Motorola Xoom SunSpider Results

2011/03/06

in Mobile

I saw a post touting the Ipad2 SunSpider JavaScript test results which mentioned wanting to run the same tests on the Motorola Xoom running Android Honeycomb.

Ipad Sunspider Results

As you can see the results of the Xoom running Android 3.0 (Honeycomb) achieving 2083ms,  is faster than the iPad2 achieving 2097ms by a very small margin.  I’ve run the test several times with a score ranging from 2046ms-2100ms to be fair, as I am not sure how how they came to the numbers provider in the original blog post.  It wasn’t mentioned how many times CNET[CNET article was pulled?] ran the test and if the results were averaged.

I took the original numbers from the post and added the Xoom to the chart for a comparison.

Sunspider results chart for Xoom and iPad

Sunspider results chart for Xoom and iPad

The Javascript Sunspider tests javascript performance on the device, which in todays Javascript heavy web applications would show a real world effect for end users.   These devices while certainly faster than previous generations,  are still slower than a laptop though.   When I ran the tests on my 2010 MacBook Pro,  the scores were in the 400-500ms range for the same tests.   It will be interesting to see what other benchmarks will compare these 2 titans of the tablet generation in the coming weeks.

I think these 2 devices are very close in performance specs, which should make the transition between ecosystems easier on application developers.  I suspect the Ipad2′s A5 shares some of the same logic of the Tegra2, since they may both be based on the Cortex A9 architecture.   Though,  for the Ipad2 being a “custom” designed chipset, and tightly coupled to the IOS software,  I would have expected it to be beat the Xoom by a greater margin, which isn’t the case here.

 

SunSpider Xoom Test Results

============================================
RESULTS (means and 95% confidence intervals)
--------------------------------------------
Total:                 2083.5ms +/- 0.9%
--------------------------------------------

  3d:                   319.7ms +/- 0.6%
    cube:               117.0ms +/- 0.6%
    morph:              105.7ms +/- 1.5%
    raytrace:            97.0ms +/- 0.9%

  access:               205.8ms +/- 0.6%
    binary-trees:        14.1ms +/- 5.6%
    fannkuch:            82.3ms +/- 0.4%
    nbody:               82.7ms +/- 1.6%
    nsieve:              26.7ms +/- 1.8%

  bitops:               159.6ms +/- 3.5%
    3bit-bits-in-byte:   14.3ms +/- 2.4%
    bits-in-byte:        28.7ms +/- 1.7%
    bitwise-and:         50.5ms +/- 10.6%
    nsieve-bits:         66.1ms +/- 0.9%

  controlflow:           13.0ms +/- 3.7%
    recursive:           13.0ms +/- 3.7%

  crypto:               124.7ms +/- 3.2%
    aes:                 48.0ms +/- 2.4%
    md5:                 39.7ms +/- 5.4%
    sha1:                37.0ms +/- 4.4%

  date:                 292.5ms +/- 1.6%
    format-tofte:        98.8ms +/- 1.2%
    format-xparb:       193.7ms +/- 2.1%

  math:                 200.1ms +/- 3.6%
    cordic:              72.5ms +/- 1.4%
    partial-sums:        94.1ms +/- 8.1%
    spectral-norm:       33.5ms +/- 2.9%

  regexp:               122.4ms +/- 1.1%
    dna:                122.4ms +/- 1.1%

  string:               645.7ms +/- 1.2%
    base64:              34.9ms +/- 5.7%
    fasta:               80.7ms +/- 3.9%
    tagcloud:           191.4ms +/- 3.0%
    unpack-code:        221.2ms +/- 0.9%
    validate-input:     117.5ms +/- 1.2%
Motorola Xoom SunSpider Java Results on Android Honeycomb

Motorola Xoom SunSpider Java Results on Android Honeycomb

Related posts:

Post comment as twitter logo facebook logo
Sort: Newest | Oldest

I ran a test in the stock browser on my Xoom and got 1449.8ms

The battery for the Xoom is 10 hours, not 8.5. And the reason the CNet tests boast that the Xoom scored worse is because they took the results of the Anandtech benchmark graphs, which are actually run on Sunspider 0.9, which results in universally higher scores by all hardware. 0.9.1 is easier/faster on the Xoom.

