I wonder how much difference which version of Gingerbread each device was running made. Because the Xiaomi Phone is the only one running Android 2.3.7 as its stock ROM with the others using 2.3.4 or earlier versions.
I too have a Milestone, but running Cronos Ginger X 1.7.0. I can't wait for the Nexus to finally come out. It is taking way too long. By the time it is readily available it might not be so great anymore. Still, it'll be a Nexus phone and that counts for a lot these days.
Can't wait to see how the Galaxy Nexus compares in this benchmark.
They also didn't include the new Motorola RAZR which run 2.3.5 and benefits from most of the new and more efficient code that will be arriving with ICS.
I think the test in comparing hardware is generally done pretty poorly here.
If you want to compare hardware you need to root them all and install the exact same version of Android. Xiaomi has a huge advantage here since it is running a far more advanced version than the other hardware. Likewise, to compare Android versions the same hardware should be used with different versions installed. Mixed the two is just bad journalism.
Oh and I tried the benchmark suite on my SE Xperia mini pro with Android 2.3.4 (rooted). The device has a Qualcomm MSM8255 (1 GHz Scorpion, Adreno 205) with 512MB of RAM.
I can't help but be impressed by the performance of the Xiaomi Phone. It lead in most of the test. This really speak the Chinese prowess, a relatively small company can build a device from the ground up in a relatively short time but with very competitive performance with well-established juggernaut and with unbelievable price to boot!
A very good job well done. Sadly, the same can't be said for American upstarts...
Are you guys sure the Tmobile Galaxy has a APQ8060 in it? Im here in Canada and have the Telus equivalent, and vellamo states its got a MSM8660_SURF. Not sure if the _SURF at the end mean anything, but I thought these two phones were identical?
My stock Nexus S got 609. Not too far behind the dual core Atrix, I guess Android's browser doesn't' take much advantage of more cores yet. I use Opera anyways, its GPU accelerated and you can really tell on image heavy pages (try zooming around engadget [desktop] while it loads on the stock browser and then Opera).
Wow, I'm shocked that both Anandtech and all of the commentators here don't seem to be familiar with MIUI.
MIUI is the second most prominent Android ROM around, second only to Cyanogenmod. It's renowned for its good performance, and its UI which is heavily revamped to both (1) have a lot of improvements in ease of use and (2) unfortunately be a bit of a ripoff of iOS.
Xiaomi is the company that's because MIUI, and they Mi-One is the first device they've launched that comes stock with MIUI. Furthermore, it has a bunch of features specifically applicable to enthusiasts that make it impossible to brick and easy to try different ROMs on.
I'm really curious how the performance changes when these phones are tested running stock (which in my experience is horrificly crap) and then also with MIUI and CM? In my experience with multiple devices, performance improved TREMENDOUSLY with CM7, to the point it was almost like an upgrade to a new device. Take the difference between the MI-One and the Sensation...they're the same chip, but with VASTLY different scores. Kinda makes you go "hmmm....".
I think by having a reputable site like AnandTech do testing for these different devices it can help highlight the affect stock ROM skins and changes are having on Android. The skins and inconsistencies are the single reason I went back to an iPhone. It's interface is consistent, and performance is consistent (if not near as fast as CM). The whole experience is just consistent...which to me makes a big difference.
I think having numbers behind the decision to differentiate their devices would be a huge benefit to consumers, and would be very eye-opening to a lot of people.
I have heard that a lot of the chinese phones are using CM7 base code right off the bat in many of their android phones, which may even explain the performance difference of the xiaomi. The good news is that most of these phones can be rooted and have custom ROMs which will improve performance. But like you are saying... wouldn't it be nice if we didn't have to rely on third parties to utilize the performance of the device.
I think that the fracturing of android via carrier and manufactuer customization will come to an end in the next few years. I get the feeling that even with android 4.0 (ice cream sandwich) google is attempting to bring the interface together in a coherent and beautiful way that most carrier will choose to adopt instead of customizing their own frontend applications (sense, touchwiz, etc...)
I certainly hope you're right. I do wish that Google was able to exert more influence over the devices, if only for consistency's sake. I don't want all the random apps that I can't uninstall, I don't want Sense or some other resource hog. If the user wants to add that, fine, but it certainly shouldn't come from the factory that way.
Just ran this on my Captivate with CM7.2 and got 846! Only single core and hanging with the big boys. See what a little optimization can do for a 1.5 year old phone?
I have HTC Desire with MIUI (2.3.7) @ 1GHz and my overall score is 680 which is better than dual core Atrix 4G. I would say that's a damn good score for such an old phone. Good job MIUI!
Lots of people are running out of time to change the original AT&T SGS2 for the Skyrocket... any quick pre-review summary on if the Snapdragon in the Hercules gets in the way in any non-game-use?
