Mario Odyssey is a super fun game on the Nintendo Switch, it would be awesome to see that running on Apple gear one day… 💫
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.
New Apple TV Reportedly Equipped With Binned A15 Chip With 5-Core CPU
- Thread starter MacRumors
- Start date
- Sort by reaction score
I upgraded from a 2021 ATV, and noticed significant performance improvements when loading high quality blu ray rips across my network.
I have a PS5, a desktop PC with a 4090, a Steam Deck, and an M1 Max (not that I frequently game on it, but when I do, it's a pleasant experience.)
I say all of that not to rub anyone's nose in it, it's not meant to be a victory lap. I point this out so people know where I'm coming from when I say "I don't personally care about gaming on Apple TV" because I have umpteenth ways of having a much more amazing gaming experience. Glad the A15 can be a better performer for those that do, though.
For anyone else that may be using their ATV as the quarterback for their entertainment system, I immediately noticed that the 2022 ATV was loading my high quality remuxes nearly instantly with Infuse. While I wouldn't have categorized the 2021 as by an means "slow", it was not going from selection -> playback instantly.
Local server is a custom built PC, with ATV & server both connected via ethernet, and a not-so-cheap router from 2021. The only thing in this setup that changed was the ATV.
I have a PS5, a desktop PC with a 4090, a Steam Deck, and an M1 Max (not that I frequently game on it, but when I do, it's a pleasant experience.)
I say all of that not to rub anyone's nose in it, it's not meant to be a victory lap. I point this out so people know where I'm coming from when I say "I don't personally care about gaming on Apple TV" because I have umpteenth ways of having a much more amazing gaming experience. Glad the A15 can be a better performer for those that do, though.
For anyone else that may be using their ATV as the quarterback for their entertainment system, I immediately noticed that the 2022 ATV was loading my high quality remuxes nearly instantly with Infuse. While I wouldn't have categorized the 2021 as by an means "slow", it was not going from selection -> playback instantly.
Local server is a custom built PC, with ATV & server both connected via ethernet, and a not-so-cheap router from 2021. The only thing in this setup that changed was the ATV.
Even with a binned chip, it has a significant improvement in performance.
They cite both actually. "...the PS5 still pulls ahead in multi-core benchmark testing [but the review said it is] starting to get to a point where Xbox One and PS4 games should... be playable on Apple TV without too many compromises."
Well, in the initial order for iPhone 14 Pro models, figure about 50% of the initial 90M units were Pro’s (even more now due to higher demand than anticipated), so did they had TSMC build only 45M chips from wafers? No, because yields are never 100%, far from it. Also, because of using the N5+ node, it was still being sorted out from a production yield standpoint.
Even if TSMC had 90% yield, which would be pretty extraordinary, not all are 100% usable in full spec. This drops the best chips yield even further to say 85%. So TSMC has to build at least 45M/85% = ~53M chips to supply initial order, leaving at least 7-8M chips that don’t meet full spec but something less which could be controlled by laser cutoff or software cutoff of the defective chip area. That’s a lot of usable chips for an application or device which doesn’t need 100% spec - like an Apple TV 4K update, maybe a new iPhone SE Model, maybe another iPad model next year. The ATV wouldn’t sell that many in a year (yet) but a new updated iPhone SE certainly could, that’s why with continued A15 production, there will always be variants available.
And yes, sometimes they even put in perfectly good chips that are constrained or throttled back just to get the product out. (happened in the remanufactured auto parts area - a big Japanese auto company was selling reman alternators at lower prices at their dealers. Demand was so brisk that they couldn’t get enough cores to rebuild. But to give the consumer choice, they decided to keep the reman part (along with the new part) in stock. How? They put new alternators in reman boxes and ate the difference. Consumers had the choice and many times still chose new over reman.)
It’s called usable chips repurposed for sale and use in an applicable device. Intel did it all the time with original Pentium and subsequent Core and iX-yyyy models with all the “steppings” which were slightly different specs of clock frequency, cores, etc. they sold them at different prices for different performance.
[[Yeah I’m thinking iPhone rejects.]]
Maybe so, I prefer to think of it as lesser performance but still quite usable chips within the A15 family. That Apple chooses not to ID them with a different public number but rather just describe them by cores is marketing and avoiding tech speak.
Consider these chips A15 variants.
Last edited:
Who in their right mind keeps dreaming that an iPhone chip is comparable to modern consoles.. Its just nonsense. The architectures are just so dramatically different it's ridiculous. Even worse..only a shameful writer would even insinuate that A PS5 is any way in the same league of Apple a15 iPhone chip. Just embarrassing.
How about bringing in a tech specialist that actually understands the stuff.
Cmon.
How about bringing in a tech specialist that actually understands the stuff.
Cmon.
Just because they can look at some numbers and say one is bigger than another means that Apple TV can play PS4 games. That's probably why there is almost zero Apple TV games that come close to a real PS4 game.
Some real amateur hour logic going on here.
We didn't need this test to confirm that the A15 GPU is now pretty close in power to the ps4 (not the Pro version), there have been plenty of others. But it doesn't need to be a match. Even the switch gets plenty of ports, even though the computing power is much lower.
The real problem is: Noone is going to take the time to port any xb1 or ps4 game to it, because that is going to be a lot of effort and i don't see them getting that money back, if no one even sees the Apple TV as a gaming device.
Apple would have to push this with financial incentives and really push for gaming. Not just add controller support and hope for the best. They need to reach out to developers, get some popular games and advertise that these can now be played on the Apple TV. And recommend a controller that the games will be optimized for or make their own.
Otherwise they will only be getting iPhone games, like they have so far. (Many devs don't even do that)
All this means is more profit for Apple with fewer manufacturing parts and zero benefit to the user. Just saying.
Dude. This is how misinformation is spread. If you don't even know the correct name, stop running imaginary benchmarks.
What's odd to me is that Apple hasnt bought a gaming studio. The M* chips on the AS macs are ridiculously capable and the A* chips on everything else arent much less so, Apple could have pretty easily outbid MS for Bethesda last year for example, and brought in a massive backlog of AAA games and ported them + had future games have mac/ios/tvos/etc versions. It's not like they don't have the cash for something like that. Hell, it's basically what MS did to launch the original xbox (remember: Halo was originally targeted at Macs by Bungie until they were bought by MS).
sure the bottom line is usually the deciding factor, but zero benefit is disingenuous. smaller size is good, fewer parts to fail is also good. lower power consumption would be a guess on my part, but also good if true.
I’m referring to the benchmarks listed in the original article. If you want to question the credibility of MacRumors and their sources, that’s your problem. As for the naming, it was a typo. If you have legitimate factual backing to say I am wrong, please explain it. I am perfectly willing to admit I’m not an expert, but I’m going to need something more than “you’re wrong” before I correct myself.
The Series X has a 12tFlop GPU. PS5 clocks in at 10.5tFlop. Nobody seems to know exactly how fast the GPU in the iPhone 13 is, somewhere around 1tFlop at most. So slower than the Xbox One (1.2tFlop) and the PS4 (1.6tFlop) and not in any way comparable to current gen consoles.