Slashdot Mirror
Geekbench Results Visualize Possible Link Between iPhone Slowdowns and Degraded Batteries (geekbench.com)
Earlier this month a post on social media which suggested that Apple might be deliberately downgrading performance on iPhone models with degraded battery was widely circulated. Benchmark Primate Labs' Geekbench has looked into the matter and is corroborating the claims. From a report: Primate Labs founder John Poole has plotted the kernel density of Geekbench 4 scores for iPhone 6s models running iOS 10.2, iOS 10.2.1, and iOS 11.2, visualizing an apparent link between lower performance and degraded battery health. The charts show that on iOS 10.2, the vast majority of iPhone 6s devices benchmarked similarly in performance. However, Poole explains that the distribution of iPhone 6s scores for iOS 10.2.1 appears multimodal, with one large peak around the average and several smaller peaks around lower scores. In other words, after iOS 10.2.1 was released last January, the performance of a percentage of iPhone 6s devices began to suffer.
fair from a technical standpoint. But why not actually tell users, or at least technical resources that might help users, about doing this?
Batteries don't just have less of a charge when they get older, their peak draw also diminishes. If, when you are using an older battery you need more than the battery can supply, the phone reboots. So Apple is slowing down the CPU on older iPhones so they do not go over the max battery draw available on older batteries.
It is preferable to have a slower iPhone than it is to have it rebooting. If the slowdown is an issue, replace the battery for $75 and the performance will be back to normal. As for those who complain "I want a removable battery", well I much prefer having the water resistance that has saved my phone a few times than a removable battery that I only need after 3 years.
Democracy is a sheep and two wolves deciding what to have for lunch. Freedom is a well armed sheep contesting the issue
Very interesting story and I have a data point to add: I bought my iPhone 6 new in 2014 and as of about two months ago the battery (which was at more than 750 cycles - Apple rates it for about 500) was getting really flakey. The reported capacity of the battery started varying between 45% of its rating and 85% with random power offs becoming common below 10%. So last weekend I finally bit the bullet and put a new battery in - fiddly but doable in about 30 mins. I then reset the phone and left it to fully charge overnight.
Since changing the battery the iPhone has been subjectively faster to launch apps and display information. So this report that Apple may have something in iOS that compensates for a failing battery doesnâ(TM)t seem implausible.
Though I would have thought that a better thing (or perhaps as well as) would be if Apple gave the user an alert once 500 cycles exceeded advising them that the battery was wearing out and linking them to the official Apple page for getting it changed.
Cheers, Chris W.
If debugging is the process of removing bugs, then programming must be the process of putting them in. --- Dykstra
It seemed pretty clear to me. Most devices run at some predictable speed, but as they age, they get "binned" into speed steps. That's why it's not just a tail to the left, but a series of peaks.
How is the Riemann zeta function like Trump rallies? Both have an endless number of trivial zeros.
Except that is not how it works. Processors constantly scale their performance based on load. If you slow down a processor to 50% speed, a task will just just run 2x as slow and keep the processor active longer, therefore using more power than if it just ran at 100% speed. Slowing down 50% does not draw 50% less power, it might draw 10% less power.
Slowing down a processor will end up using more power in the long run and further reduce battery life.
I think a Kernel Density Function is just a probability density
https://en.wikipedia.org/wiki/...
In statistics, kernel density estimation (KDE) is a non-parametric way to estimate the probability density function of a random variable. Kernel density estimation is a fundamental data smoothing problem where inferences about the population are made, based on a finite data sample. In some fields such as signal processing and econometrics it is also termed the Parzen--Rosenblatt window method, after Emanuel Parzen and Murray Rosenblatt, who are usually credited with independently creating it in its current form.
So the X axis is the Geekbench score and the Y axis is the percentage of phones.
Now if you look at iPhone 6s devices with 10.2.0 they're one peak - i.e. all the phones cluster around some mean performance. Looking at devices upgraded to 10.2.1 you see some smaller peaks to the left and with 11.2.0 those peaks grow. It seems like iOS knows if it running on an older device - perhaps it checks the battery health. If the battery is worn out it can run in a low power/low performance mode to keep the battery life acceptable.
Now personally I'd never buy a device with non removable battery. But if you're selling an OS which auto upgrades and runs on devices with non removable batteries it sort of makes sense. It also pushes users to buy new hardware. Imagine if you had an iPhone 6s with 11.2.0. If your device is on the right it'll seem OK. If its one of the devices in the low power/low performance modes it will seem sluggish and you'll probably go out and buy a new one. What I'm not sure about is what happens if you replace the battery? Does that reset the device into high performance mode or are you still stuck in slow mode?
echo -e 'global _start\n _start:\n mov eax, 2\n int 80h\n jmp _start' > a.asm; nasm a.asm -f elf; ld a.o -o a;
That would be a fair point if (a) this wouldn't happen when the device is plugged in [spoiler: it is still dog-slow] and (b) if the user had a choice between longer battery run time and smooth video playback on his phone that is (gasp) almost two years old.
Since neither is the case, the balance of probabilities shifts towards deliberate planned obsolescence through sabotaging user's devices rather than prudent battery life management.
Hahahaha in an Apple device? How are you not marked +5 Funny already?