Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!seismo!nbires!hao!hplabs!pyramid!decwrl!glacier!navajo!rokicki
From: rokicki@navajo.STANFORD.EDU (Tomas Rokicki)
Newsgroups: net.micro.amiga
Subject: Re: File random access times.
Message-ID: <810@navajo.STANFORD.EDU>
Date: Sat, 30-Aug-86 20:11:38 EDT
Article-I.D.: navajo.810
Posted: Sat Aug 30 20:11:38 1986
Date-Received: Mon, 1-Sep-86 18:58:21 EDT
References: <8608281956.AA19132@cory.Berkeley.EDU>
Organization: Stanford University
Lines: 19
Summary: Short seeks

In article <8608281956.AA19132@cory.Berkeley.EDU>, dillon@CORY.BERKELEY.EDU (Matt Dillon) writes:
>             . . .  Additionaly, if you are making small relative
> seeks, buffered I/O wins because most of the time the data is already in
> the buffer and it doesn't have to make a system call.

It would be nice if this were true; unfortunately, the fseek() call in
the Manx library doesn't seem to have been written with this in mind.
I had to rewrite fseek() it for a few programs I have to make things
reasonably fast.  (fwrite() and fread() also have a bug and require
rewriting.)

Does anyone have any explanation for the random inconsistencies in the
data Ali Ozer presented on access times?  Things are wierd, there.

Also, I want a hard drive with fast transfer rates!  When loading an
executable of > 100K (as a lot of Amiga executables are), that five
to ten second delay can be very annoying.  Any pointers?

-tom