Q3 benchmarking roundup shows Pale Moon strong

[Moonchild]

An interesting benchmarking roundup shows Pale Moon 12.3 to be a strong player, even if it is currently lacking some performance enhancements that made it into the Firefox 13 and 14 code base: http://www.ngohq.com/news/22169-web-browser-performance-roundup-for-q3-2012-a.html

[stravinsky]

i saw that page a few days ago. Didnt even bother reading the contenders. Just went to the conclusion which said "OMGWTF CHROME IS THE BEST EVER!!!1! "
read that, forgot it a few minutes later and continued using PM. These type of roundups are a dime a dozen. You can find a roundup on every site with "tech" in its name.

The only one worth reading IMO is the "tomshardware Web Browser Grand Prix" . That has many many more benchmarks that do tend to average out the weaknesses and strengths and show a better overall picture of a browser. Too bad they wont admit PM as one of the contenders, sticking to vanilla FF.

Just bring on teh PM15

Edit : most of these 'roundups' show chrome as the winner.
is Chrome more popular because these large number of 'tech' sites show chrome as winner or these sites show chrome as winner because its more popular (and hence will attract more pageviews, + google secretly ranks these pages higher in its rankings ) ?

[Moonchild]

There is no conspiracy
If you don't believe the scores you can benchmark the browsers yourself.

That being said, however:
The massive leads Chrome scores on some tests don't make me think it is that much faster, but rather that there is something very specific done in the Chrome browser that offsets the scores (one faulty result can pull the entire score table up). Remember when I talked about "using javascript to measure itself" and how it is not scientifically sound to do so? That will be exactly the issue here.
People shouting "Chrome cheats" are, in a way, right - but it's not necessarily on purpose. And they go by comparing the benchmark scores against their real-world experience comparing the different browsers where Chrome isn't as fast as the benchmarks say it should be.

I know the Firefox code well, I know that some things in javascript are really implemented the fastest way possible (nanojit is an awesome piece of code) even going down to directly plugging in in-line assembler code for time-critical paths and not leaving it up to the compiler to generate code for it. That in mind, you simply can't get a higher score for some things unless you are measuring something wrong. A processor simply needs x number of cycles to perform certain instructions like math, register loads, memory operations, etc. No amount of coding can bypass that unless you are in the Twilight Zone or one of King's novels.

As an aside: As a practical example of benchmarking not being the end-all: Pale Moon 12.3 and 12.3r2 scored equal in benchmarking. Yet r2 is notably smoother in operation. Though not the end-all, benchmarks are still an indication if you look at overall global results and filtering out odd spikes. I'm therefore not dismissing them at all, but don't take the scored results without a grain of salt when comparing browsers based on different engines.