Comment by manuelabeledo
25 days ago
I think it’s fair to say that a SoC should perform better at higher wattages, so my comment is definitely relevant.
Regardless, I don’t understand how you can say that I’m moving goalposts when I mention performance per watt, which is absolutely relevant when talking about smartphone SoC performance, and then you bring up battery capacity, which is not.
Your initial question was "How are they on par SoC-wise?"
They are on par because they now sometimes beat Apple top of the line A-chips on performance be it single core, multicores or GPU and do so within a power budget which allows the phone they ship in to be competitive screen-on time wise.
Apple doesn't have a one generation lead anymore which is a huge change compared to only three years ago.
You are moving the goalposts because the discussion was always about the gap between Apple and its competitors and you have entirely shifted to peak consumption when it was clear the conclusion would not be the one you want/expect.
The whole claim that Qualcomm is on par with Apple predicates upon results from benchmarking tools, which stress CPU and GPU and thus induce peak power consumption.
If we were to look at more thorough reviews, e.g. Geekerwan, they always include TDP and power consumption, because that gives the necessary context to understand the results.
And obviously I’m not denying that Mediatek and Qualcomm have massively improved their designs, but they aren’t on par when we account all the things that matter.
Your argument is that, since manufacturers are putting larger batteries in phones, SoC power consumption shouldn’t matter. That is moving the goalpost, because you introduce a variable that should be irrelevant to SoC performance testing to dismiss my observation.
> And obviously I’m not denying that Mediatek and Qualcomm have massively improved their designs, but they aren’t on par when we account all the things that matter.
Hardly, you are intentionally looking at peak consumption because it suits the answer you want after being proved wrong. This is nicely highlighted by how you want to casually dismiss benchmarks which don't support your point.
> Your argument is that, since manufacturers are putting larger batteries in phones, SoC power consumption shouldn’t matter. That is moving the goalpost, because you introduce a variable that should be irrelevant to SoC performance testing to dismiss my observation.
No, that is looking at the actual experience of using the phone. Peak consumption is a useless metric. Nobody cares. People care about their phone feeling smooth and how often they have to charge. From this point of view, Apple has no lead whatsoever provided by their SoC while they used to which is the point I have been making from the start.
1 reply →