From: kunst AT prl DOT philips DOT nl
Subject: Re: Disk I/O rates with DJGPP
To: S DOT D DOT Smith AT ecs DOT soton DOT ac DOT uk (Sean Smith)
Date: Wed, 2 Nov 1994 13:43:52 +0100 (MET)
Cc: djgpp AT sun DOT soe DOT clarkson DOT edu (DJGPP users list)

S D Smith writes:
> 
> A colleague of mine who is also using DJGPP has commented that the disk I/O
> (low level) for DJGPP objects is substantially slower than native 
> Windows or DOS I/O. The comment is that e.g. Zortech C/C++ can produce
> binaries that give block reads and writes of large blocks at up to
> 400kbs-1 while despite the chosen blocksize DJGPP low level reads and
> writes reach a ceiling at around 30kbs-1. Obviously there is some
> h/w dependency here, but I guess the ratios are comparable whatever the
> hardware.
> 
> I'm just off to look at the sources for read/write now, but does anyone
> have any comments on this?
> 

GO32 uses the (real mode) DOS routines for disk I/O.

DJGPP programs that read from/write to disk have their (protected mode)
data transfered to the real mode region using a 4096 byte transfer buffer.
(The exact buffer size might have been altered in recent GO32 versions.)
Then the DOS disk I/O routines are called to do the actual disk access.
There is thus a speed penalty due to data transfer and switching from/to 
protected mode for each chunk.

And no, using setbuf() doesn't help you much...

Pieter Kunst (kunst AT prl DOT philips DOT nl)