Comment by andai | Hacker News Reader

Comment by andai | Hacker News Reader

Parent Full thread

Latency. Just tested with a "hello world" page (18ms ping).

Chrome takes 102-200ms to render the page, Firefox takes 150-400ms. 1.5 to 2x slower -- and that's after all the performance upgrades they've done with Rust and Quantum.

Network inspector says Firefox is spending 150-270ms "blocked" before it even initiates the connection. On Chrome the block is 3-18ms (15-50x (!!) faster).

There's also UI lag. When I press Ctrl+N, I expect a new window to open immediately. Chrome and Chromium based browsers do that. I cannot even perceive the delay between pressing Ctrl+N and a new window appearing. With Firefox it takes about 0.5-1 seconds before anything happens at all. This seems to be another 10-20x difference (though I'd have to record my screen and measure the difference).

Curiously, Re: UI lag, I think the actual time until the navbar is usable is about the same. But Firefox waits until the whole window is built before it displays it, while Chrome opens the window immediately and fills it in. From a user experience (I do command, I expect feedback) this makes a huge difference.

I switch back to Firefox roughly once a year after reading how fast it's gotten. But I have been disappointed on this front every single time. I want Firefox to succeed, I really do. I am donating to Mozilla every month. But as it is right now, I can't tolerate such lag, for a trivial action I perform thousands of times a day.

> Chrome takes 102-200ms to render the page, Firefox takes 150-400ms. 1.5 to 2x slower -- and that's after all the performance upgrades they've done with Rust and Quantum.

How did you actually measure this? I wonder whether these small timing differences may just be differences in how the respective dev tools report the timing info.

FWIW I also don't think the relative difference will be this large for bigger sites - I'd be very surprised if a site that takes 1s to load in Chrome would take 1.5-2s in Firefox.

There was a noticeable increase in GMail loading time (when I switched to FF a while back), but once it actually loaded it was the same speed (slow af on both lol). As an aside, it's gotten even more bloated in the meantime (for some Google PM's promotion), I just use Basic HTML now. Loads instantly :)

So once the page is loaded, JS performance seems to be about on par, but page loads are faster in Chrome.

To measure I just used dev tools. But there is quite a lot of randomness involved -- sometimes the engine scheduling planets align, and sometimes they are way out of alignment. You'd need thousands of samples or more to get an accurate reading. I just refreshed a few times to get ballpark numbers.

I am using Firefox now for almost 7 years as my main browser (Chrome still installed for some pages not properly working in other Browser). But the number you are mentioning here are just not true with current versions of both browsers. Overall I agree that FF still has potential to improve it’s rendering performance (like every other browser). But as an end user it’s not noticeable anymore. Quite the opposite is the case for me since FF always used way less resources than Chrome on my notebook.

This is always the answer for me. I stayed on Firefox long after most of my friends had switched to Chrome back when it came out, but it was eventually impossible to deny the difference. Every time I've tried to switch back over the years, the difference in latency is still noticeable.

I'm using Firefox and getting navigation times on par with Chromium when loading simple documents (tested with example.com).

From pressing enter to the browser initiating a fetch was around 16ms.

This is useful feedback for building non browser uis. Instant feedback followed by a progressive loading rather than build it all then render.

Also performance issues, especially on Linux. Fan works in ~50% of my laptop usage.

Reply on news.ycombinator.com