Path: utzoo!utgpu!news-server.csri.toronto.edu!rutgers!cs.utexas.edu!uunet!mcsun!hp4nl!kunivv1!atcmpe!jc
From: jc@atcmp.nl (Jan Christiaan van Winkel)
Newsgroups: comp.arch
Subject: comments to benchmark and results
Message-ID: <609@atcmpe.atcmp.nl>
Date: 2 Jun 90 14:50:45 GMT
References: <543@iss-rb.SanDiego.NCR.COM>
Organization: AT Computing, Nijmegen, The Netherlands
Lines: 21

From article <543@iss-rb.SanDiego.NCR.COM>, by jan@ivory.SanDiego.NCR.COM (Jan Stubbs):
] In article <15049@cit-vax.Caltech.Edu> fritz@vlsi.caltech.edu (fritz nordby) writes:
] which measures Unix kernel performance. 
] 
]  for (i=0;i<=10000;i++) {
]   	/*do seek, write, seek, read, read, read. */
]  	lseek(fd,0L,0);		
]  	count = write(fd,buf,500);
]  	lseek(fd,0L,0);
] 
]  	for (j=0;j<=3;j++) 
] 	 	count = read(fd,buf,100);
NO. You do seek, write, seek, read, read, read, read (4 reads!)
What are your results based on? 4 reads or 3 reads?

BTW:
results of run (with program as posted) on Altos 2086, 80286 @ 8 MHz, 
Xenix 3.3as1: real 75 sec, user 2 sec, system 72 sec.
-- 
Jan Christiaan van Winkel              Tel: +31 80 566880     jc@atcmp.nl
AT Computing       P.O. Box 1428       6501 BK Nijmegen       The Netherlands