This is a good comparison of the stock phone experience as a whole but it does not give good data on the true power of the platforms.
I realize that only a small portion of the population uses a custom ROM for their phones. But standardizing the OS/skin across the phones will make the performance number more relevant to hardware benchmarks. Of course build optimization varies across devices running CM (or any AOSP ROM).
I have a Sensation with Cyanongenmod 7 and it's leagues faster than Sense.
My HTC Rezound which I just got on release day from VZW on 11/14/2011. Dual core...MSM 86xx 1.5ghz with 1GB RAM running Sense 3.5... but the 1280x720 S-LCD resolution MAY slow the graphics a bit on the Adreno 220...I love this phone! And the iBeats do sound great.
Well, that may change once Google's take-over of Motorola's mobile business is complete. In theory, that should lead to "the best" stock Android experience.
It's crazy to see how much a ROM/Kernel makes a difference. The T-Mobile SGS2, Sensation, and the Xaimoi Mi-One practically have the same SoC, but the Sensation performs much worse than the other two...and I don't think the difference in clock speed would make that much of a difference. HTC needs to step it's game up in that department.
Yeah, don't worry about the 300Mhz difference, lol.
I just ran this on my Sensation (rooted, ARHD 3.6, Bricked kernel) at 1.5Ghz and surprise surprise, it bested the SGS2 with a score of 1055. No, it's not stock kernel but that's a significant improvement; too much for just the kernel alone.
How fair do you think this benchmark really is Brian? I'm a little worried by the fact that it's made by someone who has a horse in the game - Qualcomm. Wouldn't that mean they will prioritize testing whatever Qualcomm chips are good at? That kind of lowers the credibility of the benchmark.
I think they did it before with Neocore, where if I remember correctly Adreno 200 was scoring better than PowerVR SGX535, and yet I never saw a game that looked better on Adreno 200 than one on PowerVR SGX535 around that time.
Besides being a Qualcomm benchmark, it also uses the default Android browser. I get the impression that noone at AT actually uses an Android device for themselves. This does seem like another Neocore- check out the results they have built in to the benchmark- the GalaxyS2 finished behind the Desire, the Galaxy 2 LTE however finishes quite a bit higher while the Xperia Arc leaves all of those phones behind. I don't know if this was supposed to be a PR piece that Brian was authoring for Qualcomm or something else entirely, but worthy of a tech site article? Certainly not.
What about the new Futuremark Peacekeeper Beta? As I understand it it is supposed to support both mobile and desktop browsers and comes with a wide range of tests.
Just ran this on my Galaxy Nexus. Unfortunately it failed all three flinger tests but it did score 896 overall. The javascript scores were very good, 2039.0 in sunspider and 158.85 in V8
I've been using Vellamo on a number of devices for testing, and at first I thought that it was great...but then I found the problem, which is similar to NenaMark (1&2), Neocore, and several other benchmarks: results are very much based on screen resolution. That is to say, if you have an Android phone with a low-res screen, you will have an inflated score.
Case in point: I have something like 7-9 handsets I'm testing right now, with screen resolutions varying from 320x480 to 720p. The low resolution LG Doubleplay manages to get an average of 810, while the Galaxy Player gets an average of 620, the HTC Rezound gets 920, and the Galaxy S II (well I had a score, but my frickin spreadsheet lost it's data while upgrading my GPU for some reason...)
As I continue testing, I'll add my findings, but it looks like the more I test the more relevant screen resolution is to the test, which inherently limits it. It may be well rounded, but the overall scores may not matter in the long run. Individual tests, however, may.
I've gone over the testing here, and after some serious testing on my own, I have to conclude that Vellamo isn't a good benchmark at all. There's a number of reasons:
1) Results don't match. I've tested across six phones now, and even on the same device results don't match up. I get 80-point spreads, which is a joke. Why does the Rezound have a low score of 856, but a high of 930? Why does the benchmark run better when other applications are running, and worse when nothing else is running? Why does it take over an hour for the scores to pick up (ie, for the phone to "warm up")? The results, on several devices (though not on all) are not accurate and don't seem to level out.
2) Results aren't accurate across tests: Vellamo uses Sunspider 0.9.1, a benchmark used here and one I use myself for testing Javascript. I like the test alot. It's accurate and very helpful. Vellamo inflates (or because the score is in less-time-is-better, deflates) scores significantly. I've run the benchmark through the default web browser and through Vellamo, and Vellamo consistently scores higher by 400+ms. If this one test which I can reproduce with minimal work isn't accurate, chances are so are plenty of others which I haven't yet gotten to. I'm actually surprised the previous Sunspider scores haven't been used as a point of comparison in this review.
