zlacker

Is this one of those synthetic benchmarks that don't mean much?

>>veave+(OP)
Yes. But being fast, even at a synthetic benchmark, is rarely worse than being slow.

>>veave+(OP)
They sure do seem to mean a lot when someone recommends Firefox and people shit on them.

>>veave+(OP)
Speedometer is one of the primary benchmarks used by the Chromium team as a proxy for real-world use of popular JavaScript frameworks: https://v8.dev/blog/speedometer-2

>>veave+(OP)
It looks to measure a bunch of little JavaScript-driven interactions. It has been around for a while, so I guess it probably measures what it says it measures.

OTOH, you might feel it doesn’t mean much depending on your priorities. I mean, JavaScript is bad anyway. So I guess anyone they cares about performance/security blocks it as much as possible and/or avoids JavaScript heavy sites.

Or, big picture, Firefox wasn’t built by a spying company, so even if Chrome had much better benchmarks switching to Chrome would be a silly thing to do.

>>veave+(OP)
It's using a fake "app", but it measures UI updates across multiple real JavaScript frameworks. In other words, probably not a good benchmark for things that are compute-bound or I/O-bound, but probably a pretty good benchmark for "snappiness".

>>veave+(OP)
Speedometer executes a bunch of TodoMVC projects, so it's a mix of JS and DOM work. FF has done a lot of cool work on their DOM and have started to focus some on making JS faster, so they are pretty fast here.

https://browserbench.org/Speedometer2.0/

If you are interested in just JS, JetStream 2 is a good metric IMO, but is perhaps less "real world" as it doesn't really do DOM stuff. FF is slower here, but has made a lot of progress recently. (going from somewhere around 35% slower to around 10% slower).

https://browserbench.org/JetStream2.0/in-depth.html

>>hajile+oc
Safari is like 30% faster than Chrome on JetStream. I have no idea why people think V8 is the fastest engine around.