>I've given up on hdparm. With my generic IDE disk, I get 8 MB/sec, and
>about 22 MB/sec if I read from the cache.
I get:
2 root@red:/# hdparm -tT /dev/hd[ab]
/dev/hda:
Timing buffer-cache reads: 64 MB in 1.69 seconds =37.87 MB/sec
Timing buffered disk reads: 32 MB in 3.29 seconds = 9.73 MB/sec
/dev/hdb:
Timing buffer-cache reads: 64 MB in 1.67 seconds =38.32 MB/sec
Timing buffered disk reads: 32 MB in 5.01 seconds = 6.39 MB/sec
>With my 7200 rpm SCSI disks (OK, SCSI-2, not SCSI-3, but
>still...), I get between 1.8 and 3.5 MB/sec (depending on wind
>direction, it seems) reading the disk, and 25 to 33 MB/sec
>reading the cache (which presumably means it's not the
>controller being slower than IDE).
My hdparm results seem very indicative of actual transfer rates
for IDE disks.
>Meanwhile my own informal benchmarks (copy from /dev/xdx to /dev/null) show
>the SCSI disks to be slightly faster than the IDE disk, which seems
>reasonable. Are the hdparm numbers from the same school of thought that
>gave us bogoMIPs ?
If you copy as described above, are you rebooting inbetween
tests, or otherwise somehow emptying any cached data? Just a
thought...
TTYL
-- Mike A. Harris Linux advocate GNU advocate Computer Consultant Open Source advocateThe DVORAK keyboard layout RULES! I memorized it in 45 minutes and I don't think I'm ever going back to QWERTY!
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/