From: MALET JL <malet.jean-luc@laposte.net>
To: Jeff Garzik <jgarzik@pobox.com>
Cc: linux-kernel@vger.kernel.org
Subject: Re: [drivers][sata-promise] TX4 has the cache enabled, it should be disabled
Date: Fri, 20 Feb 2004 14:47:09 +0100 [thread overview]
Message-ID: <40360FDD.1070607@laposte.net> (raw)
In-Reply-To: <40354871.1060009@pobox.com>
[-- Attachment #1: Type: text/plain, Size: 1697 bytes --]
Jeff Garzik a écrit :
> MALET JL wrote:
>
>> hello,
>> I've a TX4 150 card and encounter the following effects :
>> ->"shorts" files (<1Mo) are copied at 35Mo/s av
>> ->"long" files (>1Mo) are "burst" copied (ie a big "burst" then
>> hangs, burst, hangs) this has the following effects :
>> 1) average fall back to 12Mo/s
>> 2) systems "hangs" (mouse, keyboard behave like a 100% used cpu) but
>> cpu usage is still slow
>> 3) mplayer reset the streams every second (cpu usage still low)
>>
>> this is 1-2 week that I have this issue, but can't find where it
>> comes from...... yesterday by chance when I loaded sata-promise I
>> noticed "write through" this remembered be some issue I go when I
>> first had the sata board, making a short research made me remember :
>>
>> DISABLE THE BOARD CACHE! on windows this is the same when cache is
>> enabled : performance drops, system interactive is awfull....
>
>
>
> There is no on-board cache on the TX4.
>
> Further, you are thinking about the SCSI layer's caching support, i.e.
> basically whether the driver supports SYNCHRONIZE CACHE (aka
> flush-cache) scsi command.
>
> Jeff
>
I relooked into manual (see promise site support, user manual) and
there are two parameters for TX4 : write through/ write back , cache on
cache off..... I don't know which cache is it about but I'll redo some
test (argh you oblige me to reboot to windows! did you thought of my
uptime! LOL )this evening but as far as I remember the optimal values
where write through /no cache.....
on windows with cache on their was small hangs too....
I noticed then that a prempible kernel is not a good idea in my case :
lot more "small hangs"......
JLM
[-- Attachment #2: Type: application/pgp-signature, Size: 250 bytes --]
next prev parent reply other threads:[~2004-02-20 13:53 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2004-02-17 9:13 [drivers][sata-promise] TX4 has the cache enabled, it should be disabled MALET JL
2004-02-19 23:36 ` Jeff Garzik
2004-02-20 13:47 ` MALET JL [this message]
2004-02-20 14:32 ` Jeff Garzik
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=40360FDD.1070607@laposte.net \
--to=malet.jean-luc@laposte.net \
--cc=jgarzik@pobox.com \
--cc=linux-kernel@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox