GPU Performance

All of our discussions around the new iPad and its silicon thus far have been in the theoretical space. Unfortunately the state of Android/iOS benchmarking is abysmal at best today. Convincing game developers to include useful benchmarks and timedemo modes in their games is seemingly impossible without a suitably large check. I have no doubt this will happen eventually, but today we're left with some great games and no way to benchmark them.

Without suitable game benchmarks, we rely on GLBenchmark quite a bit to help us in evaluating mobile GPU performance. Although even the current most stressful GLBenchmark test (Egypt) is a far cry from what modern Android/iOS games look like, it's the best we've got today.

We'll start out with the synthetic tests, which should show us roughly a 2x increase in performance compared to the iPad 2. Remember the PowerVR SGX 543MP4 simply bundles four SGX 543 cores instead of two. Since we're still on a 45nm LP process, GPU clocks haven't increased so we're looking at a pure doubling of virtually all GPU resources.

GLBenchmark 2.1—Fill Test

GLBenchmark 2.1—Triangle Test (White)

GLBenchmark 2.1—Triangle Test (Textured, Fragment Lit)

Indeed we see a roughly 2x increase in triangle and fill rates. Below we have the output from GLBenchmark's low level tests. Pay particular attention to how, at 1024 x 768, performance doubles compared to the iPad 2 but at 2048 x 1536 performance can drop to well below what the iPad 2 was able to deliver at 10 x 7. It's because of this drop in performance at the iPad's native resolution that we won't see many (if any at all), visually taxing games run at anywhere near 2048 x 1536.

GLBenchmark 2.1.3 Low Level Comparison
  iPad 2 (10x7) iPad 3 (10x7) iPad 3 (20x15) ASUS TF Prime
Trigonometric test—vertex weighted
35 fps
60 fps
57 fps
47 fps
Trigonometric test—fragment weighted
7 fps
14 fps
4 fps
20 fps
Trigonometric test—balanced
5 fps
10 fps
2 fps
9 fps
Exponential test—vertex weighted
59 fps
60 fps
60 fps
41 fps
Exponential test—fragment weighted
25 fps
49 fps
13 fps
18 fps
Exponential test—balanced
19 fps
37 fps
8 fps
7 fps
Common test—vertex weighted
49 fps
60 fps
60 fps
35 fps
Common test—fragment weighted
8 fps
16 fps
4 fps
28 fps
Common test—balanced
6 fps
13 fps
2 fps
12 fps
Geometric test—vertex weighted
57 fps
60 fps
60 fps
27 fps
Geometric test—fragment weighted
12 fps
24 fps
6 fps
20 fps
Geometric test—balanced
9 fps
18 fps
4 fps
9 fps
For loop test—vertex weighted
59 fps
60 fps
60 fps
28 fps
For loop test—fragment weighted
30 fps
57 fps
16 fps
42 fps
For loop test—balanced
22 fps
43 fps
11 fps
15 fps
Branching test—vertex weighted
58 fps
60 fps
60 fps
45 fps
Branching test—fragment weighted
58 fps
60 fps
30 fps
46 fps
Branching test—balanced
22 fps
43 fps
16 fps
16 fps
Array test—uniform array access
59 fps
60 fps
60 fps
60 fps
Fill test—Texture Fetch
1001483136 texels/s
1977874688
texels/s
1904501632
texels/s
415164192
texels/s
Triangle test—white
65039568
triangles/s
133523176
triangles/s
85110008
triangles/s
55729532
triangles/s
Triangle test—textured
56129984
triangles/s
116735856
triangles/s
71362616
triangles/s
54023840
triangles/s
Triangle test—textured, vertex lit
45314484
triangles/s
93638456
triangles/s
46841924
triangles/s
28916834
triangles/s
Triangle test—textured, fragment lit
43527292
triangles/s
92831152
triangles/s
39277916
triangles/s
26935792
triangles/s

GLBenchmark also includes two tests designed to be representative of a workload you could see in an actual 3D game. The older Pro test uses OpenGL ES 1.0 while Egypt is an ES 2.0 test. These tests can either run at the device's native resolution with vsync enabled, or rendered offscreen at 1280 x 720 with vsync disabled. The latter offers us a way to compare GPUs without device screen resolution creating unfair advantages.

Unfortunately there was a bug in the iOS version of GLBenchmark 2.1.2 that resulted in all on-screen benchmarks running at 1024 x 768 rather than the new iPad's native 2048 x 1536 resolution. This is why all of the native GLBenchmark scores from the new iPad are capped at 60 fps. It's not because the new GPU is fast enough to render at speeds above 60 fps at 2048 x 1536, it's because the benchmark is actually showing performance at 1024 x 768. Luckily, GLBenchmark 2.1.3 fixes this problem and delivers results at the new iPad's native screen resolution:

GLBenchmark 2.1—Egypt (Standard)

GLBenchmark 2.1—Pro (Standard)

Surprisingly enough, the A5X is actually fast enough to complete these tests at over 50 fps. Perhaps this is more of an indication of how light the Egypt workload has become, as the current crop of Retina Display enhanced 3D titles for the iPad all render offscreen to a non-native resolution due to performance constraints. The bigger takeaway is that with the 543MP4 and a quad-channel LP-DDR2 interface, it is possible to run a 3D game at 2048 x 1536 and deliver playable frame rates. It won't be the prettiest game around, but it's definitely possible.

The offscreen results give us the competitive analysis that we've been looking for. With a ~2x die size advantage, the fact that we're seeing a 2-3x gap in performance here vs. NVIDIA's Tegra 3 isn't surprising:

GLBenchmark 2.1—Egypt—Offscreen 720p

GLBenchmark 2.1—Pro—Offscreen 720p

The bigger worry is what happens when the first 1920 x 1200 enabled Tegra 3 tablets start shipping. With (presumably) no additional GPU horsepower or memory bandwidth under the hood, we'll see this gap widen.

