Comments Locked

70 Comments

Back to Article

  • tipoo - Wednesday, March 30, 2011 - link

    Any idea why the iPhone 4 does worse than even the 3GS in the second page benchmarks? And would it be possible to throw an iPad 2 in the comparison, as the dual core A5 chip is more like what this will be competing against?

    Thanks
  • Brian Klug - Wednesday, March 30, 2011 - link

    The reason for the iPhone 4 lagging iPhone 3GS is display resolution, which unfortunately right now we can't test at anything other than native. GLBenchmark 3.0 will change a lot of that as we'll be able to render off-screen at an arbitrary (common) size.

    I didn't include the A5 because I didn't include any tablets at all in the comparison.

    -Brian
  • FATCamaro - Wednesday, March 30, 2011 - link

    iphone 4 has 4 times the pixels to draw to as opposed to the 3GS. The iPad 2 isn't in the comparison because it would blow everything out of the water.
  • DJMiggy - Wednesday, March 30, 2011 - link

    wat?
  • choirbass - Wednesday, March 30, 2011 - link

    http://www.anandtech.com/show/4225/the-ipad-2-revi...
  • B3an - Wednesday, March 30, 2011 - link

    wateva.

    GLBenchmark 2.0 Egypt:
    iPad 2 = 44
    This phone (Adreno 220) = 38.3

    GLBenchmark 2.0 Pro:
    iPad 2 = 57.6
    This phone (Adreno 220) = 93.6
    Or Adreeno 220 with vSync = 58.8

    Yeah the iPad 2 has a higher res screen, but it's not that much higher. The Adreeno 220 certainly competes. Not bad for a device thats a lot smaller.
  • snoozemode - Wednesday, March 30, 2011 - link

    Not that much higer? The iPad 2 has twice the amount of pixels compared to this phone.
  • choirbass - Wednesday, March 30, 2011 - link

    within 6-12 months or so, there should be closely competitive gpus again... always a refresh and release cycle.
  • jalexoid - Wednesday, March 30, 2011 - link

    It's not all about the pixels.
  • therealnickdanger - Thursday, March 31, 2011 - link

    If you are rendering in native resolution, then yes, yes it is.

    iPad = 1024x768 = 786,432 = 100%
    iPhone = 960x640 = 614,400 = 84%
    Adreno = 800x480 = 384,000 = 48%

    Adreno doesn't even have to work half as hard as the iPad. While impressive for 480p, Adreno has work to do to catch up to Apple's GPU for higher resolutions.
  • metafor - Thursday, March 31, 2011 - link

    Keep in mind that as you scale to higher resolutions, you won't be clipping the 60fps limit as much.

    So if Adreno is capable of going higher but isn't due to the fps cap, then it won't be slower at higher resolutions.

    This isn't true of Egypt from what it seems though.
  • Alexstarfire - Thursday, March 31, 2011 - link

    I would agree when you're strictly comparing GPUs, but we're almost comparing platforms in here. The problem is that you can't run most things on mobile platforms at any resolution you want. You're stuck with native resolution in just about every case.

    If you're trying to figure out which GPU is better then you certainly need to take resolution out of the equation. Can't really do that with mobile devices though.

    You also have to remember that performance isn't necessarily linear across resolutions, especially if there is an FPS cap.
  • michael2k - Thursday, March 31, 2011 - link

    What? 840x480 is half the iPad's resolution! 1024x768!

    The iPhone is 960x640, also much higher.

    The iPad->iPad 2 difference of 6.4->44.8 or 17.6->57.6 tells us the iPhone 4 should be quite competitive with the dual core Snapdragon; 5.9->42 for Egypt and 50 for GL Pro 2, and at a much higher resolution to boot.

    Any phone that uses the Adeno 220 and a higher resolution (like via HDMI out), will see their scores drop in half, while the iPhone should chug along without much difference.
  • B3an - Thursday, March 31, 2011 - link

    How can the iPhone 4 remotely compare? Dont you not mean iPhone 5?

    Because the Motorola Atrix 4G has about the same screen res (540 x 960) yet it performance more than 2x as well as iPhone 4 in most benches. The Atrix's CPU and GPU are not even as fast as this new Scorpion SoC.
  • michael2k - Thursday, March 31, 2011 - link

    You're right, I meant A5 on the iPhone 5.
  • kade1878 - Saturday, April 2, 2011 - link

    So lets compare the iPhone 4 to the Motorola Atrix 4G:

    iPhone: 960x640 = 614400
    Atrix: 960x540 = 518400

    So the ratio is ~1.19 or the iPhone 4 has 19% more Pixels than the Atrix.
    Now compare that to the Benchmark result:

    iPhone: 5.9 fps x 614400 pixels/frame = 3.624.960 pixel/s or ~3.6 MPixel/s
    Atrix: 18.7 fps x 518400 pixels/frame = 9.694.080 pixel/s or ~9.7 MPixel/s
  • flipside101 - Saturday, April 2, 2011 - link

    its slightly off topic but does anyone hve a handle on why no one else is using the PowerVR SGX543MP2 . its not apple tech but is it possible that they have an exclusive license ?

