Chrome OS Benchmarked Against Moblin, Ubuntu Netbook, More

An anonymous reader writes "Using the latest build of Google's Chromium OS source code, Phoronix built it out to run on a Samsung netbook and ran sixteen benchmarks, putting it up against Moblin 2.1, Ubuntu Netbook Remix 9.10, openSUSE 11.2, and Fedora 12. They ran some of their usual desktop benchmarks (encoding, video, etc..), but more interestingly they ran a number of battery, CPU usage, and memory consumption tests under different settings that show some of the advantages and disadvantages for each of the Linux distributions, and spotted a few bugs along the way."

If you really care about Linux performance...

[Silverlancer]

There's been an enormous improvement in the Linux scheduler in recent months [multimedia.cx]--in some cases the performance improvements are as high as 60-80% with simple multithreaded apps like video encoders. The instant 2.6.32 comes out officially, expect to start seeing some completely absurd results in stupid "comparisons between Linux distros" like these, where the distros that happened to update to .32 trash the ones that haven't yet.

"benched"

[hey]

In sports "benched" usually means taken off the field. I like this tech usage (benchmarked) -- which is new to me.

Re:Feh

[h4rr4r]

Phoronix is a Unix, Linux, Solaris site, they never bother with testing windows. The results would be of little interest to their target audience.

Re:Shocking.

[Enderandrew]

Does the Phonorix test suite even run on Windows yet? I don't think that's released yet.

Phonoroix does benchmark against the Mac all the time.

Re:Shocking.

[fuzzyfuzzyfungus]

Given the number of times that Ext4 is mentioned in Chrome's publicly released design docs, I'd be inclined to suspect that they just haven't bothered to configure it with Ext4 yet, at least not in the publicly released build.

Either that, or the left hand and the right hand need to have a little sit down and chat in the immediate future...

Re:The most boring benchmarking ever.

[Enderandrew]

So basically no offline (relatively speaking) functionality and can boot in 3 seconds. Yay?

All the web apps are using Gears, which caches everything for offline use. So all the apps will be usable offline. Your data is stored in the cloud, but also cached to the SSD. So it isn't fair to say it doesn't function offline.

Yes, it looks like Chrome.

I can't imagine I'll pass up KDE for this myself, but it passes the Grandma test. They know how to use a web browser already. You plug in a camera, a little overlay shows the camera. You literally drag and drop a photo from the camera to a Google Talk overlay, and it sends to the picture to that person.

Passing the Grandma test is important. Put Grandma (even if she knows XP) in front of 7 and Chrome. See which one she prefers.

Not really. I would barely call this Linux.

Moblin has a new UI, and people aren't saying Moblin doesn't count as Linux. Embedded Linux still counts as Linux. Headless servers still count as Linux.

Why is this not Linux, simply because there is a different UI?

This is more than a thin client, since a thin client can't be used offline. And apparently it is a bit of a misnomer to say you can't install other apps, since they installed the Phoronix test suite.

Orange Micro

[tepples]

Back when Macs still used 68000 series CPUs, Orange Micro [wikipedia.org] made single-board PCs that plugged into a NuBus slot on a Macintosh II, allowing Mac owners to run MS-DOS and Windows and their apps.

Re:Shocking.

[RanCossack]

Notice they didn't bother comparing any of them to either Windows 7 or Mac OS X. They wouldn't want open source to look bad, would they? LOL.

I know, right? Moblin's boot time can't hold a candle to Windows 7's, but the real powerhouse is Vista -- a boot time score higher than Moblin, Ubuntu, Suse, and Fedora's all *combined*... and then *squared*.

Now, if you'll excuse me, I'm off to dominate a game of golf.

Re:How?

[fuzzyfuzzyfungus]

TFA mentions that they mounted the ChromeOS image on another system, and made some adjustments. At least at this stage, it is pretty much stock Ubuntu with a chromey face on top, so anybody who has the image mounted rw on another system should be able to bodge on anything Ubuntu can do with fair ease.

Re:Let's stop calling it "Chrome OS".

[Runaway1956]

Yes, you're probably right. He has, after all, demanded that Linux/Slackware rename itself, as well as Linux/Debian, and Linux/Suse and ...

Oh wait, that hasn't really happened, has it?

Re:The most boring benchmarking ever.

[SanityInAnarchy]

their documents are all Google documents ("what? what if it's a Microsoft Word document?" ...) and you need to type that in the address bar...

Neither of these are required. I just watched a demo in which someone plugged an SD card into a laptop running Chrome OS -- it popped up a filesystem browser. They found an excel document, clicked it, and it opened in the Web version of Microsoft Excel.

So, they're working on it, but there is definitely a model for running applications that happen to be webpages.

Re:The most boring benchmarking ever.

[SanityInAnarchy]

Bejeweled [vx4.com]. MS Word [officelive.com]. MSN Messenger [live.com].

isn't caused by running familiar apps slowly, it's caused by being forced into inferior and unfamiliar facsimiles.

You could make that case about Bejeweled -- the official web version is deliberately crippled compared to the desktop version. However, the other two are very serious attempts by Microsoft to expose that functionality online. And I haven't covered alternatives -- Meebo and Google Docs come to mind.

Re:Let's stop calling it "Chrome OS".

[DragonWriter]

Repeat after me. I want my applications locally, where I can use them regardless of having a network or not.

The intent is for that facility to be provided by web apps with offline functionality, something that has been important to Google since before Chrome OS was conceived (or, at least, announced as something they were working on.)

The whole world could burn down and I could still do my work with my PC (as long as I survived).

The only initial barrier to do that with Chrome seems to be that the first time a user logs on, network connectivity and a Google Account is required, although Google has stated that that initial limitation is just that: they want to work with other authentication sources, specifically they've cited having an OpenID alternative as a goal. While this still requires network connectivity, something as simple as a home LAN with an lightweight server doing authentication would work. And, since Chrome OS is targetted for netbooks, a certain degree of network dependency is not as critical as it would be if it was intended to be a general-purpose desktop OS.

Now, you may have an issue with the entire idea of a netbook-specific OS and prefer just a general purpose desktop OS with slight visual adaptation to the limit screen real estate -- and that's certainly a reasonable preference. But I don't think that all netbook users are going to share that preference.

I don't think a great deal of thought actually went into it.

I think that its pretty clear from reading the pages on the design and plans that a great deal of thought has gone into it.

I think its equally clear that what has been released has not realized all of the things that are planned for it, and that it is not intended to be a production release, a release candidate, a beta, or even an alpha release, but more an opening up of the development code base and the work-in-progress plans to public view and comment.

Re:Shocking.

[oxfletch]

We're using ext3 because that's what the bootloader supports right now. The plan is to move to ext4.

Re:How?

[oxfletch]

This is a confusion about what "native application" means - Linux apps vs native code. It does support Native Client - http://code.google.com/p/nativeclient/ [google.com]