Chrome 80 Is 30% Slower Than Chrome 75
github.com
github.com
V8 could choose to look at both microbenchmark performance and real web performance, as both are useful signals for how V8 performs in general.
The only reason to not use microbenchmarks is because you think they will somehow lead you astray and cause you to optimize things such that user experience actually gets worse. But, frankly, I find that incredibly unlikely, and I'd love to see actual examples of such a thing.
Otherwise, this just seems like, and now I'm going to return to that idiom: throwing the proverbial baby out with the bathwater, here, simply because someone somewhere decided that microbenchmarks are universally "bad" while website performance tests are universally "good".
Of course, this presupposes that this is, in fact, what the V8 team is doing.
That's not the only reason. It's about putting the effort where it's matters. Sure gaining 30% speed in synthetic benchmark seems great, but if in real world scenario theses things doesn't happen that often (and spoiler alert, they don't), it will means actually much less than 30% of the speed gains. That's not a good metric to go toward.
It's simply about investing the development time that you got, where it matters for the user.
If your usage is closer to synthetic benchmark and thus you lost close to that 30% of performance since Chrome 75, then take your real world scenario and send that to the V8 teams, I'm sure they'll be happy to consider it in the test suite.
I doubt it would, great pattern aren't just good for humans, they are also much more predictable, and that's quite good for JIT compiler.
If it does, I'll just refer to my last sentence which still apply:
> If your usage is closer to synthetic benchmark and thus you lost close to that 30% of performance since Chrome 75, then take your real world scenario and send that to the V8 teams, I'm sure they'll be happy to consider it in the test suite.
Progress is only counted towards the real world sites, but if you make performance markedly worse on the regression test, then that's a bug.
Not at all, it's about not investing time on what doesn't matter. For sure it will means you spend less time on something else and that will slowly get worse.
That's why feature get deprecated and removed. Supporting something not used has a cost.
I'll repeat my last paragraph:
> If your usage is closer to synthetic benchmark and thus you lost close to that 30% of performance since Chrome 75, then take your real world scenario and send that to the V8 teams, I'm sure they'll be happy to consider it in the test suite.
If you do need that speed, show them.
For example, if you have a caching scheme that was designed to make the microbenchmark better, it could very easily make real world performance worse. Like if the microbenchmark used 100 sized X blocks, so you pre allocated those on start-up, great microbenchmark speed up. But in the real world that hurts start-up time, and the cache is probably not fully utilized.
A lot of times it's good to use microbenchmarks as a signpost, to indicate if some performance tuning causes a massive degradation that is unexpected.
[1] https://bugs.chromium.org/p/chromium/issues/detail?id=104449...
I think the outlier cases are interesting. The Gamespot test[2] loads 2x slower and with 2x page size using default uBlock. Strange glass-card websites these days.
Devtools are great, performance is great, battery life on MacBook 16in is great, Nightly is awesome (has its ups and downs).
I honestly only ever use Chrome to double-check if my CSS styling is cross-browser. Although I have been using Edge a lot more for this use case, because I can't stand Google.
I did this myself using userChrome.css :-)
THANK YOU.
I had no idea this existed, I will be looking into this.
You have to enable it manually nowadays, but it opens up a ton of possibilities - you can theme virtually everything (find out which elements to theme using the browser toolbox).
Found the link for anyone else interested: https://www.userchrome.org/firefox-changes-userchrome-css.ht...
That's a powerful machine most users don't have, you are less likely to notice the difference than most.
Hah, part of my happiness in my return to using Firefox more as my daily driver lately has been the square tabs. Rounded tabs and tabs that shrink way too much turned out to be my big reasons I disliked the Edge to Edgmium "upgrade".
I use them in Firefox/Chrome to signal that they're tabs I definitely want to always keep open, and which I expect to open a lot of other tabs from. It's a whole different use case from wanting them available on every window.
Like GMail is awful[https://medium.com/@boriscoder/peeking-under-the-hood-of-red...] and chrome is on its way.
My best guess to understand it is:
- On the horizontal axis, we have consecutive Chrome releases.
- On the vertical axis, we have total time of benchmark execution, relatively to Chrome 75
The source of the data is the "js-framework-benchmark"[0] run in a controlled environment in Chrome[1]
[0]: https://github.com/krausest/js-framework-benchmark
[1]: https://krausest.github.io/js-framework-benchmark/2020/table...
E.g. this https://www.blackrock.com/uk/intermediaries/products/230286/...
Doesn't start from zero, but is very clear that it doesn't start from unity either.
This type of chart is extremely common in benchmarks.
And how are finance numbers, in what I linked to, static?
I have no problem with normalizing the number. That seems like the obvious best way. That's why they also do it in the finance graph I linked to.
What I linked to doesn't show the fund price, it shows the value normalized. Just like benchmark values usually are. It's not the normalization I (and others) am saying is bad, it's the graph.