    ta
  • dagamer34 - Saturday, April 2, 2011 - link

    TI and Samsung are the only other major companies besides Apple that usually use PowerVR chips. This year, Samsung stuck an ARM-based GPU into their chip, the Exynos (ARM GPU is the Mali 400) and Ti's OMAP 4430 has so far only been used in the Blackberry Playbook. All smartphone uses of the OMAP 4xxx platform are to be launched in the 2nd half of this year.
  • GnillGnoll - Wednesday, April 6, 2011 - link

    I'm sure Intel and Renesas would disagree with your definition of "major companies", among others. And the LG Optimus 3D smartphone (with OMAP4430) will be available by the end of the month.

    Renesas announced a SoC with SGX543MP2, so it's certainly not Apple exclusive tech.
  • scofflaw - Wednesday, March 30, 2011 - link

    The original MDP8655 which I have has VEE display backlight control which when enabled can really lower the battery drain when running certain apps. Does this new MDP8660 have it as well?
  • Brian Klug - Wednesday, March 30, 2011 - link

    There's no auto brightness option in the display brightness dialog box, at least on this build. I'm not entirely certain whether it's absent however. Is there some better way to check?

    -Brian
  • Mike1111 - Wednesday, March 30, 2011 - link

    Not true, there is a Motorola Xoom in the first Quadrant Benchmark.
  • Brian Klug - Wednesday, March 30, 2011 - link

    Oops, I had the results correct but didn't regenerate the graph after fixing things. Should be right now.

    -Brian
  • Exelius - Wednesday, March 30, 2011 - link

    It's not really ironic that the prototype is running a newer version of android than most retail phones are. That's kind of the idea of a prototype. Companies have no incentive to update android on their phones after they're released, so that shouldn't come as a shock at all.
  • metafor - Wednesday, March 30, 2011 - link

    A lot of things aren't really ironic.

    Like rain on your wedding day or a free ride that you bought anyway.

    Doesn't stop people from calling it ironic though.
  • nermie - Wednesday, March 30, 2011 - link

    Choosing lines from a song called "Ironic" as an example of things that aren't ironic is pretty ironic.
  • metafor - Thursday, March 31, 2011 - link

    No more ironic than a song called "Ironic" that contains nothing that's ironic.
  • marc1000 - Thursday, March 31, 2011 - link

    this is really a lot of irony... =D
  • metafor - Wednesday, March 30, 2011 - link

    I believe the MDP comes with a plethora of profilers and hardware plugs available to software. One of these measure aggregate power/current and graphs it. It'd be interesting to see how much power the SoC is eating during CPU and/or GPU intensive tests.

    Since other devices don't have these profilers, there wouldn't be much in the way of comparing but having absolute numbers would be interesting in and of itself.
  • efeman - Wednesday, March 30, 2011 - link

    Performance is getting pretty impressive. I still don't see using my phone as a true mobile gaming platform in the near future; the battery life just doesn't cut it when I need the phone for actual phone uses. I wonder if they'll ever drop down smartphone battery drain enough (or develop insanely better batteries) to allow for this kind of usage for extended periods of time.
  • joshv - Wednesday, March 30, 2011 - link

    "isn't it ironic that prototype devices are running newer builds of Android than most smartphones you can buy retail right now"

    No - why wouldn't you expect prototypes to be running newer builds than release hardware?
  • Johnmcl7 - Wednesday, March 30, 2011 - link

    I thought that comment odd as well as I'd expect a prototype to be running a newer OS build than current released hardware
  • sleepeeg3 - Wednesday, March 30, 2011 - link

    What is the point? I thought the benefit of these faster processors was allegedly better battery life? I really don't care if I get 1000fps on Angry Birds...
  • Conner_36 - Wednesday, March 30, 2011 - link

    Do the developers have something against Mac? What's the reason why they don't have an iOS benchmark?
  • bplewis24 - Wednesday, March 30, 2011 - link

    What is iOS? What is Mac?
  • Conner_36 - Thursday, March 31, 2011 - link

    The computer platform that has 90%+ market share of computers sold over 1000 dollars and the smartphone platform that has over 50% of all smartphone profits.
    Ring a bell?
  • ilkhan - Thursday, March 31, 2011 - link

    You can't benchmark hype and magic.
  • Sterman - Thursday, March 31, 2011 - link

    Thank you sir,

    May I have another? LOL
  • softdrinkviking - Thursday, March 31, 2011 - link

    it's a lot harder to play in a closed software ecosystem, and a company like futuremark probably believes that it's more trouble than it's worth.
  • mfergus - Thursday, March 31, 2011 - link

