POST A COMMENT

64 Comments

Back to Article

  • warisz00r - Friday, December 17, 2010 - link

    The graphics shown on them looks quite... erm... frosted butts? Reply
  • zebrax2 - Friday, December 17, 2010 - link

    It looks like to me that they made the characters super shiny but because of the low polygon count of the character themselves it created chunks of colors instead of a nice transition. Reply
  • luigi777 - Friday, December 17, 2010 - link

    i have nexus one i didn't experinced the "supershiny" effect. the woman's skin looks much better Reply
  • B3an - Friday, December 17, 2010 - link

    Pretty disappointing graphics for something that's meant to show them off. The textures and modelling are not exactly great quality either. Reply
  • nnkirk - Friday, December 17, 2010 - link

    Brian, where is n900 benchmarks ? i read the software suite supports maemo also,..

    Nnkirk
    Reply
  • Brian Klug - Friday, December 17, 2010 - link

    Oops, that actually was the N900. I don't know why I put N8, it's because I've been staring at that phone a lot lately. Fixed!

    -Brian
    Reply
  • solnyshok - Friday, December 17, 2010 - link

    wrong resolution for Nokia n8. Reply
  • kagenokurei - Friday, December 17, 2010 - link

    yup..N8 has 640x360(nHD)..

    and it definitely doesnt have a 600MHz OMAP..neither a PowerVR SGX..looks like its supposed to be the N900...

    i'll be running the test on my N8 now..just to see where it sits w/ its 680MHz ARM core and Broadcom GPU...
    Reply
  • kavanoz - Friday, December 17, 2010 - link

    My N8 is not able to run some of the tests (ie: Egypt) Reply
  • Brian Klug - Friday, December 17, 2010 - link

    That should've been the N900, my mistake!

    -Brian
    Reply
  • NickDG - Friday, December 17, 2010 - link

    Can we see an OMAP3 SoC running at 1ghz? Droid X perhaps?

    Thanks
    Reply
  • Brian Klug - Friday, December 17, 2010 - link

    I had to send ours back, but I'm trying to find one to include when we launch the smartphone bench. What I have now represents what I had on hand, but I'm equally interested in seeing Droid 2 or Droid X scores at 1 GHz. I'm trying to snag one. ;)

    -Brian
    Reply
  • sarge78 - Friday, December 17, 2010 - link

    Just behind the new snapdragons.

    Egypt. 13fps

    Pro. 23fps

    (Motorola defy @ 1ghz)
    Reply
  • sarge78 - Friday, December 17, 2010 - link

    ^
    The defy is running 2.1 btw (it shouldn't make a huge difference though?)
    Reply
  • zorxd - Friday, December 17, 2010 - link

    Isn't this a 320x480 device? The Droid X/2 would also be interesting. Reply
  • sarge78 - Friday, December 17, 2010 - link

    It's got the usual Motorola 854x480 FWVGA screen (essentially the same phone as droid 2/X etc)

    Just like the new Qualcomms they'll be SGX530 or memory bandwidth bound.
    Reply
  • deputc26 - Friday, December 17, 2010 - link

    Good call, that's a fairly glaring omission. Reply
  • probedb - Friday, December 17, 2010 - link

    Seriously? Is this right? Same GPU but the iPhone 4 has a faster CPU and yet it's slower? Reply
  • Chloiber - Friday, December 17, 2010 - link

    Resolution and same GPU. Reply
  • choirbass - Friday, December 17, 2010 - link

    yep.

    3gs is .1536mp @ 320x480 600mhz
    most androids are .384mp @ 800x480 1ghz
    ip4 is .6144mp @ 960x640 850mhz
    ipad is .786432mp @ 1024x768 1ghz

    so the ip4 is driving 4x the resolution of ip3gs, most androids 2x and ipad just over 5x

    hopefully that should put the benchmarks more into perspective, because otherwise there is a huge unexplained disparity like was mentioned.
    Reply
  • Shadowmaster625 - Friday, December 17, 2010 - link

    We need to see the benchmark results adjusted for screen resolution. The simplest way to do it is take the screen resolution and multiply the X and the Y and then multiply that by the FPS score then divide by 1 million.

