Hi, On Thu, Aug 04, 2016 at 02:25:26PM +0100, James Graham wrote: > On 04/08/16 12:24, Olaf Buddenhagen wrote: > > On Wed, Aug 03, 2016 at 06:22:51PM +0200, Till Schneidereit wrote:
> > > I'm not concerned about code complexity, but about memory usage. Memory > > > usage in many-tab scenarios is one of the measures where Firefox is still > > > vastly superior to the competition, and I think we should aim for roughly > > > matching that. > > > > I'm surprised that I keep hearing this claim. It doesn't at all match my > > real-world experience from not too long ago doing my everyday browsing > > on a serverely memory-constrained system... > > Fortunately we have data! Well, it would be really good to know why exactly the memory usage for the "tabs open" point is twice as high with more content processes. If anyone knows why it *has* to be that way, please chime in -- otherwise I have to keep assuming there is just some unnecessary duplication that can be fixed... Unfortunately, you haven't posted the "tabs closed" numbers for other browsers: but for Firefox, it's clearly visible that the number is actually *lower* for many content processes than for just one! This seems to confirm the point about memory fragmentation I keep harping on. The test used doesn't reflect real world scenarios very well: typically, the browser is sitting idle most of the time, with only occasional user interaction. Also, when user interaction *does* happen, there is mostly just one active tab at any time that loads pages -- but often dozens of them in succession, sometimes even hundreds. The active tab is switched or opened/closed less frequently than pages are loaded. Typically there are a couple tabs sitting mostly idle in the background; or for "power users", who run very long sessions -- or use automatic session restore -- usually dozens of them or more. I suspect that in such a scenario the memory fragmentation problem becomes much more pronounced... Now regarding the personal experience I mentioned, in my case memory usage on startup -- after restoring about a dozen or two open windows -- was only slightly lower in Firefox (single-process) than Chromium (automatically using 4 content processes on this system, plus a couple more for extensions). Admittedly that might not be an entirely fair comparision, as I think GPU acceleration might have been disabled on Chromium but not on Firefox; and I probably had slightly more extensions on Firefox. (Plus this was two or three years ago, so it might be a bit different today...) But that's not really all that relevant: because the *real* difference was that while the memory usage of Chromium was almost constant -- rising only very slowly to perhaps 150% of the initial value after several days of usage -- Firefox was quickly exploding (even without any user interaction at all) to more than twice the initial value within a few hours, with no slowing in sight. So in practice, Firefox was *way* less usable. -antrik- _______________________________________________ dev-servo mailing list dev-servo@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-servo