Silicone is for breast implants - silicon goes in integrated circuits
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Report: Apple Silicon iMac With Custom GPU to Launch in Second Half of 2021
- Thread starter MacRumors
- Start date
- Sort by reaction score
This makes me very excited.
So far Apple's engineering team hasn't missed - what they have accomplished on tablets is simply amazing. Combine that with the gaming market's explosion since Apple adopted Intel architecture and I fully expect Apple to make a push into the gaming sphere with AS Macs, even if people don't expect it now. This is the most exciting time for Apple in the last 5 years, very excited to see how these computers will turn out.
So far Apple's engineering team hasn't missed - what they have accomplished on tablets is simply amazing. Combine that with the gaming market's explosion since Apple adopted Intel architecture and I fully expect Apple to make a push into the gaming sphere with AS Macs, even if people don't expect it now. This is the most exciting time for Apple in the last 5 years, very excited to see how these computers will turn out.
The current Retina 21.5" iMacs don't have a Radeon Pro 450 in them either. The "edu" , non-retina iMac is a cruel joke with an MBA processor in it. It isn't even close to the same league the modern 21.5" models are in.
Which Intel GPU? Intel is about to drop Gen 12 ( Xe-LP ) iGPUs in "Tiger Lake" .
Comparing back to GPUs designs from 2-4 years ago isn't saying much.
MBP 13" 2020 Gen 11 graphics ( 10nm 10th gen ) is 10519
2020 MacBook Pro 13-inch gets a boost from an eGPU
real world speed test results for performance minded Macintosh users
barefeats.com
That is better than the Radeon Pro 450 too.
Gen12 (Xe-LP , ) likely would be generally 30-40% faster still. ( very good chance Apple won't do a laptop with one. And slim chance Intel does a discrete add-in-card GPU card drivers for macOS for Gen 12 SG1 card. Intel should (because mainly good as a content video transcode card. But Intel and Apple may not see eye to eye on importance/priority. And lots of "break up" drama. )
IHMO it is doubtful that Apple is aiming at the 74K and up benchmark range on their GPU. They'll just use AMD in that space. The GPU in the 5700XT will be superceded by second half 2021 with something even faster at the same bill of materials cost range. Apple doing silicon to cover that doesn't make much sense at the unit volume that GPU die sells at in the Mac space now.
A14 era graphics should move up from A12Z scores, but Intel is dramatically moving up right now also. They may not "huge gap" Gen 12 at all for the "entry" Mac SoC. ( and hide that in Mac-to-Mac benchmarks because not going to use Gen12. )
This is going to be a really interesting product to see. If they can make a really powerful desktop CPU/GPU, that's gona cause some waves for the industry. cantwaitcantwaitcantwaitcantwait....
I thought there were slim chances Apple were going to use Intel Xe GPU. Looks like they are going All-In.
The text below the chart in the first thread post says.
"... n contrast to Kuo's prediction, today's report claims Apple's first Arm-based Mac will be a super-lightweight 12-inch MacBook, codenamed "Tonga," that will launch by the end of 2020. ..."
Apple doing a "one or two port wonder" MacBook rebirth wouldn't be all that surprising. Especially, if they left off Thunderbolt in this rebirth also. Put their "generation 4" butterfly keyboard on it and float that "thinnest of thin " system out there once again. That system really wouldn't be that much different than an iPad Pro (which pragmatically supports two ports now when extra fancy keyboard is attached. One for keyboard data and one external board. )
IMHO, I suspect Apple is going to want to 'prove' the MacBook wasn't the 'wrong path' and use Apple Silicon to take another whack at it with more electronics flexibility internally.
That isn't necessarily the same Mac SoC that would go into a MBP 13". Apple can get a "beefier" SoC the same way the A14X is bigger/beefier than the A14. Just add bigger system cache and more GPU cores of same basic design family.
(or GPU cores turned on versus off if they don't run into package size issues. But probably different die in different package like the 14X is off the baseline. Bigger system cache can perhaps crank the clocks a bit higher too on the CPU cores (without having to shift to different microarchitecture. ) )
Apple could be doing three closely related SoCs.
