home *** CD-ROM | disk | FTP | other *** search
- Path: news.primenet.com!pdtaylor
- From: pdtaylor@primenet.com (Daniel L. Taylor)
- Newsgroups: comp.sys.amiga.hardware
- Subject: Re: Amiga vs. PC
- Date: 10 Mar 1996 18:43:02 -0700
- Organization: Primenet Services for the Internet
- Sender: root@primenet.com
- Message-ID: <4i00f6$msh@nnrp1.news.primenet.com>
- References: <rdingem.4k6k@grafix.xs4all.nl>
- X-Posted-By: pdtaylor@usr1.primenet.com
-
- rdingem@grafix.xs4all.nl (Ruud Dingemans) writes:
-
- >In a message of 06 Mar 96 Dave Haynie wrote to All:
-
- > DH> Truth is, in most applications you don't need a 64-bit processor (though
- > DH> I'm sure some bozo said that about 32 and 16 bit processors years ago).
- > DH> The larger word size doesn't buy you any extra speed doing most personal
-
- >Eh, Dave, I'm afraid I don't quite get this. I can see the
- >non-usefulness concerning the 4+ GB address space, but if it makes
- >a difference shoveling 32 bits of data in and outta memory at a
- >time (instead of 16), why shouldn't 64 bit instead of 32 make a
- >real difference?
-
- This doesn't require Dave to answer. What do most personal computers, any
- flavor, spend most of their time doing? Waiting for user input. Whether
- I wait for user input with a 2 MHz 6809 or a 150 MHZ Pentium doesn't make
- much difference, does it? Beyond that, as long as the computer connected
- to the keyboard/mouse/monitor in front of you can present responses to
- your direct input in a timely (somewhat personal perference) manner, it's
- fast enough. For the relatively few who do a lot of floating point
- math, including rendering, numerical analysis, etc., the wider data bus is
- a win.
-
- The other limiting factor in desktop computer use is I/O performance.
- Just as waiting for user input doesn't require much of a processor,
- neither does waiting the 9-14 milliseconds for a disk to seek (or
- 135-350 ms for a CDROM drive), and even a 6809 can keep up with the
- transfer requirements of all but the fastest disks, much less the
- lower throughput and overhead of 28800 telecom, if the peripheral port
- itself can handle the traffic. Also, very few peripherals have a
- "funnel" to adapt port width to processor width. An ATAPI port is
- 16 bits; serial ports, except a few, are 8 bits; SCSI may be 8 or 16;
- DMA-style devices for disk may access your memory at processor bus
- width (the A3000 does, bus-master VESA does), but most don't, which
- is part of Dave's point. Desktop computers spend a lot of their CPU
- cycles services narrower I/O peripherals.
-
- This is why the 16-bit Amiga's can still provide seriously useful
- service in an era when 64-bit busses are becoming readily available.
- --
- -------------------------------------------------------------------------------
- pdtaylor __ __ ____ ___ ___ ____
- pdtaylor@primenet.com /__)/__) / / / / /_ /\ / /_ /
- / / \ / / / / /__ / \/ /___ /-------------------------------------------------------------------------------
-