Re: All the processes

2017-03-07 Thread Kris Maglione
On Tue, Mar 07, 2017 at 02:14:48PM +0200, smaug wrote: What you mean with "chrome script"? Any chrome JS? There is frame script overhead per tab in child processes (around 340kB on my machine) but then we have also tons of jsms. I see 94 chrome compartments in a child process with one tab. And

Re: All the processes

2017-03-07 Thread smaug
On 03/07/2017 05:29 AM, Ben Kelly wrote: On Mon, Mar 6, 2017 at 5:42 PM, Nicholas Nethercote wrote: On Tue, Mar 7, 2017 at 9:22 AM, Ben Kelly wrote: These measurements are for full content processes. Many of the processes in the above do not need all the chrome script we load in content pr

Re: All the processes

2017-03-06 Thread Bill McCloskey
On Mon, Mar 6, 2017 at 7:29 PM, Ben Kelly wrote: > In general I think processes that rely on native code vs chrome script will > do better here. My impression is that the chrome script js heaps make it > more difficult for the processes to share memory compared to native code > pages. I could b

Re: All the processes

2017-03-06 Thread Ben Kelly
On Mon, Mar 6, 2017 at 5:42 PM, Nicholas Nethercote wrote: > On Tue, Mar 7, 2017 at 9:22 AM, Ben Kelly wrote: > >> These measurements are for full content processes. Many of the processes >> in the above do not need all the chrome script we load in content processes >> today. >> > > That's good

Re: All the processes

2017-03-06 Thread Nicholas Nethercote
On Tue, Mar 7, 2017 at 11:27 AM, wrote: > > Intuitively I don't grasp how each content process can add that much more memory that it would become a "major problem" jumping from 4 to 8 Simple: lots of stuff gets duplicated in each process. Efforts have been made to share data between processes to

Re: All the processes

2017-03-06 Thread Kartikaya Gupta
On Mon, Mar 6, 2017 at 7:08 PM, Eric Rahm wrote: > I assume WebRenderer will have it's own process, but maybe that just gets > lumped in with the GPU process. WebRender will live in the GPU process, if there is one. The UI process otherwise. Cheers, kats _

Re: All the processes

2017-03-06 Thread Andrew McCreight
On Mon, Mar 6, 2017 at 4:45 PM, Eric Rahm wrote: > I'd suggest looking at the memory overhead of Chrome's individual processes > as compared to ours, it's pretty impressive. My blog posts on our own e10s > memory usage [1] and comparison to other browsers [2] have further details. > I'm planning

Re: All the processes

2017-03-06 Thread Eric Rahm
On Mon, Mar 6, 2017 at 4:27 PM, wrote: > Just about the "4 good, 8 bad" part, it seems quite arbitrary -- Wouldn't > that be hardware-dependent? > I would think users with "only" 1GB may have different needs and > expectations from users with 16+GB. > It's more about acceptable memory usage in c

Re: All the processes

2017-03-06 Thread gsquelart
Just about the "4 good, 8 bad" part, it seems quite arbitrary -- Wouldn't that be hardware-dependent? I would think users with "only" 1GB may have different needs and expectations from users with 16+GB. Intuitively I don't grasp how each content process can add that much more memory that it wou

Re: All the processes

2017-03-06 Thread Eric Rahm
It should be pretty easy to measure at a high level with ATSY [1], but I agree coordination on overall memory requirements before adding new processes would be useful. For more detailed breakdowns of memory usage we can depend on about:memory reports for content processes, we added support for mem

Re: All the processes

2017-03-06 Thread Nicholas Nethercote
On Tue, Mar 7, 2017 at 9:22 AM, Ben Kelly wrote: > These measurements are for full content processes. Many of the processes > in the above do not need all the chrome script we load in content processes > today. > That's good to know. But it would still be good to get measurements of these slimm

Re: All the processes

2017-03-06 Thread Bobby Holley
On Mon, Mar 6, 2017 at 2:22 PM, Ben Kelly wrote: > On Mon, Mar 6, 2017 at 5:12 PM, Nicholas Nethercote < > n.netherc...@gmail.com> > wrote: > > > Now for the reason I raised this: the major downside of using multiple > > processes is that it increases memory usage. Recent-ish measurements > showe

