August 3rd, 2009
Quickly on from Part I and Part II, here is part three of the great browser shootout. The browsers we are looking at are:
All the tests are done on a fully updated Windows 7 RC on a IBM ThinkPad R52.
The first tests I do is to start up the browsers and check the time till they are fully open and I am able to view a site. All the browsers were set to a blank home page so that this would not affect the load time. After they are open, I check the memory usage.
Then I load one by one the following sites:
For gmail and hotmail, I don’t measure the time to load the login page, but rather the time from login to the time I can see the inbox. In the case of hotmail, this is not the inbox but the main “activities” screen.
All the times taken before this are in seconds and you may take +/- 1 second error.
Following this I run the Acid2 and Acid3 tests. Considering that all the browsers here are Acid2 compliant, only the Acid3 has any meaning. Acid tests are a measure of the browsers ability to support HTML and CSS standards as defined by the W3C. Acid3 requires support for JavaScript as well.
Then I loaded up four different JavaScript test sites. At this point I again checked the memory usage. For Chrome and Internet Explorer 8, I will give the total usage of all the processes in memory.
The JavaScript tests were then run:
The JavaScript test from the V8 Browser comparison is a numerical score – higher the better; the other JS comparisons are in milliseconds – lower the better. The mootools test is actually a test to compare different JavaScript libraries, Mootools own as well as JQuery, Prototype, YUI and Dojo. This test acts as an indicator of real-world usage, since many sites use one of these libraries to run their JS. Note: IE8 failed the Dojo test completely.
Without further ado, here are the results of the tests.
Of course, IE8 handily wins this one. The slow time of Opera may be due to the fact that it is still in beta.
This is a logarithmic scale, otherwise it would be difficult to show the usage. IE8 is a huge memory hog, using 824MB when loaded, nearly 4 times the usage of the next highest which is Safari 4. Chrome easily wins this with a usage of only 127MB at max.
Most sites load pretty much the same on each browser, except on IE8 which consistently takes more time than the other four. Even hotmail, Microsoft’s own, is fastest in Firefox! Gmail also loads fastest in Firefox, but only by one second, well within the margin of error. Overall, Firefox comes in handily in first place, followed by Chrome and Safari 4 almost neck and neck. But these three browsers are easily the fastest and the difference in times to load these sites is so negligible in real usage, you can’t tell the difference.
The Acid3 test is very close. Opera, Safari and Chrome all do the 100/100. Firefox gets 93, but IE8 gets only 20! Although there is a problem with Chrome where you may see a “Linktest Failed” for a few seconds even after everything is complete until the page clears up to meet the reference.
The v8 suite is Google home-grown to torture test the JavaScript engines of the browsers, but considering the lead that Chrome has over the other browsers, one wonders if it is actually designed to highlight Chrome, considering the other JS tests don’t show such increases.
Again this graph is a logarithmic scale. All the times are in milliseconds. Here you find that IE8 is visibly the slowest in all tests, while Firefox, Chrome and Safari all are bunched together. If you go in for specifics, except for the SunSpider test (won by Chrome by a short head) and the Prototype test (won by Opera surprisingly), Safari 4 wins all the JS tests – whew!
So that’s the end of the tests. Finally, it comes out that all the browsers are pretty much equal when it comes to performance, except IE8. But even IE8 is head and shoulders above IE7 and IE6, so you should upgrade to this even if you are not willing to move to any alternative browser.
Which browser do I use? Primarily, Firefox because of the extensions I use – but I often use Chrome and Safari. I’m personally not a fan of Opera but that’s just me – it’s a very nice browser. Of course, there are still some sites which are nothing but IE compatible – so I still haven’t gotten the icon out of my quick launch.
Do drop me a line with any questions or queries.
Categories: browsers, Comparison, reviews, technology |
Tags: browser, chrome, Comparison, firefox, internet explorer, opera, review, safari, shootout, test | 2 Comments
August 3rd, 2009
Following on, after a long break, from my browser shootout – Part I can be found here – I finally got around to doing the tests themselves.
Before I go on to the shootout, let’s look at some of the standout features of the browsers here.
