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.
The final blow for real gaming on the Mac, or the beginning of a new era?
- Thread starter Frankied22
- Start date
- Sort by reaction score
Earlier in either this thread or another there were talks about Apple being able to match the TFLOP numbers of the new stuff coming out by just scaling how many GPU cores are present. Nevermind actual gaming performance (because Nvidia tends to have lower TFLOP count but better rendering performance than AMD).
I know this forum isn't as graphics geeky as Beyond3D is, but it seemed like some here would appreciate the die shot.
I guess will see if Apple is willing to make/pay for that much silicon for their GPU's in their products, considering that chip is basically 3x the size of their current largest Apple Silicon chip the A12x....
End of dedicated GPUs = end of any hopes for AAA gaming. Sorry. You know this is true. But no one can forbid you to live in your fantasy world, though.
What do you guys think about this quote/post in B3D Forum?
forum.beyond3d.com
Apple (PowerVR) TBDR GPU-architecture speculation thread
They're unlikely to license any microarchitecture which the recent press release talked about due to its lack of "multi-use" language, it's probably again a wide-range architecture and IP license as an extension of what they're already had for half a decade. Also note people saying Apple broke...
forum.beyond3d.com
What do you guys think about this quote/post in B3D Forum?
Apple (PowerVR) TBDR GPU-architecture speculation thread
They're unlikely to license any microarchitecture which the recent press release talked about due to its lack of "multi-use" language, it's probably again a wide-range architecture and IP license as an extension of what they're already had for half a decade. Also note people saying Apple broke...
I think that professional users and tech enthusiasts do a great job convincing regular users to use computers.
If they are left out of the process, the idea that Apple devices are a toy will only be strengthened.
Apparently, even with the new GPUs, Apple devices are not as impressive as Apple paints them to be, so...
What do you guys think about this quote/post in B3D Forum?
Apple (PowerVR) TBDR GPU-architecture speculation thread
They're unlikely to license any microarchitecture which the recent press release talked about due to its lack of "multi-use" language, it's probably again a wide-range architecture and IP license as an extension of what they're already had for half a decade. Also note people saying Apple broke...
CAD developers who care about performance have transitioned to DX long time ago. Metal is unlikely to support mesh shaders any time soon as they don’t do much for TBDR (although Apple could hack around it I suppose), but Metal is perfectly capable of generating geometry on the GPU.
Regarding ML or HPC, of course it’s capable of doing it. There is just no tooling. Apple’s ML accelerators are state of art, but they focus on the common app developer needs, so the API is limited. There is no doubt that frameworks like CUDA offer a vastly superior infrastructure, simply because there are so many more people using it.
I believe that the most important adoption factor is going to be performance. If Apple manages to pack a mid-range GPU equivalent in their popular laptops across the board, third party support will follow.
I think it's pretty reasonable to suspect gaming on an AS Mac will at least be better than gaming any other Mac besides a Mac Pro. When you consider the performance of games on an iPad, I just don't understand how Apple is going to make more powerful ARM chips with better cooling in a bigger chassis, and then we're not going to see better performance than that. An iPad Pro 2018 could run Fortnite at 120fps (when it was available 
), so at the very least I think you will see good performance on these mid-range GAAS games that are well-optimized and not super demanding.
I think though, considering the dedicated GPU iMac rumor, that older Mac gaming line rumor, and the WWDC developer videos, that Apple may announce gaming Macs in the next 1-2 years as part of a new initiative. There's so much money to be made and it's the one major market that Apple hasn't made a huge push into yet.
), so at the very least I think you will see good performance on these mid-range GAAS games that are well-optimized and not super demanding. I think though, considering the dedicated GPU iMac rumor, that older Mac gaming line rumor, and the WWDC developer videos, that Apple may announce gaming Macs in the next 1-2 years as part of a new initiative. There's so much money to be made and it's the one major market that Apple hasn't made a huge push into yet.
As a whole, B3D seems to be down on TBDR for desktop. I wish someone there could give better insights as to why.
You know, the definition of mid range is quite flexible. Personally, I would classify 3070 as high-end. I was talking about something more like 2060. This is the bracket of higher-end Apple (consumer) desktops. Only the Mac Pro has faster GPUs, but they come with their quirks anyway, and they certainly don’t come close to the super high end gaming units.