Brandon

"Though, for the Ipad2 being a “custom” designed chipset, and tightly coupled to the IOS software, I would have expected it to be beat the Xoom by a greater margin, which isn’t the case here."

What you didn't consider was that Honeycomb is also custom tailored for Tegra II chipset, so it too is optimized.
Another reason why I'm excited about iPad 2 is, in most things, its better than the Xoom (Battery - 8.30 vs 10 hrs, Display - LCD vs IPS, Thickness and Weight, Look and Feel, Apps etc), and Apple still has their biggest trump. iOS 5. If that brings in Notification, Widgets and improves the Browser, the Xoom will have nothing going for it other than a USB port and (possibly) more Ram

Kunal,

While Honeycomb is most likely optimized to run on the Tegra platform, its not going to be the only platform it runs on. We already have the rumblings of Dual core snapdragon, and Omap processors coming down the pipeline, which will also be running Honeycomb. IOS only has 1 chipset to run on, so I would expect it to be more finely tuned and perform better.

The hardware of the iPad2 is very nice, and very compelling, but it's the same IOS in new clothing I'm afraid. That's actually a good thing for many, since the iPad ecosystem is currently the largest and has the most apps...today. The bad thing is, IOS has become stale, and hasn't really evolved much to take advantage of modern concepts and larger screen real estate of the iPad. Multi-Tasking is tacked on, and application dependent, the home screen is still just rows of icons (and now folders...of icons), and most of all the silly notification system that really needs to be overhauled.

All of these may be fixed in IOS 5, but I suspect Apple isn't ready to make such dramatic changes yet to their IOS system. It will happen at some point like you said. At this point it's a waiting game... "If and when IOS evolves" is not something I felt like playing anymore, so I'm happy to take the chance on the Android ecosystem evolving for my needs quicker.

At least the Ipad2 will ignite other hardware designers to keeping evolving their hardware for Android to run on, so competition is definitely good for both ecosystems. :)

Thanks for posting these numbers. Given the results, can one can infer anything about the amount of RAM in the iPad 2?

J,

It seems the consensus is that it has 512mb, but I don\'t think we will know for sure until the first tear down when it launches into peoples hands next week.

I\'m not surprised it has only 512mb if it does, since IOS has made multitasking secondary, so they may not put as much emphasis on needing to have more apps/processes open at the same time.

Expect to see more info next week.

Jef

Trackbacks

  1. [...] a XOOM ran the test and scored ~2100MS. Not reliable, but it makes sense. Found this site though: http://jeftek.com/1942/motorola-xoom-sunspider-results/ __________________ I'm a [...]

  2. [...] nicht getestet werden konnte. Ein Testergebnis für das in den USA erhältliche Xoom gibt es aber auf einer anderen Seite: Dort liegt das Motorola-Tablet 14 Millisekunden vor dem iPad 2 – also gleichauf, denn diese [...]

  3. [...] comparison, a Motorola Xoom seems to grasp ever-so-slightly faster formula on a same benchmark: about 2050-2100 ms. As we can see from a above numbers for a iPad 1, though, SunSpider is apparently unequivocally [...]

  4. [...] folks over at jeftek.com also did a SunSpider test to compare with the iPad 2 and found the Xoom quicker, though not by [...]

  5. [...] iOS 4.3, под началом которой работает iPad 2, принес мобильный браузер Safari с таким же, как в настольном веб-обозревателе для Mac OS X новым JavaScript-движком Nitro, якобы ускоряющим JavaScript-обработку вдвое. Действительно, в тестах SunSpider, разработанных создателями открытого браузерного WebKit-движка, второй Apple-планшет справляется с эталонным набором JavaScript-сценариев за 2097 мс, что в четыре с лишним раза быстрее, нежели это делает за 8321 мс первый iPad под эгидой iOS 4.2. Motorola Xoom, кстати, ничуть не отстал и даже незаметно лидирует со своими 2084 мс. SunSpider: скорость обработки JavaScript-сценариев зависит в большей степени от программного обеспечения; меньше — лучше. (Изображение: CNET.) [...]