Inquisitive Insomniacs, Part 2: Browser Benchmarks
There are times when I cannot sleep. In those days, I like to do things which sate my curiosity without being excessively unproductive. The last time I was this sleepy, I decided to go ahead with making some themes. This time, I decided to finally figure out which browser was the most suitable for my needs.
The Setup
As regular readers might be aware I am a PhD student, and that involves a fair amount of reading. Long hours, which may require me to be away from a charging port, mean that I need to have a device which has a good battery life. Frequent coding means I need to have a good dev environment which biases me towards *NIX based OSes. All in all, this is to say that I went ahead and got a Macbook Pro 14inch for myself (the M2 Pro kind). But you said you needed a big screen so why purchase the 14 and not the 16? Well, because I am a student and money matters. At the time of purchase, the 16 was about $500 costlier than the 14… and that’s a lot.
To solve my delimma with the reading situation, I decided to purchase the Pixio PX277 Pro which is a decent 27inch monitor that does not break the bank. Sure, QHD is not the most ideal but all 4K monitors with features this one offered like KVM, a built in USB dock, and a 165Hz refresh rate cost roughly 1.5x this one. And I am a student, you know.
So that’s the setup: A Macbook Pro 14inch (M2 Pro) connected to a 27inch QHD external monitor in clamshell mode.
The Benchmarks
The latest browser benchmarks I could find for a Mac with Apple Silicon were done about a year ago and I figured it wasn’t the worst thing to do my own testing for things I use my device for. Three benchmarks were tried, each with multiple runs and time between runs to reduce variance and get a more accurate measure:
Each test was run on a single tab in a new private window with no other applications running and with no extensions enabled to provide as repeatable and unbiased a result as can be. Five web browsers were tested: Chrome, Edge, Brave, Safari, and Firefox. Yes, I know that the first three browsers are based on Chromium and should show similar results but even so there are differences in what exactly each of them include from Chromium.
The Results
The averaged scores (rounded to the nearest integer because it doesn’t really matter) for the three benchmarks are as follows
This is all very interesting huh. There are some clear things to note:
- Safari is far and away the best on the tests run by Speedometer. This should cover a lot of daily use cases which do not involve web apps.
As long as you are on a regular website which uses a fair amount of CSS and JS but does not really go full on heavy Web App (such as Spotify), Safari is by far the most performant.
2. When it coems to JS and WebAssembly heavy workloads, tested by JetStream, Firefox gets totally obliterated and Chromium browsers reign supreme. Safari isn’t too far behind though :)
3. For graphical on screen performance, Safari and Chromium browsers are almost at par with a narrow win for Chromium browsers. Firefox, once again, gets utterly obliterated
Firefox is a fair choice in 2024 only if your use case involves basic websites. For anything requiring intense computations or graphic display, Firefox is miles behind the competition.
The Conclusion
For funsies, I tried making two meta-scoring systems to evaluate these results. These systems have different results but share the same logos: consider each benchmark as equally important. This may not be true for you so ymmv.
Method 1: For each test, rank individually and score as 6-Rank
This means that we make 3 separate rank lists, one for each benchmark, and give points inverse to your rank, i.e. the highest scorer gets 5 points and the lowest scores gets 1 point. With this system, our rank list looks like this and Chrome takes the cake with Safari and Edge sharing a close second place.
Method 2: For each test, scale points to 10 and add
This means that we convert the number values provided by each benchmark and normalize individually for each test such that the highest score is considered a 10. In this system, a maximum score of 30 can thus be attained and the idea behind this is that ranks are too sharp of a cutoff while actual benchmark numbers are much much closer so this method should provide a more reasonable scoring. The score table is similar, but not quite the same, as
Overall, my conclusion for this is that web browsers are amazing in 2024 and if you have trouble choosing one for yourself, maybe keep the following in mind
If privacy and slightly better battery life are not super important, Chrome and Edge are the browsers of choice with their superior performance, extension support, features and being consistently close to, if not at, the top.
If you use your laptop as a laptop, switch to Safari to get the maximum battery life.
If you don’t mind being slightly subpar and having more crypto for the superior built-in adblock, paywall bypass and privacy, Brave is your best bet.
If you hate yourself, use Firefox.