Thanks for the reply. I appreciate the die shot too. I'm into all that stuff also. I just wanted to know how it related to the topic at hand which you answered quite nicely.
It seems to be a decent look at why Macs have fallen behind in 3D applications. Even if a Mac had a capable GPU, if the frameworks that this software needs aren't there, then the point is moot.What do you guys think about this quote/post in B3D Forum?
Apple (PowerVR) TBDR GPU-architecture speculation thread
They're unlikely to license any microarchitecture which the recent press release talked about due to its lack of "multi-use" language, it's probably again a wide-range architecture and IP license as an extension of what they're already had for half a decade. Also note people saying Apple broke...
Maybe this is why Apple seems to have a renewed focus on expanding Metal. Who knows if anyone's gonna adopt it or not but it's at least a strategy to expand your own API with frameworks.
well let us hope by the rumored October event Apple decides to reveal a list of software developers they are working with to insure their is a vibrant selection of software to select from and that it includes game companies, both that have games on Mac now and those that do not.
I don't need a faster chromebook and that is what this will be to me if I have to have a separate machine to play games when I want to relax.
I don't need a faster chromebook and that is what this will be to me if I have to have a separate machine to play games when I want to relax.
I don't think the comments from this poster are unbiased. The guy appears to dismiss TBDR and fails to note its advantage in perf/W or perf/TFLOPs. He says that the direct renderers already work optimally, but they don't, hence why TBDR exists. Apple seems to think that TBDR is simply better and is the future, I suppose they know better than this guy.What do you guys think about this quote/post in B3D Forum?
Apple (PowerVR) TBDR GPU-architecture speculation thread
They're unlikely to license any microarchitecture which the recent press release talked about due to its lack of "multi-use" language, it's probably again a wide-range architecture and IP license as an extension of what they're already had for half a decade. Also note people saying Apple broke...
He implies that it would be hard for Apple to implement ray tracing, but I can easily see Apple adding ray-tracing hardware. I'm not sure what the problem with "render passes" is, but it's certain not the reason why most of the pro software don't use Metal. It's simply a question of user base.
He also lists some niche usages, like particular ML inferences, that Metal doesn't currently do, but as other have pointed at, these usage are mostly irrelevant to Apple.
The thing about TBDR is that it is an unknown entity on desktop. Immediate renderers have proven themselves well, and they became much more efficient since they started implementing a limited form of tiling. TBDR is also not directly compatible with newer techniques such as mesh shaders since TBDR requires the geometry information to be stored until the rasterization phase, which takes away the advantage that mesh shaders bring in the first place (although I suppose Apple could partially circumvent it by implementing larger dedicated on-chip geometry buffers). People also worry that TBDR won't be able to scale with more complex geometry, since it in some way trades geometry processing efficiency to win in the fragment processing efficiency department.
That said, TBDR also bring some very interesting benefits. For me the main story is not even about more efficient execution, but about the fact that TBDR makes strong guarantees about memory locality. Thing enables features like Apple's programmable on-chip tile cache, which makes it very easy — and fast — to implement some pretty advanced things.
Yeah, that is a weird thing to argue about given that Imagination already has a hardware-based raytracer for their TBDR GPUs.
That said, TBDR also bring some very interesting benefits. For me the main story is not even about more efficient execution, but about the fact that TBDR makes strong guarantees about memory locality. Thing enables features like Apple's programmable on-chip tile cache, which makes it very easy — and fast — to implement some pretty advanced things.
Yeah, that is a weird thing to argue about given that Imagination already has a hardware-based raytracer for their TBDR GPUs.
My litmus test for an AS Mac is two fold.
Catalina compatible Steam titles and Steam come across with similar or better performance
Apple Arcade offers titles of similar renown and complexity that Microsoft Game Pass does
If I have to buy a separate machine to game; frankly I think anyone who suggests that and still buying a Mac is more caught up in image than reality; I will just buy a different solution entirely. A computer with these supposed capabilities should be something everyone wants to write software for and people want to buy for said software because its price and performance is on par or better than the competition.
