Print 110 comment(s) - last by Smilin.. on Sep 15 at 12:05 PM

What browser is the best? We have the information you need to decide

In the last segment of our next generation browser benchmarking and comparison, we looked at user interface features and installation details.  We also benchmarked install times and application launch times.  We now will turn our attention to CPU and memory usage in this segment.  We'll also briefly contrast security in the next gen browsers.  This segment will be followed by a third and final installment in which we'll examine performance in popular benchmarks and standards support.

4.  Resource Usage

One of the most critical aspects of a program is the amount of resources it uses per the amount of work it does.  We measured memory and CPU usage for each browser with ten tabs open and loaded -- DailyTech, AnandTech, CNET, CNN, Sports Illustrated, Gamefaqs, Google, Yahoo, Bing, and Facebook (logged in).  We then took measurements after 15, 20, and 25 minutes of operation.

When it comes to memory, Firefox really shows its worth.  This may be surprising to some as early in its development the Firefox browser was known as a memory hog, due to memory leaks.  This has completely turned around and it is now the slimmest entry. Namoroka uses significantly more memory than 3.5.2, but hopefully this is just one of the rough edges that are to be expected of an alpha release.

Looking at the rest of the pack, Opera deserves an honorable mention for a close second in memory usage.  Safari and Chrome, on the other hand, were both memory hungry.  However, no application was quite as bad when it came to memory as Internet Explorer 8, which used nearly twice the memory of its closest competitor.

Turning to the CPU, Opera was in the lead for least average use.  Opera 9.6 also led for the lowest maximum observed CPU use.  Opera 10.0 did show a rather high maximum usage.  This is due to a brief, rather uncharacteristic, spike.  This appears to be a rather isolated occurrence, but nonetheless we kept the result.

Chrome, Safari, and Internet Explorer were all rather poor when it came to CPU use.  Chrome 4 ate up the most CPU, topping at an unpleasant maximum of 64 percent.  Firefox, on the other hand, showcased low usage (with no add-ons installed), though 3.6a1 was a bit more CPU hungry than 3.5.2.  Again, hopefully these issues will be resolved before release.

5.  Security:

Having looked at the resources used, its also important to look at what is being done with them.  We already concluded that Opera provides the most built in features (non-security) in our first review (though Firefox wins when add-ons are considered).  But what about security features?

The below table illustrates some highlights of these browsers' track record:

Browser Tab/Process Isolation Private Browsing Mode Popup Blocking Ad-Filtering (JS, Flash) Anti-Phishing Malware Blacklist Unpatched Security Flaws, Secunia Unpatched Security Flaws, SecurityFocus
Opera 9.6 No No Yes Yes, click required Weak Weak 0 2
Opera 10.0 No No Yes Yes, click required Weak Weak 0 2
Firefox 3.5 No Yes Yes Via add-on Moderate Moderate 0 0
Firefox 3.6 No Yes Yes Via add-on Moderate Moderate 0 0
Chrome 2 Yes Yes Yes No Weak Weak 0 0
Chrome 3 Yes Yes Yes No Weak Weak 0 0
Chrome 4 Yes Yes Yes No Weak Weak 0 0
IE 8 Yes Yes Yes Yes (via InPrivate Filter) Strong Strong 2 16
Safari 3 No Yes Yes Via add-on Weak Weak 0 0
Safari 4 No Yes Yes Via add-on Weak Weak 0 0

As you can see, security is a rather confusing topic to rate the browsers on.  On one hand, IE 8 offers an excellent private browsing mode, tab isolation, and great blacklisting of malicious sites.  On the other hand, its InPrivate Filter doesn't catch all ads.  IE 8 is also the most frequently attacked and exploited browser, though Microsoft puts great effort into patching as quickly as possible. 

Despite this, IE 8 for the very inexperienced/naive user is probably the best bet as it blocks more blatantly malicious sites than the rest of the field.  Microsoft-sponsored research puts this block rate at 81 percent versus the next closest competitor -- Firefox -- at 27 percent.  This may be a bit of an exaggeration, but Microsoft deserves praise for its progress on this front. 

Chrome offers good overall protection with tab isolation, a private browsing mode and less vulnerabilities, but it is victim to probably the most ads of any of the browsers.  Firefox is a close runner up to IE 8, especially when add-ons are considered.  However, it lacks tab isolation.  Opera and Apple have both put a fair deal of thought into their security efforts, but they just aren't as strong or focused as those of Microsoft, Mozilla, and Google.

Note: All benchmarks were performed in 32-bit Vista on a Sony VAIO laptop with 3 GB of RAM, a T8100 Intel Processor (2.1 GHz), and a NVIDIA 8400 GT mobile graphics chip. The number of processes was kept consistent and at a minimum to reflect stock performance.

Comments     Threshold

This article is over a month old, voting and posting comments is disabled

No - Just No
By GeorgeH on 9/8/2009 3:00:22 PM , Rating: 2
I've got 4GB+ of RAM, and I want my programs to use it. I've got a 2GHz+ CPU, and I want my programs to use it.

Saying "Less is Better" is completely moronic, unless you think Netscape Navigator 4.0 is the shiznit. You can even see why this is dumb in your own data by looking at each individual browser - or do you honestly think each one is getting progressively worse?

RE: No - Just No
By gstrickler on 9/8/2009 4:27:08 PM , Rating: 2
Saying "Less is Better" is completely moronic
No, it's not, at least not in the context in which he used it. If a program can deliver similar (or better) performance while using less RAM and/or less CPU, that is better. Computer resources are limited, no matter how much you have. There are also limits on how much CPU, RAM, HD, GPU you can install in a computer before you have to replace the whole machine and/or OS. Programs that make better use of those resources will allow you to keep more programs and/or documents open concurrently with less impact on performance. The more responsive your computer is, the more work you can get done. They may also extend the usable life of your computer, thus lowering your costs.

While efficiency is always better, there is a point of diminishing returns, so it's actually about balance. Unfortunately, you can't show "balance" in one chart because everyone has different computer resources (CPU speed/cores, available RAM, HD size/speed, etc.), so you have to show different scenarios that allow an educated reader to choose the combination that works best for their particular situation. That's the fundamental problem with a single "benchmark", and it's the reason that most reviewers show results from a range of benchmarks. On each individual benchmark, lower (or higher) is "better". If that benchmark doesn't represent a scenario you'll encounter in your usage, then you can safely ignore it. Choose the product that best suits your needs.

I want software to use the resources I've got efficiently, not just to use them because they're there. If a program can provide a significant performance increase by using more RAM, great, but if it needs 2x the RAM to deliver a 10% performance increase, it might not be worth the trade-off. Of course, that depends upon how much 10% is in user time and how much 2x RAM represents of your total available resources. A decrease from 10 mins to 9 mins using 2MB RAM rather than 1MB RAM is probably a good trade-off for all users, but decreasing .10 seconds to .09 seconds for going from 1GB to 2GB would probably not be acceptable to most users.

Intel ran into that problem with the Pentium 4 line, which is why it was a dead end. It's why the teams working on the Core derived CPUs have to show that new CPU "features" provide more performance improvement than they cost (in # transistors and/or power usage). All engineering is about trade-offs, about picking the combination that best fits ALL of the customers' needs.

RE: No - Just No
By GeorgeH on 9/8/2009 7:02:21 PM , Rating: 2
Way to miss the point. Windows 98 and Vista are both operating systems, they both run browsers that display webpages, and they both let me edit my documents. Would you say "less is better" when it comes to resource usage while performing identical tasks on both operating systems?

That's not to say resource usage is irrelevant, just that the value judgment "better" is completely ridiculous. As with 98 and Vista, browsers are not doing the same things, even when accomplishing the same tasks.

To finish with the OS analogy, if you tell me you run 98 because you've got a PII and 256MB of RAM, fine. If you tell me you run 98 because it uses less resources (which is better!) I'll call you a moron.

RE: No - Just No
By gstrickler on 9/8/2009 7:34:38 PM , Rating: 2
I didn't miss the point, and your example clearly demonstrates my points, which I've summarized below:
1. Not all computers have the same resources available.
2. There are both hardware and OS limits to the available resources.
3. From the perspective of each individual benchmark, lower/smaller (or larger/faster) is better and is accurate, but which is "better/best" for a given situation may vary.
4. By publishing a range of benchmarks, you provide the information necessary for an educated reader to choose the product that best fits their needs.

In the context used in the article, each individual benchmark has a "more optimal" result and a "less optimal" result, and it's perfectly valid to call the more optimal result "better" within the context of that one specific test.

No one would claim that a 10 second launch time is better than a 2 second launch time for an application. Clearly one is preferable to the other, so 2 seconds is BETTER than 10 seconds. That doesn't mean the application with a 2 second launch time is "better" for any given situation, it may be unreliable, insecure, incompatible, too slow, or use to much RAM to be useful in any given environment, but it's still valid to claim that a 2 second launch is better than a 10 second launch.

"This is from the It's a science website." -- Rush Limbaugh
Related Articles

Copyright 2016 DailyTech LLC. - RSS Feed | Advertise | About Us | Ethics | FAQ | Terms, Conditions & Privacy Information | Kristopher Kubicki