Re: All the processes

2017-03-06 Thread Kris Maglione
On Tue, Mar 07, 2017 at 09:12:55AM +1100, Nicholas Nethercote wrote: Now for the reason I raised this: the major downside of using multiple processes is that it increases memory usage. Recent-ish measurements showed that for e10s-multi we could probably go up to 4 content processes without blowin

Re: All the processes

2017-03-06 Thread Ben Kelly
On Mon, Mar 6, 2017 at 5:12 PM, Nicholas Nethercote wrote: > Now for the reason I raised this: the major downside of using multiple > processes is that it increases memory usage. Recent-ish measurements showed > that for e10s-multi we could probably go up to 4 content processes without > blowing

Re: All the processes

2017-03-06 Thread Nicholas Nethercote
Thank you for all the responses. Here's an updated list: - main process - content process(es): 1 on release for most users; 2 on Nightly - plugin process(es): just for Flash now? (Win32 involves two processes for Flash) - GPU process (bug 1264543, in Fx53) - Gecko Media Plugin process: one p

Re: All the processes

2017-03-06 Thread Jared Hirsch
It would be great to see this info on MDN or the public wiki :+1: On Mon, Mar 6, 2017 at 6:00 AM, wrote: > > > > -Original Message- > > From: dev-platform [mailto:dev-platform- > > bounces+jmathies=mozilla@lists.mozilla.org] On Behalf Of Nicholas > > Nethercote > > Sent: Friday, Marc

RE: All the processes

2017-03-06 Thread jmathies
> -Original Message- > From: dev-platform [mailto:dev-platform- > bounces+jmathies=mozilla@lists.mozilla.org] On Behalf Of Nicholas > Nethercote > Sent: Friday, March 03, 2017 6:15 PM > To: dev-platform > Subject: All the processes > > Hi, > > I want to understand all the different

Re: All the processes

2017-03-06 Thread Evelyn Hung
We plan to introduce JSPlugin processes, one for PDFium and one for Pepper Flash. The JSPlugin process is a kind of content process, which loads a remote iframe as a projection of or tag. It will then ask main process to spawn a plugin binary process to run PDF (or Flash) binary. Both JSPlugin pr

Re: All the processes

2017-03-05 Thread Chris Pearce
We also have the Gecko Media Plugin process type. It's not the same as Flash plugin process. There are at least one per origin doing EME and one shared by all users of WebRTC that need to encode/decode H.264. The media playback team are also working on moving the code that interacts with the pl

Re: All the processes

2017-03-03 Thread Ben Kelly
On Fri, Mar 3, 2017 at 7:15 PM, Nicholas Nethercote wrote: > Do I have any of these details wrong? Have I missed any? > We plan to ship a "worker" process that will run ServiceWorker (and eventually SharedWorker) threads as part of: https://bugzilla.mozilla.org/show_bug.cgi?id=1231208 In the f

Re: All the processes

2017-03-03 Thread Andrew McCreight
On Fri, Mar 3, 2017 at 4:15 PM, Nicholas Nethercote wrote: > Hi, > > I want to understand all the different processes that we can and will have > in Firefox. Here's a list I constructed off the top of my head. > > - main process > > - content process(es): 1 on release for most users; 2 on Nightly

Re: All the processes

2017-03-03 Thread Chris Peterson
On 3/3/2017 4:15 PM, Nicholas Nethercote wrote: - plugin process: just for Flash now? On 32-bit Windows, there are multiple plugin processes because Firefox runs Flash in a plugin process and then Flash spawns one (or more?) of its own "Protected Mode" processes. IIUC, the Protected Mode proc

Re: All the processes

2017-03-03 Thread Kris Maglione
On Sat, Mar 04, 2017 at 11:15:29AM +1100, Nicholas Nethercote wrote: Hi, I want to understand all the different processes that we can and will have in Firefox. Here's a list I constructed off the top of my head. - main process - content process(es): 1 on release for most users; 2 on Nightly -