On Wed, Jul 19, 2017 at 09:40:12PM -0400, Ben Kelly wrote: > On Jul 19, 2017 8:57 PM, "Mike Hommey" <m...@glandium.org> wrote: > > On Wed, Jul 19, 2017 at 08:48:45PM -0400, Ben Kelly wrote: > > On Jul 19, 2017 6:20 PM, "Mike Hommey" <m...@glandium.org> wrote: > > > > > What would be the rationale behind this choice? > > > > Smaller memory footprint, which, you'll admit, when you're on a machine > > with (less than) 2GB RAM, makes a difference. > > > > > > I thought we had data that showed OOM (small) due to VM fragmentation > still > > outweighed OOM (large) on these machines. If that is the case 64-bit is > > strictly better. > > I don't believe you can exhaust the address space and have VM > fragmentation be an actual problem when you have less than 2GB RAM. Of > if you do, everything is slow to a crawl before that happens. > > > I don't understand why that would be, but if so it should show in > crashstats as fewer small OOMs on these devices. Does the data actually > show that?
I don't know. Can that be filtered? > Also, that tells nothing about people that never hit OOM (a lot of > people even close their browser well before that could happen). > > > You're saying the size difference between 32-bit and 64-bit is so great > users will start manually shutting down? I find that hard to believe. If > they are shutting down apps to save memory they will probably do it with > both 32-bit and 64-bit. Again, though, in theory we have usage hour > telemetry we could look at to confirm (although maybe only for beta?). I'm not saying they shut down because of memory. I'm saying a surprising lot of people have browser sessions that don't last more than 5 minutes (I've heard multiple times in the past years that we found that out from our data). Those people are not going to see OOM. Mike PS: your mailer likes to not put quotation marks, making your messages hard to follow. I thought that was because it's sending both a text and an html version, and the text version sucks, but it appears it's not sending an html version... _______________________________________________ dev-platform mailing list dev-platform@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-platform