But yes, stocks are usually shown graphed with their ticker value. Funds are usually not.
Alternatively, you can go to a log plot, where constant factors produce constant shifts. But even then, one would typically not put the x-axis origin at 1 on the y-axis.
> Bonus mess points for choosing 1 instead of 0 as the starting point on the vertical axis...
What purpose would a block of whitespace at the bottom serve?
Charts that I'm used to have units like FPS or MS on an axis that starts at 0.
If you were on a breakpoint and you reloaded the page, the code execution was resumed. Didn't realize it did that until I saw network traffic that was generated by code after the breakpoint.
to call these microbenchmarks (especially when viewed collectively) for the purpose of dismissing the results is dishonest at best.
each datapoint on the graph represents a lot of internal metrics (GC, JS, DOM, layout/css). these aren't some one-liner JS loops. when you have a diverse set of "microbenchmarks", and they all show the same trend, you can be pretty sure the results are not "contrived".
keep in mind that canary builds might not be compiled with all the typical release optimization flags turned on.
So I started trying out some new browsers recently, and really like Brave for the out-of-the-box privacy focus. The crypto stuff is novel and I'd be interested to see where that goes.
Also tried Vivaldi and absolutely love it. Reminds me of the old Opera (absolute favorite), it's got configurations for everything. Vertical tabs are a built-in!
I use Firefox from time to time too, but find the interface less polished (at least on MacOS).
My OpenLayers mapping application suddenly got VERY slow. tl;dr: Chromium team had made an optimization change to canvas `getImageData` which slowed it to ~1/40th performance in some cases. It was frustrating to have to convince them this is a problem at all.
But what frustrated me more was how they came to decide to sacrifice performance in one way to benefit it in another. I'm sure they came to the conclusion somehow, but it ended up just utterly breaking apps that are already out there. One teacher struggled with a school-board-wide course that just didn't function because the main software was broken by this.
https://bugs.chromium.org/p/chromium/issues/detail?id=100184...
Two other scenarios that will often change in performance is canvas-to-canvas draws and canvas<->WebGL interop, because how those are implemented is complex. In the past particular behaviors would instantly disable hardware acceleration for a given canvas, tanking your performance.
My money would be on that it improved performance on something google related: maps, youtube, etc
That said, I still prefer Firefox because Firefox isn't Blink.
You probably believe in curtailing certain personal freedoms for other reasons and you are morally justified to do so. Again, nothing personal.
> You probably believe in curtailing certain personal freedoms for other reasons and you are morally justified to do so. Again, nothing personal.
The key difference is that I don't want to limit personal freedoms for a specific class of people. That's when it does get personal.
That's not "a theory" I have, that's a sincere belief that religious person might have. Had you been born a century earlier, it probably wouldn't have seemed so crazy to you.
> The key difference is that I don't want to limit personal freedoms for a specific class of people.
You probably do, but you'd just arbitrarily define what is and isn't a (protected) class of people, so I'll leave it at that.
Of course the steps from 'insane belief' and 'act perceived as immoral' can involve specific steps of reasoning which seem morally correct within the context of the insane belief. That's not the point, not a good justification for the act, and not something that we as a society should tolerate.
Some kind of incompetence, delusion, or misunderstanding is very often the cause of an immoral act.
[1] https://en.wikipedia.org/wiki/LGBT_demographics_of_the_Unite...
Gay men and women who had worked alongside and under Eich for years at Mozilla were appalled at the controversy and said he had always treated them kindly and was a great friend.
I see how this thread maybe isn’t helpful — except that it demonstrates that people still have hard feeling towards him (though I doubt anyone who knew him personally would, gay or not). But it’s bringing up the old controversy yet again. Sorry.
Can you not imagine that it's possible to disagree with someone on social issues without hating them?
How so? Has Brave introduced security related features that are not in Chromium?
That would indeed be impressive, considering the amount of developers working on Chrome.
So there's a bit of insider knowledge there, yes.
Oh, and there's my desire to push back against browser engine monoculture, which I believe is bad for the Web.
Please, more people switch to Firefox. Chrome is breaking the internet.
The analytics is most likely blocked by common adblock rules. And present-day Firefox users are quite likely to be advanced users, running adblock.
Thus Firefox ends up being under-represented in the analytics.
Or have user agents devolved to the point where every browser lists every other browser and you genuinely can't tell them apart anymore?
Seems like you should still be able to identify the browser without a fancy analytics package.
Yes
Yes.
On the other hand, if you're running adblock, many sites probably don't care how and if the site works for you.
And by the way: it is estimated that 60% of all internet users block analytics. So take your stats with a grain of salt.
Source?
It is totally possible that the "problem is on my side" but they still lost my business. There's only so much effort I'm willing to put into purchasing from a company if I can't use their website. Most places work fine, both of these didn't. I normally just find a company that has a better website to give my money to instead.
