Mar 04 AT 1:17 PM Taylor Wimberly 65 Comments

Tegra 2 CPU and GPU benchmarks, Motorola Atrix 4G vs LG Optimus 2X

The first dual-core Android phone has finally hit the market and more are on the way. NVIDIA promised big performance gains with their Tegra 2 system-on-a-chip (SoC), so we put two of their superphones through a series of benchmarks to see how they stacked up. Check out the results after the jump and then download the benchmarks to see how your Android phone compares.

The test lineup

We have a lot of phones laying around the A&M Labs, but we just went with six models for this round of testing. Devices we tested include both the Tegra 2 phones that are currently available (Atrix 4G and Optimus 2X), both phones with Android 2.3 (Nexus S and Nexus One), the T-Mobile G2, and the popular Motorola Droid.

Each device was given a clean reboot before testing and they are all running unmodified, stock versions of Android. An active WiFi connection was turned on to simulate real-world experience.

Neocore

AboutNeocore is an OpenGL-ES 1.1 graphics performance benchmark for Android devices. It shows off some of the techniques that are possible on accelerated platforms such as 1-pass light maps and bump mapping. [Download]

Neocore is one of the oldest GPU benchmarks and measures OpenGL ES 1.1 performance. The LG Optimus 2X easily came out on top with a score of 77.2 frames per second. The Motorola Atrix 4G fell behind in this test because it has the highest resolution (960 x540), which means it has to process the most pixels and do more work.

Nenamark1

AboutSet your GPU on fire with NenaMark, an OpenGL ES 2.0 benchmark! NenaMark is a benchmark/demo of OpenGL ES 2.0, using programmable shaders for graphical effects such as reflections, dynamic shadows, parametric surfaces, particles and different light models to push the GPU to its limits. [Download]

Nenamark1 is another popular benchmark that has been around awhile which measures OpenGL ES 2.0 performance. The Optimus 2X once again smokes the competition and is followed by the Nexus S.

Smartbench 2011

About: Smartbench 2011 is a multi-core friendly benchmark application that measures the overall performance of your smartphone. It reports both Productivity and Games Index to suit both productivity users and 3D gaming users. [Download]

Smartbench 2011 is a new benchmark that has been updated to be multi-core friendly. Note that the Tegra 2 offers about 3x the productivity performance over all the single core phones. In the games score, the Optimus 2X is once again the leader.

GLBenchmark 2.0.3

About: GLBenchmark 2.0 has been designed from the ground up to demonstrate and measure the true potential of OpenGL ES 2.0 Hardware. The built-in shader code (GLSL) generator enables real-time performance tuning and de-compositing. This is an in-valuable feature for OpenGL ES 2.0 Hardware vendors and Handset manufacturers. [Download]

No surprises here. The Optimus 2X leads the competition when it comes to GPU performance. The Nexus S is able to once again eclipse the Atrix 4G thanks to its lower display resolution.

Linpack

About: Speed test your Android device and ROM with this standard CPU benchmark. Check the speed of your Android device and compare it to other Android devices. Results in millions of floating point operations per second (MFLOPS). Save results or post to the website to beat the best times. [Download]

In Linpack we see no performance advantages for the Tegra 2 devices. This test does not appear to be multi-threaded and all the 1 GHz cores perform about the same. Qualcomm’s Snapdragon chips generally perform well in this test thanks to their FPU (SIMD/NEON/VFP) instructions. The real world performance between a G2 and Nexus S is pretty similar, so I’m not sure how relevant this test really is.

Sunspider Javascript 0.9.1

About: This benchmark tests the core JavaScript language only, not the DOM or other browser APIs. It is designed to compare different versions of the same browser, and different browsers to each other. This test mostly avoids microbenchmarks, and tries to focus on the kinds of actual problems developers solve with JavaScript today, and the problems they may want to tackle in the future as the language gets faster. This includes tests to generate a tagcloud from JSON input, a 3D raytracer, cryptography tests, code decompression, and many more examples. [Download]

Both the Tegra 2 devices lead the pack in Javascript performance.

Conclusions

Benchmarks are not the end-all, be-all of real-world performance, but they can help us measure how different platforms stack up. It is really hard to provide an apples-to-apples comparison of the GPU performance since there are so many different display resolutions, but we can see that the GeForce GPU inside the Tegra 2 offers the fastest performance for what’s currently available on the market.

