P4 1.5G running XP in NTFS, no problem here:

sequential read: 10.9
sequential write: 5.8
random seek: 2.9
disk mark 51.6

Wait, I just compared it to David's... I guess it's not good then. I don't feel it slow though.

try running a 100mhz compaq every day on win98

Just as a side note: I have had an IBM Deskstar die on me previously. First hard drive to crash on me ever.

But in the meantime...

I got the following pieces of hardware in today to test out.

A Cheetah 10k RPM SCSI drive

Adaptec 19260 (I think) SCSI controller connected to a 64bit PCI slot.

A promise Ultra100 (Ultra ATA/100) controller

A WD Caviar.

Here's where I'm at:

I have installed the Cheetah onto the Adaptec. Took awhile in setup (SCSI can be such a pain in the rear with the # of cables - foolishly purchased from buy.com who sent just the drive, no cables, and it took awhile to get the whole termination thing right). Anyway, up and running and benchmarked at 35 MB/s (a 10X performance improvement over the Deskstar on the onboard controller).

Then I hooked up the Promise controller to the Deskstar and rebenched it. Performance is still poor on the one NTFS partition but it went up to about a 5.5 (from 3.5). The other partition tested much better though at 14.6 (from around 5.5 previously). The difference in the partitions is that one partition has clusters of 512 bytes, the other is 4K. The 512byte clusters (which occur when you use the convert command) are a huge performance hit. But still, 14.6 MB/s is pretty unacceptable as well which implies that the Deskstar, at least here, is not performing as well as it should.

I'll be trying out the Caviar soon.

That's the set up that I have, A Promise (Ultra ATA/100) controller with a WD Caviar 40GB 7200RPM ATA100 as master and a Quantum 12GB 5400RPM ATA33 as slave on channel one.

Just a note, on the WD drive, if its going to be the only drive on that channel, the jumper setting is NO jumpers on it, that had me going in circles for two days.

I've never benchmarked it, looking forward to see your outcome.

It's not necessarily what ya got, but what ya do with it

OK....my three disks....
C drive....16 gig Fat32 ATA33
sequential read: 13.7
sequential write:11.2
random seek: 4.0
disk mark 76.4

D drive....10 gig NTFS ATA33
sequential read: 14.5
sequential write:11.0
random seek: 3.9
disk mark 77.6

And...
H drive....40 gig NTFS ATA 66 [Promise IDE expansion card]
sequential read: 22.9
sequential write:15.2
random seek: 5.6
disk mark 115.4

Not too bad....P111/500 slot1 1gig PC133 ram [running at 100] Abit BX6 MoBo....

Where's this going, and can I join. I only gave half the story above, now this system is a growing thing, and this reflect its current state

Asus P3B-F mobo, PIII/700 slot1, 576mb PC133 ram (running at 100), Promise Ultra ATA100 controller with two channels, Drive C MASTER WD Caviar 40gb ATA100 7200rpm, Drive D SLAVE Quantum 12gb ATA33 5400rpm, Drive E MASTER WD Caviar 20gb ATA100 7200rpm, Drive F SLAVE Quantum 6gb ATA33 5400rpm,
Drive G on the mobo sec IDE channel Plextor 8/4/32A Writer Master, Drive H SLAVE AOpen DVD drive slot feed.

Well there it is, as of now.

Frogboy, Have you tried loading the Intel Application Accelerator, aka Intel IDE driver? Might be worth a shot.

Try seeing if theres an Intel Inf Utility that works under XP as well

Haha, my p75 is even worse:

sequential read: 2.4 M/s
sequential write: 2.4 M/s
random seek/RW: 1.0 M/s
Disk mark: 15.4

That's an old Seagate UDMA66 on this six year old p75.

fat32 by the way.

Frogboy, purchase a new Apple G4 and create a MacCustomize site!

And a proggy called 'Apple Peels'...

hehehe surelly will be the competence to Kaleidoscope (a great windows skinning app for Mac, that skins all, all, all objects on Mac GUI, and don't crash). Unfortunatelly, the Macos 9 crashes unexpectedly! maybe the macosx. Two years ago, when my MagnetoOptic was crashed, also crashed the Hard Disk!!!!!! 6 months of work at trash scan!!!!!! porca miseria

This PII 300mhz, 96meg of ram, 2 four gig hard drive, 3dfx Voodo3 vidio card and a 17inch monitor is out dated about 2 years ago. But it still works. One of these days!