Return to BSD News archive
Path: sserve!manuel.anu.edu.au!munnari.oz.au!sgiblab!swrinde!news.dell.com!natinst.com!cs.utexas.edu!asuvax!ncar!noao!amethyst!organpipe.uug.arizona.edu!afthree.as.arizona.edu!tom
From: tom@afthree.as.arizona.edu (Thomas J. Trebisky)
Newsgroups: comp.unix.bsd
Subject: Re: DMA disk controllers
Message-ID: <1992Oct26.200340.29659@organpipe.uug.arizona.edu>
Date: 26 Oct 92 20:03:40 GMT
References: <1732@optigfx.optigfx.com>
Sender: news@organpipe.uug.arizona.edu
Organization: University of Arizona, Tucson, AZ
Lines: 21
In article <1732@optigfx.optigfx.com> mrm@optigfx.com (Mike Murphy) writes:
>Reading news, .... , I came upon an article that said "ISA bus
>DMA is so slow, nobody oughtta use it". Being a skeptic, I get out an AT(tm)
>Tech. Ref. manual, and go through the numbers. DMA channels 5-7 are 16-bit
>and should be able to do about 1.6MB/sec.
>Then I look further at the circuit for the IBM(tm) hard disk adapter.
>No hard disk DMA. Only floppy disk DMA on channel 2 (8-bit, BTW).
>Awww, nuts, cheapskates,
The fact of the matter is that the CPU can move data faster than DMA
via a "rep stosb" kind of thing. This is really only true when considering
LSI DMA chips (like the 8237). Now if you build a DMA controller with
PAL's or some such, you can do better most likely, but then you are talking
about a pretty fancy (and pricey) disk controller. In the PC volume market
I guess someone decided this was not a viable product.
--
Tom Trebisky ttrebisky@as.arizona.edu
....."There's no sense in being precise when you don't even
..... know what you're talking about."
- John von Neumann