Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
They have to do multiple chips to handle multiple RAM sizes and multiple GPU cores. That's going to make their SKU management more difficult, unless they're so cheap that they can make tons of them and keep them around.
 
You’ll end up with thermal issues... that’s the constraint you hit when just throwing more cores in, I think.

Bigger chips and more cores will obviously consume more power and produce more heat... but Apple has a massive advantage here of requiring less than half power compared to their competitors. An 8-core Apple GPU draws 10 watts at it's peak. An 128-core Apple GPU will therefore draw around 150-180 watts. That's 16384 shading units!

Will Apple be reentering the dedicated server market?

Apple's senior chip designers have left the company precisely since Apple didn't intent to enter the server market.

Intrigued what they'll use for their high-end GPU options. Is it even possible/practical to cram 32 graphics cores on the same SOC as (up to) 24 Firestorms & Icestorm cores? Or would it need to be a custom discrete GPU?

They could probably fit 32 core GPU, a 12+4 core CPU and a large shared cache on a 300mm2 die, which is not too large, compared to the current GPUs.


The current 8-core GPU gives us a metal score of >20.000 and >2.5 Tflops. Does this mean a 128-core GPU wil score >320.000 and >40 Tflops?

A 128-core Apple GPU would be faster than anything currently on the market.


An Apple GPU that's NOT on a discrete graphics card seems unlikely to be performance competitive with a 3rd-party AMD/NVidia/Intel discrete graphics card GPU. Therefore: a new 2022 Mac Pro with an ARM CPU (instead of an Intel CPU) would likely still require a graphics card of some kind.

What is your basis for this statement? Apple should be able to make relatively large chips while still being commercially viable (even if the yield rate is low, it is still going to be cheaper than buying them from Intel+AMD). And there is always the chiplet route if they feel that a single die is not large enough.

I am certain that Apple will not reintroduce the traditional dGPU in it's new Macs — those break unified memory, and low-level latency data exchange between CPU and GPU are one of the Apple Silicon key selling points.
 
I think this is great for the future of the Mac line. I will be upgrading my MacBook Pro 16 in 2 years. By then Apple will have all the hardware bugs for SOC located and Applications will then be native ASIC, not having to run with Rosetta II or a bloated universal binaries.

Can not wait until December 23 when Microsoft updates Microsoft Flight Simulator 2020 with Virtual Reality. VR has been running great under Windows 10 and bootcamp with a Oculus Quest 2 VR headset on the MacBook Pro 16.
 
Not at all likely. Do you really think Apple engineers are that much smarter than AMD or NVidia GPU designers? If that were true, Apple would begin designing graphics cards that could also be used in the other 90% of the PC market held by Windows machines.

I don't know who has smarter engineers, but Apple has more money, they have smarter GPU technology and they don't need to compete in the IHV market. Apple is free to implement new features without any fear of ecosystem fragmentation simply because they do not have any competition.

And no, Apple will not sell GPUs for the rest of the market. As they have said multiple times, they sell the product, not the chip.
 
Yes, the M1 is top notch...and based on tech...from there we will see a linar upgrade...think about what an 20 core imac with apple custom gpu can do next late year

Think about it...the pros who wants RAW power...using windows will switch...because PRO are not limited by OS but limited by time...for PROs time is relevant...so, the macs now with impressive battery life and impressive power...can steal a lot of - both the pros and the casual users from windows and linux users
From 4 years using both windows oem and mac...now im going full mac starting next year
Whoa, go easy on the ellipses.

Pros need the tools to be available though, in my case a lot of the tools that are use are limited to Windows and will never be ported to Mac - of any flavour. therefore I fully rely on being able to run Win10 in a VM and with the M1 chip that is currently not possible. hence I cannot stay with Apple - which I prefer for the rest of my workflow.
 
Whoa, go easy on the ellipses.