    For example, the iPad 4 scored 4.6 and has a resolution of 1024x768. So the "real" score would be 1024 x 768 x 4.6 / 1000000 = 3.62. And the 3GS is 320x480x14.5/1000000 = 2.23. By doing this we will easily see in the graphs that the iPad is 62% faster than the 3GS.
    Reply
  • 7Enigma - Friday, December 17, 2010 - link

    The problem with this (unlike desktop/laptop computers) is that the phone is going to run native resolution ALL the time. You can't simply lower the res and get better performance, so while the numbers are skewed, I'd argue they are accurate.

    What obviously has to happen then is games made for a higher resolution device such as the iPhone4 will have to have lower-quality textures/effects to maintain playable while the lower res phones can have higher-quality textures/effects.

    Similar to the current (or maybe a bit ago) desktop gaming where people with >24" monitors were having to keep some of the eye candy off (high levels of AA for instance) while those of us gaming on 19" screens could have everything maxed out. It's personal preference whether bigger (with less quality) is better.
    Reply
  • vol7ron - Friday, December 17, 2010 - link

    I totally agree that the numbers should be kept as is, since that is how they will be viewed. It's not theoretical figures we're trying to see, it's actual performance.

    It's important to note that certain devices have higher resolution, thus the fps may be affected, but the visual quality should seem better.

    Though, it would be nice to have an indexed theoretical comparison as mentioned. Something that uses the lowest common denominator so we can compare the performance in terms of resolution. Of course, there are many factors that would affect this index (memory size, memory speed, gpu, cpu, available nand, screen type, screen resolution, etc), but since we're lumping the memory/gpu/cpu all into one category, the resolution is the only other variable that will mostly affect the performance. - I'm curious to see what kind if a correlation truly does exist.
    Reply
  • therealnickdanger - Friday, December 17, 2010 - link

    From the first test, using your methodology, it certainly does shake up the list a bit...

    Samsung Fascinate 8.3, Google Nexus S 8.0 top out the list.
    Reply
  • therealnickdanger - Friday, December 17, 2010 - link

    Dumb spam filter...

    T-Mobile G2 6.6, T-Mobile myTouch 4G 6.3 seem comparable.
    Reply
  • therealnickdanger - Friday, December 17, 2010 - link

    Apple iPhone 4 3.6, Apple iPad 3.6 are identical... big surprise! Reply
  • therealnickdanger - Friday, December 17, 2010 - link

    OK, apparently I can't post any more of my "spam".

    Anyway, it would be good of AT to normalize the benchmarks for resolution. It seems like the only fair way to compare them. Of course, the raw data is important as well, but quadrupling resolution is certainly an important distinction.

    I'm running the full suite on my LG Optimus S right now and it's doing pretty well - much better Egypt score than the Optimus One.
    Reply
  • therealnickdanger - Friday, December 17, 2010 - link

    Sprint LG Optimus S - root, but stock ROM, no OC:

    Egypt: 14.6fps
    Egypt FSAA: 12.8fps

    But then it crashed trying to upload results. Oh well, the site thinks my phone is a Optimus One (similar, but not identical), so it's probably best that I don't upload my scores.
    Reply
  • zorxd - Friday, December 17, 2010 - link

    It actually is. The iPad has a better CPU so should be faster. However it has less RAM. But it's surprising that they get the exact same score. Reply
  • Shadowmaster625 - Friday, December 17, 2010 - link

    lol does it think you are trying to sell something? Like that chinese spammer who kept posting t-shirts 5, shoes 9 bla bla bla Reply
  • metafor - Friday, December 17, 2010 - link

    That really only makes sense if the benchmark is completely fill rate bound. Which it doesn't appear to be. There's more to 3D graphics than just filling the pixels on screen. Reply
  • marc1000 - Friday, December 17, 2010 - link

    yep, they kept the same GPU but increased the resolution A LOT. this thing is looking bad for apple, because the new hummingbird is so much faster... if only the developers for android could leverage performance.... Reply
  • ET - Friday, December 17, 2010 - link

    The NookColor looks like it's going to be a good Android tablet to buy once it gets Android 2.2 and app store access next month. Can't help but wonder what kind of 3D performance it has. Reply
  • kilkennycat - Friday, December 17, 2010 - link

    Would be very interesting to add the GLBenchmark 2.0 results for the upcoming Tegra2-based LG Optimus 2X phone to this collection. See:-

    http://www.dailytech.com/article.aspx?newsid=20411

    Likely to blow away all the competition? Dual-core 1GHz A9 Cortex-ARM plus nV graphics.

    Also, it now seems pretty obvious why Steve Jobs is running scared from the latest GPU-accelerated Adobe Flash 10.1/10.2. The iPhone and iPad are true graphics slugs...
    Reply
  • marraco - Friday, December 17, 2010 - link

    How pathetic is the iPad. Reply
  • xype - Friday, December 17, 2010 - link

    Infinity Blade plays nicely on the iPad and the 3GS I have. As does Rage HD.

    But then, I haven't seen any comparison where those two were running on Android, so who knows, right?
    Reply
  • silverblue - Friday, December 17, 2010 - link

    The lack of cross-platform games isn't helping here.

    For one, I'd like to see how Asphalt 5 performs, but it seems frame-capped on or under-optimised for the Galaxy S. Give it a benchmark mode and you could compare the iPhones and Android devices in at least one title.

    A benchmark for Angry Birds would likely be a waste of time as it's smooth enough.
    Reply
  • headrush69 - Friday, December 17, 2010 - link

    Obviously these tests are resolution dependent since the iPhone 3GS with a slower CPU and same SGX 535 beats the iPad and iPhone 4.

    As a result comparison direct comparison between Android and newer iOS devices isn't exact but it appears iOS devices are lagging.

    One wonders if the SGX 540 used in new iOS devices will change that or if it's a poor Open GL ES stack implementation from Apple, same as on OS X.
    Reply
  • marraco - Saturday, December 18, 2010 - link

    1024x768 does not justify the difference.

    And the fruity company sells the screen size, so it should sell a fully functional unit.

    It pretends to "sell a Ferrari", but without the Ferrari engine.
    Reply
  • prdola0 - Friday, December 17, 2010 - link

    Hello AT,
    the screenshots in the article show some serious rendering glitches. My guess is that part of the graphics memory is damaged. This is NOT how the benchmark should look. On which device did you get these results?

    Best regards,
    Prdola
    Reply
  • Devo2007 - Saturday, December 18, 2010 - link

    I saw these same issues on my Galaxy S (Froyo). so it's possible that is the phone they grabbed the screenshots from. Reply
  • overzealot - Saturday, December 18, 2010 - link

    Have you ever seen a rendering with damaged memory before?
    That video is nothing like it. Nothing AT ALL.
    Reply
  • LordSojar - Friday, December 17, 2010 - link

    Why no Tegra2 benchmarks? LG's Optimus 2X utilizes it, as does an upcoming phone from Motorola. And that's excluding the tablets that current are or soon will be using it: Viewsonic G, RIM Playbook, Notion Ink ADAM, Samsung's Galaxy S-2 Tab, and Motorola's tablet that was shown running Honeycomb.

    So.. why no Tegra2 results. We already know it will likely slaughter everything, but I'd like to know by how much.
    Reply
  • A5 - Friday, December 17, 2010 - link

    They have to have hardware to test, and as far as I know they haven't reviewed any Tegra 2 hardware. Reply
  • Roland00Address - Friday, December 17, 2010 - link

    They do have a tablet running Tegra2 hardware. It is the viewsonic g1 tablet running Tegra2 and 1024x600

    We got a first look here
    http://www.anandtech.com/show/4054/first-look-view...
    Currently they are getting it ready for a review. Sadly the software it comes with stinks so they have to root it to make it useful.
    Reply
  • srench - Friday, December 17, 2010 - link

    lordSojar wrote: "So.. why no Tegra2 results. We already know it will likely slaughter everything, but I'd like to know by how much."

    You should wait for the hardware and benchmarks before calling the result. Nvidia have very little experience in low power SoC.

    Most of the players listed are well into next gen graphics chips. I expect the later versions of SGX will appear in ipad2 and possibly MP versions.