In the couple of weeks that I have spent with both Tegra 2 phones, I can say they generally feel much faster than the single-core competition. Not many apps are optimized to take full advantage of multi-core processors, but there is definitely a noticeable difference in the overall responsiveness of Tegra 2 devices when multiple apps are running.

Even though the LG Optimus 2X has half the RAM of the Atrix 4G, it easily out-performs the competition in GPU benchmarks thanks to its lower resolution. The qHD resolution of the Atrix 4G is a nice spec to have, but there is a performance hit in most games since the GPU has to do more.

We can clearly see the performance gains of Tegra 2 in benchmarks that are multi-threaded like Smartbench 2011. This should be a good indication of what Tegra users can expect in the future as the Android OS and applications are optimized to take full advantage of dual-core processors.

For everyday use, the Nexus S can still hold its own against newer dual-core phones. I expect the Nexus S will continue to be competitive for the rest of 2011 thanks to it always having the latest version of Android. Look for Tegra 2 (and other dual-core CPUs) to take the lead in the second half of this year as the software catches up to the hardware.

Taylor is the founder of Android and Me. He resides in Dallas and carries the Samsung Galaxy S 4 and HTC One as his daily devices. Ask him a question on Twitter or Google+ and he is likely to respond. | Ethics statement

    Most Tweeted This Week

  • http://www.bewareofraj.com Raj

    I don’t know much about phone displays but I’m surprised at how much the resolution made such a big difference. To me, I didn’t think there was that big of a gap in resolution.

    • http://Website Dave

      The Motorola Atrix 4G is 540 x 960 = 518400 pixels
      At 55fps * 518000 pixels = 28490000 pixels per second

      The LG Optimus 2X is 800 x 480 = 384000 pixels
      At 77fps * 384000 pixels = 29568000 pixels per second

      29568000 / 28490000 = 1.038.

      So the screen resolution explains most of the difference but the Optimus is still ~4% faster.

      • http://Website watbetch

        Did you take the PenTile subpixel matrix into account for the Atrix?

        • http://Website Dave

          The PenTile subpixel matrix lowers power consumption and looks different (some say worse), but I wouldn’t expect it to affect the rendering speed. It’s part of the display, not the renderer.

          I’d guess the 4% difference comes from texture caching being more effective at lower resolutions. For example, a polygon might fit into 1 texture memory page in the lower resolution, but use 2 texture memory pages at a higher resolution causing the gpu to swap texture pages.

          • http://sean-the-electrofreak.blogspot.com/ Sean the Electrofreak

            Dave is right, PenTile algorithms are done by the display itself, not the GPU.

      • http://Website SynT

        Who care?

        100-500FPS? 20-10FPS? There is no point to do benchmark if all new games require Nvidia….

        … Even on tablet with much less performance…. still if you got tegra you can play. On Galaxy S NOOO

      • http://hom rulda

        OMFG LG 2x doesnt have FPS CAP !!!!! artix yes !!!

    • http://Website DivinuSfiliuS

      Oh yeah, it is a big difference for a CPU, to you might seem like small numbers, but things really add up like the number of polygons and texture resolution, shaddows etc.

  • http://Website troy g

    Nexus s still hanging with the big dogs!!

  • http://upanddowns.tumblr.com Jason

    why not use the mytouch 4g instead of the G2. The mt4g has more ram

    • http://Website Mikey

      Not suprising none of the six are a second gen snapdragon. This site wouldn’t want their precious tegras embarrassed.

      • http://androidandme.com Taylor Wimberly

        Um the G2 is a second-gen Snapdragon. And the Tegra 2 will only be the leader in performance until the dual-core Snapdragon comes out later this year.

        • http://Website Mikey

          Excuse me, I meant second gen 1ghz.

        • http://Website Mikey

          Leader in performance? Tegra didn’t exactly blow the comp outta the water.

          • http://Website lovo

            Yea but keep in mind the os isn’t built for them it’s Kik.da like a car with a rev limter speedometer says hr can do 160mph but its limited to 110 just wait till the new os. Shows up! I think gingerbread will support dual core no? Nd will the LG come to the states ?

        • http://Website DivinuSfiliuS

          …and so later this year there will be Tegra 3 quad core – don’t you read your tech news, nVidia is over a year ahead of everyone, they are going all out lol.

