Path: utzoo!utgpu!news-server.csri.toronto.edu!rutgers!usc!elroy.jpl.nasa.gov!jarthur!uunet!virtech!cpcahil
From: cpcahil@virtech.uucp (Conor P. Cahill)
Newsgroups: comp.unix.sysv386
Subject: Re: Disk benchmark (long)
Message-ID: <1990Oct15.233517.19388@virtech.uucp>
Date: 15 Oct 90 23:35:17 GMT
References: <1990Oct9.015948.551@jdyx.UUCP> <1990Oct12.231159.23105@ico.isc.com> <1990Oct13.032355.3176@jdyx.UUCP> <4190@segue.segue.com> <1990Oct15.091904.5439@jdyx.UUCP>
Reply-To: cpcahil@virtech.UUCP (Conor P. Cahill)
Distribution: usa
Organization: Virtual Technologies Inc., Sterling VA
Lines: 27

In article <1990Oct15.091904.5439@jdyx.UUCP> shawn@jdyx.UUCP (Shawn Hayes) writes:
>    I've tried modifying the benchmark to write/read from the raw-disk      
>interface, but I had some problems with it.  I kept getting phys-io errors
>in my tests.  If you've got a sample of code that handles raw-disk io I'd be

I didn't look at the benchmark that you posted, but to read/write from a
raw device all you need to do is write the data in whole blocks to whole
block boundaries.  For most systems the block size is 512.  For some (like
a gould powernode) it is 1K.  If you always write it in multiples of 1K
you satisfy both systems.

There is also an upper limit for the size of a single i/o on the raw deivce
on some systems (Perkin-Elmer has a limit of around 400K).

The part about "write to whole block boundries" means that a given write
must begin at the start of a disk block.  You cant do the following:

	lseek(fd,5,0);
	write(fd,block,blocksize);

Unless, of course, the block size for that device is 5.


-- 
Conor P. Cahill            (703)430-9247        Virtual Technologies, Inc.,
uunet!virtech!cpcahil                           46030 Manekin Plaza, Suite 160
                                                Sterling, VA 22170