Question About IDE Raid

[cbman]

Maybe someone has... I'll look for some tests when I get home... the only problem is all the hardware sites testing new hardware never seem to go raid they just use big fast drives.

[Ribbit]

The point is that most applications/circumstances are not STR-limited. The access time of the drive(s) has a far bigger effect on performance. I think this is the StorageReview article dZeus was referring to: http://faq.storagereview.com/tiki-in...leDriveVsRaid0

For a doubling of STR, three out of four simulated-workload benchmarks gain less than 10%. The fourth is the WinXP bootup simulation.

[cbman]

I'm not looking for the performance increase for gaming purposes... I am going to be doing alot of file transfers between the two raid setups if I get it going at its full potential but my problem isn't that I'm not seeing an improvement... its the unimprovement that has me concerned... even in that test they have two fairly recent drives and they are recieving an increased STR total... I'm not... I've actually been set back by about 10MB/s... I guess I'll just have to get the PC WINBench99 that they used and see what numbers I garner.

[Kurt]

Originally posted by dZeus
lol... nope.

Read up on the RAID performance comparison article on storagereview.

If the PCI bus is the limitation, performance will drop as the disks perform unnecessary rotations. When the PCI bus is fast enough, you do not incur such penalties and your drives can work at full speed. _Of course_ it doesn't mean you'll get amazing increases in performance all round, it just means they won't be slowed down.

Also, with the advent of SATA, RAID0 might see the I/O per sec. count raise even more (thanks to less signal skew, smaller overhead, full duplex ops, etc). It's a tad early to say as most SATA solutions use bridges...