  • http://Website B2L

    It was nice to see this comparison. I love my Nexus S more everyday. It does everything I need it to.

  • http://Website Russ

    I I hope the release the 2x in the states soon, I what that phone.

  • http://Website brookmarin

    by year’s end we’re supposed to also see some quad-core tegra chips on the market. So the question is, will developers write code to utilize as many cores as the chip has? or will they just limit writing to only two cores?

    i’m waiting to hear what the competition has to say about nvidia’s aggressive schedule of releases… and what they will do to compete!

    • http://Website jonathan

      i think theyll do something like apple is doing with there A4 & A5 soc’s, basically theyre making games that would have less polygons & lighting effects on an A4 but when it detects the A5 it will add more lighting effects & polygon counts. its like PC games where you can pick the setting of the game to be high details or less detailed & depending on the GPU card youd have in your pc would depend on what effects you would see in your game. but since its only 2 gpu in apples gadgets then it will be easier to make games that will look better depending on what hardware you have. the same would be with the tegra 2 & tegra 3.

  • http://google.com Alex

    Why you guys always comparing with HTC G2? there is exist HTC Desire HD witch is more powerful than G2

    • http://androidandme.com Taylor Wimberly

      Yeah the Desire HD is a 200 MHz faster CPU and 256 MB more ram. It’s still a second-gen Snapdragon and on par with the G2, but a little faster.

      Short answer – no Desire HD in the USA, and I don’t have a myTouch 4G

      • http://Website Oliver

        You forgot about the HTC Inspire. It’s the Desire HD in US and all the root methods and roms are compatible from Desire HD if i might add.

      • http://google.com Alex

        oh yeah.. i completely forgot. It’s just because i’m from Holland we have here Desire HD

  • http://Website jay

    so all those people putting down LG should think twice. I was hoping moto would have the better performer but the LG looks great

  • http://Website jorge

    LG x2 FTW!

  • Brian Campbell

    I am still using my OG myTouch 3G with 2.2…..I’m so sick and tired of the phone, but hopefully the LG O2X will be out by the time my two year upgrade happens. I know its going to be the first dual core phone, and it will have better competition in about 3 months after it comes out. But everything about this phone is amazing, it kills the mycrapTouch 3G 1,000,000 times. I just want this phone to come to the US so I don’t have to order it from Europe.

  • http://Website Jim

    Maybe I’m missing something, but why not do a Quadrant test, too?

    • http://sean-the-electrofreak.blogspot.com/ Sean the Electrofreak

      Because it’s a crappy benchmark. It’s insanely popular because all of the Snapdragon phones perform so well on it because of the way the benchmark was written, and the HTC (Snapdragon) phones have a strong development community that likes benchmark tools. Of course, they’re going to pick the benchmark that has pretty graphs and gives their phones big numbers.

      It’s also a fairly simple matter to fake scores on Quadrant. I got a Quadrant score of 2597 out of my Epic 4G without doing anything other than run a ROM that had a RAM-drive hack.

      The smartphone industry is slowly starting to realize this however, and benchmark apps like Smartbench are quickly replacing it (thankfully!)

  • http://Website Waltah

    For those complaining about the models tested don’t forget they are testing phones they have in front of them.

  • http://Website Hobo Stu

    Something fishy here? How did you get stock Android on the Atrix?

    • http://Website d.

      ‘Stock’ means ‘out of the box’ not ‘AOSP’ The only phones sporting that are Nexii & the G series on T-Mobile. I have the G2 & it has ‘stock’ Android as-in AOSP Android (+ crapware like Photobucket Mobile).

  • http://Website Dave

    The first commenter is right. You need to normalize for resolution otherwise this is a farce.

  • http://Website Albert

    Great review. Not sure what to make of it. I have to pick one. Do I go with the higher res screen(atrix) or the speedier device.(g2x) . There’s even a third option, I could wait for the HTC pyramid with qhd screen and the dual-core snapdragon. Decisions! Decisions!

  • http://Website Richard

    All of this is great…But means nothing when the Htc Evo 2 come out with the Dualcore Snapdragon Processor. The Evo 2 WILL BE KING

  • BiGMERF

    I just fell in love with my Nexus S all over again :)

  • http://Website Miguel

    I am glad A&M didn’t give breaks for screen resolution. If that’s the resolution the manufacturer went with then that’s what they should be rated with. In games you will need to run the phone as it is so if game performance takes a hit due to resolution than that’s good to know.

