Speedometer browser benchmark: show me yours and I'll show you mine


That’s interesting. Wumpus’ PC is clocked about 23% higher than a 4770K (max 3.9 GHz), and scored about 18% higher on the benchmark. Which OS are you using Scott?


Windows 10 (forgot to mention).


My older i7 went from ~96 to 108, and I noticed the error bounds are much less without extensions (with ±15, without ±2)


I got 78.7

So what is this test? Looks to me at first glance that it’s running exercising Selenium on a bunch of different versions of the sample ToDo app, implemented a bunch of different ways (vanilla JavaScript, EmberJS, BackboneJS, ReactJS, Angular… others too I’m sure).

I’m not convinced this is much of a benchmark.

If you open up the developer tools, you’ll also see that it’s blowing chunks throughout the entire test:


I just ran it on my windows 10 system using chrome 59. I have an i5 6600k running at 4.5k, 16 gigs of ram. I scored 138.4.


Around 130 at work on Chrome 58, 112 on Chrome 59. i7 4790@3.6Ghz on the latest Windows 10 build. So that’s super odd.


83.7 on iPhone SE (64 mb) in Safari. The browser always seemed snappy enough…


Wasn’t there a different thread? I could have sworn I posted my score in that one. Wow, you guys can really run Qt3 much faster than me. I’m really slow at Qt3. Luckily games still run fine on my first generation i5 from 2009.


Uh, how long is this benchmark supposed to run? I let it thrash the same to-do list for a couple minutes, gave up, then had to pound the back button about 17 times to get back here.


I’m reasonably sure the numbers you get from this benchmark mean fuck-all re: actual PC gaming, so I wouldn’t put much stock in whatever numbers you get (which also seem to have quite a lot of external factors from Chrome version, to whether or not you just restarted your PC before running the bench).


Fixed that for you. :)


Actually, I’m pretty sure it’s probably a good indicator for Qt3 browsing speeds. I’m sure Discourse probably loads a certain number of milliseconds faster if I had a faster CPU and therefore did better on this test. After all, the new Qt3 discourse site does run slower for me than the vBulletin site we used to have. I’m sure that wouldn’t be the case if I had a faster CPU, surely?

As for games, yep, I know. Games haven’t been CPU limited in a long time.


I have this, not overclocked, and got 183. Close enough.


Interesting. I got a 48 on my 64GB iPhone 6S, which shares the same A9 processor as your SE.


It’s of niche usefulness. It’s basically just about OS/browser specific tuning of a few functions. For instance I have a dual Xeon E5-2690 workstation and just got a 31.2 in my main linux install and 79.4 in a windows VM that actually has a CPU max clip set on it. But he’ll keep posting it over and over again.

It’s funny because @wumpus keeps quoting Benedikt Meurer but he somehow keeps not reading his articles to the end:

One closing comment: Don’t use traditional JavaScript benchmarks to compare phones. It’s really the most useless thing you can do, as the JavaScript performance often depends a lot on the software and not necessarily on the hardware, and Chrome ships a new version every six weeks, so whatever you measure in March maybe irrelevant already in April. And if there’s no way to avoid running something in a browser that assigns a number to a phone, then at least use a recent full browser benchmark that has at least something to do with what people will do with their browsers, i.e. consider Speedometer benchmark.

It’s really the most useless thing you can do


There are political reasons for him to say that, because Qualcomm is so far behind on CPU speed, and the whole Android ecosystem is completely dependent on Qualcomm. As you can see from our Intel results, this scales directly with CPU speed. As long as the browser is the same version across devices, you will see consistent results.


I have 58.0.3029.110 on both my linux install and my windows VM. One was literally more than twice the other. It’s not even across devices, it’s the same device. And the faster one is the one with the execution cap of 90% set in the VM.


58 isn’t the correct (latest) version. You should have

Version 59.0.3071.86 (Official Build) (64-bit)


What does that matter? If I have the same version across the same device it is supposed to be consistent. It clearly isn’t.

It’s really the most useless thing you can do


I dunno but you are fucking this up, somehow – 58 is not the current version.