*BSD News Article 16072


Return to BSD News archive

Xref: sserve comp.windows.x.i386unix:1560 comp.os.386bsd.questions:2448
Path: sserve!newshost.anu.edu.au!munnari.oz.au!news.Hawaii.Edu!ames!elroy.jpl.nasa.gov!usc!howland.reston.ans.net!zaphod.mps.ohio-state.edu!caen!hellgate.utah.edu!cc.usu.edu!ivie
From: ivie@cc.usu.edu
Newsgroups: comp.windows.x.i386unix,comp.os.386bsd.questions
Subject: Re: XFree1-2 + 386BSD performance
Message-ID: <1993May12.211743.67745@cc.usu.edu>
Date: 12 May 93 21:17:43 MDT
References: <1993May12.025731.29769@latcs1.lat.oz.au> <1993May12.144311.14744@sifon.cc.mcgill.ca> <1993May12.182849.29379@cm.cf.ac.uk>
Organization: Utah State University
Lines: 18

In article <1993May12.182849.29379@cm.cf.ac.uk>, paul@isl.cf.ac.uk (Paul) writes:
> When doing compiles that access disk a lot my system hangs until the
> disk activity stops. I think what's happening is that disk interrupts are
> crowding out everything else. Is there some way of preventing this, it
> would make it a lot easier to run background compiles. At the moment
> it's like running a single tasking system because when the disk is being
> accessed nothing else happens.
> 
> Incidentally, my drive is an IDE, I get the feeling that this doesn't
> happen with SCSI which is why not everyone sees it.

Welcome to the wonderful world of IDE! That's right, IDE _doesn't_ do DMA;
the CPU has to block-transfer the data to and from the data register. Why?
Because on the stock 6 MHz AT, the block output instruction is significantly
faster than 16-bit DMA (the DMA controller runs at _3_ MHz, you see).

Roger Ivie
ivie@cc.usu.edu