The main problem with DK is that it doesn't fully defrag everything (to save time apparently) and it doesn't intelligently place files on the disc so fragmentation occurs more quickly after the defrag process. I found that DK always left gaps in the files on my disk (it didn't move everything up to remove small gaps). The net effect is that the file allocation system places files in these gaps and they are automatically fragmented all over the disc surface. I wrote and complained about this to DK (and the lack of control over file placement) and their response was "we know best". Not impressed.
Yes, DK does not consolidate all the free space all the time, since they claim it does not really provide enough benefit to justify the work the drive has to do to move everything into one chunk everytime. I saw a more detailed discussion of this at some other site when I was trying out the different defraggers. It's bookmarked on another PC and I'll post it later when I have access to the PC.
Personally, I am on the fence when it comes to full free space consolidation. It seems to my n00bish reasoning that pushing up all the files into one contiguous block will not help delay fragmentation. Because, as soon as some of those files are modified
(1) they have no space to expand, resulting in file fragments placed after the block of consolidated files
(2) reduced file sizes result in space between the consolidated files leading to further fragmentation during writes.
Maybe if some 'locked' free space is left at the end of frequently modified files to allow only those files to exclusively expand without fragmenting, then that would be a better solution? How it could be done, I have no idea since it would be dificult to predict how much a file is going to expand
On a similar note, DK has a feature that allows you to enlarge the MFT to prevent fragmentation, but AFAIK none of the other defraggers have this.
That said, DK does consolidate free space reasonably well, atleast from what I can see in the graphical reports. Pity it does not give a detailed log like some other defraggers do.
PD has all the scheduling functions of DK (including screensaver mode which I like best) but it also allows you to run an extra intensive defrag whicjh removes all gaps on the disk. You can also set number of days for each drive to determine files that don't change, files that change occasionally and files that change regularly. This allows the files to be moved during defrag so that the files that don't change much can be in an area where they are blocked together and left undisturbed (similarly for the other file groups). This minimises the need for future defragamentation and makes them faster too.
Actually, DK's automatic mode is quite different from the scheduling function built into the other defraggers. It is an 'always-on' thing that defrags if (1) fragmentation has increased beyond some threshold (what is the value? No idea!) (2) enough idle resources are available to defrag without interfering with concurrently running apps. So, it's truly set-it-and-forget-it. That's what drew me to DK ultimately, it was a complete no-hassles solution with great results.
There is a conventional scheduler too, but I never use it.
I never understood how the placement method of PD worked in conjunction with the free space consolidation idea. Does it move the frequently modified files to the beginning of the drive and the rarely modified files to the end of the drive, leaving a large block of free space in the middle? Or does it bunch up all the files towards the beginning of the drive, but the order in which they are arranged is such that the most frequently modified are at the very beginning to provide the performance boost from faster access times?
But, that would also mean that a block of most frequently modified files stuffed together will fragment the quickest too, right? Or am I looking at it the wrong way?
Windows defrag is based on an old version of DK in pure brute force mode. Basically it just shoves everything up to the start of the disc in any old order. It also doesn't support boot time defrag which means that system files and the MFT get badly fragmented and lots of disk space is lost in excess fragments which it can't reclaim.
I find the biggest speed difference occurs after you do a boot time defrag followed by a full defrag. Without the boot time defrag all the system files are left in a fragmented mess which especially forces Windows to startup slowly.
That's very true. A good boot-time defrag is necessary to complete the process. Incidentally, I have my paging file on a separate, dedicated partition on a different drive to minimize fragmentation. My OS has it's own dedicated partition too. Infact, my drives are all partitioned to hell