Of course, it is supposed to speed up disk access. But long passed the times when absence of disk cache made computers literally unusable. These days cache is largely for mere user convenience – it reduces continual annoying small waits during operator activities. However…
I noticed, every time Windows XP comes back from hibernation it gets its System Cache virtually empty – right after wakeup it is bare 100 MB left of 1000 MB disk cache, usually accumulated across the previous workday.
It is understandable that to restore disk cache from the
hiberfil.sys is almost ridiculous – that’s virtually same price in I/O terms as to read the cache’s data just on-demand when, and if, they’ll be actually needed. Therefore, to discard cache data during hibernation seems a sensible motion providing faster times going into and out of hibernation.
Yet I doubt the idea is absolutely brilliant.
I’ve just doubled the amount of memory in my workstation – not for running more programs – for faster overall responsiveness. That is, I’ve plunked cash for an additional memory module specifically for cache. Now, Windows says my money-spending concept was wrong.
See, I don’t care if my workstation startup time increases from 30 seconds to 40 – I can be patient there – in order to read up the “futile” cache. (I don’t care at all about the shutdown time.) But I want it to fire up Word, or show All Programs menu for that matter, instantly always the first time I do it (and I start Word really seldom more than once a day for a quick edit).
I wish they’d have an option for cache inclusion in
hiberfil.sys in place.