Andrea Arcangeli (arcangeli@mbox.queen.it)
Thu, 18 Jun 1998 00:45:58 +0200 (CEST)
On Wed, 17 Jun 1998, Tim Waugh wrote:
>Here are the results of that test you asked for.
Hug, they are not very good ;-). At least the CPU usage should be minor.
Some time ago I had an HD that goes very faster with dma disabled
(according to hdparm -t) but the overall performances was highter with dma
enabled since even if the raw disk throughtput was lower, the CPU was more
free (I _hope_ that this will apply also to the current interrupt lp ;-).
>Tim. */
>
>---------- Forwarded message ----------
>Date: Wed, 17 Jun 1998 16:24:05 +0200
>From: Stanislav Brabec <utx@k332.feld.cvut.cz>
>To: tim@cyberelk.demon.co.uk
>Subject: lp speed next check
>
>Hallo Tim Waugh,
>
>I have checked 2.1.105 for printing speed.
>
>The speed was originally cca 130 kB/sec with polling and 93 kB/sec with IRQ.
>
>After tunelp /dev/lp0 -c 10000 printing speed have decreased to
>88kB/sec with IRQ.
>
>My machine is 6x86 MX200.
>
>My parport is configured in BIOS to IRQ7/DMA3 (it can use DMA).
>
>Now I'm upgrading to 2.1.106, so next test will be done on it.
>
>-- Stanislav Brabec
>
Andrea[s] Arcangeli
-- To unsubscribe, send mail to: linux-parport-request@torque.net --
-- with the single word "unsubscribe" in the body of the message. --
This archive was generated by hypermail 2.0b3 on Wed 30 Dec 1998 - 10:17:52 EST