Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Fragmentation, and another question

  1. Feb 20, 2007 #1

    Simfish

    User Avatar
    Gold Member

    Does BitTorrent tend to cause a lot of fragmentation on the hard disk? The very nature of BitTorrent tends to distribute files according to fragments, and my defrag doesn't seem to really reduce fragmentation percentage. On the other hand, it's not really a problem since fragmentation really only affects the files that are fragmented, right?

    Another has to do with USB microdrives and any medium of data storage: is it possible for a USB microdrive or a CD to also become fragmented? Say, you transfer a fragmented file from hard drive to microdrive? And does this fragmentation then transfer from microdrive to new hard drive?

    ==

    The other question I have relates to RAM and CPU speed. The amount of RAM on my computer is 854 MB, which is less than the 1 GB of ram specified by the packaging [I use a Gateway, Windows Vista]. What happened? The processor speed is also lower than stated on the package.
     
  2. jcsd
  3. Feb 20, 2007 #2

    russ_watters

    User Avatar

    Staff: Mentor

    Yes, downloading multiple files at the same time leads to extrordinary levels of fragmentation.

    Microdrives hold many fewer files and they are smaller, but yes, they can become fragmented too.

    No, fragmentation does not transfer between drives.

    Your processor speed should be as advertised - your ram is too, but some is being used for video processing (your computer does not have a discrete video card). 854 is an odd number, though - 1024-128=896 is more typical.
     
  4. Feb 20, 2007 #3

    AlephZero

    User Avatar
    Science Advisor
    Homework Helper

    Not quite right. If the free space on the drive is also fragmented (i.e. lots of small areas of free space where files have been deleted) then new files may be fragmented when they fill up the fragmented free space.

    See the previous comment.

    Processor speeds can be reported incorrectly - for example if the CPU used variable clock speeds for power saving, you may get the current speed reported, not the maximum speed. That happened to me with an AMD processor. If you really want to check out your system performance, download something like Sisoft's SANDRA program (easy to find with Google, and free).
     
  5. Feb 20, 2007 #4

    verty

    User Avatar
    Homework Helper

    A nice way to handle fragmentation is if you have another drive with plenty space, then you can copy stuff across so you aren't left with much on the first drive, the defrag the first drive, then copy the stuff back. Since drives are 500GB+ nowadays, there is no excuse not to have space free.
     
  6. Feb 20, 2007 #5

    russ_watters

    User Avatar

    Staff: Mentor

    One thing I forgot - on laptops, cpu speed is not constant, it varies with load, so if you see it reported "incorrectly", that may just be because you aren't using it.
     
  7. Feb 26, 2007 #6

    Simfish

    User Avatar
    Gold Member

    Thanks for the responses!
    Funny - I found that my processor was running at 798 MhZ under Power Saver - but at 1.60 GhZ when it's under full power


    Model : AMD Turion(tm) 64 X2 Mobile Technology TL-50
    Speed : 798MHz
     
    Last edited: Feb 26, 2007
  8. Mar 8, 2007 #7
    Why copy files all over the place? Just defrag the drive as it is..

    One random preference of mine: Norton's speed-disk is pretty nice, you can choose what files you want to be stored closest to the center of the drive for fastest access speeds.
     
  9. Mar 9, 2007 #8

    verty

    User Avatar
    Homework Helper

    BoredNL, you can save a stack of time doing that.
     
  10. Mar 9, 2007 #9
    if you set your pre-allocate your download files, i don't think they'll get fragmented in the sense that each downloaded bit of data is put in a different sector.
     
  11. Mar 9, 2007 #10
    How is that saving time?

    You're taking files from one hard drive to another. When you defrag, aren't you using your hard drive's buffer and your RAM to store temporary information that is being rewritten? Wouldn't moving files from one hard drive to another and then back be like using that hard drive as your temporary space? I can't imagine how this would save anyone time. Maybe the defrag would not take as long, but that time + the time it took to move the files to another hard drive and back would be a lot lot longer. And not only that, but you have to wait till the files are moved before you can run the defrag, and then you have to wait till the defrag is done to move the files back. Why not just leave your computer up one night running defrag and be done with it?
     
  12. Mar 10, 2007 #11

    verty

    User Avatar
    Homework Helper

    The two relevant factors seem to me to be the efficiency of the defragging algorithm and your priorities. For those who would prefer to leave the computer to sort itself out, that is fine. There is much less seeking when copying files to another drive because the access is sequential so you get an average throughput of half the bandwidth of the drive. If the drive can transfer at 20MB/s then you'll get an average of 10MB/s. So it depends on whether your defrag program can perform at an average of 10MB/s. If it can't, moving the files to another drive will be quicker but will need more input as you say.

    I'm just putting the information out there.

    Oh, thinking about it, no defragging program could go at 10MB/s in that scenario because it would be reading/seeking/writing, so necessarily less. QED. :)
     
    Last edited: Mar 10, 2007
  13. Mar 12, 2007 #12
    This can be tested if you have Raid 1 configuration and some fragmentation on your drive. Just split the hard drives, then use method 1 and then method 2 and time it (then use the raid utility to mirror one of the drives back to the other). That'd be interesting.

    I still don't see how it would be faster moving files around though, even if you get a faster speed while moving them. If you keep up with defragging your hard drive, it might only take 20 minutes to perform a defrag. However just moving humongous files from one hard drive to the other might take that long, and then there would be the defrag time + moving the files back time. That makes no sense to me. Seeing as how you are moving ALL data to move large files, when a defrag only moves fragmented data around and only some files that aren't fragmented for optimal defragmentation. I could only see this being possibly faster on an insanely fragmented drive, but I doubt it even then.

    Not only that, but if you do this, you will be putting the large files farther from disk index and increasing the seek time. I imagine most people would want these files to have a faster seek times if they are working on them.. Especially because of how large they are.

    Just food for thought.

    Phil
     
    Last edited: Mar 12, 2007
  14. Mar 12, 2007 #13

    russ_watters

    User Avatar

    Staff: Mentor

    Yes....
    Well ram is super-fast, so sure you take a hit when using the hard drive as ram, but the situations aren't analgous.

    The relevant issue is that a single drive can't read and write simultaneously. It reads a chunk, then writes it in a different place. If you are defragging, though, you need to read off one spot, then read off another to free-up space, then write both sections.

    It could be faster, but it really isn't worth the effort when you could just start a defrag before going to bed - it should be finished by morning.
     
  15. Mar 13, 2007 #14
    This is a pretty moot point isn't it? lol
     
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook

Have something to add?



Similar Discussions: Fragmentation, and another question
Loading...