The graphics seems to be well behind the CPU, at least compared to the competition. Both those GFXBench scores and 3DMark scores aren't far off from this year's Snapdragon or Exynos. In some cases they're even behind.
I'm curious to see if Apple can push that further in future chips now that the entire graphics portion is in-house.
Yes, the iPhone's GPU is definitely behind the Android competition as far as the benchmarks tell. I'm not sure how much this means in the real world. It definitely helps Android devices in terms of bragging rights. I don't know why the iPhone's GPU is weaker than Qualcomm's Snapdragon GPU. I'm sure Apple will be able to fix that eventually but if it's not going to help iPhone sales, then it probably isn't even worth the effort to have a slightly better GPU which most consumers won't even notice.
And the Android BS pushers have arrived right on queue...
Put down your Tiki Torch and go look at the benchmarks for yourself.
I don’t even know what that junk you posted means. One score, for what exactly? And you can’t compare Geekbench 4.0 scores to Geekbench 3.1 scores.
Junk? They're the official websites for Futuremark and GFXBench.
Why are you bringing up Geekbench?
Except that they are not comparable, they aren’t using different versions of the software for the tests. We know very well that doing that produces different results, and the makers of GFXBench particularly make a point of that, as do the makers of Geekbench. future ark doesn’t use metal in their testing that that has OPenGL available, and ApL,e doesn’t use the latest OpenGL, as they don’t really support it that much anymore. But iOS developers support Metal.
So, basically the same GPU performance as the 7's taking into account the die shrink from 16-10nm. Rather than being world beating or ground breaking this 'entirely Apple designed GPU' looks VERY similar indeed to the PowerVR GPU in the 7's.
Maybe they should have licensed Furian ߘ⦬t;/div>
Well, you guys don’t really understand this, do you? The previous GPU from Imagination, which, according to anandtech’s extensive tests, which are the only really believable ones out there, said that Apple’s technology was pushing Imaninations out for two years. But, this new GPU uses 3 cores, whereas the older one, based on Imagination’s used 6 cores.
so 30% better performance from half the cores, despite what you might think, is pretty damn good.
Until we know the die area (scaled to the fab) of the GPU, the core count really doesn't matter, it's not like a CPU where doing more with fewer cores is almost always preferable.
Each GPU manufacturer counts cores differently, Intel has 8 shaders to an EU, AMD has 64 ALUs in a Vega CU, Nvidia has 128 ALUs in a SM...In all likelihood Apple changed the ratio of ALUs per "core", as increasing the performance while actually having less ALUs is unlikely. They're just sorting them into larger groups, one front end dispatching to more ALUs. Would potentially have better performance per millimeter, but not to the point where it's like 3 cores is necessarily half the size of 6 (again, scaling things to be even between the two fabs)
It's a great GPU, mind, but "more performance with half the cores" is misunderstanding how GPUs work, it's the ALU count per core and in total that will be the big question.
You don’t know any of that. It’s just assumptions. I’ve also asked whether Apple’s cores are larger, and we won’t know that until we see a die shot. Apple did show us more of their die this year in the presentation than ever before, but it’s not enough.
Comments
You don’t know any of that. It’s just assumptions. I’ve also asked whether Apple’s cores are larger, and we won’t know that until we see a die shot. Apple did show us more of their die this year in the presentation than ever before, but it’s not enough.