    • http://Website dave

      Your point would only make sense if the comparison was between phones. It wasn’t. It was about the chips themselves. Its like comparing two engines by comparing the 0 to 60 speeds of the car they are in. Except one car weighs 2500 pounds and the other 4000.

      • http://Website dave

        well I re read the article and it is about the phones and the resolution is mentioned. So my mistake. You are 100% right. I was going off the inaccurate summary on bgr.

    • http://Website NickK

      Good point, at least for current games, since they tend to run at the phone’s native display resolution. But these 3D benchmarks should start rendering to a common default size, otherwise score comparisons between mobiles with varying resolutions are meaningless. In order to deal with performance fragmentation, future 3D android games will likely render to a fixed size (WVGA) and scale to the screen res in the final pass. On a 4 inch display, you’ll hardly notice the difference. Almost every PS3, 360 game now renders some part of the visuals at lower res and scales to 1080, so this is pretty common. Otherwise, game developers will need to offer a resolution option, such as in the other highly “fragmented” gaming platform, PCs.

  • http://Website Raptor

    Great that Tegra2 got its way to the smartphones too and not just to the iPad-like and Xoom-like slates as Nvidia initialy made the impression.

    Out of topic: One of the best Android sites on the web, but to my humble opinion, all Android-like icons on this site need some artistic rebuild. They look so ChiLdiSh , LOL, ROFL, as a result there are so many fifth-graders commenting.

    • http://Website Rayyan

      Actually, I think the icons are cool, but I’d prefer having disqus-powered comments

  • http://Website revs

    nice !

    not that impressed wheres the mt4g ! we all know thats what these phones need to be compared to
    mt4g still king ???????

  • http://Website Jojo

    What´s interesting in my opinion is the performance of the Exynos/Orion in the SGS2. I did the Neocore Benchmark (other din´t work) today an I only got 52,5fps.

    Does the 50% more subpixels have a effect on the performance?

  • http://Website adam

    The Nexus phones always seem to be one year ahead of their time, rather the other phones are 1 gen behind. Can’t wait to see what google has in their bag for the next nexus phone running 2.4 or what have you.

    • http://sean-the-electrofreak.blogspot.com/ Sean the Electrofreak

      Actually I’m doing to disagree with that statement. The Nexus One used a fairly revolutionary 1st-Gen Snapdragon, but the Nexus S used the Hummingbird, which, while it’s a well-tweaked Cortex-A8 with a decent GPU, is essentially the very last of last-gen hardware. A few more months and a Cortex-A9 would have made for a much more impressive flagship phone.

      That said, I’m glad to see that Google didn’t stick with Snapdragon; mixing it up this way helps to see Android optimized for a wider variety of platforms. I hope that their next phone is an OMAP SoC.

  • http://Website ACR

    Neocore is not fair for Nexus S, since it is capped at 56fps. Neocore is using the old OpenGL 1.1 that can’t challenge these new GPUs. Nenamark1 is more up to date with OpenGL 2.0.

    • http://Website Worsty

      My Nexus S can go up to 66.5 FPS on Neocore when you switch the stock kernel with Netarchy’s of example. With the same kernel at overclock 1400ghz, i’m able to do scores of 3300 for smartbench gaming index.

      That’s the draw of having unlocked bootloaders and also why i’m staying with the Nexus line for all my future phones unless the other manufacturers decides to be more flexible.

    • http://Website Worsty

      My Nexus S can go up to 66.5 FPS on Neocore when you switch the stock kernel with Netarchy’s for example. With the same kernel at overclock 1400ghz, i’m able to do scores of 3300 for smartbench gaming index.

      That’s the draw of having unlocked bootloaders and also why i’m staying with the Nexus line for all my future phones unless the other manufacturers decides to be more flexible.

  • http://Website Spencer

    My Motorola xoom of even higher resolution and tegra 2 got a sun spider score of 2154… I’ll benchmark the rest later.