3) Most of the tests are frankly pointless and don't show useful results. As someone who will use the benchmark scores to chart data, points are fine. But they don't really say anything besides what the points are. The tutorial that shows what each test is and why is great, but ultimately the data is what's important, and the data is flawed from the start. In an attempt to simplify it, the data just doesn't make any actual sense.
4) As I mentioned in a post earlier, scores rely heavily on display resolution, which is not good. If smartphone displays are going to vary as they do now - from 320x480 to 720p - then the scores aren't going to match well in any way. I have a few phones scoring higher than others solely because of this. The LG Doubleplay, for instance, a dumbphone by nearly all regards except that it runs Android 2.3.5, averages 200 points higher than the Samsung Galaxy Player in an average of 10 tests. What the hell.
I'll keep playing with it, but Vellamo thus far doesn't look like a benchmark worth using, at least not in the long run.
Hi! I am from Shanghai China.I bought a Xiaomi one week ago.It costs me ￥1999(about $316.64).I have used it for one week.It's really excellent！Hope you who wants one could get and enjoy Xiaomi as soon as possible. My e-mail is [email protected]u can get more informatio from me.
Hi! I am from Shanghai China.I bought a Xiaomi one week ago.It costs me ￥1999(about $316.64).I have used it for one week.It's really excellent！Hope you who wants one could get and enjoy Xiaomi as soon as possible. My e-mail is [email protected]. You can get more informatio from me.
I'm surprised that such a well respected site like Anandtech would fail to mention that Vellamo is actually written by surprise surprise QUALCOMM!, So there's little wonder the scores are so well regraded for the Qualcomm chipsets. Seriously if Anandtech continues to use this benchmark in the future then their scores would mean nothing and be a waste of time with such a high bias to Qualcomm chips with a Benchmark written by Qualcomm.
We’ve updated our terms. By continuing to use the site and/or by logging into your account, you agree to the Site’s updated Terms of Use and Privacy Policy.
Facebook
Twitter
RSS
82 Comments
Back to Article
jalexoid - Wednesday, November 16, 2011 - link
Well, looks like China is ready for some home grown brands. I bet we will start seeing more and more marvels like Xiaomi come out of China.Samus - Wednesday, November 16, 2011 - link
Look at HTC. The G1 is still as revolutionary as the iPhone. It's design at the time was groundbreaking.stevessvt - Wednesday, November 16, 2011 - link
1009 overall score.Gunde - Wednesday, November 16, 2011 - link
I wonder how much difference which version of Gingerbread each device was running made. Because the Xiaomi Phone is the only one running Android 2.3.7 as its stock ROM with the others using 2.3.4 or earlier versions.Gunde - Wednesday, November 16, 2011 - link
Should also add that I tried this on my quite antiquated Motorola Milestone running CM7.1 (2.3.7) @ 1 GHz (OMAP 3430) and I got a score of 487.I suppose that means that I should stick to browsing mobile versions of sites... the Galaxy Nexus can't come soon enough!
3DoubleD - Wednesday, November 16, 2011 - link
"the Galaxy Nexus can't come soon enough! "I too have a Milestone, but running Cronos Ginger X 1.7.0. I can't wait for the Nexus to finally come out. It is taking way too long. By the time it is readily available it might not be so great anymore. Still, it'll be a Nexus phone and that counts for a lot these days.
Can't wait to see how the Galaxy Nexus compares in this benchmark.
mcnabney - Wednesday, November 16, 2011 - link
They also didn't include the new Motorola RAZR which run 2.3.5 and benefits from most of the new and more efficient code that will be arriving with ICS.I think the test in comparing hardware is generally done pretty poorly here.
If you want to compare hardware you need to root them all and install the exact same version of Android. Xiaomi has a huge advantage here since it is running a far more advanced version than the other hardware. Likewise, to compare Android versions the same hardware should be used with different versions installed. Mixed the two is just bad journalism.
TheMan876 - Wednesday, November 16, 2011 - link
The reviews are generally done as reviews of the product presented to the consumer, including tweaks the manufacturers have done to the browser.warisz00r - Wednesday, November 16, 2011 - link
I live in Malaysia, and I can get the Xiaomi imported in for cheaper than the rest of the lineup.warisz00r - Wednesday, November 16, 2011 - link
Oh and I tried the benchmark suite on my SE Xperia mini pro with Android 2.3.4 (rooted). The device has a Qualcomm MSM8255 (1 GHz Scorpion, Adreno 205) with 512MB of RAM.My score was 763.