Pros need the tools to be available though, in my case a lot of the tools that are use are limited to Windows and will never be ported to Mac - of any flavour. therefore I fully rely on being able to run Win10 in a VM and with the M1 chip that is currently not possible. hence I cannot stay with Apple - which I prefer for the rest of my workflow.

Agreed on the need for software ports - but Windows is already prepping windows and office to work on M1 and there are VM companies who are porting as well. As Apple's chip progress - I don't see that being an issue long term.
 
Here’s hoping “next year“ holds and they’re able to deliver on time. Today’s Apple is far from perfect, but the Apple Silicon Macs make me more excited about computing than I’ve been since the release of the original OSX. Honestly I think a lot of WinTel people are still in denial about how groundbreaking this jump will be.
 
  • Like
Reactions: jmgregory1
All great news! looking forward to this. hope they can pull it off, and it should give plenty of time between now and then for dev's to update their software accordingly.
 
I think I'll wait for the M-16.
Seriously though, if there ever was an argument for not buying a new Mac Mini, this is it.
They basically just told us that the M1 is a Beta chip.
 
  • Disagree
Reactions: NetMage
Great news overall, bummer news for me - I have to buy a new computer in 2020 to replace my 2015 27" iMac.

Do I get the last Intel iMac?
Or the new shiny M1 Mac Mini + Big Monitor?

Both are an improvement, but which will let me stick with my 5-year upgrade plan?
 
The PowerPC transition was similarly fast and the first chip was short lived, but for the opposite reason.

The 601 was not suitable for laptops and an under-performer doing “rosetta” tasks like 68k emulation.

Within a year, the 603e was in laptops and the 604 was shipping in Quadra (now powermac) computers.

One thing that encourages developers to transition faster was that there were PDS slot cards for older macs, and even logic board swaps from Apple, to turn older macs into PowerPC macs. And those macs would also have longer shelf life because RAM was both soldered on AND expandable. (My IIsi was maxed out with memory, a sonnet PPC card, and a video card on a dual slot PDS riser).

Those early PPC chips sipped power, at the same low wattages of today’s M1 and lower. But as they pushed for higher speeds, they heated up. By the G5, they were so hot and power hungry there was no chance for putting them into laptops and the MacPro was liquid cooled...

I hope Apple doesn’t run into the same issues of heat and power trying to add more cores and GPUs. Possibly a dual chip config in the iMac and Pro would help like the dual 604s and G4s and G5s.

And I beg them to offer the old fashioned design of soldered RAM + Ram slot(s). If they are afraid the extra RAM would be slower, than make it for SWAP only. A 16GB soldered with 32GB swap seams like it would be a fast solution for the way the Mchips and OS11 operate.
 
I don't know who has smarter engineers, but Apple has more money, they have smarter GPU technology and they don't need to compete in the IHV market. Apple is free to implement new features without any fear of ecosystem fragmentation simply because they do not have any competition.

And no, Apple will not sell GPUs for the rest of the market. As they have said multiple times, they sell the product, not the chip.

spot on - Apple went through all this effort so they could control their product and software design and schedule without having external constraints from Intel/AMD.
 
The PowerPC transition was similarly fast and the first chip was short lived, but for the opposite reason.

The 601 was not suitable for laptops and an under-performer doing “rosetta” tasks like 68k emulation.

Within a year, the 603e was in laptops and the 604 was shipping in Quadra (now powermac) computers.

One thing that encourages developers to transition faster was that there were PDS slot cards for older macs, and even logic board swaps from Apple, to turn older macs into PowerPC macs. And those macs would also have longer shelf life because RAM was both soldered on AND expandable. (My IIsi was maxed out with memory, a sonnet PPC card, and a video card on a dual slot PDS riser).

Those early PPC chips sipped power, at the same low wattages of today’s M1 and lower. But as they pushed for higher speeds, they heated up. By the G5, they were so hot and power hungry there was no chance for putting them into laptops and the MacPro was liquid cooled...