MacBook ( no Thunderbolt. And just use excuse didn't have it before as to why they skipped it here again. Mac Boot support as an option but pretty close to the 14X die if not a 'feature a , b, c' turned on/off separation. )
MacBook Air ( bigger die and bigger package , 2 Thunderbolt port ( one discrete controller so wider PCI-e lane support ) , Bigger GPU core count (but not max ). )
MacBook Pro ( same die but bigger package, more PCI-e line pad/pins out to flush out more ports, higher clock count , Max GPU core count for this family of GPUs. )
Apple has been on the Captain Ahab trek after the "thinnest of thin" laptops for a long while. Mac specific silicon is exactly what they need if they want to continue that epic quest. Apple knows that the large bulk of their laptop buyers don't want this, but there is a large enough subset that do and Apple wants to do it.
People shouldn't completely "freak out" at the first Mac that Apple ships with Apple Silicon. It is highly unlikely that it is going to be completely represent what the rest of the Mac line up is going to be constrained with. More likely it will be the better match silicon for that specific product (not for the whole line up. Other systems will get better matches to them. And that is also going to a take a decent amount of time to roll out over the whole line up. )
It won't be very surprising though if Apple keeps one of the Mac SoC pretty close to the iPad Pro's. It would be a way of doing much higher volume of the exact same die. The iPad Pro has zero processor variation in the new product line up ( 11" versus 12.9" screen .. exact same die). Apple adding another product to crank even higher volume on that die would entirely within their common practices for component reuse. Other Mac products would share SoC dies between themselves too. ( but the very bottom edge of Mac performance range, overlapping with the iPad Pro probably isn't a big deal. )
An Xe GPU on the iMac ? Probably no chances at that even if there were not switching.
Other than super odd-ball non- retina model which is so lame they could just leave it alone. I suspect when Apple goes to a more mainstream 4K display panel baseline that they non-Retina model will just disappear. ( "Retina" should be any sort of price premium at this point in the feature evolution timeline. It is just a mainstream apple feature. Even the MBA has it.
Desktop Xe GPU would be waiting on the "Frankenstein" , Rocket Lake products from Intel if shooting for 2020.
And would be an odd ball fit to what is left of the iMac line up still waiting.
Android's fumbles with Tablet experience enhancing aspects is nothing like Windows on x86-64 ( or even ARM ).
Software has hobbled alternative tablets to some extent.
Apple's silicon engineering team has spend its time shooting at just one SoC in every product grouping at a time. (if that.)
Bleeding edge iPhone gets one single SoC. (rest of entire line up gets "hand me downs")
iPad Pro gets a SoC periodically (not every year). Rest of iPad line up gets "hand me downs".
Apple Watch SoC tweak most years.
AppleTV " hand me down".
HomePod old "hand me down"
iPod comatose
One of the reason they have been successful is that most of their competition delivers 4-15x as many SoCs as they are on any given cycle. For Apple it is very often same SoC die is just reused in multiple products to get coverage. With a very long lead in front and close system requirements ( hardly any ports . one and only one screen , etc. ) that works. If coming from behind ( trying to match status quo) and competitors are working at a 12 month cycle pace ( e.g., AMD now and Intel of several years ago) that path probably won't work as well.
Part of Intel's problems is that they probably are spread too thin, but have built of base of making "everything for everybody" in the CPU package space.
Eh? Driver here is Intel or games on iOS flowing back to macOS ?
This GPU for an iMac has a pretty good chance of being the GPU for the smaller screen option. That hasn't been a "super gamer" system historically. It probably won't be one going forward either.
Apple is probably looking to sell lower priced GPUs at higher volume in some systems. (e.g., MBP 16" , iMac 21-24" ). Not max BTO 27" model or anything in the iMac Pro to Mac Pro range. Quite unlikely that they are going to try to do a full broad spectrum, low to ultimate gamer GPU line up at all
Mostly iGPUs on die and some supplemental dGPUs where stretching the on die GPU gets a bit painful internal bisection bandwidth wise ( raster display output competing with computation core count data stream competing with AI/ML compute stream ) .
Yep, just go for that gaming PC separately if you want one. You'll get more perf per buck anyway.
I hope not, I don't want to loose all the connectors so its thinner and more water proof.
Then get ready for disappointment:
("No direct booting == no bootcamp")
...although we don't know yet whether that means that Apple will be actively blocking that or simply not supporting it - but even then there's no guarantee that any existing OSs will be able to run "bare metal" on an AS Mac - there's more to computer "architecture" than just the processor instruction set and, at the very least, an OS will need Windows and/or Linux drivers for Apple Silicon graphics, storage, sound, networking... Bear in mind that Intel Macs basically are PCs whereas we don't know what AS Macs are going to have in common with other ARM systems.