  • http://Website Speencer

    Just finished benchmarks on my xoom tablet. Neocore didn’t run. Nenamark1 gave me 35fps. Smart bench 2011 – 3548, 2489. GLBenchmark 2.0.3 required a login to see the results, however it ran very smoothly, much more so than on my droid x. Linpack gave ≈ 37 on average. Sun spider scored 2154… pretty impressive considering its pushing so many more pixels…

    -Xoom is 1280*800 (1024000)
    -Atrix 4G is 540*960 (518400)
    -Optimus 2X is 800*480 (384000)

  • http://Website noobnl

    the test is not fair.
    the fps limited on most devices is 60 fps
    the optimus x2 has uncapped fps rate

    • http://Website JohnB

      This depends on tests.

      If I’m not mistaken, Smartbench 2010/2011 artificially boosts up the polygon counts so that there are many more polygons to render than it appears. Even on Nexus S, it was barely hitting 30-40fps range.

  • http://Website Nate

    it’s nice to see that even after over a year my nexus one can still pull second in one of these.

  • http://Website Gunnar

    would love to see how these Android phones matchup against the iPhone

    • http://sean-the-electrofreak.blogspot.com/ Sean the Electrofreak

      Closest match to the iPhone hardware is the Nexus S; the Hummingbird and the A4 have some similar roots via Intrinsity and Samsung, and the GPU is similar, though weaker if I’m not mistaken.

      That said, Apple optimizes the iOS specifically for their phones, and it tends to be a bit more efficient than Android. It’s the price we pay for the power of choice, and a price I’m more than happy to pay.

      • http://Website Derek

        iOS is a “real” OS. Its apps are written in Obj C, not java. Its apps run natively on the cpu. Android apps are java and run inside a java virtual machine (the Dalvik java machine). Thats why games are always going to run so much better on iOS. I have an old iphone 2G, and it plays 3D games better and smoother than my Galaxy S captivate, even though the Captivate hardware on paper smokes whats in the old, original iphone 2G. Its the OS, iOS is sooo much better in that one regard.

        • http://Website Greg

          Wow. Congratulations on having an elementary grasp on mobile operating systems. Go back to Yahoo answers with that crap.

        • http://Website Dags

          You have the choice of writing Android apps in Java (SDK) or C/C++ (NDK) or both. High performance games make use of the NDK and perform every bit as well as iOS games.

  • http://atrixphoneblog.com/ Ricky Fantana

    We all know that Motorola Atrix is far from being a perfect Laptop or Desktop PC. But is there any smartphone that can do even half of what this device can do? Motorola is still on the development stage when it comes to merging the smartphone and laptop/desktop. Sooner or later (hopefully this year), things will improve and we will no longer be needing a CPU. Just an opinion guys!

  • http://Website counsel

    2.3.3 has really slowed down my captivate… I’d have liked this better with all using 2.1.1…

  • http://Website ashok

    I feel something is missing in this,, Thats Samsung galaxy s2,, comparing with that … and show the result

  • http://Website observer

    the problem i have with this test is that the tegra2 is being compared to last gen tech which it is of course going to prefomr better than comparing the tegra to the original droid is obvious what the outcome will be. the test that should be done here is the tegra 2 vs the qualcom snapdragon soc with the adreno 220 and the exenos soc with the mali 400 and then see who wins this currents test show nothing. also with these benchmark they do not take full advantage of the avaliable hardware that is inside the new processor they state how much better one of dual core processors is to the last gen

  1. RajGuest 4 years ago

    I don’t know much about phone displays but I’m surprised at how much the resolution made such a big difference. To me, I didn’t think there was that big of a gap in resolution.

    • DaveGuest 4 years ago

      The Motorola Atrix 4G is 540 x 960 = 518400 pixels
      At 55fps * 518000 pixels = 28490000 pixels per second

      The LG Optimus 2X is 800 x 480 = 384000 pixels
      At 77fps * 384000 pixels = 29568000 pixels per second

      29568000 / 28490000 = 1.038.

      So the screen resolution explains most of the difference but the Optimus is still ~4% faster.

      • watbetchGuest 4 years ago

        Did you take the PenTile subpixel matrix into account for the Atrix?

        • DaveGuest 4 years ago

          The PenTile subpixel matrix lowers power consumption and looks different (some say worse), but I wouldn’t expect it to affect the rendering speed. It’s part of the display, not the renderer.