I hope Apple doesn’t run into the same issues of heat and power trying to add more cores and GPUs. Possibly a dual chip config in the iMac and Pro would help like the dual 604s and G4s and G5s.

And I beg them to offer the old fashioned design of soldered RAM + Ram slot(s). If they are afraid the extra RAM would be slower, than make it for SWAP only. A 16GB soldered with 32GB swap seams like it would be a fast solution for the way the Mchips and OS11 operate.
Sound like over complicating things and making the SKUs more complicated.
 
  • Like
Reactions: NetMage
I can imagine doubling the High performance cores to 8, beyond that, will that even still be beneficial on a desktop !? Apple will then have problems to limit the SOC to one chip, this also concerns the space problem of even more unified memory, this will cause a trade-off. My priorities are optimised CPU-multithreading and more advanced GPU's. 4 (8) Cores should be sufficient for any (high-end) user.
 
  • Disagree
Reactions: NetMage
If I wrote this information in my small blog would I get a macrumors article about it? Passing this off as inside info :rolleyes:
 
The more Apple Silicon news/rumors the better! I don't care if this is "common-sense" news (it is), I just like hearing about Apple Silicon. It sparks a different side of my imagination, and I havent felt that way about Apple, or computing in general, since the PowerPC days of my youth. This is a very exciting time for Apple and some of its user base.
 
The time for the Mac X is finally coming.

It's the rumored "smaller Mac Pro", which will be considerably cheaper without Intel Tax and the need for hyper-cooling their steampunk Xeons.
Intel Tax?
not doubting Intel may be more expensive, just questioning whether Mac will be cheaper as a result. That’s not exactly an Apple thing to do, why wouldn’t they have done it with M1 Macs then?
 
  • Disagree
Reactions: NetMage
Great news overall, bummer news for me - I have to buy a new computer in 2020 to replace my 2015 27" iMac.

Do I get the last Intel iMac?
Or the new shiny M1 Mac Mini + Big Monitor?

Both are an improvement, but which will let me stick with my 5-year upgrade plan?
Not sure why you would even consider getting an Intel based Mac at this point, if you can live with some of the initial software compatibility issues. The M1 Mini is a fantastic option and will be usable as a daily work device for the next 5+ years, if you need to make it last that long. Worst case, you end up needing / wanting to upgrade to a future M2+ Mac before a new Mini has been released, you could buy a MacBook Pro and then simply use the larger monitor / keyboard / trackpad with it, likely spending about the same as you would have in total to buy just one 27” Intel iMac.
 
  • Like
Reactions: iwbyte
We're going to have to wait and see if Apple can produce these chips in quantity as the cost per chip increases as the size increases. It's all about yield. The consensus is that for non-laptop you need a desktop class CPU which means getting rid of these so-called efficiency cores. That's dead space in desktop CPUs which don't have thermal restrictions like most laptops.

In addition, Apple needs to resolve the issues with the small memory footprint available in current M1 machines to get them viewed as competitive to the Intel models.
 
  • Like
Reactions: torncanvas
Intel Tax?
not doubting Intel may be more expensive, just questioning whether Mac will be cheaper as a result. That’s not exactly an Apple thing to do, why wouldn’t they have done it with M1 Macs then?
They did. The M1 Mac mini is $100 less expensive than the Intel Mac mini. The two M1 Mac notebooks kept the same price point but had many advances included that wouldn't have been possible at that price point had they kept the Intel chips.
 
The time for the Mac X is finally coming.

It's the rumored "smaller Mac Pro", which will be considerably cheaper without Intel Tax and the need for hyper-cooling their steampunk Xeons.
As exciting as it is, I remain very pessimistic about the price. I don't believe the Mac Pro costs what it costs because of Intel. It costs that because of Apple. Intel has no bearing on the $1000 monitor stand.

IMHO, a Mac Pro mini or whatever the concept is, should aim to start at $1,999 and BTO up from there.
 
  • Disagree
Reactions: NetMage
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.