Re: IDE drive data transfer rates, kernel optimization?

Meelis Roos (mroos@tartu.cyber.ee)
Fri, 7 Aug 1998 15:56:21 +0300 (EEST)

> I'd be suprised if you can measure it. The real reason for the IDE unmasking
> is that a few very old controllers choked when they didnt get a continual
> bitstream for the entire IDE transfer. Basically instead of blocking while
> the interrupt service routine caused a delay they seemed to carry on making
> up bytes for themselves

On my current hardware I can't measure it. But on an older computer
(100 MHz AMD 486 w/ PCI MB and 1.2G WD hard disk) I can:

root@krahv:~>hdparm -u /dev/hda

/dev/hda:
unmaskirq = 0 (off)
root@krahv:~>hdparm -t /dev/hda

/dev/hda:
Timing buffered disk reads: 16 MB in 12.42 seconds = 1.29 MB/sec
root@krahv:~>hdparm -u1 /dev/hda

/dev/hda:
setting unmaskirq to 1 (on)
unmaskirq = 1 (on)
root@krahv:~>hdparm -t /dev/hda

/dev/hda:
Timing buffered disk reads: 16 MB in 15.41 seconds = 1.04 MB/sec

This is repeatable. And the test was done when there were no users logged
in on any serial lines. And I think I stopped using irqtune a while ago
because I saw no improvement so this doesn't disturb the test.

---
Meelis Roos (mroos@tartu.cyber.ee)

-
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.altern.org/andrebalsa/doc/lkml-faq.html