Live data from Hacker News

Scalability

gregor-wagner.com

61–70 of 165 posts

Re: Scalability

#61
post #5

Performance is the side-effect, not the cause. The cause is the fact that chrome uses separate processes both for security, and so if one tab crashes, you don't lose them all. The fact that it uses more memory is a design tradeoff (although with shared text pages it's not as bad as one might think). The real question is whether you would ever have that many tabs open, and how much memory do you have on your desktop?

200+ tabs open, 12 GByte RAM (I so could profit from a 64 bit Firefox).

Re: Scalability

#62
post #60
post #48

Earlier quoted context omitted.

> The real question is whether you would ever have that many tabs open, and how much memory do you have on your desktop? I have 60 tabs open now (in FF), and this is after closing most of them, on a 32-bit system with 4GB of RAM. I regularly hit over 200 tabs when browsing.

200 tabs.. how is that possible? Do you ctrl+click on every possible links on every page?! I'm actually really surprised. I, too, can't stand more than 5-6..

Let's say you're on 4chan and there's a creative thread with a bunch of images. Threads can reach 150 images before they max out. You could use the Linky extension to "Open All Image Links in Tabs" and you would instantly be up over 100 tabs.

That's just one use case, but there are certainly many others.

Re: Scalability

#63
post #19
post #5

Performance is the side-effect, not the cause. The cause is the fact that chrome uses separate processes both for security, and so if one tab crashes, you don't lose them all. The fact that it uses more memory is a design tradeoff (although with shared text pages it's not as bad as one might think). The real question is whether you would ever have that many tabs open, and how much memory do you have on your desktop?

The limit is found pretty soon on older machines; while the OP tests on a Mac with 8GB of Ram, on my main machine I have only 2GB of Ram, and Chrome becomes unresponsive after ~10 tabs opened. I didn't think Firefox would be better but if it is, I might go back to it!

I have 6GiB of RAM, and Chromium eats most (typically, about 70-80%) of it. When I have too many tabs open with a lot of browsing history (for reference, currently, I have 61 tab open) Chromium eats all available memory, so, obviously, system chokes on swap.

Still, I prefer manual memory usage control (I have a meter in my panel, near the clock). This feels easier than less-predictable browser-wide hangs where I can't even switch the tab. I've tried Firefox 4 when it was released and I was unsatisfied with its responsiveness, for example, when I open a lot of flash-heavy tabs. While I've heard Mozilla devs did move plugins to separate process, in practice there were cases that the browser did hung for several seconds.

Re: Scalability

#64
post #24

Anyone care to share how their usage patterns lead to having hundreds of tabs open at a single time? How is it even feasible to find a specific tab without doing a time consuming linear search? I rarely have more than 4-5 tabs open at a time, so I'm honestly curious.

Go to 4chan and find a funny or creative thread with a bunch of image replies. Use the Linky Firefox extension to "Open All Image Links in Tabs". Voilà.

Re: Scalability

#65
post #42
post #15

Earlier quoted context omitted.

What you mean is not in a mandatory fashion and at the C code level and that's true. But what I meant is that you can do it for the objects you manage, such as JS code and the rest of the webpage. It doesn't prevent an issue in the Firefox interpreter itself.

Every browser does that. Virtually all the problems that matter are internal to the browser.

thats not true. FF 3.6 and before that did not compartment it per tab.

Likewise for IE, but they're per process like Chrome now.

Not sure about other webkit browsers and Opera, it's likely other wekbit browsers (except safari) do not do that.

Re: Scalability

#66
post #5

Performance is the side-effect, not the cause. The cause is the fact that chrome uses separate processes both for security, and so if one tab crashes, you don't lose them all. The fact that it uses more memory is a design tradeoff (although with shared text pages it's not as bad as one might think). The real question is whether you would ever have that many tabs open, and how much memory do you have on your desktop?

In the Gregor's blog post, Chrome performance is drastically worse even when it uses only one renderer process, not one per tab. In the demo, if one tab crashed, they all would.

Re: Scalability

#67
post #54

Interesting. This doesn't match my experience at all. When I have lots of tabs open in Firefox, there are occasional pauses where everything, including the browser's chrome, freezes. I've never noticed these pauses in Chrome. And yes, I do open hundreds of tabs. (As a side note, Chrome's far smaller minimum tab size is a major bonus for my usage, simply because I can see more tabs without scrolling the tab bar).

you need Firefox 8 to feel the improvements

Re: Scalability

#68
post #65
post #42

Earlier quoted context omitted.

Every browser does that. Virtually all the problems that matter are internal to the browser.

thats not true. FF 3.6 and before that did not compartment it per tab. Likewise for IE, but they're per process like Chrome now. Not sure about other webkit browsers and Opera, it's likely other wekbit browsers (except safari) do not do that.

I can't follow what you're trying to say here. I'm probably just misunderstanding you. Someone suggested that Firefox had memory segmentation, and just did it with threads. That's obviously not accurate. Then you said, "they have it, just with javascript objects". What mainstream browser allows javascript in a tab to access objects in any other page, let alone another tab?

Re: Scalability

#69
post #56
post #18

Opera's doing just fine. As I am writing this, Opera has all 150 tabs open and still feels snappy. ~1.5GB memory and ~30% CPU (currently). I could not measure the time but it was well below 10min.

Opera has always been a performer when low specs and lots of tabs were involved, cruising with 50+ tabs when Firefox and Chrome choked my system at 10+ tabs.

The major issue with opera is that you can't easily automate the tests, that's why it's often missing when people test IE and Chrome and FF. So it's hard to benchmark stuff and most don't take the time to do it.

That said opera is known to be rather fast.

Re: Scalability

#70
post #23

Windows, Linux or OSX? I'm curious because on Linux, threads and processes are essentially the same thing - they're the same data structure in the kernel. But as far as I know, that is not true on Windows. I also find his conclusion a bit disingenuous: I never have more than ~10 tabs open. I'm more interested in the performance around my actual usage, not an order of magnitude more.

How is it disingenuous at all? It clearly specifies the conditions and re-emphasizes in the conclusion that the analysis was done based on which browser handles having a large number of tabs open better. Also, the plural of anecdote is not data, but I routinely have more than 100 tabs open. So this analysis is clearly relevant and interesting to me. Edit: proof-reading.

I guess we all have different definitions of "many" in the statement of "many tabs." I didn't think 100 open tabs is realistic, but enough people here seem to do it.
Post reply on HN