silis

can this get a sticky?? seeing as how quite a few home pc's run windows... and most people I know want the best performance they can get...

---

Xie

If your gonna sticky this you should also sticky this same damn post that we already did this in (that I also linked to).

</rant>

---

LeeJend

GoNzO - I'm surprised you had a problem with Battlefield / DC or Vietnam at 1 gig RAM. Mine was a royal pain at 512 but when I went to 768 it's been smooth as silk and that's with a 768 fixed size pagefile.

Perris
I can read a defrag screen and when I had a pagefile under system control, auto sized by windows it got fragmented all to hell! Had to kill it, defrag the disk and recreate it. That's why I switched to fixed size.

As for setting it too small causing the fragmentation that corroborates what I said. If you make it adjustable you will eventually cause it to expand and start fragmentation. I'd just as soon windows had to purge it often enough to keep it small.

---

perris

leejend, read my paper, everything you just asked is addressed

of course if the pagefile is resized, the new extent is in a differant location (for that episode only).

BUT the new extent is eliminated on reboot...the original extent is in the exact condition and lacation as it always was (obviously)...there cannot possibly be fragmentation of the original pagefile on reboot

further, the reason your pagefile resized in the first place is because your initial minimum was too small, (obviously, otherwise it wouldn't have expanded)

system managed pagefile will start the pagefile at 1.5 memory...USUALLY 1.5 is more then enough, IT IS NEVER TOO BIG, SOMETIMES (for power users) IT'S TOO SMALL. NO PERSON NEEDS TO LOWER THE DEFAULT, THERE IS NEVER A GAIN FROM LOWERING THE PAGEFILE. period

for some reason, people have been missguided, and told to lower the pf from the default...then they have trouble, then they complaign about the trouble they caused themselfs

you DID NOT have to do any of that to the pagefile to get it contiguous again, it would have been back to it's original condition by merely rebooting...if it was contiguous before the expansion, it HAS to be contiguous when the added parts are discarded on the next boot.

BUT, your initial minimum was obviously too small, otherwise it wouldn't have the need to expand, and it will have to expand again whenever memory goes under pressure until you raise the initial minimum to a proper value.

the fact that your pagefile resized at all means that you had set your pagefile too small in the first place.

this usually only happens to people that have the missplaced idea that a pagefile smaller then the default is somehow better then what the operating system sets up for itself

now, as you thinking that making it adjustable causing it to expand

no

there is only one thing thatcauses the pagefile expand, one thing and one thing only;

when the commit charge reaches the commit limit the pagefile needs to expand, nothing else causes it to happen.

for those that don't know what those terms mean;

when the memory your use is requesting, (commit charge) is more then the memory that's available (commit limit).

so you make the point by the fact that you expanded at all.

NEVER go lower then the default for your initial minimum, some users need to go higher, no user needs to go lower.

then, if your pagefile ever expands where once it didn't, it's NOT because you've set it dynamic.

it's because you've become more of a power user then you once were, you've loaded more programs or more sophisticated programs, or you're doing more, etc.

again, if your pagefile expanded, THE ONLY WAY THIS CAN HAPPEN IS IF YOU ARE REQUESTING MORE MEMORY THEN THE OS CAN DELIVER, and your pagefile is obviously too small.

now, the situations we are talking about, where the os needs more memory then available will almost NEVER happen to people that understand that a small pagefile is a bad idea...do not ever go lower then the defualt, and for most users, the pagefile will never expand

but to those that it has ever happened to in any situation, the expanded pagefile helped when it happened, because if that didn't happen when your commit charge reached your commit limit, the os would unload a dll or exe that has recently been referanced and shouldn't be unloaded, and if it can't do that, you will freeze, fail, crash, any of the above

leejend, read my paper that xie posted, good reading, and everything you mentioned here is addressed

GoNz0

sorry m8, was problems back in 512 days, got 1gb and all problems went away

Xie

gaming and such w/ a 1GB or more is much nicer

---