Originally posted by: Griswold
Originally posted by: BD2003
Other than your last sentence, you're entirely wrong. The cached entry he provided is *NOT* the superfetch cache. Superfetch disk cache = free memory. Furthermore, the cache will not be paged out to disk, it's a disk cache after all, so what would be the point of paging to disk things that come from disk in the first place? Disk cache in memory is treated like free memory - it's dropped like it's hot when another program needs it.
Look it up
here Superfetch does exactly what I said, it loads apps into RAM and drops it from RAM when RAM is needed - be it to the swapfile or just reloads it from disk, doesnt really matter - the disk is accessed again to load it. I'd really like to know how you come to the conclussion that the free memory entry is superfetch.
Superfetch isn't the free memory entry, but as far as windows is concerned, it's treated as "free" memory. It's instantaneously dropped when needed, it's contents are not paged to disk. Where are you reading that it is? Writing to disk what's already on disk is pointless, which is why it doesnt bother to do it. The disk is *obviously* accessed again to reload it, because it's a disk cache - everything in it came from the disk in the first place.
Superfetch WILL make the disk thrash after you exit, but it's on a low priority i/o thread so theoretically it shouldnt get in your way. But most of the thrashing your experiencing after exiting isnt superfetch, it's a lot of stuff that WAS resident before you launched COH and it made room for it, which you now need again.
Look at taskman and/or performance and reliability monitor after exiting a memory intensive app or game and you will see (in taskman) that cache is being filled again (again, what makes you think the "cached" entry is not the application cache?), and to take it even further, you can also check what resources have been swapped to disk to free memory and are being loaded again.[/quote]
Believe me, I'm on top of it. The "cached" entry you're referring to in taskman is
absolutely *NOT* the superfetch disk cache - read the other thread about it near the top of the forum to understand.
The resources that were swapped to disk and are being reloaded again is not the superfetch cache, it's program memory that was written out to make room.
Much of this memory is part of what windows considers the "system cache" in the taskman, which is being refilled, but again...this is *NOT* the superfetch disk cache, it's something else entirely.
No argument that the superfetch disk cache is going to be refilled after exiting, it absolutely is. But that comes second to paging back in resident code memory (the "system cache").