          I’d guess the 4% difference comes from texture caching being more effective at lower resolutions. For example, a polygon might fit into 1 texture memory page in the lower resolution, but use 2 texture memory pages at a higher resolution causing the gpu to swap texture pages.

      • SynTGuest 4 years ago

        Who care?

        100-500FPS? 20-10FPS? There is no point to do benchmark if all new games require Nvidia….

        … Even on tablet with much less performance…. still if you got tegra you can play. On Galaxy S NOOO

      • ruldaGuest 4 years ago

        OMFG LG 2x doesnt have FPS CAP !!!!! artix yes !!!

    • DivinuSfiliuSGuest 4 years ago

      Oh yeah, it is a big difference for a CPU, to you might seem like small numbers, but things really add up like the number of polygons and texture resolution, shaddows etc.

  2. troy gGuest 4 years ago

    Nexus s still hanging with the big dogs!!

  3. JasonGuest 4 years ago

    why not use the mytouch 4g instead of the G2. The mt4g has more ram

  4. B2LGuest 4 years ago

    It was nice to see this comparison. I love my Nexus S more everyday. It does everything I need it to.

  5. RussGuest 4 years ago

    I I hope the release the 2x in the states soon, I what that phone.

  6. brookmarinGuest 4 years ago

    by year’s end we’re supposed to also see some quad-core tegra chips on the market. So the question is, will developers write code to utilize as many cores as the chip has? or will they just limit writing to only two cores?

    i’m waiting to hear what the competition has to say about nvidia’s aggressive schedule of releases… and what they will do to compete!

    • jonathanGuest 4 years ago

      i think theyll do something like apple is doing with there A4 & A5 soc’s, basically theyre making games that would have less polygons & lighting effects on an A4 but when it detects the A5 it will add more lighting effects & polygon counts. its like PC games where you can pick the setting of the game to be high details or less detailed & depending on the GPU card youd have in your pc would depend on what effects you would see in your game. but since its only 2 gpu in apples gadgets then it will be easier to make games that will look better depending on what hardware you have. the same would be with the tegra 2 & tegra 3.

  7. AlexGuest 4 years ago

    Why you guys always comparing with HTC G2? there is exist HTC Desire HD witch is more powerful than G2

    • Yeah the Desire HD is a 200 MHz faster CPU and 256 MB more ram. It’s still a second-gen Snapdragon and on par with the G2, but a little faster.

      Short answer – no Desire HD in the USA, and I don’t have a myTouch 4G

      • OliverGuest 4 years ago

        You forgot about the HTC Inspire. It’s the Desire HD in US and all the root methods and roms are compatible from Desire HD if i might add.

      • AlexGuest 4 years ago

        oh yeah.. i completely forgot. It’s just because i’m from Holland we have here Desire HD

  8. jayGuest 4 years ago

    so all those people putting down LG should think twice. I was hoping moto would have the better performer but the LG looks great

  9. jorgeGuest 4 years ago

    LG x2 FTW!

  10. I am still using my OG myTouch 3G with 2.2…..I’m so sick and tired of the phone, but hopefully the LG O2X will be out by the time my two year upgrade happens. I know its going to be the first dual core phone, and it will have better competition in about 3 months after it comes out. But everything about this phone is amazing, it kills the mycrapTouch 3G 1,000,000 times. I just want this phone to come to the US so I don’t have to order it from Europe.

  11. JimGuest 4 years ago

    Maybe I’m missing something, but why not do a Quadrant test, too?

    • Because it’s a crappy benchmark. It’s insanely popular because all of the Snapdragon phones perform so well on it because of the way the benchmark was written, and the HTC (Snapdragon) phones have a strong development community that likes benchmark tools. Of course, they’re going to pick the benchmark that has pretty graphs and gives their phones big numbers.

      It’s also a fairly simple matter to fake scores on Quadrant. I got a Quadrant score of 2597 out of my Epic 4G without doing anything other than run a ROM that had a RAM-drive hack.

      The smartphone industry is slowly starting to realize this however, and benchmark apps like Smartbench are quickly replacing it (thankfully!)

  12. WaltahGuest 4 years ago

    For those complaining about the models tested don’t forget they are testing phones they have in front of them.

  13. Hobo StuGuest 4 years ago

    Something fishy here? How did you get stock Android on the Atrix?

