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.
Major gaming news at WWDC: Metal and New games
- Thread starter Homy
- Start date
- Sort by reaction score
I know Apple would never do this but let's just do a theoretical exercise. The Steam machine is said to be stronger than the steam deck which puts it around 9 tflops. If Apple actually wanted to create an Appletv Pro or gaming console hybrid would they need to stick a M2 Max or M3 pro or something similar into this machine? How would the pricing be?
I'd say the conversation we were having was the interruption for the thread, yours is the return to regularly scheduled programming
You know even I fall into this trap myself. In a post earlier in this thread and in other musings elsewhere I said Apple needs to release a certain piece of hardware with capabilities X to make inroads into gaming. But really that's not true. I mean don't get me wrong, better hardware for PC games so enthusiasts can play at decent to high graphics settings would be welcome but what separates mobile gaming from PC/consoles is culture not hardware. Take the Nintendo Switch - an incredibly successful console with great first party content and decent 3rd party content. Absolute potato of a chip inside. It is projected that the Switch 2 will be even more successful with only very modest upgrades in the hardware to the point that an A19 Pro is very likely to be a more powerful chip, especially when the Switch 2 is in handheld mode.
Basically Apple could make an "Apple console" with a base Mx or even A19 Pro*, sold with game controllers, and marketed as a gaming console. The hook would be, does it have a great slate of games that attracts the people who play such games. In other words, the actual cost would be getting great gaming exclusives with both an initial slate of games and a full calendar of games to come with a great marketing plan about the games not just the hardware to grow the audience on Apple hardware that plays those games. Then developers want access to it and will port. At the moment, Apple appears to confuse the total addressable market for Mac games with the total market of people who have mac/tv/iPad/iOS despite that most people who bought those devices didn't do so with the purpose of buying and playing PC/console games even though a large fraction of those devices (especially in the last couple of generations) technically could on a hardware level. They bought something else for that purpose.
*admittedly the weaker the hardware the better the games have to be, Nintendo "gets away with it" primarily because of the strength of their first party library which would be a difficult if not nearly impossible advantage to recreate. So maybe an A19 Pro is exaggerating things a bit.
However, to answer your query and to match the Steam machine one thing to keep in mind is that TFLOPS is not a great way to compare GPU architectures unless they are very similar. RDNA 3 has AMD's weird dual issue units that are only really dual issue under very specific circumstances. So is the 9 TFLOPS number under dual issue or normal conditions? If it's under normal conditions, then yes an M3 Pro could work (Apple would want the RT units as a point of advantage over RDNA 3). If it can only do 9 TFLOPS under dual issue, then a base M4/M5 would be enough. I suspect looking at the specs and other RDNA 3 GPUs the 9 TFLOPS you are quoting is under normal conditions, so a Pro-level GPU would indeed probably be needed to match.
Last edited:
To answer the pricing side, we are expecting the Steam Machine to be at least as expensive as the Mac Mini, probably closer to the PS5 Pro. Valve has not been shy to talk about how it is a computer as well as a way to play games.
Really Steam is the reason why it will sell any numbers whatsoever. Apple doesn't really have the same draw even if it (Apples version of a console) played iOS games that had controller support.
Lets use Horizon Steel Frontiers as an example of a game that is coming to iOS and PC but no mention of macOS (especially considering it should be trivial to get the iOS version running on macOS). Apple should be going to NCSoft/Sony and asking for them to also bring it to the Mac. But they won't, as it isn't their style to ask for such things.
I've been musing about this myself and got to the exact same conclusion. Apple has all the pieces except the software. They could very well update the Apple TV 4K with an A19 Pro that can run Resident Evil 4 perfectly, bundle a DualSense controller, and market that as a game console. They only need a slate of AAA games to give some credibility, and they are good to go. I fail to see how they are not pursuing this.I'd say the conversation we were having was the interruption for the thread, yours is the return to regularly scheduled programming
You know even I fall into this trap myself. In a post earlier in this thread and in other musings elsewhere I said Apple needs to release a certain piece of hardware with capabilities X to make inroads into gaming. But really that's not true. I mean don't get me wrong, better hardware for PC games so enthusiasts can play at decent to high graphics settings would be welcome but what separates mobile gaming from PC/consoles is culture not hardware. Take the Nintendo Switch - an incredibly successful console with great first party content and decent 3rd party content. Absolute potato of a chip inside. It is projected that the Switch 2 will be even more successful with only very modest upgrades in the hardware to the point that an A19 Pro is very likely to be a more powerful chip, especially when the Switch 2 is in handheld mode.
Basically Apple could make an "Apple console" with a base Mx or even A19 Pro*, sold with game controllers, and marketed as a gaming console. The hook would be, does it have a great slate of games that attracts the people who play such games. In other words, the actual cost would be getting great gaming exclusives with both an initial slate of games and a full calendar of games to come with a great marketing plan about the games not just the hardware to grow the audience on Apple hardware that plays those games. Then developers want access to it and will port. At the moment, Apple appears to confuse the total addressable market for Mac games with the total market of people who have mac/tv/iPad/iOS despite that most people who bought those devices didn't do so with the purpose of buying and playing PC/console games even though a large fraction of those devices (especially in the last couple of generations) technically could on a hardware level. They bought something else for that purpose.