tananiki - Thursday, February 2, 2012 - link
Hi Waris, i'm from Malaysia too. How much you can get? Less than official price USD310?DanD85 - Wednesday, November 16, 2011 - link
I can't help but be impressed by the performance of the Xiaomi Phone. It lead in most of the test. This really speak the Chinese prowess, a relatively small company can build a device from the ground up in a relatively short time but with very competitive performance with well-established juggernaut and with unbelievable price to boot!A very good job well done. Sadly, the same can't be said for American upstarts...
tynopik - Wednesday, November 16, 2011 - link
from one OEM another -> OEM to anotherappliation -> application
are attempt -> are an attempt
how long it that takes -> it takes
Brian Klug - Wednesday, November 16, 2011 - link
Thank you, my bad, those are fixed now!-Brian
Pino - Wednesday, November 16, 2011 - link
Who uses stock Android browser?Would be better to have the platforms compared using the same third part browser.
tipoo - Wednesday, November 16, 2011 - link
It doesn't let you choose which browser to run in.edsib1 - Thursday, November 17, 2011 - link
completely agree with u. My Desire scores around 3535 in sunspider using stockbrowser, yet with firefox it scores 2419. (840 Vellamo)I'm afraid Anandtech is starting to fall behind the times.
The whole point of Android is that you are not stuck with the pre-installed apps. You install whichever is the best/your favourite
balamasti - Wednesday, November 16, 2011 - link
Where can we buy "Xiaomi Mi-One" online. Is there any good website?dwh2008001 - Wednesday, November 16, 2011 - link
china www.xiaomi.com 哈哈psyclist80 - Wednesday, November 16, 2011 - link
Are you guys sure the Tmobile Galaxy has a APQ8060 in it? Im here in Canada and have the Telus equivalent, and vellamo states its got a MSM8660_SURF. Not sure if the _SURF at the end mean anything, but I thought these two phones were identical?metafor - Wednesday, November 16, 2011 - link
It's the same chip. The APQ simply has the modem fused off.psyclist80 - Wednesday, November 16, 2011 - link
so then I have the modem in mine? or is vellamo reporting it incorrectly? I just like to know whats under the hood is all :)metafor - Wednesday, November 16, 2011 - link
The chip ID doesn't change. So what you have was a MSM8660, but the modem has been fused off.Brian Klug - Wednesday, November 16, 2011 - link
Modem inside the T-Mobile SGS2 is MDM8220.http://twitpic.com/7cyfrx
-Brian
Brian Klug - Wednesday, November 16, 2011 - link
Absolutely positively certain it is an APQ8060 and not MSM8660, you're just seeing remnants of that string from the software port. :)-Brian
ganjha - Wednesday, November 16, 2011 - link
My Samsung Galaxy mini running CM7.1, MSM7227@ 600MHz scores 319 if anyone cares.Kraszmyl - Wednesday, November 16, 2011 - link
I have an HTC Sensation, tis one of the 1.5ghz models and the score for it isSee the sun canvas 25.29
Deep sea canvas 17.8
Aquarium Canvas 154.56
Pixel Blender 105.66
SunSpider uh didnt catch the run time but the score is 112.35
V8 105.51
Ocean 38.15 , oddly lower than the 1.2
Image 20.39
Text 32.41
Network 81.16 , also oddly lower
Overall score 1023
Wieland - Wednesday, November 16, 2011 - link
My Sensation at 1.5GHz with CM7 Alpha 10 and the final Faux Kernel came up with a similar overall score, but the breakdown is much different.Sea The Sun Canvas: 51.08
Deep Sea Canvas: 56.95
Aquarium Canvas: 147.43
Pixel Blender: 130.33
Surf Wax Binder: 121.66
Sun Spider (Online): 123.33
V8 Benchmark (Online): 116.42
Ocean Flinger: 74.17
Image Flinger: 38.46
Text Flinger: 58.74
Network Loader: 113.69
Overall Score:1032
Wieland - Wednesday, November 16, 2011 - link
Sun Spider Run Time was 2301.Wieland - Wednesday, November 16, 2011 - link
See The Sun Canvas FPS: 15.02Deep Sea Canvas FPS: 8.65
V8 Benchmark Total: 717
Ocean Flinger FPS: 44.5
Image Flinger FPS: 23.08
Text Flinger FPS: 35.24
admiralpumpkin - Wednesday, November 16, 2011 - link
I'd like to see how the iP4S compares.Is that comparison possible?
