Do you defragment your computer?

Post Reply New Topic

That is correct. The quality of the memory cells vary and so does the technology. Example: OCZ Vertex drives have fast IO chips and quality memory to achieve max SATA III speed. OCZ Agility drives use IO chips with data compression and lesser quality memory. It is a cheap way to boost performance and you get good value for your money. But if you ask me - crap is still crap.

Just because I've never bothered looking, I checked what it was on my desktop PC's main partition: ext4, 27GB, currently 84% used. This was created in May 2009 and has just been thrashed every day since then as my primary workhorse partition.
The result was 1,266 non-contiguous files (0.4%) from a total of 343,036 files.

Yep, but it's about as close as makes no difference.

^ wait... how can you have a PC with a main ext4 partition?

When people talk about "PC vs mac" I assume PC's are strictly Windows.

Ah well - I just take PC to mean "Personal Computer", which is what I believe it originally meant when IBM first introduced the IBM PC before the Redmond Virus took hold.
Maybe these days "computer" or "machine" would make more OS-neutral sense, while still referring to the same hardware.

Although... on this machine I do have a copy of XP installed on another partition, so I guess that would fulfill your definition too.

That would be Linux read/write supported partition. But since 2009? Ok that would have been ext4dev right? I'm still happy using xfs after all these years. The only let down is that it doesn't come with a shrink tool, but with drives always growing it doesn't really bother me. the de-fragmentation tool that comes with it does a good job.

$ uptime
11:10:40 up 1385 days, 21:06, 3 users, load average: 0.00, 0.00, 0.00

will have to shutdown this box, and might not come back up

Occasionally, I defrag for the hell of it, but I know it doesn't really do much when it comes to modern computers.

If your computer is slow and relatively modern, the slow speed is very likely to be due to something other than fragmentation.

No - ext4 went stable in 2008.

I have my defragmenter run on a schedule, once a week, so I don't even have to worry about that bit of maintenance. Back in the Win95 days, though, I did sometimes indeed stare, slack-jawed, at those little boxes lining up.

I have used Linux in the past and I never had a fragmentation problem with Ext2, 3, and 4. The developers are incredible with the allocation algorithm (I am a geek with file systems). XFS is like ext4 with delayed allocation, I don’t know anything about reserves (never bothered). Now JFS, can fragment heavily. This is because it has very few block groups (like 65) and it has a bad allocation algorithm as well. The allocator doesn’t bother looking at how much free space the regions have. JFS also locks a block group when a file is opened inside it. While good for the file to grow contiguously, it does mean other files will have their fragments spread out to other block groups, both increasing head seeking time, and fragmentation.

HFS+ (Apple mac os x) is also good at avoid fragmentation, but only when the file system is new. This is because the HFS+ allocation algorithm avoids reusing recently freed space for as long as possible. While good in the short run, it is bad in the long run. This allocator algorithm hurts it in the long run, because as the larger free space become exhausted, it will have trouble filling the smaller spaces as I found HFS+ always prefers to allocate space in the largest free space gap.

FAT32 obviously fragments.

Now NTFS, I did tons of observations in Win XP to Windows 7 and I have found that NTFS fragments significantly less in Windows 7 than it does with XP. I can’t say for sure why, but I do have a possible theory on why this happens. The windows up allocator will allocate files into the small free space gaps, which drives fragmentation up. On windows 7, it is more intelligent by allocating in the bigger free space gaps, which is why, my windows 7 disk has over 70000 free space extents. I think the windows 7 allocator is similar to HFS+ in which they both avoid the small gaps of free space. The problem with HFS+ is that you can’t defrag the free space, but on windows 7 if the free space fragmentation gets bad at least you can defragment both the files and the free space. This theory is built because Microsoft dosent have any powerful tools avaible for NTFS that allow me to do the crazy observations I would do with EXT2/3/4 and HFS+.