Gizmos Freeware Reviews  

Go Back   Gizmo's Freeware Forum > Debating Chamber > Internet, Web Apps and Networking

Reply
 
Thread Tools Display Modes
Old 20. Feb 2010, 11:38 AM   #1 (permalink)
Senior Member
 
Ritho's Avatar
 
Join Date: Apr 2008
Location: Planet Earth
Posts: 1,379
Lightbulb My Browser Tests - Choosing a New Browser

In light of discussions in the recent thread here, I decided to do some testing of browsers for myself. I started small and kept throwing in more browsers as I went along, so a couple of the tests don't include all the browsers I tested. All test were carried out using the latest version as of 2/18/10

I was looking at three issues. 1. Browser to Network Interaction by Testing Throughput (Download/Upload Speeds) 2. Browser to Hard Drive Access Frequency. 3. Website Rendering Speed.

___________________________________________

1. Throughput Test:

Since the original tests in the thread were done with www.bredbandskollen.se I decided to continue using their speedtest, although afterward I decided to do a second round of testing with www.speedtest.net. Both, however, produced nearly identical results so I am only going to list one set here.

Methodology: I retested each browser 10 times and then took an average. Any one test that was way out of the normal range of the others was thrown out as an anomaly and the test was run again.

Results: (My max internet speed is 12/2Mbps [Down/Up])
  • Firefox ----------- 11.40/1.94 Mbps
  • SeaMonkey ---------11.33/1.94 Mbps
  • Internet Explorer ----11.12/1.93 Mbps
  • K-Meleon -----------11.48/1.93 *(1.67) Mbps
  • Chrome ------------11.14/1.93 Mbps
  • Opera -------------11.27/1.89
  • Safari --------------11.12/1.92 Mbps

Conclusion: There seems to be very little to no difference in throughput speeds from browser to browser. With these very slight differences I do not think this will in anyway affect my choice of a new browser. *Strangely K-Meleon's upload speed was consistently slower with bredbandskollen.se, even when I tested on two different occasions, but with speedtest.net it tested equal with the other browsers. I assume that this was a peculiarity with the bredbandskollen servers, and not K-Melon.

___________________________________________

2. Hard Drive Access Frequency Test:

In the previously mentioned thread, the debate arose over the frequency of which browsers read/write to the hard drive while the browser is left sitting idle. Constant access could cause extra wear on the drive and shorten its life span.

Methodology: For my first round of tests I decided to use Window's own built in Performance Monitor. After testing it on three different browsers I realized that the graph display was not sensitive enough to display very short reads and writes and there was no way to differentiate which process was actually accessing the drive, so I abandoned testing with it. Still since I had done the work I decided to post the graphs anyway. (see thumbnail bottom of page) After some research I found a nice little freeware utility to monitor hard drive access frequency, with which I could also tell when the browser was the "guilty party" and not another process. This utility is called "Activity Indicator" and can be found at http://www.eserver.gr/limboonline/ac...indicator.html Before finding this utility, I had wanted to use Sysinternals Process Explore" or "Process Monitor" since you can monitor individual process I/O with it, but I could find no way to discriminate between hard disk read/writes and network traffic. Activity Indicator however worked brilliantly for this test, and it was very enlightening to see everything that was going on with the hard drive. Before performing the tests I parked each Browser on a webpage with no dynamic or updating content, disabled any browser plugins, or live bookmarks, and left it sit idle for several minutes before beginning to monitor the hard disk activity.

Results:
  • Firefox: Writes briefly (about one second) to the hard drive in exactly two minute intervals to update its sqlite journal.
  • SeaMonkey: Same behaviour as Firefox.
  • Opera: Quiet "talkative" for first few minutes, then becomes very quite - very little to zero access.
  • Chrome: After writing in random intervals for the first couple of minutes drops into exact five minute update intervals of its journal. (Although it does seem to update its safe browsing bloom journal quite often as well, which leads to a lot of hard drive access. Plus it seems that Chrome creates an Google update task that runs even when the browser is not running. Somewhere in all of this I would suspect is Google's internet usage tracking process also accessing the hard drive, which combined with the others adds up to a lot of disk reads and writes)
  • SRware Iron: (Based on Chrome or more specifically Chromium, and does not contain Chrome's usage tracking.) It is a lot less talkative that Chrome, exhibiting only the brief five minute updates of the journals.
  • Safari: Seems to drop to no little or no activity after the first run.(On the first run, it seemed to take an extraordinarily long time (40min to an hour??) to update its safe browsing journals with literally thousands of writes and deletes. I don't understand why it had so many deletes in files it had just created since I had just installed it.)
  • Internet Explorer: Almost immediately drops to zero access after loading a web page.
  • K-Meleon: Like IE immediately after loading a website drops to zero access.

Conclusion: IE and K-Meleon are the clear winners here. Browsers with a safe browsing database seem to be the most talkative. Chrome seemed to be the worst offender, its sibling SRware Iron, however, was much better.
___________________________________________

3. Website rendering speed:

I tested using numion.com's stopwatch, and loading the front page of Lifehacker. Then I used Futuremark's browser benchmarking tool at http://service.futuremark.com/peacekeeper/index.action

Methodology: With numion.com I did two sets of tests on each browser. 1. Testing the page load speed 5 times after clearing the cache in between each test. 2. Testing the page reload 5 times without clearing the cache. In this test I also threw out any anomalous readings that were way out of the normal range, and retested. Below are the averages from these tests.

Numion Stopwatch Results:
  • Firefox: ------------------ Clean 4.91 Cached 4.67
  • SeaMonkey: ------------- Clean 3.56 Cached 2.62
  • Opera: ------------------ Clean 6.65 Cached 2.85
  • Chrome: ----------------- Clean 2.56 Cached 2.62
  • SRware Iron: ------------ Clean 2.50 Cached 2.722
  • Safari: ------------------ Clean 6.64 Cached 2.85
  • Internet Explorer: ------- Clean 6.67 Cached 4.61
  • K-Meleon: -------------- Clean 7.80 Cached 6.55

Futuremarks Peacekeeper Results:




Conclusion: SRware Iron & Chrome, come out on top in both tests, followed by Safari and SeaMonkey with varying results between the two testing methods. These tests may be somewhat subjective depending on your processor and graphics card, but my results seem to line up with the tests of others. Something to note about Safari is that it wins out as being one of the fastest browsers when it comes to rendering Java script. You can test browser's ability to render specifically java script here http://www2.webkit.org/perf/sunspide...sunspider.html. Since I am concerned more about overall browsing ability, the results of tests like the ones I made with Stopwatch would weigh more heavily in my choice of a browser that those from Peacekeeper.

Final Conclusions: At this point SRware Iron is looking the most favorable to me, as having the best balance between all the tests. So while I may not bid full farewell to my current browser immediately, if I find the user experience of Iron agreeable it is likely to become my browser of choice.

Ritho
Attached Images
File Type: jpg Hrdskgraph.jpg (32.3 KB, 7 views)
__________________
The smallest good deed is better than the greatest intention.
Ritho is offline   Reply With Quote
Old 20. Feb 2010, 11:58 AM   #2 (permalink)
Super Moderator
 
Anupam's Avatar
 
Join Date: Jul 2008
Location: India
Posts: 15,332
Default

Very interesting tests and results there Ritho . We should feature this as an article on the main site... what do you all think?

Ritho, in the PeaceMaker test, the Firefox version is 3.5.8. So, did you run the tests with 3.5.8, or 3.6?

About the speed tests, the results show, all the browsers have more or less the same speeds... then why is Bengt getting a difference in his speed tests, any clue?

About the hard drive frequency, I also performed that test today with Firefox and K-Meleon... although with not as technicality as you.

First, I started K-Meleon, and just kept it open with a blank page. As you reported, with K-Meleon open, the activity is zero.

With Firefox, every few seconds, the hard disk is accessed, with a blank page open. I then started Firefox in its safe mode, with all the add-ons disabled. Then for long interval of time, Firefox would keep quiet... but still performs activity, after some interval, but for most of time, it remains quiet in that mode. So, I guess, the frequent hard disk activity is because of an add-on. I have Mozilla WeaveSynch, and XMarks... maybe because of them? I have other add-ons too, but I think the above two might be responsible.

About the speed, I find Firefox 3.6 very fast. I have stopped using NoScript because of this .

Well, I am OK with Firefox as my browser of choice, but this hard disk activity is sort of bugging me. Otherwise, I have no problem at all with Firefox, am very very happy with it.

I have few things to say about K-Meleon though, but I will say that in the other thread.

Good test Ritho.
__________________
Anupam
Anupam is offline   Reply With Quote
Old 20. Feb 2010, 12:03 PM   #3 (permalink)
Super Moderator
 
Anupam's Avatar
 
Join Date: Jul 2008
Location: India
Posts: 15,332
Default

Ritho, about Firefox, how are you sure, that in every two minutes, it writes into the sqlite database? Can the hard disk activity be, because of any other reason in Firefox?
Does Firefox behave the same way in its own safe mode?
__________________
Anupam
Anupam is offline   Reply With Quote
Old 20. Feb 2010, 12:11 PM   #4 (permalink)
Senior Member
 
Ritho's Avatar
 
Join Date: Apr 2008
Location: Planet Earth
Posts: 1,379
Default

Quote:
feature this as an article on the main site
I thought about it, but I would need to to a little more work on it I think.

Quote:
did you run the tests with 3.5.8, or 3.6
Shoot! I missed that. I thought 3.5.8 was the latest stable version.

Quote:
With Firefox, every few seconds, the hard disk is accessed
I was not using Firefox in its safe mode, but as I stated I had all of its plugins disabled. There was a point in which its safe browsing database was being updated that there was a lot of activity, but it only seems to access the dive for about a second every couple of minutes.
__________________
The smallest good deed is better than the greatest intention.
Ritho is offline   Reply With Quote
Old 20. Feb 2010, 12:14 PM   #5 (permalink)
Senior Member
 
Ritho's Avatar
 
Join Date: Apr 2008
Location: Planet Earth
Posts: 1,379
Default

Quote:
how are you sure, that in every two minutes, it writes into the sqlite database?
The utility I mentioned, Activity Indicator, records a timestamp, and the exact file that was changed. I.e. C:\Documents and Settings\....\Application Data\Mozilla\Firefox\Profiles\.....\places.sqlite-journal
__________________
The smallest good deed is better than the greatest intention.

Last edited by Ritho; 20. Feb 2010 at 12:16 PM.
Ritho is offline   Reply With Quote
Old 20. Feb 2010, 12:19 PM   #6 (permalink)
Super Moderator
 
Anupam's Avatar
 
Join Date: Jul 2008
Location: India
Posts: 15,332
Default

Quote:
Originally Posted by Ritho View Post
I thought about it, but I would need to to a little more work on it I think.
It would be good to feature this on main site. Its a good test.

Quote:
Originally Posted by Ritho View Post
Shoot! I missed that. I thought 3.5.8 was the latest stable version.
LOL, thought so . Please repeat the test for Firefox 3.6, because FF 3.6 is so much better in terms of speed than its previous versions. Also, uninstall NoScript when you perform the test, because NoScript affects the speed, a lot.


Quote:
Originally Posted by Ritho View Post
I was not using Firefox in its safe mode, but as I stated I had all of its plugins disabled. There was a point in which its safe browsing database was being updated that there was a lot of activity, but it only seems to access the dive for about a second every couple of minutes.
Yes, FF does perform a lot of activity sometimes, and that I think is because of the malicious website database. FF used to perform a lot of activity with the live feed folder. After removing that folder, the activity reduced significantly. But, it still does keep accessing the hard disk. Although the activity is not a lot, like thrashing or something, but still activity is there, even when I had unchecked the option to block malicious websites. Add-ons are responsible for that.
__________________
Anupam
Anupam is offline   Reply With Quote
Old 20. Feb 2010, 12:20 PM   #7 (permalink)
Super Moderator
 
Anupam's Avatar
 
Join Date: Jul 2008
Location: India
Posts: 15,332
Default

Quote:
Originally Posted by Ritho View Post
The utility I mentioned, Activity Indicator, records a timestamp, and the exact file that was changed. I.e. C:\Documents and Settings\....\Application Data\Mozilla\Firefox\Profiles\.....\places.sqlite-journal
Ah, so that's how. Thanks for sharing that Ritho. That software looks good, I will also download it.
__________________
Anupam
Anupam is offline   Reply With Quote
Old 20. Feb 2010, 12:22 PM   #8 (permalink)
Senior Member
 
Ritho's Avatar
 
Join Date: Apr 2008
Location: Planet Earth
Posts: 1,379
Default

Quote:
Also, uninstall NoScript when you perform the test
Does NoScript come pre-installed?
__________________
The smallest good deed is better than the greatest intention.
Ritho is offline   Reply With Quote
Old 20. Feb 2010, 12:23 PM   #9 (permalink)
Super Moderator
 
Anupam's Avatar
 
Join Date: Jul 2008
Location: India
Posts: 15,332
Default

No, no, NoScript does not come pre-installed with FF... I just told, so that in case you have it installed, you disable or uninstall it.
__________________
Anupam
Anupam is offline   Reply With Quote
Old 20. Feb 2010, 05:38 PM   #10 (permalink)
Senior Member
 
deya's Avatar
 
Join Date: Oct 2009
Location: UK
Posts: 1,518
Default

Ritho, interesting stuff you've posted there. You mentioned in your original post that a couple of the tests don't include all the browsers you tested. Could I just ask you if by any chance you tested Maxthon 2.5 ?
deya is offline   Reply With Quote
Reply

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off



All times are GMT +1. The time now is 07:51 PM.


Powered by vBulletin® Version 3.8.7
Copyright ©2000 - 2020, vBulletin Solutions, Inc.