Yes, I know it's an Android app at the moment. How about an iOS app? Or a web app version?
tipoo - Wednesday, November 16, 2011 - link
You can do the individual benchmarks like Browsermark and Sunspidertipoo - Wednesday, November 16, 2011 - link
My stock Nexus S got 609. Not too far behind the dual core Atrix, I guess Android's browser doesn't' take much advantage of more cores yet. I use Opera anyways, its GPU accelerated and you can really tell on image heavy pages (try zooming around engadget [desktop] while it loads on the stock browser and then Opera).tipoo - Wednesday, November 16, 2011 - link
*Opera mobile, not mini.tiredad - Thursday, November 17, 2011 - link
For info: my Nexus S with MIUI scored 879 which surprised the hell out of me.tipoo - Saturday, January 14, 2012 - link
Wow, 890 on cm7 with Trinity kernel...makes you wonder why the stock kernel can't be faster, mines not even overclocked.The Saint - Wednesday, November 16, 2011 - link
Wow, I'm shocked that both Anandtech and all of the commentators here don't seem to be familiar with MIUI.MIUI is the second most prominent Android ROM around, second only to Cyanogenmod. It's renowned for its good performance, and its UI which is heavily revamped to both (1) have a lot of improvements in ease of use and (2) unfortunately be a bit of a ripoff of iOS.
Xiaomi is the company that's because MIUI, and they Mi-One is the first device they've launched that comes stock with MIUI. Furthermore, it has a bunch of features specifically applicable to enthusiasts that make it impossible to brick and easy to try different ROMs on.
Legendarydust - Wednesday, November 16, 2011 - link
Tested my MyTouch 4g running CM 7.1 running stock 1ghz and overclocked 1.5ghz scores are belowStock 1.0ghz/Overclock 1.5ghz
See the Sun Canvas: 6.98 / 9.26
Deep Sea Canvas: 4.54 / 7.35
Aquarium Canvas: 101.69 / 122.85
Pixel Blender: 56.33 / 58.41
Sun Spider: 3142 / 2163
V8 Benchmark: 448 / 624
Surf Wax Binder: 81.31 / 112.35
Ocean Flinger: 32.23 / 41.32
Image Flinger: 10.7 / 12.74
Text Flinger: 31.9 / 40.27
Networking Loader: 82.43 / 108.15
Overall Score: 682 / 879
Not bad overall for an Android device that is over a year old
fabarati - Wednesday, November 16, 2011 - link
Just ran on my LG Optimus 2x (GR4/HP SR1), on Edge when testing.See the Sun Canvas: 16.46
Deep Sea Canvas: 10.81
Aquarium Canvas: 43.99
Pixel Blender: 35.0
Surf Wax Binder: 110.33
Sun Spider: 2706
V8 Benchmark: 296.0
Ocean Flinger: 43.69
Image Flinger: 13.59
Text Flinger: 24.76
Networking Loader: 97.82
Overall Score: 731
Link to a screenshot of the charts : http://i.imgur.com/IWVU0.jpg
It seems that where my phone falls down is rendering. I'm guessing bad browser optimization?
fabarati - Wednesday, November 16, 2011 - link
And at 1.2GHz the overall score is 926 points. Still with rendering issues,See the Sun Canvas: 13.38
Deep Sea Canvas: 13.45
Aquarium Canvas: 45.53
Pixel Blender: 41.75
Surf Wax Binder: 139.11
Sun Spider: 1996
V8 Benchmark: 718
Ocean Flinger: 49.41
Image Flinger: 14.29
Text Flinger: 26.95
Networking Loader: 155.54
Overall Score: 926
http://i.imgur.com/GFyad.jpg
weiln12 - Wednesday, November 16, 2011 - link
I'm really curious how the performance changes when these phones are tested running stock (which in my experience is horrificly crap) and then also with MIUI and CM? In my experience with multiple devices, performance improved TREMENDOUSLY with CM7, to the point it was almost like an upgrade to a new device. Take the difference between the MI-One and the Sensation...they're the same chip, but with VASTLY different scores. Kinda makes you go "hmmm....".I think by having a reputable site like AnandTech do testing for these different devices it can help highlight the affect stock ROM skins and changes are having on Android. The skins and inconsistencies are the single reason I went back to an iPhone. It's interface is consistent, and performance is consistent (if not near as fast as CM). The whole experience is just consistent...which to me makes a big difference.
I think having numbers behind the decision to differentiate their devices would be a huge benefit to consumers, and would be very eye-opening to a lot of people.