    • d.Guest 4 years ago

      ‘Stock’ means ‘out of the box’ not ‘AOSP’ The only phones sporting that are Nexii & the G series on T-Mobile. I have the G2 & it has ‘stock’ Android as-in AOSP Android (+ crapware like Photobucket Mobile).

  14. DaveGuest 4 years ago

    The first commenter is right. You need to normalize for resolution otherwise this is a farce.

  15. AlbertGuest 4 years ago

    Great review. Not sure what to make of it. I have to pick one. Do I go with the higher res screen(atrix) or the speedier device.(g2x) . There’s even a third option, I could wait for the HTC pyramid with qhd screen and the dual-core snapdragon. Decisions! Decisions!

  16. RichardGuest 4 years ago

    All of this is great…But means nothing when the Htc Evo 2 come out with the Dualcore Snapdragon Processor. The Evo 2 WILL BE KING

  17. I just fell in love with my Nexus S all over again :)

  18. MiguelGuest 4 years ago

    I am glad A&M didn’t give breaks for screen resolution. If that’s the resolution the manufacturer went with then that’s what they should be rated with. In games you will need to run the phone as it is so if game performance takes a hit due to resolution than that’s good to know.

    • daveGuest 4 years ago

      Your point would only make sense if the comparison was between phones. It wasn’t. It was about the chips themselves. Its like comparing two engines by comparing the 0 to 60 speeds of the car they are in. Except one car weighs 2500 pounds and the other 4000.

      • daveGuest 4 years ago

        well I re read the article and it is about the phones and the resolution is mentioned. So my mistake. You are 100% right. I was going off the inaccurate summary on bgr.

    • NickKGuest 4 years ago

      Good point, at least for current games, since they tend to run at the phone’s native display resolution. But these 3D benchmarks should start rendering to a common default size, otherwise score comparisons between mobiles with varying resolutions are meaningless. In order to deal with performance fragmentation, future 3D android games will likely render to a fixed size (WVGA) and scale to the screen res in the final pass. On a 4 inch display, you’ll hardly notice the difference. Almost every PS3, 360 game now renders some part of the visuals at lower res and scales to 1080, so this is pretty common. Otherwise, game developers will need to offer a resolution option, such as in the other highly “fragmented” gaming platform, PCs.

  19. RaptorGuest 4 years ago

    Great that Tegra2 got its way to the smartphones too and not just to the iPad-like and Xoom-like slates as Nvidia initialy made the impression.

    Out of topic: One of the best Android sites on the web, but to my humble opinion, all Android-like icons on this site need some artistic rebuild. They look so ChiLdiSh , LOL, ROFL, as a result there are so many fifth-graders commenting.

  20. revsGuest 4 years ago

    nice !

    not that impressed wheres the mt4g ! we all know thats what these phones need to be compared to
    mt4g still king ???????

  21. JojoGuest 4 years ago

    What´s interesting in my opinion is the performance of the Exynos/Orion in the SGS2. I did the Neocore Benchmark (other din´t work) today an I only got 52,5fps.

    Does the 50% more subpixels have a effect on the performance?

  22. adamGuest 4 years ago

    The Nexus phones always seem to be one year ahead of their time, rather the other phones are 1 gen behind. Can’t wait to see what google has in their bag for the next nexus phone running 2.4 or what have you.

    • Actually I’m doing to disagree with that statement. The Nexus One used a fairly revolutionary 1st-Gen Snapdragon, but the Nexus S used the Hummingbird, which, while it’s a well-tweaked Cortex-A8 with a decent GPU, is essentially the very last of last-gen hardware. A few more months and a Cortex-A9 would have made for a much more impressive flagship phone.

      That said, I’m glad to see that Google didn’t stick with Snapdragon; mixing it up this way helps to see Android optimized for a wider variety of platforms. I hope that their next phone is an OMAP SoC.

  23. ACRGuest 4 years ago

    Neocore is not fair for Nexus S, since it is capped at 56fps. Neocore is using the old OpenGL 1.1 that can’t challenge these new GPUs. Nenamark1 is more up to date with OpenGL 2.0.

    • WorstyGuest 4 years ago

      My Nexus S can go up to 66.5 FPS on Neocore when you switch the stock kernel with Netarchy’s of example. With the same kernel at overclock 1400ghz, i’m able to do scores of 3300 for smartbench gaming index.

