Date: Fri, 30 May 1997 08:54:52 -0400 (EDT) From: "Mike A. Harris" Reply-To: "Mike A. Harris" To: OpenDOS discussion list Subject: Re: 16 vs. 32-bit performance In-Reply-To: <199705151828.OAA00071@delorie.com> Message-ID: Organization: Total disorganization. MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Precedence: bulk On Thu, 15 May 1997, Jonathan E. Brickman wrote: > Most programs still read in files byte-per-byte, Programs read byte for byte or int for int, but they certainly don't read it directly from disk sectors, they read them from disk buffers in the operating system. The OS reads in several sectors at a time and buffers them. It is impossible to read a single byte from the hard disk. The interface does not work that way. The whole 16/32 bit argument is retarded. If 16 bit code was faster than 32 bit code, then all games and other software would remain 16 bit. Extending the argument even further - 8 bit would be faster than 16 bit. Both are obsurd. Another point is that in 32bit Pmode 32bit ops and 8 bit ops run fast, but 16 bit has a penalty. This makes 16bit code even slower. Likewise in Real mode, 32bit registers, etc have the penalty, but the process of doing 32 bits at a time WITH penalty is still faster than 2 separate 16 bit operations in almost all cases. I refer you to Michael Abrash's "Zen of Assembly Language", and "Zen of Code Optimization", and "Zen of Graphics Programming" which go into optimization techniques in much detail. 32bit OS's and programs are in whole much faster than their 16bit counterparts. Nuff said. No reply needed, nor expected. Mike A. Harris | http://blackwidow.saultc.on.ca/~mharris Computer Consultant | Coming soon: dynamic-IP-freedom... My dynamic address: http://blackwidow.saultc.on.ca/~mharris/ip-address.html Email: mharris at blackwidow.saultc.on.ca <-- Spam proof address URL: Digital Equipment Corp. http://www.digital.com