The ultimate showdown - 3DMark releases new cross-platform benchmark

[Zodiark1593]

19 hours ago, WereCatf said:

I'll have to brag a little about how my phone is #2 on the list of mobile-phone scores:

 

Yes, that is my literal phone, not just someone else's same model

The demo looks pretty immensely impressive to be running at decent frame rates on a mobile device. I’d wager that, like the XS series, the GPU is ran at a very high clock speed for a short time though. According to Anandtech, the XS drove the gpu clocks high for compute tasks. I’d imagine the successors do the same as well. 
 

I’m curious to see how older cards like my GTX 960 or a 750 TI compare. 

[Bombastinator]

There’s a thing I am noticing about some of the graphs:  the graphs I am seeing for Apple phones seem to have similar “hook tip spikes” that weren’t all that dissimilar looking to the graphs I saw on userbenchmark.  I’m wondering if there’s a bit of a push there.  There’s still too much difference in power to account for, but there still may be a bit of “help”  going on.

[Zodiark1593]

15 minutes ago, Bombastinator said:

There’s a thing I am noticing about some of the graphs:  the graphs I am seeing for Apple phones seem to have similar “hook tip spikes” that weren’t all that dissimilar looking to the graphs I saw on userbenchmark.  I’m wondering if there’s a bit of a push there.  There’s still too much difference in power to account for, but there still may be a bit of “help”  going on.

In an Anandtech article for the Apple A12, it was commented that Apple drove the clock speeds well beyond a sustainable level for a short time (the difference is pretty ridiculous). I suspect that the A13 probably exhibits similar behavior, and the short benchmark falls within the thermal threshold. 
 

https://www.anandtech.com/show/13392/the-iphone-xs-xs-max-review-unveiling-the-silicon-secrets/7

[Warin]

Just ran it on my iPad.

Pretty good results, I would say!

[Zodiark1593]

Here ya go. 

[Mihle]

Have any run it on newer Intel integrated graphics for example name compared?

 

I am really more curious about x86 vs ARM real CPU proformace than GPU performance, but I don't know how accurate a comparison between those you can do.

[Zodiark1593]

3 hours ago, Bombastinator said:

There’s a thing I am noticing about some of the graphs:  the graphs I am seeing for Apple phones seem to have similar “hook tip spikes” that weren’t all that dissimilar looking to the graphs I saw on userbenchmark.  I’m wondering if there’s a bit of a push there.  There’s still too much difference in power to account for, but there still may be a bit of “help”  going on.

https://hothardware.com/news/ul-benchmarks-3dmark-wildlife-stress-test-benchmark-iphone-android
 

Seems I’m not far off the mark as far as theory goes. The thermal stability, or lack thereof, indicates that Apple employs a very aggressive burst clock speed.  Sustained performance appears to be much closer to Android phones, with that Asus RoG phone edging out a win over the iPhone 11s slowest run. 

[Bombastinator]

I’m not buying that iPhones are twice as fast and have consistently been so.   My suspicion is that there is something inaccurate about the test.  Time was mentioned.  Perhaps the test needs to be longer, or there needs to be another longer test or something to at least figure out if the one minute thing is causing an inconsistency. 

[Mihle]

My phone does like 600 or something.

But it's good enough for me as I do not game and stuff on my phone.

Day to day usage I have never been in a situation where I notice my phone isn't the fastest, no lag or anything.

[Bombastinator]

17 minutes ago, Mihle said:

My phone does like 600 or something.

But it's good enough for me as I do not game and stuff on my phone.

Day to day usage I have never been in a situation where I notice my phone isn't the fastest, no lag or anything.

Whereas I did more than 3 times that with an iphone7.  I could buy that iPhones are fast.  That theyre THAT fast seems a bit much.  iOS should be lighter weight than android.  There’s no need for THAT much speed. It will probably get worked out. Could explain why iPhones last so much longer than android phones I suppose.  Spare horsepower.  Still, it’s a bit nuts.

[Mihle]

1 hour ago, Bombastinator said:

Whereas I did more than 3 times that with an iphone7.  I could buy that iPhones are fast.  That theyre THAT fast seems a bit much.  iOS should be lighter weight than android.  There’s no need for THAT much speed. It will probably get worked out. Could explain why iPhones last so much longer than android phones I suppose.  Spare horsepower.  Still, it’s a bit nuts.

My phone is Snapdragon 710 so weaker than any flagship android tho but yes.
My previous phone, Nexus 6p did not feel slow or lag anything with my usage when I switched from it last year either, thats why I didnt get anything more powerfull.

[Zodiark1593]

10 hours ago, Bombastinator said:

Whereas I did more than 3 times that with an iphone7.  I could buy that iPhones are fast.  That theyre THAT fast seems a bit much.  iOS should be lighter weight than android.  There’s no need for THAT much speed. It will probably get worked out. Could explain why iPhones last so much longer than android phones I suppose.  Spare horsepower.  Still, it’s a bit nuts.

From what I’d researched, gpu compute is actually fairly commonplace on iPhone, in which case would merit very high (albeit unsustainable) burst clock speeds on the GPU. Run the 20 loop stress test, and iPhones come back down to Earth pretty quickly. 
 

I ran the 20 loops with the phone out of the case, propped up on a stand, with a fan blowing on it. My bedroom was about 68 degrees Fahrenheit. Pretty much best case, within reason scenario here. 
 

It’s worth noting that despite being on the standard 1A charger the entire run, I still lost battery power.