks and memory specifications though!).

Since Apple Silicon is very power efficient, scaling it up should not provide significant engineering challenges. These are solved problems. So yes, 16x as many cores (128 cores, 16384 compute units) would put the hypothetical Apple GPU on roughly the same level as the 3090.



Some might (especially people relying on legacy plugins or software), many will not.

I read somewhere than GPUs can scale easier than CPU, so if Apple has an efficient GPU then could just add more GPU cores until it is at

Wow, excellent info, now to see what ends up happening. Since power draw and heat are not an issue, they could cram a LOT of GPU cores in a Mac Pro. Now, based on what customers need, it will be interesting to see what actually gets made.
 
My biggest issue is the Pro market is very GPU heavy. ML, Metal, OpenCL, Video Rendering, Color Grading, 3D rendering, ML/AI all require the latest and greatest in GPUs and will always want more power.
I hate the term "pro" in these discussions because it is really twisted.

For most of the "pro" market current M1 is more than enough and actually a bit overkill.


This is coming from a "pro" user as in I'm using my M1 mac to do professional tasks as in what I earn my living from. In my case MS office (mostly excel), mail, browser, AutoCAD, PDFs etc.


It is the niche market that do rendering and stuff like that that aren't as well served with the M1.
 
Apple Silicon for iPhone and iPad (A Series):

I have not seen this mentioned anywhere. I don't think we've seen the last of "X" series A chips. I think the stage is set for the iPhone Pro and Pro Max to have X series chips. It will further separate the devices, and it wouldn't surprise me at all to see this happen.
🤔 I never thought about this before, but now that you mentioned it. I think we’ve see the very last AX chip. From now on normal iPads will have A series chips and iPad Pros will have M series chips. Despite the name Pro all iPhones fall basically into the same category of display performance needs. The Pro Max might have more RAM, but the same number of CPU and GPU cores.

A-series: iPhone, iPhone Pro, iPad
M-series: iPad Pro, MacBook Air, iMac, Mac mini
MX-series: MacBook Pro, iMac Pro, Mac Pro
 
Those systems have massive tradeoffs such as eMMC storage at either 32 or 64GB RAM, 2-4GB RAM, lower resolution displays, and significant keyboard/trackpad flex.
Those are all chromebooks, and *aren't* enough to run Windows. The recently deceased Windows X might have been capable of running on them, but I wouldn't want to do it.
 
Honestly these days something like the MacBook Air or Pro running M1 is enough for me. The only reason I didn't get one is because it doesn't supports dual displays alongside the lack of a redesign.

If I do find the Air on a good promo I'll get it to carry it with me everywhere.
 
Those are all chromebooks, and *aren't* enough to run Windows. The recently deceased Windows X might have been capable of running on them, but I wouldn't want to do it.
They aren't all Chromebooks - there are Windows laptops at Best Buy with those specs right now. And yes, they run like absolute garbage, to the point where they can't even install Windows updates because there isn't enough free storage space.
 
Round 2 pushes up to the Pros.
And M2Z for the larger iMac (presumably now called the iMac Pro so we don't have to say 4K and 5K or 4.5K and 5.5K, but just iMac and iMac Pro), (I won't speculate about hardware beyond this: 27" minimum, 32" maximum, 30" likely, 5.5K? who knows)..
I don’t understand why we think a redesigned larger iMac will be called the iMac Pro. We had 2 sizes of iMac for a decade which the larger model was popular with a wide mix of user types, not to be just for pros, but anyone wanting a larger screen and better performance. The reason the iMac Pro was end of life was that regular 27” iMac could be configured to equivalent performance, no longer needing to be around.
Reference this site
  • At launch, the iMac Pro was the most powerful iMac, but it never received a major update following its 2017 launch and has been discontinued as of March 2021. At the time of its discontinuation, the standard 27-inch iMac offered a much better value, while the most demanding pro customers should look to the Mac Pro.
I am generally annoyed by this site not properly separating the M1 based 24” iMac and the current intel 27” iMac from its buying recommendations. We all know a 30” or 32” model is forthcoming and is due 2nd half 2021.

There is also the possibility of a mid-performance desktop without monitor with similar configurations discussed like a low cost MacPro offering.

None of the current 27” iMac users consider the capability of 24” M1 as adequate for a number of reasons especially GPU intensive software. Even my old 2015 model is significantly faster with OpenGL/Metal scores. So yes pretty much any newer iMac 27” model is that much faster utilizing discrete GPUs.
 
Last edited:
  • Like
Reactions: MacCheetah3
They aren't all Chromebooks - there are Windows laptops at Best Buy with those specs right now. And yes, they run like absolute garbage, to the point where they can't even install Windows updates because there isn't enough free storage space.
No way there's a Windows laptop with only 2G of RAM. They tried 32G of disk once, but that's not big enough to do updates.

I just checked Bestbuy and it's as I thought, there are no Windows laptops smaller than 4G RAM and 64G disk. (at least in the U.S.)
 
I hate the term "pro" in these discussions because it is really twisted.

For most of the "pro" market current M1 is more than enough and actually a bit overkill.

This is coming from a "pro" user as in I'm using my M1 mac to do professional tasks as in what I earn my living from. In my case MS office (mostly excel), mail, browser, AutoCAD, PDFs etc.

It is the niche market that do rendering and stuff like that that aren't as well served with the M1.
Yet businesses are still buying 16" MBP's and 27" iMacs because they need the larger screen size and its overall performance for business/engineering usage. If you do a survey here, I bet you find most of those users, not satisfied with the present 8 core M1 performance. They want a more capable Apple Silicon powering what they buy next. Yes the "pro" term is thrown out as a excuse to say everyone should like the M1 except pro's. Bah! :D
 
No way there's a Windows laptop with only 2G of RAM. They tried 32G of disk once, but that's not big enough to do updates.

I just checked Bestbuy and it's as I thought, there are no Windows laptops smaller than 4G RAM and 64G disk. (at least in the U.S.)
Wait a month - when Back to School starts, you will see multiple Lenovo laptops that will be in the 11"-13" screen range and include one year of Office 365 with 2GB RAM and 32GB eMMC storage. I worked there for over four years, so I've seen (and tried to stop) people buying those cheap pieces of crap. What's worse is that some of those models will be carried through the Christmas shopping season since some people think cheaper is better when it comes to laptops.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.