Just my thoughts.
cbutters - Wednesday, November 16, 2011 - link
I have heard that a lot of the chinese phones are using CM7 base code right off the bat in many of their android phones, which may even explain the performance difference of the xiaomi. The good news is that most of these phones can be rooted and have custom ROMs which will improve performance. But like you are saying... wouldn't it be nice if we didn't have to rely on third parties to utilize the performance of the device.I think that the fracturing of android via carrier and manufactuer customization will come to an end in the next few years. I get the feeling that even with android 4.0 (ice cream sandwich) google is attempting to bring the interface together in a coherent and beautiful way that most carrier will choose to adopt instead of customizing their own frontend applications (sense, touchwiz, etc...)
weiln12 - Thursday, November 17, 2011 - link
I certainly hope you're right. I do wish that Google was able to exert more influence over the devices, if only for consistency's sake. I don't want all the random apps that I can't uninstall, I don't want Sense or some other resource hog. If the user wants to add that, fine, but it certainly shouldn't come from the factory that way.ThisWasATriumph - Wednesday, November 16, 2011 - link
Just ran this on my Captivate with CM7.2 and got 846! Only single core and hanging with the big boys. See what a little optimization can do for a 1.5 year old phone?mortyg99 - Wednesday, November 16, 2011 - link
just tried it on my t-mobile g2 running cm7.1 and got 636nnitko - Wednesday, November 16, 2011 - link
I have HTC Desire with MIUI (2.3.7) @ 1GHz and my overall score is 680 which is better than dual core Atrix 4G. I would say that's a damn good score for such an old phone. Good job MIUI!s44 - Wednesday, November 16, 2011 - link
Lots of people are running out of time to change the original AT&T SGS2 for the Skyrocket... any quick pre-review summary on if the Snapdragon in the Hercules gets in the way in any non-game-use?TooYellow - Wednesday, November 16, 2011 - link
This is a good comparison of the stock phone experience as a whole but it does not give good data on the true power of the platforms.I realize that only a small portion of the population uses a custom ROM for their phones. But standardizing the OS/skin across the phones will make the performance number more relevant to hardware benchmarks. Of course build optimization varies across devices running CM (or any AOSP ROM).
I have a Sensation with Cyanongenmod 7 and it's leagues faster than Sense.
Exodite - Wednesday, November 16, 2011 - link
My stock, orginal Xperia arc using a 1.0GHz MSM8255 and firmware 4.0.2.A.0.42 got a result of 845.Not too bad.
georgekn3mp - Wednesday, November 16, 2011 - link
My HTC Rezound which I just got on release day from VZW on 11/14/2011. Dual core...MSM 86xx 1.5ghz with 1GB RAM running Sense 3.5... but the 1280x720 S-LCD resolution MAY slow the graphics a bit on the Adreno 220...I love this phone! And the iBeats do sound great.See the Sun Canvas: 56.66
Deep Sea Canvas: 87.68
Aquarium Canvas: 83.59
Pixel Blender: 104
Surf Wax Binder: 127.02
Sun Spider: 2316
V8 Benchmark: 696
Ocean Flinger: 49.46
Image Flinger: 18.1
Text Flinger: 36.54
Networking Loader: 82.72
Overall Score: 883
6th overall of 31 tested running 2.3.4
TareX - Wednesday, November 16, 2011 - link
As expected, my stuttering, slow, disappointing Atrix was remarkably inferior to the other dual cores. Never, ever, buy a Motorola phone.phoenix_rizzen - Wednesday, November 16, 2011 - link
Well, that may change once Google's take-over of Motorola's mobile business is complete. In theory, that should lead to "the best" stock Android experience.whsmnky - Wednesday, November 16, 2011 - link
1167 on my cm7 atrix. Not bad considering its one of the original dual core phones. Don't count it out just yet.whsmnky - Wednesday, November 16, 2011 - link
with the 1.45oc that is.B166ER - Friday, November 18, 2011 - link
Oops, reply'd too soon, I got it.B166ER - Friday, November 18, 2011 - link
How'd you get 1167? My Atrix, CM'd of course, got 835!mfenn - Wednesday, November 16, 2011 - link
etc. is not an acronym. It is an abbreviation for et cetera, so stop writing e.t.c.B166ER - Friday, November 18, 2011 - link
Yeah, you wouldn't write it as etc, either. Try et c. Note the space!psyclist80 - Wednesday, November 16, 2011 - link
alrighty thanks guys!aNYthing24 - Wednesday, November 16, 2011 - link
It's crazy to see how much a ROM/Kernel makes a difference. The T-Mobile SGS2, Sensation, and the Xaimoi Mi-One practically have the same SoC, but the Sensation performs much worse than the other two...and I don't think the difference in clock speed would make that much of a difference. HTC needs to step it's game up in that department.catnaps - Thursday, November 17, 2011 - link
Yeah, don't worry about the 300Mhz difference, lol.I just ran this on my Sensation (rooted, ARHD 3.6, Bricked kernel) at 1.5Ghz and surprise surprise, it bested the SGS2 with a score of 1055. No, it's not stock kernel but that's a significant improvement; too much for just the kernel alone.
catnaps - Thursday, November 17, 2011 - link
1.2Ghz score: 884.So no, it's not the kernel, it's the 20% speed increase.
