UPDATED POLL: Do you believe defragmenting an NTFS volume produces a significant performance increase?

[flexy]

its a nobrainer for me that, especially under NTFS a fragmented MSFT, pagefile etc. can decrease performance.

Also...bootfile fragmentation plays a role, for me fast boottimes are essential.

Using PerfectDisk i can see a CLEAR perofrmance gain when it comes to bootimes.

I keep my HD defragged (with the right tool, MANY are not up to my standards, i i just like Perfectdisk and how it does its job)...and do another defrag run THEN eg. once i installed major new drivers (chipset, graphics etc.) or a bigger app/game befcause i just "like the thought" of it all being neatly together

I didnt really bother measuring exact load-times etc. while running apps/XP...but as said since i see a very clear perfrmance gain booting (ie. loading drivers and everything i have at startup) i do NOT have a reason to say i do NOT want to defrag

 

[flexy]

Originally posted by: ArtVandalay
If it's been awhile since a defrag and the drive has gotten very fragmented, I'll notice a difference. As someone else said though, you usually end up reformatting before then anyway.

thyen i would've reformatted my HDS already hundreds of time

What's up with that "reformat is cure for evreything" talk all the time anyway ? I reformat exactly ONCE.....that's when i get a new sys/PC and install windows.

Rest is keeping it up well and good maintenance as long as it doesnt end up done too excessively...there is certainly no need to defrag every day or every couple of days. Do it when idle every couple of weeks, maybe even each month is totally fine. But then there's also no need to "reformat" just because it's an odd date and friday and "people say so".

 

[Nothinman]

If so then you will realize that if the data on the platters is spread out, then it will effect preformance. On files like movies etc. This won't make a difference because the data is buffered. But on something like games, large video files, OS files etc, It can cause slower data reading simply because the drive has to move a longer distance back and forth to retrieve data.

The affect will be minimal no matter what, games may take an extra second to load or the machine might take an extra second to boot. But otherwise the extra seeks caused by fragmentation aren't going to hurt much. Nearly all OS files will stay in memory after boot, large video files will be cached long before the data is actually put to the screen and games cache everything up front anyway.

 

[flexy]

>>>
The affect will be minimal no matter what, games may take an extra second to load or the machine might take an extra second to boot. But otherwise the extra seeks caused by fragmentation aren't going to hurt much.
>>>

well thats like saying its not worth overclocking because you might have a hard time seeing a real diff in real life.

But then..regarding the "defragging hurts the HD because of the strain".....wrong logic.

The MAIN reason for a defrag and free space consolidation is to reduce HD head movements. You will put some strain on the HD while defragging - no question. But then i dont think its something out of specs or an average HD cant handle. I defrag since years and always look that my filesyste is defragged..and never ever (knock on wood) did i have problems with an HD !

ACTUALLY - after the defrag: in your every day apps and work HD head movement will be reduced - so you will have less strain OVERALL than having a REAL fragmented HD and using it every day with all the unneeded HD head movements.

go for an hour or so heavy defrag..and then enjoy many more hours every day work with less HD seek-movements: Profit

 

[BespinReactorShaft]

Do you think the effect of fragmentation (and thus importance of defrag) is more pronounced back in the past (e.g. 486 days) when we had lower hard drive performance?

 

[networkman]

I use Diskeeper both on my systems at home as well as my servers at work. As someone mentioned previously, for those systems with lots of small files, yes, I do note a significant performance increase, while with fewer larger files, the difference is negligible.

As to what one considers "significant" - a 10% boost in performance is, to me, significant - for others, it may not be so. The one exception to the large files scenario is databases, SQL in particular. Performance gains from doing an external defrag first, and then a re-index of the database afterwards can be quite huge!

 

[Nothinman]

well thats like saying its not worth overclocking because you might have a hard time seeing a real diff in real life.

And I agree that's true for the most part, especially with the added risk of running your CPU above spec. But the difference is that turning the clock rate up yields a measurable advantage. You can easily time how long it takes to encode an MP3 with 2 different clock rates and see how much of a difference there is, but benchmarking a hard disk or filesystem is much more difficult because there's so many more variables involved.

ACTUALLY - after the defrag: in your every day apps and work HD head movement will be reduced - so you will have less strain OVERALL than having a REAL fragmented HD and using it every day with all the unneeded HD head movements.

Depends on your workload and for most people, it's not likely that a defragged drive will result in less seeks. Infact a properly fragmented drive would be much better because most file reads are in very small chunks, very rarely is an entire file loaded in one contiguous read request. This is why XP purposely fragments it's bootfiles in an attempt to speed up the boot process.

Do you think the effect of fragmentation (and thus importance of defrag) is more pronounced back in the past (e.g. 486 days) when we had lower hard drive performance?

