Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!seismo!rutgers!mit-eddie!genrad!decvax!decwrl!pyramid!batcomputer!braner
From: braner@batcomputer.tn.cornell.edu (braner)
Newsgroups: comp.sys.atari.st
Subject: Re: I/O speed (was: PD C_compiler & Csh)
Message-ID: <822@batcomputer.tn.cornell.edu>
Date: Sun, 26-Apr-87 14:42:28 EDT
Article-I.D.: batcompu.822
Posted: Sun Apr 26 14:42:28 1987
Date-Received: Mon, 27-Apr-87 00:48:21 EDT
References: <387@sbcs.UUCP> <1935@ihuxy.ATT.COM> <859@viper.UUCP>
Reply-To: braner@tcgould.tn.cornell.edu.UUCP (braner)
Organization: Cornell Theory Center, Cornell University, Ithaca NY
Lines: 11
Summary: Use Fread() with a 9K buffer

[]

I have also achieved much better disk I/O speeds using Fopen, Fread, etc
(instead of the C library fopen, etc), together with a large buffer.
My experiments show that you do _not_ need a 'monster' buffer.  The speed
rapidly levels off after about 4.5K (the size of one SS track) or 9K
(one DS track).  I use 18K in my microEMACS, 4.5K in some utilities (e.g.
the tiny uuxxcodes).  (In 'more' I used 64K, but only to enable going back
to former screens.)

- Moshe Braner