*admittedly the weaker the hardware the better the games have to be, Nintendo "gets away with it" primarily because of the strength of their first party library which would be a difficult if not nearly impossible advantage to recreate. So maybe an A19 Pro is exaggerating things a bit.
However, to answer your query and to match the Steam machine one thing to keep in mind is that TFLOPS is not a great way to compare GPU architectures unless they are very similar. RDNA 3 has AMD's weird dual issue units that are only really dual issue under very specific circumstances. So is the 9 TFLOPS number under dual issue or normal conditions? If it's under normal conditions, then yes an M3 Pro could work (Apple would want the RT units as a point of advantage over RDNA 3). If it can only do 9 TFLOPS under dual issue, then a base M4/M5 would be enough. I suspect looking at the specs and other RDNA 3 GPUs the 9 TFLOPS you are quoting is under normal conditions, so a Pro-level GPU would indeed probably be needed to match.
Really? Can you provide some proof that a mobile chip used in the iPhone is capable of running AAA games?
Thanks, that's what I was asking. I'm actually surprised that its on there but hey - learn something new
and the A19 Pro has subtantially better GPU performance. They could easily pull this off, I guess the numbers don't add up to them...The software is the most important part in a games console. Nintendo has repeatedly shown this with big successes on horrible hardware that's slow from day one in comparison to competitors (Game Boy, Wii, Switch). You need exclusives or some hook to be enticing. I don't think an A19 Pro ATV 4K would be enticing for gaming. The hardware isn't better than a dedicated console like a PS5 or Xbox Series S even, so unless it's really cheap why get one? Marginally smaller box under TV?
For gaming they'd have to either boost the ATV storage considerably or make it easy to add external drives and I just don't see Apple changing to an A19 Pro, upping storage to like 512 GB and adding a controller for less than $400. To play games about as well as a PS4...
Fair points but I do think they could target a 299 price and play same game on iPhone, iPad and macOS to lure people into their ecosystem.
As much as I would like native ports, Apple as a corporation has its goal of earning a profit for its shareholders. If Apple hasn't yet paid these companies for ports, it's because Apple thinks it's not worth it.
Rather then pay these companies, Apple should make it easier to port to the Mac.
We don't know if Apple paid for the current ports. Plus business strategy can change.
Do we have any graphics programmers on here? What time of performance boost do you think metal 4 command pipeline changes with have on performance. For instance I ran the metal performance on cyberpunk and it returned these insights:
Frame Encoding
Frame Encoding
Shader Compilation
Frame Encoding
Would Metal 4 help resolve some of these issues and result in a big performance boost?
Frequent Render Target Change
Detected high number of render passes with similar attachments. Render passes are expensive on Apple GPUs, consider merging them or adopting color attachment mapping.Frame Encoding
Interleaved Blit Command Encoders
Detected high number of interleaved blit command encoder and other types of encoders. Consider batching resource updates or adopting MTL4ComputeCommandEncoder.Frame Encoding
Shader Compiler Activity
Compiled 4 shaders. Total compilation time is 99.74ms. Consider adopting MTL4Compiler for more fine grained compilations.Shader Compilation
GPU Command Encoding - MTL4CommandBuffer
Detected high CPU encoding cost with encoders spending an average of 100.00% of frame time encoding. Consider migrating to MTL4CommandBuffer.Frame Encoding
Game Mode Support
LSSupportsGameMode key not found in Info.plist. This key will become required to be eligible for Game Mode in a future OS release.Would Metal 4 help resolve some of these issues and result in a big performance boost?
Yes paying a company to port to the Mac doesn't make much sense. Either there is money to make or not. No company would halt the roadmap of their next big hit because they've been given dollars by Apple.
What Apple can do is indeed to make it easier to port to the Mac, to provide engineering ressources to help optimise games for the Mac, but mostly make the Mac gaming market profitable.
I think that Apple should lead by example and setup an internal Apple Game Studio to release original games showcasing their hardware and bringing awareness to Mac users and publishers that their computers are great at running great games.
If apple doesn't think there's money to be made, why would publishers? To put it another way, if Apple is not willing to put their money where their mouth is, why would those same publishers risk their own money?
They didn't have to "do anything" to get gaming on iOS to become the behemoth it currently is. Why would they have to do anything on macOS?
Why?
Their current strategy of doing nothing has netted them exactly that - nothing.
Uh, it's clearly very different circumstances. They've been doing nothing for macOS gaming for 30+ years and it's only gotten them nowhere.
I've seen enough posts defending, praising apple and at times worshiping Apple, that I cannot assume anything anymore.
Valve's ambitions and strategy of doing "everything" has only got them/Linux 0.94% higher Steam market share than Mac (Steam deck included, Mac App Store not included) after all these years so apparently it's hard to break into the Windows dominance.
Last edited:
Only?
Doing a bit of googling it seems since its release, Valve has sold an estimated 4 million units, that's pretty impressive imo.
Valve never intended this to be a primary gaming device but rather a complementary product. They were not trying to take market share away from windows