Of course. If you have a 15K RPM drive that takes an average of 3ms to do a seek and compare it to even a recent 5400 RPM drive that takes on average 10ms per seek, you're going to notice big difference no matter how the files are organized.

The one exception to the large files scenario is databases, SQL in particular. Performance gains from doing an external defrag first, and then a re-index of the database afterwards can be quite huge!

Which is why it's possible to configure unix databases like Oracle to use a raw device for storage. There's no filesystem to get in the way and hinder performance.

 

[corkyg]

Why not address the subtle distinction between "defrag" and "optimize?"

 

[Fistandantilis]

think of it this way, if you just put all your socks in no particular order in your sock drawer when you got to get a pair of socks you have to match them up, but if you put them pair them up you can find a pair much faster.
defragging your drives is good good good. less search time for your drives = better and faster performance.

 

[SoundTheSurrender]

Originally posted by: Crism
I've got 2x WD 80GB drives on a RAID 0 config. I'm also a heavy gamer with CounterStrike Source so I need my hard drives in tip-top shape. I usually defrag once a week to keep the performance at peek. Also, after I install a large program or put a 1GB CompactFlash full of pictures on my computer I defrag to get those files in order. I've got 3 hard drives in my system...the RAID then 1 additional 80GB for non-gaming storage. I used to back up this information to the RAID drives. However, now I just transfer it all across my network to my mother's computer. Reason for this being that my RAID became fairly full and slower as I put the 50GB of data on it. I noticed I wasn't getting the same FPS in games. So I deleted all 50GB of data, defragmented the hard drive, and I then noticed a performance boost in the games! So pretty much, filling up your hard drive will decrease performance . Remember to defrag boys and girls.

Sounds like a waste of time. Defragging only puts files that are spread across on the harddrive back in order. Whats the point if there already together?

 

[kingsrook7]

yes

 

[redbeard1]

I've seen very noticeable performance increases after drives were defragged. These drives had not been defragged since the computers were new, something like two or three or four years. Until I had these systems show this, I did not put much stock in stories of faster performance. It does not make much of a difference if it has been ran regularly.

I run a defrag every three or four months. Usually I'm bored and just happen to remember.

 

[flexy]

people who use Diskeeper should really give PerfectDisk a try..maybe your opinions will change

 

[boomdawg]

As other people have stated, if it's never done it helps a lot the first time you defrag, if you defrag regularly it doesn't make much of a difference. I work repairing peoples computers. It truly is amazing the difference a defrag can make when a drive hasn't been defragged ever. Even windows can install fragmented. Not only can it increase performance but when one guy came in with less than 300MB of free space on his 40GB drive the defrag freed up nearly 600 MB (I didn't think it was supposed to be able to do that).

 

[Nothinman]

Why not address the subtle distinction between "defrag" and "optimize?"

Because most people consider them the same thing. I mean why waste time moving data blocks around that you don't ever see, if not to help performance? The OS hides the physical location of the data from the user, so there's no other conceivable benefit to defragging a filesystem.

 

[TGS]

Originally posted by: boomdawg
As other people have stated, if it's never done it helps a lot the first time you defrag, if you defrag regularly it doesn't make much of a difference. I work repairing peoples computers. It truly is amazing the difference a defrag can make when a drive hasn't been defragged ever. Even windows can install fragmented. Not only can it increase performance but when one guy came in with less than 300MB of free space on his 40GB drive the defrag freed up nearly 600 MB (I didn't think it was supposed to be able to do that).

That's due to the wasted space on files under or over the cluster size.

Defragging a drive will yield almost insignificant gains on I/O the real benchmark in how you gauge hard drive performance period. The actuator is better utilized using command queuing to make more logical movements, as seek times are effected by head movement over the disk.

Unless the disk is ridiculously fragmented, and not using command queuing of some sort(read home users not on hardware RAID controllers) then will rarely ever notice the defragged results.

Now reindexing a database to optimize how the database is parsing the data will make a more pronounced impact on I/O. Most DB systems have an online "defrag" or indexing system to do this on a daily basis.

The reasoning behind HDD command queuing and hardware cards with cache, is that the hardware controller and HDD has a much better idea on how the data needs to be written and read off for optimal results. The OS may not provide as good a job the HDD itself, which is why you address HDDs as a logical device.

 

[spyordie007]

Defragging a drive will yield almost insignificant gains on I/O the real benchmark in how you gauge hard drive performance period. The actuator is better utilized using command queuing to make more logical movements, as seek times are effected by head movement over the disk.

:thumbsup:

Now reindexing a database to optimize how the database is parsing the data will make a more pronounced impact on I/O. Most DB systems have an online "defrag" or indexing system to do this on a daily basis.

Databases are a whole differant story; they typically do a lot more than just simply putting the files in contigous order...

Most database systems do other very importance maintenance near/during their online defrag process such as commiting data from transaction logs or compacting unused space (since while they are online they never shrink).