Log in
Log in
Forgot your password ?
or
Create an account
or
Log in using a Facebook account
NotificationsNo new notification
< Back to Computer Music
Enlarge
Computer Music
Audio & music gear
Computer Music
Products 29,707 29,7k
Actu 12,957 13k
Articles 529 529
Media 310,391 310,4k
Tutorials Tutorials & Tips 28 28
Reviews 15,441 15,4k
Classified Ads Classifieds 3 3
Forums 2,743 2,7k

Thread RAID 0 for Large Sample Libraries?

  • 0 comment
  • 1 participant
  • 4,503 views
  • 0 follower
Sujet de la discussion RAID 0 for Large Sample Libraries?
Hey everyone. I'm new to this board and have a quick question... Does anyone know if a RAID 0 (striping) array would work well for storing the libraries for large sample-based VSTi's like Synthogy Ivory, Drumkit From Hell Superior, or any of the large Quantum Leap packages? I know it's best to store the libraries on a seperate audio drive so all of the drive's read speed can be used for the samples (instead of wasted on running programs), and that a faster drive is better. So I would assume a RAID 0 array would be ideal, because it shares the work between 2 (or more) drives and thus reads and writes a lot faster. My question is about specifics... Does anyone know what a good stripe size and cluster size for the array would be? The drives are both 160GB Western Digital SATA drives. I was thinking that probably a larger stripe and cluster size (the max stripe size in my RAID utility is 128kb, and from what I've read the max cluster size in XP is 64kb) would be best, since from my reading on the subject it sounds as though a larger stripe size gives better positioning speed (like, the speed of the drive at finding files, rather than reading and writing them) which seems like it would be important for sample libraries (ie finding the right wave file right away when you press a key). And for cluster size, from what I've read larger cluster sizes yield better speeds, but at the expense of some wasted space if you're storing a lot of small files on the drive (which wouldn't be the case for sample libraries). So I'm leaning toward just using the max stripe and cluster size, but I don't know for sure, and I'm not sure if there should be some ratio between the stripe and cluster size (like if having them the same would be good, for example). Anyone with experience at this? Any help would be very greatly appreciated.