Defragmentation

I've been following the Defrag shootout with interest. Members have expressed their preferences on which defragmenter they prefer, I've got my own too. Assuming that you use these defraggers, it would be interesting to know whats your take on claims that defragmenting is a myth and that no real performance gain can be noticed or that its defragging that actually causes the wear and tear of the HDD rather than the fragmentation.

 

Anecdotal evidence such as this is usually not a good foundation upon which to base assertions that generalize beyond one's own experience. Don't get me wrong, I think you have alot of know-how with regards to computers, but it would seem that there are simply too many variables that can affect hard drive performance and longevity. The immense variation between harddrives alone is enough to make relying only on anecdotal evidence illadvised. I'd be interested in some technical discussion about the notion that defragging doesn't really improve performance and may even be counterproductive in regards to the integrity of the drive.

Again some technical discussion is in order. Are there any tests that attempt to scientifically assess defragging's effect on performance? Or are there white papers or such that discuss the correlation between defragging and performance?

 

Pete~ While I agree that wear and tear is a negligable consideration (unless one defrags their disks a few times every day!), I don't see why you feel that because your disks are only about 15% full that reduces potential performance benefits when defragging them.

If one tends to maintain a static software environment, then defragging will be of minimal value, whereas smart* defragging should definitely improve performance in a dynamic software environment, and even more so if directories are scattered all over the platters!

*By smart defragging I'm alluding to any defragger that allows you to place drectories on the outer tracks and (preferably) adjacent to the MFT tracks.

Just my 2-cents worth. ~pv

 

My drive is about 4 years old and I also defrag at least once a week. I have also only use about 15% of my drive. I have found that when I burn copies of my dvd (home movies), that if I don't defrag at regular intervals my times for reading and burning the movies are almost 30% longer then if I where to defrag first.
I have found this to be consistant. If I forget to defrag and I see that the movie is going to take a long time ( the estimated time by the program I am using) I cancel and defrag my drive. I use JKdefrag ( faster then windows defrag) and because I give copies of the home movies to my offspring ( 3 copies) I have found that it is worth the time to defrag first.
I have been wondering for a long time if I am alone here or if others have the same experience. With ordinary use the time difference would probably be so small, I probably wouldn't notice.

 

To get the the freedom to place any files everywhere on disk is something what makes defrag less boring,ou have to think about what to place on the outer tracks and what goes to the spindel area,every file placing on the outer tracks is't good, rarely used data stays in the way there so you better move these stuff to the inner tracks,and i repeat if some app. gives you these opportunity then defrag get some little exiting to do.

 

In general useage disc access is random, only small files are read completely in one go, pagefile is accessed a lot (check filemon by sys internals),
also things like the cache on disc, system file cache all help reduce performance loss to normal fragmentation.

The NTFS filesystem driver has a basic/unintelligent placement policy, which means files usually fragment as they age (*nix file system drivers group related files and scatter groups across the discs, so that if a file does get fragmented the fragments will be located nearby, vastly reducing the potential performance loss of fragmentation).

Anyhow, there will be a gradual increase in file and fragmentation with age in general use (web surfing, office apps etc), but IMHO its not unless you run low on free space (10-20% depending on drive size, file size and write intensity), will you get any noticable peformance loss (that is the time it takes to defrag will be gained back in read/write performance) OR you are doing heavy write activity - downloading lots of big files simulatanously (know to cause fragmentation even on most *nix filesystems), or generating large files in things like multitrack music and recording movie streams.

Spin up and down is the limiting factor of HDD life, its how HDD manufactuers measure HDD life.

If you spin up your drives specifically to defrag them you will reduce life, if they are already spinning there will be no significant increase in wear.

 

I am myself quite sceptical about the theory that defragmenting can reduce HDD life. As i see it, there cant be much difference in the stress that a HDD undergoes if you compare say 30 minutes of defragmenting everyday (at the max) with playing a game or movie on a drive thats fragmented already, can there?