- Mar 24, 2011
- 15,781
- 1,812
A growing number of users around the globe use their Androids not only to call and text but also to connect to the web. Among those who do take their mobile web browsing seriously, the choice of browser app is very crucial.
What makes users prefer one browser to another? Several subjective considerations include feature set, intended use, and esthetics. Another crucial factor is speed and performance. In this post, we will be discussing this particular factor.
We subjected 11 of the best Android browsers to a series of tests and benchmarks to find out which one is the top performer. Which browser is the fastest Android browser? Continue reading to find out.
Aims, method, and limitations
Our aim for conducting these test was to provide extensive (but not necessarily exhaustive) objective and measurable data helpful in deciding which Android browser to favor for daily use.
We tested the following specific browser versions:
- Baidu Browser 3.1.0.2
- Boat Browser 7.1
- Google Chrome 32.0.1700.99
- Dolphin Browser 10.2.3
- Mozilla Firefox 26.0.1
- Maxthon Browser 4.1.5.2000 build 2860
- Naked Browser Pro 1.0 build 25
- Next Browser 1.16
- Opera Browser 18.0.1290.67495
- Puffin Web Browser Free 3.1.10679
- UC Browser for Android 9.5
Test device used
The browsers were tested using a Google Nexus 4, a Google Experience Device designed and optimized to let you experience Android as Google intended it. The operating system running on the test device at the time of testing was stock Google Android 4.4.2 KitKat.
The Nexus 4 was factory reset before use for testing. No other app was installed on the phone, apart from updates to the pre-installed apps, the browser apps to be tested, and a system monitor app for determining memory use. We wanted to limit the possibility of the results being influenced by stray, leftover, or inessential processes.
Reboots
Before each test, the Nexus 4 was rebooted to ensure that the tests ran on a clean slate. Browsing data, history, cookies, and other browser app data were also cleared before each test was run, except in the cached page loading test. We wanted to minimize the possible influence of other apps or data upon the test results.
Also, to improve data accuracy, we ran each test or benchmark three times and computed the arithmetic mean or simple average of the three recorded results.
Areas tested
Our series of tests covered these general areas:
- JavaScript and overall browser performance
- Page loading speed
- Memory consumption
To test JavaScript and overall browser performance, we used the following popular and well-known benchmark suites:
- Sunspider 1.0.2
- Mozilla Kraken 1.1
- Browsermark 2
- Peacekeeper
SunSpider 1.0.2
We first conducted the SunSpider 1.0.2 JavaScript Benchmark, designed to test the core JavaScript abilities of the browser. Measured in milliseconds, a low score means faster/better performance.
Next Browser and Baidu took third and fourth places, respectively. Naked Browser Pro, Boat, Opera, Maxthon, and Firefox stayed within the vicinity of 1,300 ms to 1,500 ms, a range that was still a far cry from Puffins superfast performance.
The two slowest browsers were Dolphin Browser (1,771.700 ms), and UC Browser (2,127.133 ms).
Mozilla Kraken 1.1
After SunSpider, we ran Mozilla Kraken JavaScript Benchmark 1.1 , which, like SunSpider, measures the browsers JavaScript performance through the use of different test cases extracted from real-world applications and libraries. The benchmark is developed by the Mozilla Foundation.
As in SunSpider, the score is in milliseconds; smaller score means better performance.
The in-between apps scores stayed within the 14,000 ms to 21,000 ms range, but their rankings changed a bit in this test. Chrome, for instance, performed more slowly than Firefox in the Kraken suite, although in SunSpider, it performed better than Firefox.
So far, for JavaScript performance, Puffin Web Browser consistently sits as king of the hill, while both Dolphin and UC Browser stay consistently at the bottom with their poor performance in the tests.
Browsermark
The third benchmark, Browsermark 2.0, measures overall browser performance. Said to be better than its earlier version, Browsermark 2.0 checks performance in these areas: general (e.g., page load time, screen resizing, etc.), JavaScript, CSS, DOM, and graphics (particularly WebGL and Canvas performance).
The race this time is not about who gets the lowest score but about who gets the higher score.
Peacekeeper
Next, we ran the Peacekeeper benchmark suite. This well-known and frequently used browser benchmark measures overall performance by testing the browsers ability to handle commonly used JavaScript functions. Its individual tests include those for rendering, HTML5 (WebGL, video, web worker, game, Canvas), data arrays, DOM operations, and text parsing.
The averaged scores are shown below. Higher score means better performance.
While Dolphin (355.333) moved up the ranks a bit besting Firefox, actually UC Browser stayed consistently the lowest (202.333).
Puffin disappeared in this round. Peacekeeper just wouldnt finish running in Puffin, despite several attempts. It ran until the web worker test, and from there, it would stop responding. We have inquired about this from the developers of Puffin, but we received no reply at all.
Puffin performed superbly in SunSpider, Kraken, and Browsermark, but it wasnt able to prove its worth in the Peacekeeper suite, giving away the top spot to Boat Browser.
So far, Puffin has bested the other browsers in all of the synthetic tests that weve run (except Peacekeeper) and has proven its promising potential for great performance. But, practical tests such as page loading and memory usage tests can also shed light on a browsers power. The next sections talk about such practical tests.
Page loading speed
To avoid the potential negative effect of network lag or flaky connections for our page loading speed, we set up a simple mock Android Authority website hosted on a local network server.
Cold Loading
In this test set we measured how fast a browser loaded our test websites homepage completely for the first time (i.e., uncached loading or cold loading). The browser cache and browsing data were cleared first, then the Nexus 4 was rebooted before opening the locally hosted website in the browser.
Google Chrome reigned with its fast average page load time of 2.550 seconds. Naked Browser Pro followed closely behind (2.584 seconds), while Opera came in at third place (2.822 seconds).
The other browsers registered average page loading times greater than 3 seconds. Dolphin finished the race last and was the slowest (6.317 seconds).
Again, we couldnt include Puffin in this set because it wouldnt load our locally hosted website. Instead, we got a connection refused message each time we attempted open the page. Perhaps this behavior had something to do with Puffins use of cloud servers for pre-processing and data compression the very same technology that made it lord over the other browsers in the other benchmarks and tests that we conducted. We reached out to the developers of Puffin in order to elicit a comment or clarification. Until today, all we got from them was deafening silence.
Hot Loading
Next, we performed the hot page loading test to see which browser hot loads our test webpage fastest. Hot loading or cached page loading usually runs faster than uncached loading. This happens because the some of the webpages elements have already been stored or cached in the browser, so they dont need to be re-downloaded anymore.
In the hot page loading test, we first opened our test webpage in the browser, then exited or force-stopped the browser without clearing its browsing data. Then, the app was run again, the test page was reopened, and the loading time was recorded. The steps were performed thrice per browser. The chart below shows the average load time per browser.
Dolphin (2.456 seconds) was the slowest in the cold loading test, but Firefox took its place in this round (3.343 seconds).
Hot loading scores for Naked, Chrome, Maxthon, Baidu, and Next all went above 2 seconds, while those for Boat, UC, Opera, and Dolphin stayed above 2 seconds but below 3. Only Firefox earned a hot loading score above 3 seconds.
Puffin has no score for this test for the same reason as already explained in the section on cold page loading.
Differences in cold and hot loading times
The chart below summarizes the cold and hot page loading scores for all browser apps (except Puffin), as well as the differences in each browsers scores.
Although it placed last in the cold loading test and second to the last in the hot loading test, Dolphin registered the biggest drop between its cold loading and hot loading times at 3.861 seconds. This could possibly mean that Dolphin made the most significant and advantageous use of its caching mechanism among the other browsers in this group. And, because of this, subsequent reloads of a previously loaded page in Dolphin will be perceptibly much faster than its first-time load.
Memory consumption
Memory consumption is another crucial factor in browser selection, especially among users of low-range and mid-range Android devices having limited memory. Our last set of tests measured how much memory was used by each browser.
No open page
We first measured how much memory was consumed by each browser running without any open tab or page.
To ensure result accuracy, we cleared the browser cache first, then rebooted the Nexus 4 before launching the browser. If the browser had popup dialogs, tours, or getting started prompts, we first dismissed those. Then, we terminated the browser and relaunched it. Only then did we take a reading of the browsers memory use. Three readings were taken for each browser, the averages of which are shown in the graph below.
Except Firefoxs, memory usage of all the test browsers fell below the 100-megabyte mark. These can be considered relatively lightweight, with Next, Naked, Baidu, and Chrome as the most lightweight (lowest usage range of 50 to 60 MB) in the group.
5 open tabs
Which of our 11 browsers uses the least memory even when multiple pages/tabs are open? This was what we wanted to find out in the next set of tests.
The procedure for this set is similar to that for the zero-tab test earlier. First, browser app data were cleared, the device was rebooted, browser startup dialogs were dismissed, and the browser was terminated and relaunched. Then, 5 real web pages were opened one by one in the browser while observing a 7-second delay between tab openings. After the last page has loaded, the memory use of the app at that point was recorded. Three readings were taken for each app and then averaged.
From top position in the previous test, Next Browser slid down to fifth place in this test. Taking its place was Naked Browser Pro (117.233 MB), which ranked second in the zero-tab test.
Conclusion
After conducting all the tests for this post, we affirmed a notion that weve always held to be true: the fastest Android browser doesnt exist in an absolute or universal sense.
Popular, solid, and stable browsers such as Firefox, Chrome, Dolphin, Opera, and UC registered variable performance scores in the different tests. They werent the fastest performers in our tests. They werent even the most lightweight in the bunch, although Chrome did show memory usage that stayed within the lightweight half of the list.
Rising stars
Maxthon, Next, Baidu, and Boat are rising stars in the browser market. Though still building up their user base, some of these browsers actually fared better than some of the highly popular ones.
Next Browser, for example, bested the others in the SunSpider test, although it showed average performance in most of the other tests. The same can be said of Baidu and Boat.
Maxthon, meanwhile, flexed its muscle in some of the other tests. For instance, it placed second in the 5-page memory usage test, outrunning even Chrome and Puffin. It also fared well in the Kraken and Peacekeeper tests.
Promising potential
Puffin caught our interest as we went through the rounds of testing. It effortlessly registered speeds and scores that left the other test browsers eating dust. It registered the fastest performance in synthetic tests such as SunSpider, Kraken, and Browsermark.
But, we were annoyed because it refused to play ball with Peacekeeper and refused to open locally hosted web pages. Those made our data incomplete. Worse than all these is that the developers never replied to our emailed inquiries.
Naked Browser Pro also caught our eye. Before working on this post, weve only heard of it twice from an obscure acquaintance, and then completely forgot about it until we saw its potential with our own eyes. It landed on top spot in two tests (hot page loading and 5-page memory usage tests) and ranked second place in three (Peacekeeper, cold page loading, and 0-page memory usage tests).
However, it stayed midway in the rankings for SunSpider, Kraken, and Browsermark. Yet, despite the browsers mediocre results in these last three tests, Naked Browser Pro seems to be worth considering for use as primary browser, especially among those who dont really need a memory-intensive browser for accessing online and locally hosted content not requiring very complex JavaScript functions.
The day of the Naked Puffin
Based purely on measurements and scores provided by the tests that we used, it appears that Puffin and Naked Browser Pro have earned enough proof for us to recommend them with caveats, of course to those who are looking for faster browsers.
But, as has already been noted earlier, browser speed is just one of several considerations when deciding which browser to stick with for a long time. In this article, we provided you with objective and extensive although not exhaustive data to help you with your decision making.
(with contributions from Alvin Ybañez)