All the browsers now work towards standards compliance, which means, hopefully, that if you write a page which is standards compliant, it will work the same in all browsers. All the browsers now have newly (re)designed JavaScript engines. This means that sites which are JS heavy, like Gmail or the new Yahoo Mail, are rendered much faster than earlier.
Two browsers, Chrome and Internet Explorer 8, maintain separate processes for their tabs, which means if one tab crashes, the whole browser does not go down.
Chrome, Safari and Opera have a speed-dial feature (sometimes called top sites or recently browsed), where when you open the browser it shows you a set of thumbnails of the sites you visited most recently. Safari’s implementation is the best, because it updates the thumbnail even before you click on it, so you have a near real-time view of the site even without visiting it.
Firefox, Opera, Safari and Chrome all support HTML 5 in some measure, including new features like support for <audio> and <video> tags. Firefox has a lot of support for HTML 5 video which you can see on this page. Unfortunately, you must you FF3.5 to view it.
Opera has a nice feature where you can see the thumbnails of your open sites rather than just the plain tab-bar, but this can be added via an extension to Firefox as well. Safari’s tab implementation is a little wonky. Normally I would expect that clicking on a link which is marked “target=_blank” (which on non-tabbed browsers would open a new window), should open the link in a new tab, but in Safari it works the old way unless you Ctrl-click. No other browser does this.
And the standout feature still is Firefox’s extensions and themes. None of the other browsers have anything close to this. The range of extensions is simply mind-boggling. In fact, you would hardly need half-a-dozen other stand-alone applications, because there would mostly be an extension for it. From chat clients to weather updates, world clocks, HTML validators, FTP tools – you name it it must be there. Or you could create one yourself!
Of course, Chrome will be adding a lot of this stuff soon, but there is a lot of catching up to do.
The tests themselves are a mix of loading speeds, memory usage and JavaScript speeds. Read about them in Part III.
Categories: browsers, Comparison, reviews, technology |
Tags: browser, chrome, Comparison, firefox, internet explorer, javascript, opera, safari, shootout | No Comments
July 9th, 2009
This post has been a long time coming. I installed Windows 7 RC the day after it was released, and Ubuntu the next day. Since I have an older laptop, I wanted to see how well it each would work.
My laptop is an IBM Thinkpad R52 – before the Lenovo buyout. It has an Intel Pentium M processor (1.7GHz), 1.25GB RAM and a 160GB hard disk. The disk is not the original disk, but is a Samsung ATA drive.
Here in India, it is important to note that many of us eke out the most we can manage out of older hardware. I remember installing Windows 3.1 on a XT PC with 1MB RAM… aaah, those were the good old days! 🙂
But back to this comparison. The laptop originally came preloaded with Windows XP Professional. Unfortunately, it also came preloaded with a whole load of crap that IBM believed was essential to a good computing experience (NOT!). So, when I upgraded the RAM with an new 1GB chip, and I put in the new hard disk – I also decided to install Vista.
I am not one of those who does not like Vista. Yes, it sucks in a lot of ways, but in many ways it doesn’t. It is the natural progression of the Windows Product Line, with a few stupid ideas thrown in. Let me not really rant about those here. But oddly enough, it ran faster than XP (maybe because of the lack of IBM crapware).
There is an interesting aside to note here, though. On all my XP machines, I have noticed that after a few weeks/months of usage, it starts to slow down. And no matter how much you try to clean up the system, defrag, regclean and everything, it still never gets its sprightly edge back! Maybe it’s just me, but with Vista, after 8 months, I still am at the same speed I started out with – it’s not fast, but hey – it’s consistent.
So, now I decided to install a triple boot system (Vista, 7 and Ubuntu).
Installation
No points for guessing what happens here – Ubuntu wins hands-down! 36 min for the Ubuntu install. Unfortunately, the first disk I burned came out corrupt – even after the verify, so I had to burn a second disk. But, even if you include the total burning time, it still came to only 50 minutes (including the time to figure out the disk was corrupt).
Windows 7 RC came in a distant second – 1 hour 23 minutes and Vista came in nearly 10 minutes later at 1 hour and 31 minutes. And of course, this was not the end of the installation for either Windows OS. The drivers, the drivers… the drivers nearly drove me MAD! In fact, there is still some device on the Windows 7 system that it hasn’t yet figured out although I’ve thrown everything I had at it. But it finally works, so that’s bearable.
But the problems don’t end with installing the drivers. This laptop has an Intel 900 GMA chipset, so there is no WDDM display driver for Vista or W7RC. And there is never going to be one. And this bug I lay squarely at Microsoft’s door. If Ubuntu and even a hacked Mac OS X can run all their graphics properly on this chip, why did MS have to raise the bar so high? So, now I don’t have all the fancy whiz-bang stuff that both these OSes are supposedly famous for!
Boot Time
Again, Ubuntu wins the day. It takes just 27 seconds to get to the login screen while Vista takes just shy of two minutes and W7RC takes 2 min 39 sec to reach the point where I can enter my password. All of them, however, just take an additional 10 seconds to get to the main desktop.
The nicest part about Ubuntu was that it gave me a proper progress bar. I always knew how much of the boot process was complete, unlike with the Windows OSes. Why they can’t get this simple thing right is beyond me.
One thing I didn’t quite figure out with Ubuntu, though, was why it always turns on the Bluetooth on the laptop when booting up. I always have to manually turn it off. Hmm…
Shutdown
It still goes to Ubuntu, but not by such a large mark. Ubuntu – 10 sec, W7RC – 11 sec and Vista – 21 sec.
Battery Life
I did this by running a simple video in VLC with all the brightness at maximum and no shutting down/standing by of anything. Windows 7 wins this one – nearly 2 hours. Vista comes a respectable second with 1 hour 49 minutes. And Ubuntu comes in last with an hour and forty minutes.
Usage
This is the part where the OSes character changes. I have certain problems with Ubuntu which prevent me from choosing it (out of the box) as my primary OS. Firstly, the fonts suck. They are far too large. Secondly, Evolution (the Outlook equivalent) has too much space occupied by the chrome and less by the email/appointment/task that I am looking at. Thirdly, I don’t want to have to download a 500 MB of updates every three days. Can’t we just collate the updates? Fourthly, the Bluetooth thing – see above. And finally, the Thinkpad keys don’t work correctly.
Now, don’t get me wrong, all of these (except maybe the Evolution problem) can be fixed with a little tweaking – but not everyone can tweak the OS that easily. Even I had trouble with the keymapping. If all these could be fixed out of the box – Ubuntu would be the best OS on my laptop. Of course, I’d have to give up Photoshop (sorry, GIMP doesn’t really cut it yet) – but that would make it possible.
Windows 7 just doesn’t do it for me, I’m afraid. Mostly because all the nice things about W7RC require that I have a laptop with a WDDM driver. So I don’t get all the fancy mouseovers, the nice popup stuff and all that. And since I don’t get all of that, it doesn’t have enough to make it more worth my while than Vista. There are all sorts of nice ideas included in W7RC like the pinned menu and the fast shortcuts and the idea of doing away with the system tray, but a lot of the stuff needs to work on my system before I could enjoy it. It works on my desktop, and I like it there, but on this system – naah. (I like the new Solitaire though – finally cleaned up after 15 years!)
So, finally we come to Vista. Which is the system I use primarily and I like it. Yes, there are a lot of things to dislike about it – the stupid implementation of the UAC, the backward compatibility breakages (especially with stuff like VNC), the lack of the WDDM drivers. But overall, it works. It works fast on this system – faster than XP at least (and since I don’t have a standalone XP license, I cannot put XP on the system). I like the new explorer – it gives a lot of flexibility and overall it has enough eye candy (even on this system) to make it nice to look at.
But I keep switching between Ubuntu and Vista. Depending on my mood, I guess. However, W7RC is relegated to the outhouse for the moment. Although, I am planning a browser test next – IE8, FF3.5, Chrome, Safari and Opera. And I’m doing this test on W7RC! Stay tuned.
Categories: Operating Systems, reviews |
Tags: 9.04, Comparison, IBM, Jaunty Jackalope, release candidate, review, Thinkpad, Ubuntu, Vista, Windows 7 | 1 Comment
