Quote from: Dthoma515 on October 05, 2021, 05:28:32
There's a very good reason why iOS hasn't been ported to commercial grade computers, automotive infotainment and or smart  [insert any device] that isn't just a basic app launcher.

Yes. The reason is that Apple doesn't manufacture such products. And Apple doesn't sell their operating systems to third parties. The system is tailored to what they want to do. It has always been that way. They weren't afraid of limiting developers in order to attain their goals. It was never meant as a general purpose OS. That was the point. It had a very specific purpose. If Apple ever builds a car, its infotainment will probably run carOS. :D

These tests are so pointless. If you can use this chipset running Android that would be a true one to one test....or vice versa running iOS with the SD.

The standard background limit process on Android is 20 before it pauses. iOS does like four then pauses....then even worse refreshes when the user hits  the preset cpu usageb cap. Essentially clearing the cache on the fly.  it's why it flys through these kind of tests. It's why there's very few 3rd party apps that can sub out the default apple apps. iOS isn't built to handle computing. Just simple tasks that your average iPhone users does.

There's a very good reason why iOS hasn't been ported to commercial grade computers, automotive infotainment and or smart  [insert any device] that isn't just a basic app launcher.

Yes we know, the known universe is not big enough for Apple

It's nice that Apple started to care about that after so many years or ignoring it. Add to that high screen refresh rates. Add to that widgets. Add to that whatever iOS users wished they had but Android has had for years. Of course, Android sucks, right?

I believe in low power mode, the 2 performance cores are not turned off. They are just clocked at lower frequency (around 1.3ghz).
This was from geekerwan's A15 review video.

I never said its not impressive at all, i'm also excited how good are Apple CPUs.  I also agree that geekbench doesnt scale well, though  difference between single and multicore is too big to assume that big cores are turned off.

As for clock speeds -  Apple is obviously dynamically changing frequencies of all cores depending on needs. Its proven way to boost energy efficiency of processor. We also know that energy required to operate on higher frequencies rises exponentially so reducing core frequencies should give better performance per watt than disabling cores.

BTW
Frequency -  1.37/3.23 = ~42%
Single Core score - 726/1742 = ~42% - coincidence? Or rather hint that 1.37 ghz actually points to high-perf core?

Thanks for the comments regarding what Apple may or may not be doing in low power mode with its chip management. I do try to qualify this, although Geekbench listing the clock speed at 1.37 GHz pointed to other cores being in use besides the two high-powered cores which are clocked at 3.23 GHz. Naturally, it is possible that Apple dynamically adjusts the high-performance cores as well. As 'Tridents' points out, even Apple's high-efficiency cores are more performant than people realise, as the A15's low power mode performance attests to.

However, I hope this point doesn't detract from the key point of the story which is that the A15 in low power mode matches the performance of the Snapdragon 888 in multi-core on Geekbench in the Galaxy Z Fold3 despite it running in its normal performance mode.

Whether you're an iPhone fan or Android fan, or neither, you have to admit that is incredibly impressive. Hopefully it continues to spur the competition to keep raising their game too!

nobody deactivate performance cores, even low power option!

I hope that in future iOS will get that option from BIOS? or from another app or from settings, but nobody can do this.

Axx can't be OC or UC.

Quote from: Mate on October 04, 2021, 12:10:39
"To the best of our knowledge, Apple deactivates the two high-performance 'Avalanche' cores in A15 Bionic when in low power mode"
You know its not true, dont you?
If single-core performance peak is at 726 points then its impossible for CPU with 4 cores like this to get over 3500 points.   More likely low power mode means  significantly lower peak frequencies for all cores(especially high performance ones).

You can confirm this if you compare results in low power mode to A15 results from your own database.
Single-Core performance 1740
Multi-Core performance 4823
4823-1740*2 = 1343  <- this number gives rough estimation how powerful Blizzard cores are.
1343/4 = 335 <- estimated performance of single Blizzard core.
So how 355 turned into 726?

It is true that you cannot achieve 3500 points with only 4 cores at 726 points.
However, you cannot estimate single Blizzard core performance like that. Geekbench doesn't scale well with more cores, you cannot just add up isolated single core performances to achieve multicore performance.
For example, in low power mode, adding 2 cores at 726 points would mean the 4 Blizzard cores would be responsible for 2050 points. This should be far closer to their real contribution than 1343 points.
Apple efficiency cores are far more performant than most people assume.

"To the best of our knowledge, Apple deactivates the two high-performance 'Avalanche' cores in A15 Bionic when in low power mode"
You know its not true, dont you?
If single-core performance peak is at 726 points then its impossible for CPU with 4 cores like this to get over 3500 points.   More likely low power mode means  significantly lower peak frequencies for all cores(especially high performance ones).

You can confirm this if you compare results in low power mode to A15 results from your own database.
Single-Core performance 1740
Multi-Core performance 4823
4823-1740*2 = 1343  <- this number gives rough estimation how powerful Blizzard cores are.
1343/4 = 335 <- estimated performance of single Blizzard core.
So how 355 turned into 726?