kakii said:
Windows XP latest SP
Hdd dedicated for page file installed, single partition.
Page file of fixed size generated on this disc.
While reviewing the partition with Windows on-board defragmentation
utility, the page file is presented in outer partition sections.
It doesn't start from partition beginning.
I guess files critical for system performance should be placed at inner
section - faster read/write access.
Does windows on-board defrag utility really show the facts?
Does this file is really placed in outer partition section?
If yes, why does windows do that this way?
Defragmentation can only make contiguous a file if there is sufficient
contiguous free space in which to store the file. One of the problems with
fragmentation is that some "system" files (which are often very small)
cannot be moved by any defragger, even the commercial ones (PerfectDisk,
Diskeeper, O&O). I forget which but one of them will actually let you see a
map of the files and that's how I found some tiny files marked "System" by
the defragger (but "system" was never well defined) that would not get moved
by a defrag operation. Because some of my files were huge in size, there
wasn't enough contiguous free space to pack the file's sectors all together
so it would always remain with 1 or 2 fragments. While it wasn't worth the
effort, I had the time and obstinence to find a solution, which was to save
an image (logical, not sector-by-sector) of the partition and then restore
that image (using a bootable CD so the OS was *not* running in that
partition). Restoring put the files together in contiguous clusters. Those
tiny "system" tagged files got slammed together instead of at random places
on the disk which limited the max size of any section.
You never mentioned how much physical RAM you have in your system. You
never mentioned how much was reported as free (since free RAM is wasted
RAM). You never mentioned the min/max size for your paging file, or if they
were set to the same value to eliminate or reduce fragmentation. You never
mentioned if all of the pagefile was in the OS partition or split to another
partition on another hard disk. I didn't run into the problem of these
system-tagged files not providing large enough space for the pagefile to
reside within one of them, only for the much larger data file. I suspect
you made your pagefile was too huge.
Why do you think the "inner section" is faster (if you mean the inside
tracks of the platters; i.e., those closest to the spindle)? Rotational
speed is the same across the entire platter but linear speed is greatest at
the outside, not inside. More sectors fly under the heads per second at the
outside than do at the inside. Cylinders, tracks, and mapped sectors are
numbered starting from the OUTSIDE of the platter. More sectors pass under
the heads per second at the outside of the platter than do at the inside,
even with variable recording. Platters are rigid. Part of them doesn't
rotate slower or faster than any other part. Spin a plate on your table.
Is it somehow elastic or disconnected so the center of it somehow spins
faster than the outside. It all rotates at the same speed. A 7200RPM disk
spins ALL parts of its platters at 7200RPM. Sectors don't get bigger in
size as you progress from the center to outside. They're all the same size.
That means the same sized segments at the inner track are spinning SLOWER
than the same sized segments at the outside tracks. More distance to cover
at the outside than for the inside within the same single rotation for it
all.
Start here:
http://www.pcguide.com/ref/hdd/geom/tracks.htm