We do know that there will be virtualisation support for ARM Linux since Apple showed Parallels running Debian Linux at WWDC. A hypervisor like Parallels can act as a bridge between guest OSs written for other ARM hardware and the "real" hardware supported by the MacOS drivers. So whether you can run Windows 10 for ARM in a VM on AS Macs will largely depend on whether Microsoft deigns to license it (last I looked, it wasn't offered as a consumer product, only bundled with a Surface Pro X). Since AS Macs will outnumber ARM-based Windows PCs approximately 5 minutes after they go on sale, I suspect Microsoft will want in unless they get into a... water-passing contest with Apple over something. Win10 for ARM has an x86 emulation (currently 32-bit only, 64 bit supposedly in the works) so that will probably be your route for running x86 Windows software (...and should be far more efficient than trying to run the whole of x86 Windows under software-emulation).
As for x86 Windows... it will be a choice between full software emulation (Ahh, back to the good old days of watching SoftWindows slowly grind away - although emulation has come a long way since then) or virtual desktop talking to a "real" x86 in the cloud.
Still, you've got another 3 years or so - before you have to jump to AS - to either kick the Windows habit, buy a PC or spin up a PC in the cloud.
I was under the impression that the A14x would, like the A12x, have a built in GPU: "the A12X was simply an 8-core GPU chip with one GPU core disabled" If the GPU core is enabled that been this thing will kick butt.So another year before an Apple Silicon iMac is released. Add another couple of years before I consider it safe to buy myself one ...
Li'fuka
Last edited:
There is a large amount of energy with computers operating 24 hours a day with screen dimming still in companies.
You're right this is likely some green initiative to make computers be a lot more energy efficient. I am interested with how this will work out was far as doing things now and with enhanced performance CPU and graphics. I am not convinced by Apple's simple demonstrations and promotion of this by WWDC demos. I will wait until this has been explored more with real world examples.
Bootcamp very probably yes. ( technological gaps ).
Windows 10 probably not if willing to accept it in a virtual machine. ( biggest gap is Microsoft licesning .. which isn't a technical problem. More a business problem ). Windows 10 on ARM has increasingly been getting better virtualization support. So putting it on a virtual machine on a Mac isn't a huge jump from the Hyper-V and other virtualization that folks have done ( stuffed on QEMU ). But Microsoft doesn't sell it to mainstream folks detached from a system. there are some specific virtual driver issues to be worked out for stuff like Apple's trackpad for Apple specific features. ( so may be some Microsoft - Apple - VM vendor finger pointing as to who suppose to fix what when there are some quirks. ) . Microsoft Azure probably is quite keen on selling time on Windows 10 on Arm virtual instances.
Windows 10 Arm on VM probably isn't a "day 1" feature. But 4-10 months in after there is several million systems out in the wild ... I wouldn't beat against it. If Microsoft can make money selling licenses they probably aren't going to be a blocking factor.
VMWare has some interesting collection of things on the fusion blog stream.
desktop hypervisor Archives
Their VM upcoming working on top of apple's hypervisor system. ( no surprise. Parallels wasn't going to be the only player.)
Their Workstation (Windows) product can play nice with Hyper-V on windows too. ( So probably not a funny quirk that the Mac variant has to deal with but the Windows side ignores. Both can run in the context of OS vendor provided Hypervisor systems )
They have pulled the graphics rendering out of needing lots of expensive low level host calls.
As long as the VM presents a virtual UEFI foundation and palatable virtual components to Windows 10, it should run pretty well virtualized on top of macOS.
Now some folks will still be disappointed with running Windows 10 only in a virtual machine context. Especially, if the GPU is highly simulated ( rather than close to 'raw' on the hardware). However, that isn't "dead on arrival". It works. just not happy with how it works.
Apple needs to do lots more work to make that much, much 'thinner" ( faster ). That probably isn't coming short term, but also isn't necessarily dead in the water either (at least if Apple isn't lazy) .
This.
Apple is putting macs out into a far off island, with the few inhabitants pleading 'come on, join us, its so much nicer out here!'.
Meanwhile the real world just shrugs and moves on.
Last edited:
Highly doubtful. For better or worse , many Thousands of Mini's are racked up in custom slots for the current form factor. For example
Cloud-hosted Bare Metal Mac
MacStadium Bare Metal is dedicated, genuine Apple hardware hosted in MacStadium data centers.
www.macstadium.com
Telling those folks they have to rip out all of their custom stuff and go to another new funky form factors will make several really big buyers unhappy. Lots of small dev shops with a mini(s) off in the corner enclosure/shelf/stacked doing something too. A major contributing reason why it didn't chance much in 2018.