The Impact of Larger Memory A5X vs. Tegra 3 in the Real World
POST A COMMENT

233 Comments

View All Comments

  • Ammaross - Wednesday, March 28, 2012 - link

    "It has the fastest and best of nearly every component inside and out."

    Except the CPU is the same as in the iPad2, and by far not the "best" by any stretch of the imagination. Hey, what's the problem though? I have this nice shiny new tower, loads of RAM, bluray, SSD, and terabytes of hard drive space. Oh, don't mind that Pentium D processor, it's "good enough," or you must be using it wrong.
    Reply
  • tipoo - Wednesday, March 28, 2012 - link

    What's better that's shipping today? Higher clocked A9s, or quad core ones like the T3? Either would mean less battery life, worse thermal issues, or higher costs. Krait isn't in a shipping product yet. Tegra 3's additional cores still have dubious benefit. These operating systems don't have true multitasking, you basically have one thing running at a time plus some background services like music, and even on desktops after YEARS few applications scale well past four cores outside of the professional space. The next iPad will be out before quad core on tablets becomes useful, that I assure you of. Reply
  • zorxd - Wednesday, March 28, 2012 - link

    I'd gladly trade GPU power for CPU power.
    That GPU is power hungry too, probably more than two extra A9 cores, and the benefit is even more dubious unless you are a hardcore tablet gamer.
    Reply
  • TheJian - Wednesday, March 28, 2012 - link

    LOL, the problem is you'll have to buy that new ipad to take advantage because YOURS doesn't have those cores now. Once apps become available that utilize these cores (trust me their coming, anyone making an app today knows they'll have at least quad cpu and gpu in their phones their programming for next year, heck end of this year), the tegra 3 won't need to be thrown away to multitask. Google just has to put out the next rev of android and these tegra3's etc should become even better (I say etc because everyone else has quad coming at 28nm).

    The writing is on the wall for single/dual. The quad race on phones/tables is moving FAR faster than it did on PC's. After win8 these things will start playing a lot more nicely with our current desktops. Imagine an Intel x86 based quad (hopefully) with someone else's graphics running the same stuff as your desktop without making you cringe over the performance hit.

    I'm not quite sure how you get to Tegra3 costing more, having higher thermals (umm, ipad 3 is hot, not tegra3). The die is less than 1/2 the size of A5x. Seems they could easily slap double the gpus and come out about even with QUAD cpu too. IF NV double the gpus what would the die size be? 162mm or smaller I'd say. They should have went 1920x1200 which would have made it faster than ipad 2 no matter what game etc you ran. Unfortunately the retina screen makes it slower (which is why apple isn't pushing TEGRA ZONE quality graphics in their games for the most part...Just blade?). They could have made this comparison a no brainer if they would have went 1920x1200. I'm still waiting to see how long these last running HOT for a lot of people. I'm not a fan of roasted nuts :) Too bad they didn't put it off for 3 months and die shrink it to at least 32nm or even 40nm would have helped the heat issue, upclock the cpu a bit to make up for 2 core etc. More options to even things out. Translation everything at xmas or later will be better...Just wait if you can no matter what you want. I'm salivating over a galaxy S2 but it's just not quite powerful enough until the shrinks for s3 etc.
    Reply
  • tipoo - Wednesday, March 28, 2012 - link

    I didn't say the Tegra 3 is more expensive or has higher thermals; I said the A5X, with higher clocked cores or more cores would be, and we all know Apple likes comfortable margins. Would I like a quad core A5X? Sure. Would I pay more for it? Nope. Would I switch for reduced battery life and an even hotter chip than what Apple already made? Nope. With the retina display, the choice to put more focus on the GPU made sense, with Android tablets resolution maybe Tegra 3 makes more sense, so you can stop attacking straw man arguments I never made. There are still only a handful of apps that won't run on the first iPad and that's two years old, "only" two cores won't hold you back for a while, plus iOS devs have less variation of specs to deal with so I'm sure compatibility with this iPad will be assured for at least two or three years. If I was buying one today, which I am not, I wouldn't be worried about that.

    Heck, even the 3GS runs most apps still and gets iOS updates.
    Reply
  • pickica - Monday, April 02, 2012 - link

    The New Ipad 2 is probably gonna have a dual A15, which means dual cores will stay. Reply
  • Peter_St - Monday, April 02, 2012 - link

    The problem here is that most people have no idea what they are talking about. It was just few years ago that we all used Dual Core CPUs on our Desktop Computers and we ran way more CPU load intensive applications, and now all of a sudden some marketing bonzo from HTC and Samsung is telling me that I need Quasd Core CPU for Tablets and mobile devices, and 2+ GB of RAM,
    If you really need that hardware to run your mobile OS, then I would recommend you to fire all your OS developers, get a new crew, and start from scratch...
    Reply
  • BSMonitor - Wednesday, March 28, 2012 - link

    If you were to run the same applications a tablet is designed to, then yes, your Pentium D would actually be overkill. Reply
  • PeteH - Wednesday, March 28, 2012 - link

    The point is made in the article is that it would be impossible provide the quad GPUs (necessary to handle that display) AND quad CPUs. Given you can only do one or the other, quad GPUs is the right choice. Reply
  • zorxd - Wednesday, March 28, 2012 - link

    was it also the right choice to NOT upgrade the GPU when going from the iPhone 3GS to iPhone 4? Reply

Log in

Don't have an account? Sign up now