Catalina compatible Steam titles and Steam come across with similar or better performance
Apple Arcade offers titles of similar renown and complexity that Microsoft Game Pass does
If I have to buy a separate machine to game; frankly I think anyone who suggests that and still buying a Mac is more caught up in image than reality; I will just buy a different solution entirely. A computer with these supposed capabilities should be something everyone wants to write software for and people want to buy for said software because its price and performance is on par or better than the competition.
But that’s the reality; if you want a Mac you aren’t really getting it for gaming at this point. Not everyone is interested in gaming on their Mac, it’s not necessary to be myopic.
I disagree.
A lot of us like the ecosystem Apple provides, including unified iCloud, password, and other features for our general computing. It’s not an image thing, it is us continuing to use the computers and devices that we already know, and like to work with. I simply can’t get that level of integration using windows with my Apple ecosystem. Therefore, I will stick with a Mac for my general uses, rather than a PC.
Of course, some things I can‘t get my Mac to do, they pretty much require a pc.
For some reason, the breath of niche software isn’t sometimes available of macOS users. I have a couple of radio scanners I use to listen to the local aviation airbands, and use to take to watch auto races to listen to various frequencies during the race. No one simply makes a macOS version of any of the scanner apps. Why? No technical reason that I know, the various software will run on a cheap windows PC stick, but not on macOS. All I can guess is that only a few people make scanner software, and they simply have no interest in porting it.
It isn’t always about power, but sometimes just the willingness of a community to develop (or refuse to port) software.
While video games aren’t as niche as the above, they also don’t historically get most games at all, if they do get some top sellers it is usually months to years later. (Even worse is situations like Bioshock for iOS, sold me a game that only worked for two months before it broke on iOS and abandoned.) While I would recommend an Xbox to most people instead of getting a gaming PC, I understand some types of games play better on the PC, or have entrenched mouse/keyboard players who refuse to pick up a controller. In that case having two computers could be worth it to them.
Lately, the biggest “game” has been Microsoft Flight Simulator 2020. I don’t think there will ever be an macOS port. The competition (Xplane 11) does run on macOS, but certain add on hardware you can purchase only has windows plug-ins and not macOS or Linux. Why? No reason, other than the manufacture didn’t think it was worth the effort.
I currently have bootcamp set up, but I know that option is going away. I will have to make some decisions once I am ready to replace my current iMac.
Of course, something like MSFS 2020 tends to be future looking, it wil struggle with any of today’s current pc hardware, it has been designed to run best for computers and graphic cards that haven’t even been released yet. Apple (except a $5k+ Mac Pro) doesn’t have any hugely expandable/replaceable component computers anymore. It simply a niche that Apple doesn’t want to fill. For flight sim enthusiests, who need the option to expand to multiple screens, graphic cards, and deal with windows only hardware, having two computers (mac for everything except flight sim, and a flight sim only computer) could make sense.
I look at is as using the right tool for the job. I wish Apple covered all of my needs, but they simply don’t. Windows doesn’t either however. It is up to each person to decide if one device only will work, or if they should have a foot in both Windows and macOS.
Hopefully, once the arm transition is complete, they will wow enough game developers and others that Mac gaming becomes a viable option alongside windows/xbox/playstation, but I have had that hope before and it never happened.
I don't necessarily disagree with that, but I think everyone here buys Macs for a myriad of other purposes, and gaming is a bonus. It all depends on how important gaming is to you, if you're a big gamer, and can only afford one machine, well then a PC is for you. That's kinda the way it's always been, for better or worse.
I suppose that one could not get rid of their Intel Macs for a while, just to do a "wait and see" on what happens with Windows on ARM, and any possible emulator/translator. Run Windows on the Intel Mac for a while until it resolves either way, and then decide if you want to get a PC, or load Windows or ARM, if it becomes avaialble.
Mainly that SeriesX and PS5 are both SoC designs with unified memory similar to Apple's A chips. It is very possible to generate an SoC with a massive amount of die space given to the GPU and get very good compute/rendering performance results.
Naturally, game consoles are the standards that most AAA games are mastered to now-a-days. So having a console design in an A series chip is possible. Of course, the likelihood of HBM2/GDDR6x memory on an A chip is very unlikely (at least on consumer macs), but might be possible for iMacs or Mac Pros. These machines could have very good compute/rendering performance.