More likely apple is going to stuff "more compute and GPU grunt" into a Mini rather than massively change the form factor. Also just run cooler and quieter at the better performance numbers.
The big gap to cover versus the DTK is to add back in the "Mac" ports. Thunderbolt 4 (or maybe 3 if Apple is dragging on first iteration. ). Headphone jack. 10GbE option.
The PHYS adapter chips for the ports will take up similar space as they do now. Handling decent power on the ports also. DIMMs for RAM and a NAND daughter card for the SSD would avoid the crippling constraints the laptops (and iOS-ipad devices have).
Apple going to an even more "micro" Mini with thinner height and/or even less ports . That probably won't buy as much as it hurts in sales. There are at least as many folks would have a bigger "mini" ( more like mythical xMac ; iMac fratricide product ) than want something smaller.
I have no idea why people keep chasing after Apple and a Mac for their gaming needs. I have owned Macs since 1989 and enjoyed some games along the way, but never deluded myself into thinking a Mac will ever be a substitute for a console or a gaming PC or thought Apple would suddenly make that market a major point of focus and investment. It’s just not them.
There is little to point to Apple completely blocking virtualizing Windows.
Apple GPUs sold in 2019 .... about 100M.
Apple GPUs sold in 2020 ... about 100M .
Game developers putting in work targeting Apple GPUs numbers in the thousands now. That isn't a tiny island. Bigger versions of those in Macs serving as system GPUs will incrementally grow the island; not make it smaller.
What Apple is largely doing is replacing Intel in the line up. There already are three major PC GPU players.
Ahead of Ampere, Nvidia Grows to 80% of Dedicated GPU Market, Analysts Report
COVID-19 may be affecting the market.
In the Mac transition to Apple Silicon, they are going primarily going after Intel's chunk. Nvidia's is already crossed off the list in the Apple ecosystem.
if want portable games that work across the whole Apple ecosystem of products then the Apple GPU is going to work better.
Now there are certainly folks who have a 180 degrees shifted world view. That looks more like this.
but folks vastly fooling themselves if think that is the way Apple looks at the world. It isn't. That is peepholing to a different room. Vast bulk of Apple Mac sales are laptops, not "boxes with slots".
Apple probably isn't gong to chase the high end gamer market on the Mac systems. It is small and lots of good reasons to leave that to AMD ( or perhaps Intel if they ever manage to get off the ground in a substantive way in that space. )
A chiplet pragmatically implies shared the same memory controllers as the other "chiplets". At that point calling it a "dGPU" rather a huge stretch. A chiplet is more like a bigger die chopped up into smaller pieces.
A dGPU should be something that has its own primary working memory system. That runs counter to Apple pounding the drum on "unified memory" CPU-GPU solution as being a key point of their objectives.
The programming model would need to shift. Multiple models for Apple GPUs will be a bit of a dual edged sword in terms of code comlexity for the apps.
In the context of an iMac variant or MBP ... a separate logic board? Errrrr no. It isn't now with the modern Intel designs so why would that change with Apple Silcon. Especially with their own custom designed GPU. If custom designed then extremely like part of the customization is to embedded it into the specific Mac it is going into.
MCM works if the thermals work. if both pretty close to being about the same. If vastly different then painted right back into the corner the Mac Pro 2013 was in. Doubtful Apple is trying to go back to the same corner.
There is a limit to how much stuff Apple can roll out on 5nm given the large volume the iPhones are going to soak up and the maturity level of TSMC's set up. (and the other customers who are lined up to buy wafer slots ). Apple gets first dibs but they can't starve out everyone else for double digit months of time.
Going to Apple Silicon is a solution in search of a problem.
Losing the ability to boot into Windows to gain what? A bit more speed.
I was there for life when Apple was off doing its own thing while Intel ruled the world. It nearly killed the company. Intel architecture still rules the chip world. Going back to that way of life is the definition of insanity.
Losing the ability to boot into Windows to gain what? A bit more speed.
I was there for life when Apple was off doing its own thing while Intel ruled the world. It nearly killed the company. Intel architecture still rules the chip world. Going back to that way of life is the definition of insanity.
Congratulations! You got everything wrong.
Careful you are not using anything that is Apple Silicon based (not the A12 thru A14 ARM used in iPhones and iPads) with existing x86 programs/drivers running Big Sur Beta. This is just your assumption that everything will be great. We can hope.
Last edited: