RSSExpect 250% Firefox speed blast after 3.5 RC release
By Scott M. Fulton, III | Published June 8, 2009, 11:15 AM
If last Friday's Release Candidate for Mozilla's Firefox 3.5 is truly indicative of the final release (last week's was not, unfortunately), then how much faster performance will Firefox users expect to see the moment they install it? When the organization first started seriously ramping up the development of its TraceMonkey JavaScript engine last year, we said that speed boost would have to be in the triple-digit range to keep up with competition, as well as to meet the high expectations Mozilla set.
Today, Betanews tests have a preliminary answer, and it's exactly what developers have been looking for: A speed score of 253% that of the Windows 7 RC -- better than two and a half times the speed of version 3.0.10 -- and 222% that of Windows Vista SP2, in tests conducted with the "Beta 99" release candidate build posted last Friday, versus the current stable Firefox release. The general public may get a chance to see that performance improvement later this week, assuming this time Mozilla releases Firefox 3.5 RC to the general public as planned.
This morning's tests give Beta 99 a 9.23 Betanews performance index score in Win7 RC, and 7.44 in Vista SP2, once again demonstrating the much-needed agility in the underlying Win7 platform compared to its predecessor. (Our index compares speeds to a relative 1.00 for Microsoft Internet Explorer 7 in Vista SP2.) In speed tests alone, Beta 99 on Win7 was 32% faster than Beta 99 on Vista.
Where is Firefox finding the speed? It's not in the rendering department, which we've noticed has actually ticked downward, perhaps in pursuit of better standards compliance (the RC still scores 93% on the Acid3 test). Our tests reveal that the latest 3.5 handles array objects twice as fast as the stable, and recovers from errors 10 times as fast. The RC accesses data in memory three times as fast, handles math operations five times as fast, processes textual strings over three times as fast, and processes bitwise logic 753% faster than version 3.0.10.
Still, it will not be the fastest Windows Web browser available. Our performance index score for the latest stable release of Google Chrome (2.0.177.1) in Win7 RC is 13.43, reflecting nearly 56% better speed over Beta 99. But the Firefox speed boost will put version 3.5 in a league with Chrome and Apple Safari 4 (still in testing) in the speed department; and with a fuller slate of add-ons and more developed outside support, there's good reason to believe it will continue to be the preferred alternative browser to Internet Explorer in the coming months.
Of course, the Back button will need to work. During what was supposed to have been Mozilla's final "crunch time" testing two weeks ago, testers found a handful of blockers that were bad enough to delay final release by at least a week. We saw one ourselves: In one of the previous daily builds for Firefox 3.5 last week, the Back button was inoperative in Windows 7 RC. Not in Vista, however; and the latest daily "Minefield" build of Firefox 3.6 didn't have that problem. We were relieved not to find it in last weekend's Beta 99 post.
Personally, and I have used Firefox since 'almost' day one,
to me, it's all about the addons......
I have 25(?) addons that I consider important/necessary/I'm glad
it's there when I want to use it and half of them won't work in 3.0+
So, 'I' don't really care if this new version displays the page 'if I think
about clicking on a link', I'll take function over speed any day and
my speed is just fine with 2.0.0.11 (2.0.0.06 on my other PCs)
Score: 0
|Well, remember that bug fixes usually take away that extra speed and what usually remains is good but hardly a blast of speed. I'd love to see it all work out much faster than expected but reality usually comes into play.
Score: 0
|aren't more people still using XP? can't we hear how it does on that?
Score: 2
|No, not really. :)
Score: -3
|I agree, in the real world, a very large number of people have skipped Vista and are now waiting for the final verdict on Windows 7. The majority of people I speak to are still using XP including me.
As for IE8 in business at least the majority of companies I deal with are still running IE6 on all their PC's( at least those that have finally moved from IE5 !! ).
Personally I use Firefox, one of the main reasons that I see sites load faster in Firefox than IEx is because I use AdBlock and FlashBlock to stop all the annoying adverts!!
Score: 0
|Scott,
once again your percentages confuse me. If you are comparing 3.0.10 with the RC as the basis for the 250% "blast", that blast would only be 112%. Remember, in percentages, when you increase something by 100%, it means you have doubled whatever it is you have been measuring. In this case, 9.23 is just about double 4.36... therefore, it can't be an increase of 250%. If that where the case, then the release candidate should be performing at a aproximately the 15.25 range.
In comparisons, your baseline is always 100% and you increase or decrease from there. When you markup prices, you don't say a s*** that used to cost 10 dollars and now costs 15 had a markup of 150%. The markup was 50%. This speed comparison is no different. If 4.36 is your baseline, then when doing your percentages, you have to decrease that amount by a 100% to take the initial baseline in consideration.
Score: 1
|You're right, soundchazer, that was a little confusing, so I tweaked my language.
-SF3
Score: 0
|i wonder... can't this test start to be measured against IE8?
IE7 is really far away from the spotlight now, and at this point it is well known and assumed that IE8 is the slowest of the main browsers, so no more IE7 please (just reckon it's 60% slower and were done), and let's have measurements made on IE8 as a basis.
The whole point is that it would be much more precise and clearly informative, putting the browsers in perspective without so much fluff of inflated numbers.
You know what i mean.
Score: 0
|Well, the thing about a sensitive index score is, you need a slow browser to measure against. IE8 isn't really as slow, so if we were to use IE8 as the index, the granularity of the scores would decrease by half. Furthermore, using IE7 gives us the opportunity to test IE8 improvements when they come around. Otherwise, we'd have to reset everybody else's scores with slower indexes whenever Microsoft posts a big patch or a service pack.
-SF3
Score: 0
|