    Nvidia would have to beat these results by miles just to stay relevant in this space.

    re ipad results - too many comments are not taking into account screen size!
    Reply
  • DeathReborn - Friday, December 17, 2010 - link

    Although not benched by me these are the scores for the Malata Zpad (Tegra 2 tablet)

    1024x600 Res:

    Egypt - 19.9 fps
    Pro - 41.4 fps

    For Comaprison the Galaxy Tab gets this:

    Egypt - 15.9 fps
    Pro - 30.9 fps

    About Hummingbird SGX540 performance with a higher resolution. Performance of Tegra 2 may increase with a Dual Core configured OS (Malata Zpad isn't running one).
    Reply
  • epol - Friday, December 17, 2010 - link

    If you are looking to add a Droid 2 to the mix, use the Droid 2 Global. This is what will be replacing the Droid 2 and it runs at 1.2 ghz. Right now they're both being sold but Moto's plan is to let the Droid 2 Global replace the Droid 2 as stock sells out. Reply
  • zhill - Friday, December 17, 2010 - link

    The disparity in actual performance between the iPhone 3GS and iPhone 4/iPad is exactly why I'm very skeptical of a *major* upgrade to the display in the next iPad. Anything approaching "Retina" resolutions (300dpi) would mean the panel would far exceed 1080p and there is no way that the current-gen or even next-gen SoCs could push enough pixels for playable framerates at that resolution--probably not even Tegra2. If 1024x768 is 0.7MP and 132ppi then to get even to 200ppi you'd need something at least 1550x1162 (more realistically it would be either 1400x1050 and approx 180ppi or 1600x1200 and 206ppi) while a full 300ppi is a ridiculous 2325x1743--almost the same number of pixels as a 30" lcd monitor! So for the sake of argument, let's say they go with 1600x1200 to give the iPad a nice 206ppi. That is a 1.92MP display and a 144% increase in pixel count! I haven't seen any verifiable performance tests of Tegra2 and current SoCs, but there is a chance that nvidia's silicon could be up to the task; however, that would mean Apple would have to abandon their own silicon, which I don't see happening. Battery life would also most likely be abysmal, and the cost of such a large high-res screen would make it infeasible. Hence, I don't think we'll see any major upgrades to the iPad screen any time soon and these benchmarks show why. Reply
  • overzealot - Saturday, December 18, 2010 - link

    Unless they decide to run a discrete GPU. It'd have to be pretty light though. Reply
  • ThePooBurner - Friday, December 17, 2010 - link

    The N900 is known to perform substantially better at almost everything when overclocked, and most of them can run at 900mhz without breaking a sweat. Could you try running it on an OC'ed N900? :) Reply
  • michael2k - Friday, December 17, 2010 - link

    Same HW, faster clock, but lower score? Even the iPhone 3GS scores higher... I'm assuming this is due to the screen resolution, then, more than anything else?

    Also, the caption for the second graph is wrong if you meant to say OpenGL ES 1.1
    Reply
  • ltcommanderdata - Friday, December 17, 2010 - link

    People have already talked about iPhone 4/iPad performance against the iPhone 3GS not looking good due to the Retina Display, but I wonder why the gap closes in the OpenGL ES 1.1 results? Perhaps the iPhone 3GS is bandwidth or CPU bound in the simpler OpenGL ES 1.1 test vs. the iPhone 4/iPad being shader and ROP bound in the OpenGL ES 2.0 test? It's interesting that the iPhone 4/iPad seem to overtake a few Android phones in the OpenGL ES 1.1 test which probably indicates different levels of OpenGL ES stack and driver optimizations.

    Personally, I think it would be useful to supplement the raw results with comparisons normalized to the device resolution. This will give an indication of raw GPU power and level of OpenGL ES stack and driver optimization between devices. This is especially useful to compare progress between Android devices running different OS versions.

    As well, for the iPhone 4, these results give credence to the suggestion that in some cases it might be beneficial to stick with the 480x320 resolution but implement AA rather than go 960x480. The iPhone 4's 64-bit memory controller with double the memory bandwidth should be very helpful in making AA viable and sticking with a common 480x320 resolution should allow a common code path between 3rd gen and 4th gen devices which is also tempting for developers.
    Reply
  • ltcommanderdata - Friday, December 17, 2010 - link

    Oh, I don't suppose the OpenGL ES 1.1 test can be run on the previous MBX Lite based devices? It would give an interesting point of comparison of how much progress has been made in the last 2 years. Reply
  • madmongoose - Friday, December 17, 2010 - link

    How is it a fair comparison if u compare each with its own native resolution

    Shouldn't they be tested at one low resolution that hopefully all support ?
    Reply
  • DeathReborn - Friday, December 17, 2010 - link

    This site lists loads of scores across various devices.

    http://www.glbenchmark.com/result.jsp?orderby=407&...

