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.
AnandTech Finds iPhone 11 Pro Has 50-60% Higher Sustained Graphics Performance Than iPhone XS
- Thread starter MacRumors
- Start date
- Sort by reaction score
When did Apple start making their own custom desktop graphics cards?????!!!
THIS is the reason I come to MacRumors! Sometimes you end up learning something exciting and new.
Link me all the info on that project. I’m super interested. I’ll join you in sending notes to the them to please put more effort into it, as soon as you share a bit more of the details!
I’ve been a reader of Anandtech for a long time. Good stuff they put out. And consistently so.
Apple doesn't make desktop processors or graphics cards. If they ever decide to, these mobile A chips are where the foundation would be laid for those desktop chips.
I just looked at your signature. I don't think you'll be leaving the Mac universe any time soon.
Are you suggesting an Apple-designed GPU for Mac? If so, I'm not sure that would satisfy the Nvidia-or-nothing crowd, any more than Apple's current use of AMD does.
Not surprising considering no process shrink this time (still 7nm) and the A12 was already very efficient, it was already probably squeezing nearly every drop of performance per watt, so if they wanted more power out they needed more power in. With 5nm next year, it should be a different story. All things considered, improving on the A12 as much as they did with no node shrink is impressive - even given slightly higher peak power draw.
Not really. Give developers lots of RAM and they get lazy, very lazy.
As a software engineer, Ive built a web applications framework. It does 100x what it did 15 years ago, and performs twice as good. A lean, mean design has always been a core goal.
There is a camp of developers out there that say to leave optimizations for last, but I disagree entirely. Keep optimal design as part of your work as you go along.
Almost all Macs. I am in the ecosystem due to the Mac, the others are easily replaceable.I just looked at your signature. I don't think you'll be leaving the Mac universe any time soon.
As long as Apple doesn't screw up the Mac, I will stay.
Android has been using ART with AOT compilation by default since version 5.0 now. It replaced the JIT compilation of Dalvik, the previous runtime.
Your other points remain valid, however.
You are correct. I was looking at an out-dated post which was released before the iPhone 11s hit the market.
In a shocking discovery it was found that a new, more modern platform is more powerful than its previous generation. Unbelievable, no-one expected that
Seriously though,I’d love to see what Apple could achieve with its A series architecture, without the thermal and power limits of a mobile device.
Imaging a desktop chip with more cores, a higher clock speed, more power draw and better thermal efficiency. Only a year or two and we’ll find out
Seriously though,I’d love to see what Apple could achieve with its A series architecture, without the thermal and power limits of a mobile device.
Imaging a desktop chip with more cores, a higher clock speed, more power draw and better thermal efficiency. Only a year or two and we’ll find out
Adding more memory doesn’t speed up a computer unless it’s memory constrained. Memory isn’t speed, it’s capacity. Android phones, lagging in speed, are focusing on memory to try to look better on spec sheets.
It’s like a truck hauling stuff: if you have more stuff than your truck can fit, you have to take more than one trip. If you have less stuff than your truck can fit, you can take just one trip - and using a bigger truck still takes one trip.
50%/60% is a really impressive leap in just one years time. To Be honest, this probably has to be the one of the most surprising things about the A13 over the A12. The A12 really was more about 60 FPS for gaming and smart HDR, but I think the trade-off is definitely favoring the A13, especially with the face ID improvements.
Last edited:
Where did I say Apple should make custom GPUs for their desktops? All I said is I wish they would show their desktop GPUs some love. They could do that by using more powerful GPUs and designing better cooling systems for their computers.
I was wondering this, so I did a very rudimentary calculation to see if the percentage increases are going down (as they're starting from a higher base) but the numerical leaps are staying quite steady generation to generation. It appears that is the case. Again, the below is completely arbitrary, using Apple's claimed percentage increase for each generation, but it shows the broad point I'm trying to make:
If you say the A7 has 100 units of CPU performance to start,
the A8 has 25% more (125 units absolute, an increase of 25),
the A9 is 70% faster than the A8 (212.5 (A7) units, an increase of 87.5 over the A8)
the A10 is +40% to 297.5, up 85 on the A9,
the A11 is +25% to 371, up 74 on the A10,
the A12 is +15% to 427 up 55 on the A11,
the A13 is +20% to 513, up 85 on the A12.*
So the A13 is 85 of our theoretical units more powerful than the A12, about the same as the A10 over the A9 and the A9 over the A8, a little more than the increases from A10-11-12, even.
* In each case where the chip has more than two cores, the increase is just for the two performance cores the chip includes, so it's comparable even despite core count increases
Years ago I asked about relative performance between mobile & laptop processors. I was told there’s no commonality for comparison. What’s changed?
I'm confident that Apple will release a laptop with a custom ARM processor. I just don't think it'll be branded a Mac or iPad.