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
What do I think? I think it confirms what I was thinking all along. We now have a 10W passively cooled, $999 ultra-thin laptop with a performance of an entry-level dedicated dGPU. This will make Mac a much more attractive target for game development.
Speaking about that BG3 demo... they played more of it in a dev tech talk, and it was running higher settings with what appeared to be a better performance than what I can get out of my 5500M. The performance promises to be surprisingly good for games that are optimized for Apple's GPU — but it will be more then decent even for older Intel-based games.
The M1 seems impressive for what it can do. Seems like the GPU will give you something between intergrated graphics and mid range discrete graphics on a low end machine. This was a chip for the lower end Macs. I would assume there will be an M1X or M1Z for Macbook Pro, iMac Pro and Mac Pro that have more CPU cores and GPU cores for better graphics performance.
Real world Benchmarks will tell the real story.
Real world Benchmarks will tell the real story.
What I found telling was how few well known companies were highlighted during this presentation. I truly don't know half of them nor can I remember who they were.
I wouldn’t read too much into the performance differences since we don’t know what build Apple has versus what we have in Early access. Larian has said that over time the requirements will probably be reduced from what shows in Steam/GoG currently.
I know, but still. I am running it at med-high settings with 1680x1050 resolution and it still stutters here and there. The M1 demo is running at 1080p and high settings — which is very impressive for a 15W chip.
I run it at the same settings on my 460 and see similar stuttering. Which leads me to believe it isn’t the GPU or CPU but likely asset loading causing hitching, which is silly cause we are all on NVMe storage. At this point I don’t understand why we still have loading longer than a couple of seconds on any Mac or iOS device. [Series S/X and PS5 load times for stuff is perception changing]
...the video embedded on a web stream “looked” low frame rate while the *developer* of the game is heard in the voiceover talking about the buttery smooth and astounding playability with that chip. Ok. ?
Ultra settings they say in this video at 6:45: https://developer.apple.com/videos/play/tech-talks/10859/
BG 3 at Ultra settings 1080p on M1 in this video at 6:45: https://developer.apple.com/videos/play/tech-talks/10859/
Ah, well then
Anyway, I should be getting my 13" Pro in a couple of weeks and I will run some tests. But I am sure someone else will do it before me.
I'm guessing the big players (if they were watching at all) were skeptical of what Apple could do, and I highly doubt Apple reached out to them in the first place.
Likewise, no big titles I can think of even have an in-house Mac port, it's usually farmed out to Aspyr or Feral.
I do wonder if they'll ever release their own discreet GPUs, maybe that wouldn't make sense. Perhaps they'll roll support for AMD on the higher-end.
While M1 looks really good for it's category, the TFLOPS are only a tenth of what is out there at the very high-end (extremely high budget and power, I know, but for those who need it / want it / can afford it).
While M1 looks really good for it's category, the TFLOPS are only a tenth of what is out there at the very high-end (extremely high budget and power, I know, but for those who need it / want it / can afford it).
I'd say no, if the Steam page says no Mac support, then there's no Mac support. Actually, even if the Steam page says there's Mac support, there's no guarantee there's Mac support. (Valve doesn't do quality control)
Rumors are that they're working on a dGPU. I'd go with that considering the surprising power of the iGPU of the m1 chips.
I’m not really an expert in that, but it seems like apple has more synergies to gain when trying to develop the best SoC in the world.
For dGPUs their business is basically limited to a few higher end Mac models. Not sure if it makes sense for them to make that huge investment.
Apple will continue using their own GPUs even in the high-end. AMD would break down the perfect development ecosystem Apple is building. Not to mention that Apple has already overtaken both Nvidia and AMD in terms of performance per Watt.
GPUs can be scaled up
I don’t really expect RTX 3090-level in Apple Silicon any time soon, but the baseline performance will see a big boost.I can't see apple making a 350-watt space heater GPU any time soon either. Am I the only person who's annoyed at the out of control power consumption that high-end GPUs are running? Likewise, the mid-and-low end GPUs are completely stagnant, with Polaris still being AMD's low-end answer 4 years on, and NVidia recommending GTX1650's still.GPUs can be scaled up
GTX1650 is Turing. So it isn’t that old. Polaris has long legs thanks to the consoles, that should be changing soon. If anything they have appeared to stick with Vega longer than it feels like they should. The Zen 3 APU should be using RDNA (probably 2) this time around.
GTX1650 is like half Turing. Somewhere between Pascal and Turing iirc. Also Vega stuck around because it was surprisingly efficient in APU form, it just didn't scale up. Either way, there's now a segregation between "high-end" GPUs which use massive amounts of power, cost ludicrous amounts of money, and are on the latest arch, and "low-end" GPUs which use reasonable amounts of power, for reasonable(ish) prices, but are on generations from years behind.
Actually, with Apple Silicon they're gonna do better than that, because they're shooting for processor updates yearly, so it's looking like each successive generation of both CPU and GPU are gonna be full-featured. Even if they fall short of 3090 territory, I think it's a better way.
AMD and Nvidia can’t secure enough production for the full RDNA2 and Ampere parts. Which is where they make the most money. We won’t see the “cut down” parts for a while.
It will be interesting to see if Apple can kick them in the pants.
Stumbled on to a reference to The future of computer: a conversation with John Hennessy (Google I/O '18) which says the same thing I was in regard to the future of the x86 - unless they fix the heat problems it is going to hit a wall even in the server side of things.