    The Malata Zpad is the Tegra 2 device listed on there, hopefully some Advent Vega's (also Tegra 2) will be on it soon.
    Reply
  • sarge78 - Saturday, December 18, 2010 - link

    Looks like a pure GPU benchmark to me, I get the same results from going from 800Mhz all the way to 1.3Ghz (OMAP 36xx/SGX530) Reply
  • DesktopMan - Saturday, December 18, 2010 - link

    Would have liked to see HTC Desire in these charts as it's very popular, at least here in Europe. Also, the correct name for Samsung Fascinate is Samsung I9000 Galaxy S as Fascinate is just a re-branded version for Verizon and a name completely unknown to global readers. Reply
  • Spazweasel - Monday, December 20, 2010 - link

    A closer look:

    Obviously, the number of pixels in the screen matters a great deal. I've calculated the number of pixels on the screen, and using the performance data given have calculated the number of megapixels per second the phones in question can push:

    EGYPT - MEGAPIXELS PER SECOND:
    SamsungFascinate SGX540 8.2
    GoogleNexusS SGX540 8
    TMobileG2 Adreno205 6.6
    TMobileMyTouch4G Adreno205 6.2
    AppleIPhone4 SGX535 3.6
    AppleIPad SGC535 3.6
    NokiaN900 SGX530 2.8
    MotorolaDroid SGX530 2.8
    HTCEvo4G Adreno200 2.8
    NexusOne Adreno200 2.7
    AppleIPhone3GS SGX535 2.2
    LGOptimusOne Adreno200 1.5

    PRO - MEGAPIXELS PER SECOND
    GoogleNexusS SGX540 16
    SamsungFascinate SGX540 15.6
    AppleIPhone4 SGX535 9.9
    AppleIPad SGX535 9.9
    TMobileMyTouch4G Adreno205 9.6
    TMobileG2 Adreno205 9.4
    NokiaN900 SGX530 4.9
    MotorolaDroid SGX530 4.6
    AppleIPhone3GS SGX535 3.7
    HTCEvo4G Adreno200 2.2
    NexusOne Adreno200 2.2
    LGOptimusOne Adreno200 1.3

    The clear winners are the Fascinate and Nexus S. The next best are the Apple iPhone4 and iPad on the Pro benchmark, while on the Egypt benchmark the TMobile products are third and fourth.

    CPU matters hugely (compare iPhone 3GS, 4, iPad, same GPU, different CPU), but the scaling is much greater on the Pro benchmark (better CPU gives almost 3x the megapixels per second) than on the Egypt benchmark (about 1.7x).

    Resolution comes at a price in framerate. But you're Anandtech readers, you knew that. Right?
    Reply
  • sarge78 - Monday, December 20, 2010 - link

    Doesn't the iPad/iP4 SGX535 run @ ~200Mhz? Almost twice the 110Mhz of the 3GS... (Largely due to the die shrink) Reply
  • ibex333 - Wednesday, December 22, 2010 - link

    Why would you benchmark Motorolla Droid, but now Droid 2 GLOBAL, and Droid X?

    Seems unfair to pit an outdated device against newer, faster ones.
    Reply
  • smile chen - Monday, December 27, 2010 - link

    When playing games and operating ipad without Reply
  • smile chen - Monday, December 27, 2010 - link

    china ipad Reply
  • Rish105 - Friday, December 31, 2010 - link

    The graphics in this app don't look to good. I think it is directly testing the shaders and not pushing fill rate at all. Which would put iOS devices at a major disadvantage, since Apple is using the same crappy OpenGL ES 2.0 stack that they use on Macs. Otherwise I'm sure that the SGX 535 is about equal to the Adreno 205 (in terms of shear fill rate and raw processing power), since it was over twice as fast as the Adreno 200 in earlier benchmarks. And the SGX 540 should completely top the chart by more than it is doing now (Samsung's fault for using crap 3D drivers like Apple)

    Also we need resolution relative scores. An iPhone 4 is pushing 4X the res of a 3GS, and 2X the res of the WVGA phones. Which would put it at 12 FPS if it scales perfectly and just about half the SGX 540 which seems realistic.

    But still the graphics in this app look terrible. I've played Real Racing 2, Infinity Blade, Rage HD, NOVA 2 and other intensive games on my 3rd gen iPod and they all run at a stable 30 FPS and they all look better than this (better shader usage, more detail, more polygons, etc.).

    We need a more comprehensive benchmark. This is a bad one.
    Reply

Log in

Don't have an account? Sign up now