Post #2467860

2026-04-13 23:34 UTC

@nepi@gts.nepi.gay Is it something to do with how the page file works under Windows, like it dynamically grows it as it needs more? I thought more recent kernels were actually trying to address this somewhat. Also, I think there's an overcommit sysctl parameter that can be tweaked as well.

Replies (1)

@nepi@nepi.gay 2026-04-13 23:42

@toroidalcore@masto.hackers.town I'm sure dynamic pagefile sizing helps, but I'm not really well-versed enough in NT or Linux to say what the difference might be. You can really get into the weeds with this stuff. If I had to hazard a guess I would say the fault lies in the way that distros will ship things that "work well enough" rather than picking an objective and optimizing for it. Reality is that in 99% of cases, people will either not be hitting memory pressure issues or are being paid to prevent them in the first place. Not a lot of incentive to make things better, I suspect. It's something that's bothered me for a long time but I've just not put in the legwork to figure it out and understand it.

Open ##2467861