I don't think GLBenchmark is an on-paper comparison. At most it's a synthetic benchmark, but it still shows the difference. There's really no other way to compare the two SoCs since they are on different platforms.
The buck doesn't stop at a useless "on paper" comparison. Where the rubber hits the road is where it counts.
LOL. So now the anti-apple crowd thinks they should dictate how apple should market their products? The hilarity continues!
The phrase is always true...haters gonna hate.
I don't get the point of doing "Offscreen" benchmarks...
If "Offscreen" benchmarks are the holy grail, then why don't we start using them in benching CPUs/GPUs instead of doing numerous resolution changes, numerous AA/AF/HDR changes, numerous game platform changes, and so on?
It seems to me that "Offscreen" benchmarks benches theoretical performance, not actual performance.
It's safe to say that Apple just took a single benchmark that had 4x performance out of the dozens of other benches they tested.
Nvidia does that a lot on the PC discrete graphics side too. In fact, they've been doing that for decades. It's an art for them.
But again, "offscreen" is not "on-paper" because the GPU still needs to work and churn out graphics during that time. "On-paper" is you take the number the clock speed, number of transistors and other hard numbers about the GPU and then calculate the result using math. There is no benchmark involved to get "on-paper" data. The only difference between regular and "off-screen" is that with "off-screen", you don't see the output of the graphics, but the GPU still has to do all of the rendering, and so on.
It remains as useless, though. We don't use our computers, tablets, and smartphones without seeing the output.
I don't get the point of doing "Offscreen" benchmarks...
If "Offscreen" benchmarks are the holy grail, then why don't we start using them in benching CPUs/GPUs instead of doing numerous resolution changes, numerous AA/AF/HDR changes, numerous game platform changes, and so on?
Outputting it won't make a difference in performance. It's like you unplug the monitor from your computer while running Crysis 2. It doesn't make Crysis 2 automagically 10x faster.
Viewing output is how we all use these devices, which is the only performance comparison that truly matters.
They shouldn't make the "4x the performance of Tegra 3" claim.
If you have to view output, given there is no other output than the built-in display, then 1024 x 768 on A5 is faster than 1280 x 800 on Tegra 3 any day. That's a no-brainer.
It stands to reason that 1024 x 768 on A5X would also be significantly faster than Tegra 3 at 1280 x 800 as well. Given that developers may still continue to do graphics at 1024 x 768 for a while yet, it's not unreasonable to assume that the iPad 3 running some 3D games at 1024 x 768 would far outperform Tegra 3.
In the unlikely scenario that Anand could hook up a display that's capable of showing 1280 x 720 to both A5 and Tegra 3, then the benchmark numbers would still be the same when you view them on that display. It makes no difference.
I'm sure if Samsung or Asus released a $500 tablet with an insane 3.1 megapixel screen and GPU equal to the Playstation Vita's (or hell, even one that managed to match the A5's), everybody here would be trumpeting it to the heavens.
Uh no, that's not what you were arguing before. You were whining that the high-res display needs a more powerful GPU, and that Apple was somehow being deceptive by not mentioning that fact. Then when someone brought up the fact that Apple does indeed mention that in their promo video, you whined some more about how it needs to be "omnipresent" in their marketing. I was pointing out how completely ridiculous that suggestion is.
Oh well. I'm sure if Samsung or Asus released a $500 tablet with an insane 3.1 megapixel screen and GPU equal to the Playstation Vita's (or hell, even one that managed to match the A5's), everybody here would be trumpeting it to the heavens.
The point is... what makes the difference is comparing the performance of one tablet versus another while viewing the output. Everything else is purely academic.
Isn't that what this entire thread is about? The SOC? This thread is about comparing SOC's, why do you keep on trying to change it to something that its not?
I'm not changing the subject, I'm essentially saying that an academic discussion about the SoC is useless. I'm commenting on the subject, not changing it.
Doesn't look that way when reading your posts. You always defend/credit Apple and rarely, if ever, criticize anything they make or do.
LOL, I know, it's awesome to watch them squirm and bitch.
I'll bench one of mine next week, hell, I might keep both of them
Great, you commented and nobody agrees with you. Now its over. So what else do you have to post in here?
The point is... what makes the difference is comparing the performance of one tablet versus another while viewing the output. Everything else is purely academic.
Anand said:If you remember back to our iPad 2 review however, the PowerVR SGX 543MP2 used in it was largely overkill for the 1024 x 768 display. It's likely that a 4x increase in GPU horsepower wasn't necessary to deliver a similar experience on games.
And as I said, if you view the output, 1024 x 768 on A5 is faster than 1280 x 800 on Tegra 3. And due to there being less pixels to work with, A5 is undoubtedly 2x faster than Tegra 3 in that case.
Also if you plug a 1280 x 720 display into A5 and Tegra 3 then run GLBenchmark, then you'll get the same results Anand got.
That's it. Tegra 3 is just slower than A5. I'm sure that has been beaten to death and back in other threads. But I do agree with you that the difference between A5X and Tegra 3 can't be 4x.
In any case, even Anand has admitted that a 4x increase in GPU horsepower from A5 is not very necessary to deliver a similar gaming experience on the iPad 3, possibly hinting that a 2x increase might just be enough to some extent.
He also went on to note that memory bandwidth would play a more significant role. But I'd guess that since the GPU core count is doubled, bandwidth should be boosted in some way to prevent bottleneck as well. If the A5X has higher clock speed, then that would surely remedy the bandwidth issue somewhat.
There are too many things we don't know about it right now, so it's hard to judge. What I do know, though, is that Apple successfully drove 2048 x 1536 at the same smoothness as iPad 2 for general interface performance. It wouldn't be a stretch to say that they tried to do the same with 3D gaming.
runawayprisoner said:But what I'm more interested in, is how much faster A5X is compared to Tegra 3 when both have to handle 2048 x 1536, and I'm also interested in how much performance A5X loses when it has to handle such a monstrous res.
... which is what I want to see, too.
Resolution, AA/AF? Sure they can be compared.Offscreen means it's resolution-independent.
It's a good metric for crossplatform benchmarking, especially when there is no way you can test the same hardware on the same resolution.
The reason why they do multiple resolutions + AA/AF for modern GPUs is because it's doable, it gives them more data, and they don't have to rework their tools to test different hardwares.
Say, how do you do an offscreen benchmark of Crysis 2? You can't. But what if you want to know how a certain GPU does against Crysis 2? You'll have to use standard resolution.
But with GLBenchmark, it's not that simple. You have tablets running at 1024 x 768 and other tablets running at 1280 x 800. Needless to say, the pixel count isn't the same, and if they don't do "offscreen", each device will have an advantage/disadvantage due to their native resolution.
But again, "offscreen" is not "on-paper" because the GPU still needs to work and churn out graphics during that time. "On-paper" is you take the number the clock speed, number of transistors and other hard numbers about the GPU and then calculate the result using math. There is no benchmark involved to get "on-paper" data.