11 ms·
This came up a month ago, and the test was faulty. They accidently measured the memory usage of child processes in Chrome, but not Safari: https://news.ycombina
by jefftk 6y ago
This came up a month ago, and the test was faulty. They accidently measured the memory usage of child processes in Chrome, but not Safari: https://news.ycombinator.com/item?id=26186114 https://news.ycombinator.com/item?id=26186114
(Disclosure: I work at Google, speaking only for myself)
- smoldesu 6y agoThank you for summarizing this. I was going to have an aneurysm after seeing this make it to the front page of HN twice in a row.
- 908B64B197 6y agoAs an aside, if someone only runs a web browser on his machine (and maybe a few other programs that aren't resource intensive) then using as much RAM as possible sounds like the correct tradeoff since unlike CPU, RAM has the same power budget whether used or not (so it is worth it to cache results). The alternative would be to page to disk, but that would hurt responsiveness. If nothing else requests the RAM you are better off keeping tabs alive.
- smoldesu 6y agoSure, but ideally we should be writing software that erases technical debt, not shifting it. To my understanding, the system you're suggesting is actually how Chrome manages it's tabs: it will let you open whatever you want until memory is constrained, at which point it will kill/compress the biggest offenders to keep things snappy.
- freediver 6y agoHere is a test that measures all child processes: https://twitter.com/vladquant/status/1375557440578539521 https://twitter.com/vladquant/status/1375557440578539521 This brings it to 2x more RAM for Chrome, compared to Safari.