Standing on my soap box...
This little bit from the Diskeeper FAQ is quite interesting. They say that
free space consolidation has no affect on performance. However, if you look
at the Diskeeper online help under Reference Information/Theory of
Operation, it states the following:
"As used in reference to Windows operating systems, disk fragmentation means
two things:
- a condition in which pieces of individual files on a disk volume are not
contiguous, but rather are broken up and scattered around the volume; and
- a condition in which the free space on a volume consists of little pieces
of space here and there rather than a few large free spaces.
The effects of excessive fragmentation are twofold as well:
- file access takes longer because a file must be collected in pieces here
and there, requiring several disk accesses instead of just one; and
- file creations take longer because space for the file must be allocated in
little pieces here and there instead of just one contiguous allocation.
In designing Diskeeper, the following goals were established:
- The product must be completely safe to use.
- It must improve Windows system performance . It is not designed to make
the disk look "pretty" - it is designed to improve disk performance and, as
a result, overall system performance.
- It should process live disks without interfering with user access to
files.
- It should run without operator intervention .
- It must defragment all possible files and consolidate free space into the
smallest possible number of large spaces."
The key things to note are they define fragmentation as also including free
space fragmentation and that free space fragmentation also causes problems.
They also state that a defragmenter must consolidate free space into the
smallest possible number of large spaces. They also state that
defragmenting must improve Windows performance. In addition, Diskeeper also
provides a Free Space Consolidation defragmentation method - further
indicating that they consider free space consolidation at least of minor
importance. However, on the other hand they try to minimize the importance
of free space consolidation. Why is that? Does anybody remember prior
versions of Diskeeper that included statistics on free space fragmentation?
Since Diskeeper 7 Second Edition (I believe), Diskeeper now longer provides
this information. Why is that when they obviously place a little bit of
importance on it?
There is a serious dis-connect here. One one hand they say that free space
consolidation is important and provide a method to try to improve free space
consolidation but on the other hand they say that it doesn't help to improve
performance.
Speaking as somebody who works for a defrag vendor, I agree that the
purpose of defragmenting isn't to make the drive look "pretty". The purpose
is to improve performance to the best possible. Simply defragmenting files
improves only read performance. However, how many Windows systems are
simply read I/O? Not many. Consolidating free space improves write
performance - meaning that total disk performance is improved - not just 1
part.
In part, you can measure the impact of free space consolidation using
something like WinBench - which (among other things) includes a benchmark
suite to measures drive performance. The nice thing about WinBench is that
it wasn't written by a defrag vendor
There are instances where drive
performance as measured by WinBench can actually decrease after
defragmentation - even though all files were defragmented. The reduction in
drive performance can be attributed to lack of free space consolidation -
which forces the file system to work harder to create new files - meaning
that performance slows down.
I'll get off my soap box now
- Greg/Raxco Software
Microsoft MVP - Windows File System
Disclaimer: I work for Raxco Software, the maker of PerfectDisk - a
commercial defrag utility, as a systems engineer in the support department.
Want to email me? Delete ntloader.
Carey Frisch said:
neat and tidy in the analysis screen, with solid blue bars
all the way across the screen (representing fragmentation-free files) and
the rest white space (representing consolidated space).
Clearly, the speed of the disk, meaning how fast you can access the data
on it, is more important than the prettiness of the display
or the consolidation of all the free space into one place. Free space
consolidation might be important if you have to create one
gigantic contiguous file, but it has no effect on performance. So
Diskeeper uses algorithms that achieve the highest speed from your
drive regardless of the arrangement of the free spaces on the drive and on
the screen. Diskeeper does so without wasting time on
excessive consolidation of free space. We simply go for the fastest
possible file access times and then stop.
Even so, you might ask why we don't continue and rearrange the files
further to get a neat display? The answer is, "Because it takes
computer power to do so." It would be wrong to consume more of your
computer's performance than we give back. So Diskeeper
defragments until the disk is in top shape Performance-wise and then
stops. Any further work is a waste of your computer resources.
