[6172] in linux-scsi channel archive
RE: how to get maximum datarate from SCSI
daemon@ATHENA.MIT.EDU (Mateu Batle i Sastre)
Thu Mar 25 03:08:23 1999
From: "Mateu Batle i Sastre" <mbatle@iname.com>
To: "Dietmar Stein" <dstein2203@t-online.de>, <linux-scsi@vger.rutgers.edu>
Cc: <dstein@phoenixcontact.com>
Date: Thu, 25 Mar 1999 08:46:09 +0100
In-Reply-To: <36F9C664.C42E4959@t-online.de>
Hi Dietmar,
> I understand your problem but still thinking it has something to
> do with the
> synchronous writing of the data, because if you have "dirty"
> buffers - which is the
> fact by writing data to the cache through an application - they
> will always be
> written immediately to disk.
I definetely This is not the case. I've not tried yet the writing.
First it must work with reading of data from the user application.
But the problem is the same for reading and writing. The cache is
harmful for this type of applications.
> But I think also I am not familiar enough with your problem - on
> the other handside
> we have an application running, which need to be buffered in
> order to get the
> highest performance (database application).
I think our case is much different. Databases need normally access
to the data transfered for processing. Our applications is just
to transfer data from SCSI disks to the memory in the device. The
CPU does not need to access anything in the data. For example in
a multimedia application that transfer video/audio from disks
to PCI card memory. One single transfer can be as large as several
megabytes.
> Never mind - what about the raw I/O patch? Where can I get it?
> And does it mean to
> use raw-devices under linux?
>
The raw patch modifies the kernel to bypass the buffer cache and
transfer directly to/from user buffers.
The patch can be found at ftp://ftp.uk.linux.org/pub/linux/sct/fs.
There are two versions one for kernel 2.2.0 (raw-19990204.tar.gz) and
the other kernel 2.1.131 (976 raw-19990204.tar.gz).
> I would beg to answer on my account at work dstein@phoenixcontact.com (I
am very
> interested in the patch) and on this discussion also.
Happy to know that.
Cheers,
Mateu
-
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@vger.rutgers.edu