    It's just a benchmark to gauge hardware. Apple doesn't design any unique hardware anyways so you can always just look for something with similar hardware. They'll probably come out with something for iOS eventually but its pretty understandable why they never bothered with OSX.
  • michael2k - Friday, April 1, 2011 - link

    Um, the benchmarks for OpenGL ES2 and Egypt run on the iPhone and iPad.
  • mfergus - Friday, April 1, 2011 - link

    I was referring to 3dmark, that was his question.
  • Omid.M - Wednesday, March 30, 2011 - link

    Except, the MSM8960 is what we should be looking for, due to the integrated LTE antenna. Right?

    Does the 8960 have the Adreno 220? I forget. I think 28nm 8960 will be substantially faster and more power efficient than Tegra 2, assuming T2 doesn't improve that much architecturally by the time 8960 is in mass produced devices.
  • jordanclock - Wednesday, March 30, 2011 - link

    The MSM8960 will perform identically, except it will have LTE. All MSM8x60s will have the same hardware, with the sole difference being the cellular antenna. Since this MDP has no cellular connectivity at all (or even WiFi) it's a moot point, anyway. They can't test any of that kind of stuff yet.
  • Omid.M - Thursday, March 31, 2011 - link

    Great point. Thanks for the clarification.

    Really impressive SoCs coming out. I just wish battery tech evolved instep with SoCs.
  • metafor - Thursday, March 31, 2011 - link

    8960 will bring a new CPU design; Krait will replace Scorpion.
  • Lucian Armasu - Thursday, March 31, 2011 - link

    Tegra 2? The 1.5 Ghz quad core Tegra 3 will be long out before the 1.5 Ghz dual core Snapdragon chip is out. I think they said it will ship around late Q4 this year
  • WizardofAwesomeness - Friday, April 1, 2011 - link

    Lucian,

    You have many postings all over the web where you're singing the praises of Tegra's graphics performance even when faced with irrefutable benchmark data, such as that from Anandtech, proving that Tegra 2 often has inferior graphics to SGX based devices like iPad2, and now ,it is clear that Tegra 2 has inferior performance to Qualcomm's 8660. Is Tegra 3 really better than future products/chipsets from Apple, TI, Samsung and Qualcomm, and do you think that these other companies are just going to stand still while NVIDIA alone produces new chipsets? Hmmm...think it through man, everybody has a killer next chip.

    Also, you've posted above a strong statement about the availability of future Tegra 3 quad core based OEM products vs. the commerical availability of those based on dual core products Qualcomm, TI and others. - again, how do you know? Just because NVIDIA says something in the press doesn't exactly make it so. Please stop being such an obvious NVIDIA fanboy, and do a little more research before you post such drivel.
  • scofflaw - Wednesday, March 30, 2011 - link

    On the 8655MDP VEE is available in the Trepn Profiler app under the Current option.

    Thanks.
  • cabbeer - Wednesday, March 30, 2011 - link

    Considering the A5 is a good gauge for whats going to be in the next iphone, I think it would be interesting to see the ipad2 benchmarks in the comparison.
  • Griswold - Saturday, April 2, 2011 - link

    Maybeso, but it will definitely run at a lower clockspeed.
  • Blacula - Thursday, March 31, 2011 - link

    The Samsung Galaxy Tab is in there too.
  • djgandy - Thursday, March 31, 2011 - link

    Looks like GLBenchmark is vsynced for the rest of the phones too.

    How exactly did you turn vsync off?
  • metafor - Thursday, March 31, 2011 - link

    As mentioned in the article, it's a feature the development phone he used allowed but most phones don't.
  • Chloiber - Thursday, March 31, 2011 - link

    ...when I look at 3DMark Mobile...my poor HTC Desire...

    Impressive performance! Looking forward to the HTC Pyramid ;)
  • duploxxx - Thursday, March 31, 2011 - link

    uggh its very ugly
  • alxxx - Thursday, March 31, 2011 - link

    TI omap4460 is sampling according to TI
    http://focus.ti.com/general/docs/wtbu/wtbuproductc...

    even though they haven't released the docs for omap4430
    but can get them via pandaboard.org
    see TRM http://pandaboard.org/content/resources/omap-devic...

    omap5430 and 5432 look interesting
    http://focus.ti.com/general/docs/wtbu/wtbuproductc...

    Hey Brian or Anand how about getting a pandaboard and doing some benching.
    You can load pandroid http://pandaboard.org/node/225/#android
    http://omappedia.org/wiki/OMAP_Pandroid_Main
    http://code.google.com/p/pandroid/

    I'd volunteer mine but a bit to far away plus not familar with android, mostly run angstrom.

    Other interesting one(for dsp people) is TI finally have their floating point dsps above 1Ghz
    with the release of the new integra/sitra chips http://focus.ti.com/dsp/docs/dspcontent.tsp?conten...
  • FITCamaro - Thursday, March 31, 2011 - link

    This is what I'm waiting for before I get a new phone. My Droid will be enough until then. I don't have 4G service in my area yet anyway.
  • PWRuser - Thursday, March 31, 2011 - link

    Hopefully the MDP has little RAM so that actual phones may perform better.

    It's early but in terms of performance, it appears that this SoC lines up somewhere between Tegra 2 and the (projected) Tegra 3 performance.

    Also it will be interesting to see a refresh of this unit when the fab switches to 28nm.
  • SheldonCooper - Friday, April 1, 2011 - link

    The Taiji test is pixel heavy since it is a closed environment whereas the Hoover test is texture heavy (lots of terrain). They are useful to compare unified shader vs dedicated pixel- and texture shader architecture.
  • kade1878 - Saturday, April 2, 2011 - link

    The iphone 4 does suck these days

    So lets compare the iPhone 4 to the Motorola Atrix 4G:

    iPhone: 960x640 = 614400
    Atrix: 960x540 = 518400

    So the ratio is ~1.19 or the iPhone 4 has 19% more Pixels than the Atrix.
    Now compare that to the Benchmark result:

    iPhone: 5.9 fps x 614400 pixels/frame = 3.624.960 pixel/s or ~3.6 MPixel/s
    Atrix: 18.7 fps x 518400 pixels/frame = 9.694.080 pixel/s or ~9.7 MPixel/s
  • Sameer254 - Sunday, April 3, 2011 - link

    What's Vsync?
  • Nick7237 - Sunday, April 3, 2011 - link

    The EVO 3D will sport a 1.2GHZ version of the above chipset but still with the Adreno 220, and when your using the final product on battery it will probably peak at 34 FPS when testing with Egypt.

    The iPad 2 which sports the Apple A5 (Dual core 1GHZ CPU and Dual Core GPU POWER VR SGX543MP2) was getting 44 FPS without AA and 44.8 with AA on using tested with Egypt but while on battery and at a higher 1024 X 768 resolution.

    Here's the link as proof: http://www.anandtech.com/show/4216/apple-ipad-2-gp...

    When the iPhone 5 is released which is going to sport the A5 but at a 960 X 640 it's going to yield a 50 FPS on the final product when tested with Egypt.

    Expect this years iPod touch to have the A5 as well

    Being that Gameloft and EA release most of their mobile games as exclusives to iOS and that Gameloft just licensed the unreal engine for 4 upcoming games with NOVA 3 being one of them, iOS will be the king of mobile Gaming this year with this Trio of devices that will outclass the competition and yield the Best looking games once again rivaling the PS3 and 360.

    When the iPhone 5 is released checkout this webiste for the benchmark results and you'll see the headline "iPhone 5 destroys the competition"
  • Hedgeson - Sunday, April 3, 2011 - link

    What about SONY's NGP? If you're talking about mobile gaming, don't forget dedicated consoles.

    Quad Core Cortex A9 and Quad core SGX543MP4. Same hardware as iPad 2 and possibly iPhone 5, except for having twice the cores.
    Gameloft and EPIC are already partner's of SONY for the NGPs launch.

    I know this website doesn't feature gaming hardware or software, but I'm curious how the NGP would compare with phones in benchmarks.
  • Nick7237 - Friday, April 8, 2011 - link

    I completely agree, I was salivating when I saw the NGP. To put it basically the NGP has 2 A5's inside. I must rephrase what I said now that you brought that up. The A5 will be the king of mobile gaming in smartphones and tablets this year. The NGP is a beast that I will buy from day one. They haven't even said what each core will be clocked at so realistically they could be clocked at 1.5GHZ each, which would make it better than two A5's. I think when the NGP comes out that Anandtech will definitely benchmark it and they'll state "NGP obliterates competition" especially the primitive 3DS.
  • 66699885 - Wednesday, April 27, 2011 - link

    very good!
  • kenshinhimura - Sunday, May 15, 2011 - link

    PossIble to bench against samsung galaxy s2, brian?
  • DeepDg - Tuesday, September 20, 2011 - link

    How can I obtain a copy of 3Dmark Mobile 2010 for Android?
  • rathnamachary - Monday, May 18, 2015 - link

    HI All,
    pls help me how to run glbenchmark by remote login(i.e.. ssh), i have ran but it is giving 'xopendisplay' error.

    mail id:[email protected]
  • rathnamachary - Monday, May 18, 2015 - link

    HI All,
    pls help me how to run glbenchmark 2.5.1 in ubuntu desktop by remote login(i.e.. ssh), i have ran but it is giving 'xopendisplay' error.

Log in

Don't have an account? Sign up now