I have tried to switch to Firefox multiple times, but I couldn't stand how much slower it felt than Brave, especially on Android. I also prefer the minimalistic interface of Chromium, and run Brave with its "Rewards" feature turned off.
YMMV of course. Firefox has a long way to go it seems on mobile.
It's possible it would have eventually worked if I started the browser in safe mode or something, but this doesn't seem reasonable to me to do to be able to give them money.
And overall the extension landscape looks highly bleak. What are the recommended adblockers? I found Ghostery Lite (which according to the reviews seems quite disappointing) and a number of paid adblockers...
From my test some while back there at least seemed to be a decent number of extensions out there but now it's different. Why?
Edit: The Bitwarden extension seems to have been installed automatically along with the desktop app.
Regarding ad blockers, my solution has just been to stop using the internet so much. I almost exclusively use it for work now. So I run into almost no websites with ads.
I heartily recommend going internet free except during work hours, and using the library for the rest. It's been such an amazing life improvement for us, we wish that on everyone.
Laughing foremost because I'd certainly be asking the wrong guy as I'd expect most HN guys to be power users. Not laughing because I recognize the huge potential of waste of time the internet brings along, that I hadn't really seriously considered until recently, had just figured it's half necessary to keep myself up to speed with things. We really don't need to be up to speed with things, outside of work at least, and if work is stressing us to the point where it becomes uncomfortable we should simply just change our work.
I do however like to be efficient in whatever I do so I would still like a convenient way of browsing the web, but... in the big picture it's not that important, especially if you leave out the time consuming stuff. Then it's mostly looking up stuff, which you can do using whatever tools.
In a couple of months I will be cutting off most of those time consuming tasks, whether that's FB, Reddit, HN, following politics, mindlessly looking for entertainment/escape. It's a life choice and I'm really looking forward to it.
For work I simply use Chrome and will keep it that way. Simple, easy, compatible, not fighting the mainstream. For private browsing though I would prefer to use something more.. personal. Brave works fine, currently trying out Cliqz. And I've tried out Next and Qutebrowser the last two days. Being a geek/ADHD I guess. There are more important things in life. Meditation helps. So does sailing. And not ranting on HN.
> In a couple of months I will be cutting off most of those time consuming tasks, whether that's FB, Reddit, HN, following politics, mindlessly looking for entertainment/escape. It's a life choice and I'm really looking forward to it.
Forgot which book it was, Sinner's Guide or Philothea, but the author said that Hercules went off into a lonely area to debate with himself what kind of life he ought to life, and he concluded that a life of virtue is the only one living. Your statement reminded me of it because (a) it's purely a choice of the will, which is the only thing we really have control over, and the only thing we own that nobody else has control over; and (b) he went off away from the busyness of the world in order to have an opportunity to think about his life and how he should live it.
It's unlikely that you will notice the difference anyway, few websites manipulate that many DOM elements at interactive rates. Those would be much better off with Canvas, which is also way slower on Firefox.
[1] https://www.phoronix.com/scan.php?page=article&item=chrome-8...
in my experience, both browsers performed very differently under Windows than under Linux (Manjaro). both mozilla and google clearly spend more time optimizing for the overwhelmingly more common user (Windows).
https://chromium.googlesource.com/chromium/src/+/master/docs...
They release often for the same reasons you would do continuous deployment. Release a little, but often. It helps them to better isolate regressions in the field and deliver non-critical security fixes quickly.
I know I say that it helps them prevent regressions and this post is talking about regressions, but it's also work keeping in mind that Chrome telemetry is insanely extensive. I'd wager that in the field, to the typical user, that Chrome 80 is actually faster than 75.
It's not a rigorous test, there's no data on other browsers or earlier versions, it's a github issue on one particular js/DOM benchmark detailing a likely regression. There aren't any details I can see on what hardware is tested. Even if you wanted to start researching chrome performance, this is a bad place to start.
So why is it up so high after just a few minutes and 54 upvotes?
https://news.ycombinator.com/item?id=16020089
A good rule of thumb for this effect is when the number of comments on a submission exceeds its score.
In this case, there is a broad ongoing argument about Google and its products [jumping the shark/being evil]. That argument is so much fun that, as I write this, almost every single top-level comment in this thread is a salvo in that argument, only tenuously connected to the OP. This article is just a convenient place to have the argument.
(I have a personal learning experience here: The first article of mine that hit the HN front page was a writeup of a PyCon talk entitled "To Make the World Better, Remove Some Javascript from it"[0]. It also shot to the top within minutes, and was [defensibly] flagged off the front page very quickly, because the title alone immediately turned it into a place to argue about JS [not] being evil, rather than discuss the Python-to-Javascript compiler I was talking about. Suffice it to say that subsequent articles have been more modestly titled!)