It's like overclocking your CPU very close to the limits of your cooling ability. You will get great benchmark scores, but your CPU will also last a lot less as well (though with CPUs it doesn't matter if they now run for 20 years or 32 years, but if you push it a lot more you might damage it. But hey, at least you got great benchmark points).
Wouldn't you also get more hardware errors, causing e.g. sudden kernel panics and stuff like that? And wouldn't that damage the brand more than the speed improvement is worth?
Yeah. One could argue that if you are getting hardware errors you are already at the red line (and not just close to it).
Regarding damage: It depends what your goal is. Some people want to get as much performance out of the CPU as possible since they maybe upgrade after a year or two anyway. Some people even try to set records by using crazy cooling systems.
Optimizing for speed instead of device survivability. Basically, instead of running at 110% for the first year then throttling to much less so the device doesn't break, they should have shipped them running at 90-100% so they can last longer. (Those numbers are random but the point is still valid)
I don't think that's a fair representation of what's happening though. The reality is that running it at 110% (for whatever that means) is using a new battery at its maximum discharge rate. As the battery ages, its maximum discharge rate decreases, and would have whether run at 110% or 90%. You need to go way out of spec to prematurely age batteries by discharging too fast. They age mostly by being exposed to high temperatures, stored fully charged, and by being fully discharged (partial charge cycles are better). Artificially limiting performance to the post-degradation level can be seen as wasteful, can it not?
Now the way Apple handled it, silently, throttling performance without a configuration option is interesting. I think the benchmarks may not be truly representative of what is happening, and it's probably closer to a TDP limit on the CPU where it runs real fast until it realizes it can't then it ... doesn't. That comes across as maximally efficient in real life but slow in benchmarks.
This is all in the marketing. Either you can look at it as "iPhone find a way to squeeze every last bit of battery life out of both new and old devices" or you can look at it as "my phone slows down over time" -- the reality of the chemistry of lithium ion devices.
IMO this should be handled with a configuration switch and a notification when performance starts to become limited, though how many would be totally confused by this? That hundreds of millions of users took years to notice should be an indication this whole thing worked pretty well IMO.
> That hundreds of millions of users took years to notice should be an indication this whole thing worked pretty well IMO.
People absolutely have been noticing, this has been a very popular and long running issue. It took years for Apple to admit it, not for people to notice.
I think this could have been better handled by throttling according to the actual condition of the battery, rather than controlling it via software updates. The former seems reasonable, the latter is pretty shady. (Maybe this is actually what is happening and the press has confused what is going on.)
> People absolutely have been noticing, this has been a very popular and long running issue. It took years for Apple to admit it, not for people to notice.
The behavior in question was introduced in 10.2.1 which was first released in Feb 2017.
So no, it was not a conspiracy going on for years. The annual slowdown was just poor unoptimized software. Occam's razor.
They are throttling it according to battery's condition (but actual throttling apparatus was added via an update). If you replace old battery with a new one, phone starts running at full speed again.