SATA drive "super" performance in FC3

Peter Teuben teuben at astro.umd.edu
Wed Nov 10 16:54:21 UTC 2004


FC3 comes with a 2.6.9-1 kernel, and on my SATA controller

00:1f.1 IDE interface: Intel Corp. 82801EB/ER (ICH5/ICH5R) IDE Controller (rev  02)
00:1f.2 IDE interface: Intel Corp. 82801EB (ICH5) SATA Controller (rev 02)

I'm measuring a rather curious ever-increasing disk performance. Repeatedly
giving the command
	hdparm -t /dev/sda
I get to see:

Timing buffered disk reads:  156 MB in  3.03 seconds =  51.41 MB/sec
Timing buffered disk reads:  304 MB in  3.01 seconds = 101.15 MB/sec
Timing buffered disk reads:  440 MB in  3.00 seconds = 146.54 MB/sec
Timing buffered disk reads:  572 MB in  3.01 seconds = 190.13 MB/sec
Timing buffered disk reads:  686 MB in  3.02 seconds = 227.11 MB/sec
...

of course this can't be right. Also, if the first estimate is correct/accurate,
it's slightly less than the 55 i measured on a 2.4 kernel with RHEL3!

I should also add that unlike in FC3-test3, i'm now getting an error message:

# hdparm -t /dev/sda
 
/dev/sda:
 Timing buffered disk reads:  862 MB in  3.02 seconds = 285.00 MB/sec
BLKFLSBUF failed: Operation not supported
HDIO_DRIVE_CMD(null) (wait for flush complete) failed: Operation not supported


which may or may not be related to the ever increasing disk i/o speed.
I suspect it's the kernel, but I would appreciate if poeple could do
this test repeatedly on their SATA, SCSI and/or IDE drives to see what
bucket isn't being reset.

- peter




More information about the fedora-list mailing list