georgekn3mp - Wednesday, November 16, 2011 - link
Well it looks like my 4G signal strength on HTC Rezound sucked at work...much better results on wifi ;) well less apps running too.4G WIFI
See the Sun Canvas: 56.66 / 63.76
Deep Sea Canvas: 87.68 / 118.46
Aquarium Canvas: 83.59 / 131.65
Pixel Blender: 104 / 108.41
Surf Wax Binder: 127.02 / 128.33
Sun Spider: 2316 / 2357
V8 Benchmark: 696 / 708
Ocean Flinger: 49.46 / 50.23
Image Flinger: 18.1 / 18.15
Text Flinger: 36.54 / 37.17
Networking Loader: 82.72 / 86.88
Overall Score: 883 / 967
6th overall of 31 tested running 2.3.4 / 2nd on list on later run
georgekn3mp - Wednesday, November 16, 2011 - link
Well it looks like my 4G signal strength on HTC Rezound sucked at work...much better results on wifi ;) well less apps running too.
4G WIFI
See the Sun Canvas: 56.66 / 63.76
Deep Sea Canvas: 87.68 / 118.46
Aquarium Canvas: 83.59 / 131.65
Pixel Blender: 104 / 108.41
Surf Wax Binder: 127.02 / 128.33
Sun Spider: 2316 / 2357
V8 Benchmark: 696 / 708
Ocean Flinger: 49.46 / 50.23
Image Flinger: 18.1 / 18.15
Text Flinger: 36.54 / 37.17
Networking Loader: 82.72 / 86.88
Overall Score: 883 / 967
6th overall of 31 tested running Gingerbread 2.3.4 / 2nd on list on later run
HTC Rezound Quadrant score - 2450
NenaMark1 - 53.3 fps
NenaMark2 - 35.8 fps
Speedtest - 4g 25685 kbps down / 12122 kbps up
wifi-N 6146 kbps down / 1173 kbps up
joemtoney - Wednesday, November 16, 2011 - link
1113 Overall Score On my captivate at 1.44Ghz, silly dual core processors1 :d 2.3.7 is better!MarkAroundTheWorld - Thursday, November 17, 2011 - link
It looks like it's more a function of which sub-version of Android than hardware platform or even perhaps OEM optimizations.I was surprised to see my 1.5 year old phone reach 2nd on the list.
Vibrant (Galaxy S) on CM7.1 @1.3GHz
Platypus Kernel + SmartAssV2 Gov
Total Score: 1003
I wonder how the Galaxy Nexus will do... launch dammit!
pSupaNova - Friday, November 18, 2011 - link
Looks like you are correct. Scroll down to section :- Powerful web browsing.http://www.android.com/about/ice-cream-sandwich/
Its a good thing MUI is already intergrating the ICS framework into it ROM.
Lucian Armasu - Thursday, November 17, 2011 - link
How fair do you think this benchmark really is Brian? I'm a little worried by the fact that it's made by someone who has a horse in the game - Qualcomm. Wouldn't that mean they will prioritize testing whatever Qualcomm chips are good at? That kind of lowers the credibility of the benchmark.I think they did it before with Neocore, where if I remember correctly Adreno 200 was scoring better than PowerVR SGX535, and yet I never saw a game that looked better on Adreno 200 than one on PowerVR SGX535 around that time.
BenSkywalker - Thursday, November 17, 2011 - link
Besides being a Qualcomm benchmark, it also uses the default Android browser. I get the impression that noone at AT actually uses an Android device for themselves. This does seem like another Neocore- check out the results they have built in to the benchmark- the GalaxyS2 finished behind the Desire, the Galaxy 2 LTE however finishes quite a bit higher while the Xperia Arc leaves all of those phones behind. I don't know if this was supposed to be a PR piece that Brian was authoring for Qualcomm or something else entirely, but worthy of a tech site article? Certainly not.Mr Alpha - Thursday, November 17, 2011 - link
What about the new Futuremark Peacekeeper Beta? As I understand it it is supposed to support both mobile and desktop browsers and comes with a wide range of tests.B166ER - Friday, November 18, 2011 - link
My Cyanogen'd (7.1b) Atrix running 2.3.7 did 835. What's that over stock?juicytuna - Friday, November 18, 2011 - link
Just ran this on my Galaxy Nexus. Unfortunately it failed all three flinger tests but it did score 896 overall.The javascript scores were very good, 2039.0 in sunspider and 158.85 in V8
juicytuna - Friday, November 18, 2011 - link
Sorry, that should have been a V8 total score of 1335.0ChronoReverse - Friday, November 18, 2011 - link
My HTC G2 is running CM7.1 with a patch that enables GPU acceleration in the web browser.With the single core CPU overclocked to 1.5GHz, my score is 1090.
Jamezrp - Sunday, November 20, 2011 - link
I've been using Vellamo on a number of devices for testing, and at first I thought that it was great...but then I found the problem, which is similar to NenaMark (1&2), Neocore, and several other benchmarks: results are very much based on screen resolution. That is to say, if you have an Android phone with a low-res screen, you will have an inflated score.Case in point: I have something like 7-9 handsets I'm testing right now, with screen resolutions varying from 320x480 to 720p. The low resolution LG Doubleplay manages to get an average of 810, while the Galaxy Player gets an average of 620, the HTC Rezound gets 920, and the Galaxy S II (well I had a score, but my frickin spreadsheet lost it's data while upgrading my GPU for some reason...)
As I continue testing, I'll add my findings, but it looks like the more I test the more relevant screen resolution is to the test, which inherently limits it. It may be well rounded, but the overall scores may not matter in the long run. Individual tests, however, may.
Jamezrp - Sunday, November 20, 2011 - link
I've gone over the testing here, and after some serious testing on my own, I have to conclude that Vellamo isn't a good benchmark at all. There's a number of reasons:1) Results don't match. I've tested across six phones now, and even on the same device results don't match up. I get 80-point spreads, which is a joke. Why does the Rezound have a low score of 856, but a high of 930? Why does the benchmark run better when other applications are running, and worse when nothing else is running? Why does it take over an hour for the scores to pick up (ie, for the phone to "warm up")? The results, on several devices (though not on all) are not accurate and don't seem to level out.
2) Results aren't accurate across tests: Vellamo uses Sunspider 0.9.1, a benchmark used here and one I use myself for testing Javascript. I like the test alot. It's accurate and very helpful. Vellamo inflates (or because the score is in less-time-is-better, deflates) scores significantly. I've run the benchmark through the default web browser and through Vellamo, and Vellamo consistently scores higher by 400+ms. If this one test which I can reproduce with minimal work isn't accurate, chances are so are plenty of others which I haven't yet gotten to. I'm actually surprised the previous Sunspider scores haven't been used as a point of comparison in this review.
3) Most of the tests are frankly pointless and don't show useful results. As someone who will use the benchmark scores to chart data, points are fine. But they don't really say anything besides what the points are. The tutorial that shows what each test is and why is great, but ultimately the data is what's important, and the data is flawed from the start. In an attempt to simplify it, the data just doesn't make any actual sense.
4) As I mentioned in a post earlier, scores rely heavily on display resolution, which is not good. If smartphone displays are going to vary as they do now - from 320x480 to 720p - then the scores aren't going to match well in any way. I have a few phones scoring higher than others solely because of this. The LG Doubleplay, for instance, a dumbphone by nearly all regards except that it runs Android 2.3.5, averages 200 points higher than the Samsung Galaxy Player in an average of 10 tests. What the hell.
I'll keep playing with it, but Vellamo thus far doesn't look like a benchmark worth using, at least not in the long run.
ickyboo - Saturday, February 4, 2012 - link
Don't forget it's also written by Qualcomm, So all scores will be biased towards their own chipsets.sunwufan - Tuesday, November 22, 2011 - link
I am proud of using it now and hopefully they could get the 4.0 ROM ready soon.branwell1 - Friday, December 23, 2011 - link
嗨,大家好，我来自中国上海。一个星期前，我买了一部小米手机，仅仅花了我1999人民币(约合316.64美元)，用了一个星期，小米非常棒。希望想要这部手机的人能够尽快得到。Hi! I am from Shanghai China.I bought a Xiaomi one week ago.It costs me ￥1999(about $316.64).I have used it for one week.It's really excellent！Hope you who wants one could get and enjoy Xiaomi as soon as possible. My e-mail is [email protected]u can get more informatio from me.
branwell1 - Friday, December 23, 2011 - link
Hi! I am from Shanghai China.I bought a Xiaomi one week ago.It costs me ￥1999(about $316.64).I have used it for one week.It's really excellent！Hope you who wants one could get and enjoy Xiaomi as soon as possible. My e-mail is [email protected]. You can get more informatio from me.ickyboo - Saturday, February 4, 2012 - link
I'm surprised that such a well respected site like Anandtech would fail to mention that Vellamo is actually written by surprise surprise QUALCOMM!, So there's little wonder the scores are so well regraded for the Qualcomm chipsets. Seriously if Anandtech continues to use this benchmark in the future then their scores would mean nothing and be a waste of time with such a high bias to Qualcomm chips with a Benchmark written by Qualcomm.Henk Poley - Sunday, March 25, 2012 - link
I wonder why the Xiaomi is not used in more recent Android comparisons here on Anandtech.