Support NeoGAF

[Zeyphersan]

Let's say you can bend that rule a bit for the now very powerful iPad Pro devices and upgrade to at least 1-2 iOS versions further.

So the A11 now benchmarks about the same single core and better multicore than the A10X in my iPad Pro, are they good to extend by another 1-2 versions further? Given that power is the determining factor for this hard and fast rule. At what point does this break down, given that the A series and the AX series go back and forth every year in terms of power?

Anyway, I'm not caring about this benchmark pretty much, though it is interesting. The better read is Anandtech, where they test newer versions of iOS on older hardware and compare real-world marks like load times and whatnot.

 

[blu]

Pretty much. I remember I had an iPad 3 which shipped with iOS 5, but iOS 6 was released shortly after and everything was peachy. Then iOS 7 dropped and tanked UI responsiveness: there was a very noticeable pause when bringing up the keyboard for the first time on each app, several animations that were silky smooth in iOS6, like rotating the home screen, became janky/stuttery and the task switcher, oh God! The task switcher caused the whole system to hang for a short moment before showing up and ran worse than the Saturn port of Daytona USA. Made me nearly stop using the iPad for most things I used it for.

The backslash was fierce so Apple eventually patched it up several point updates afterwards so the slow downs (but not all) were greatly reduced, but by then my iPad was no longer part of my routine.

I'm pretty certain a synthetic CPU and GPU benchmark would also run almost the same on the iPad 3 on iOS6 versus iOS7. Things have to go horribly wrong for an OS update to negatively affect the speed at which the CPU can crunch numbers and how many pixels the GPU can push. That has absolutely nothing to do with Apple rewriting their animation code so it runs like utter crap on it, deliberately or not, and knowingly letting the update be irreversibly applied to devices they know can't handle it acceptably (the alternative being nobody at Apple tested it on such devices).

It's hard to work against your own interests.

 

[ArtHands]

I thought it was for each point release...

To be honest that makes it even worse! The iPhone 7 CPU performance has dropped ~15% in 13 months. The 6 and 6S will have even greater drops from release.

This data to me clearly shows a performance decay rather than refuting it.

...I think you are turning yourself into a joke here

 

[nekkid]

Synthetic benchmarks are meaningless if the user experience is worse.

Ios11 has been okay-ish on my iPhone 6. It does get hitchy now and then and there are very noticeable bugs but for me the large decrease in battery life is the biggest issue. The last version of iOS10 was fairly solid so moving to 11 feels like a alpha build.

I’m finding the camera takes ages to launch on mine now.

 

[Dingens]

I mean... car manufactures taught their cars how to detect "benchmarks" so they'd perform better. Shouldn't be that much harder with a phone.
and as others have said, synthetic benchmarks aren't exactly representative for real-life performance anyway

 

[Ovid]

I never thought that Apple deliberately made their older phones slower. I just thought it was shitty on their part by not informing users (with older phones) how much of a performance hit they would take by updating to the latest version iOS on those devices.

In that sense since performance decreases dramatically with later updates it's a kind of a planned obsolescence. Nobody wants a phone that's slow.

Also, not everyone is tech savvy, so when they push the latest version of iOS on those devices, people with usually update it.