      That’s the draw of having unlocked bootloaders and also why i’m staying with the Nexus line for all my future phones unless the other manufacturers decides to be more flexible.

    • WorstyGuest 4 years ago

      My Nexus S can go up to 66.5 FPS on Neocore when you switch the stock kernel with Netarchy’s for example. With the same kernel at overclock 1400ghz, i’m able to do scores of 3300 for smartbench gaming index.

      That’s the draw of having unlocked bootloaders and also why i’m staying with the Nexus line for all my future phones unless the other manufacturers decides to be more flexible.

  24. SpencerGuest 4 years ago

    My Motorola xoom of even higher resolution and tegra 2 got a sun spider score of 2154… I’ll benchmark the rest later.

  25. SpeencerGuest 4 years ago

    Just finished benchmarks on my xoom tablet. Neocore didn’t run. Nenamark1 gave me 35fps. Smart bench 2011 – 3548, 2489. GLBenchmark 2.0.3 required a login to see the results, however it ran very smoothly, much more so than on my droid x. Linpack gave ≈ 37 on average. Sun spider scored 2154… pretty impressive considering its pushing so many more pixels…

    -Xoom is 1280*800 (1024000)
    -Atrix 4G is 540*960 (518400)
    -Optimus 2X is 800*480 (384000)

  26. noobnlGuest 4 years ago

    the test is not fair.
    the fps limited on most devices is 60 fps
    the optimus x2 has uncapped fps rate

    • JohnBGuest 4 years ago

      This depends on tests.

      If I’m not mistaken, Smartbench 2010/2011 artificially boosts up the polygon counts so that there are many more polygons to render than it appears. Even on Nexus S, it was barely hitting 30-40fps range.

  27. NateGuest 4 years ago

    it’s nice to see that even after over a year my nexus one can still pull second in one of these.

  28. GunnarGuest 4 years ago

    would love to see how these Android phones matchup against the iPhone

    • Closest match to the iPhone hardware is the Nexus S; the Hummingbird and the A4 have some similar roots via Intrinsity and Samsung, and the GPU is similar, though weaker if I’m not mistaken.

      That said, Apple optimizes the iOS specifically for their phones, and it tends to be a bit more efficient than Android. It’s the price we pay for the power of choice, and a price I’m more than happy to pay.

      • DerekGuest 4 years ago

        iOS is a “real” OS. Its apps are written in Obj C, not java. Its apps run natively on the cpu. Android apps are java and run inside a java virtual machine (the Dalvik java machine). Thats why games are always going to run so much better on iOS. I have an old iphone 2G, and it plays 3D games better and smoother than my Galaxy S captivate, even though the Captivate hardware on paper smokes whats in the old, original iphone 2G. Its the OS, iOS is sooo much better in that one regard.

        • GregGuest 4 years ago

          Wow. Congratulations on having an elementary grasp on mobile operating systems. Go back to Yahoo answers with that crap.

        • DagsGuest 4 years ago

          You have the choice of writing Android apps in Java (SDK) or C/C++ (NDK) or both. High performance games make use of the NDK and perform every bit as well as iOS games.

  29. Ricky FantanaGuest 4 years ago

    We all know that Motorola Atrix is far from being a perfect Laptop or Desktop PC. But is there any smartphone that can do even half of what this device can do? Motorola is still on the development stage when it comes to merging the smartphone and laptop/desktop. Sooner or later (hopefully this year), things will improve and we will no longer be needing a CPU. Just an opinion guys!

  30. counselGuest 4 years ago

    2.3.3 has really slowed down my captivate… I’d have liked this better with all using 2.1.1…

  31. ashokGuest 4 years ago

    I feel something is missing in this,, Thats Samsung galaxy s2,, comparing with that … and show the result

  32. observerGuest 4 years ago

    the problem i have with this test is that the tegra2 is being compared to last gen tech which it is of course going to prefomr better than comparing the tegra to the original droid is obvious what the outcome will be. the test that should be done here is the tegra 2 vs the qualcom snapdragon soc with the adreno 220 and the exenos soc with the mali 400 and then see who wins this currents test show nothing. also with these benchmark they do not take full advantage of the avaliable hardware that is inside the new processor they state how much better one of dual core processors is to the last gen