New
#21
No, no, you weren't disrespectful at all. On the contrary. That was more for my benefit, to proactively seek forgiveness for my personal tendency towards being incessantly argumentative on a technical level :)
I agree, it's not likely to be"dangerous", but I would question whether the additional risk - slight though it may be - should be recommended or undertaken without first understanding the environment-specific usage patterns. After all, you'd never advise a corporate client to make pagefile size changes without first examining their servers' stats. Why would home users be more likely to benefit from making changes just because they now have some arbitrary amount of RAM?
No, that's a misunderstanding between us. The pagefile is certainly utilised way before RAM "runs out", even at relatively low pressure levels. The magic part of the Mm, and the reason why many of the smartest developers working on any OS are frequently to be found on that team, is the way it reaches a decision as to what is appropriate to page out and what's not. On a simplistic level, the scenario you portrayed would be a bug because pages which were getting repeatedly "touched" were forced out for no good reason. The trimmer algorithm ought to be a lot smarter than that.
People who claim knowledge without demonstrating it generally rub me the wrong way, so I will resist my initial impulse to confine myself to "believe me"
Diagnosis: placebo effect.
Greetings to you too. I have some excellent friends in Amsterdam. You remind me a lot of the way they operate. That's a compliment :)