From owner-freebsd-questions Tue Oct 14 23:34:53 1997 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.7/8.8.7) id XAA21854 for questions-outgoing; Tue, 14 Oct 1997 23:34:53 -0700 (PDT) (envelope-from owner-freebsd-questions) Received: from shell.futuresouth.com (shell.futuresouth.com [207.141.254.20]) by hub.freebsd.org (8.8.7/8.8.7) with ESMTP id XAA21845 for ; Tue, 14 Oct 1997 23:34:50 -0700 (PDT) (envelope-from fullermd@futuresouth.com) Received: from shell.futuresouth.com (mail.futuresouth.com [207.141.254.21]) by shell.futuresouth.com (8.8.5/8.8.5) with SMTP id BAA00200; Wed, 15 Oct 1997 01:32:24 -0500 (CDT) Date: Wed, 15 Oct 1997 01:32:23 -0500 (CDT) From: "Matthew D. Fuller" To: Don Wilde cc: Marty Leisner , Jacques Hugo , questions@FreeBSD.ORG Subject: Re: mmx or ppro In-Reply-To: <3443CA23.2697@PartsNow.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-questions@FreeBSD.ORG X-Loop: FreeBSD.org Precedence: bulk On Tue, 14 Oct 1997, Don Wilde wrote: > Since no software is written for MMX extensions, a PPro will work better > than a Pentium/MMX, but a PentiumII will beat either, since a) it has > larger caches and b) it has faster clock rates. Even if the software > were compiled with MMX, only the X/graphics code would see improvement. > There are very few situations where the MMX instructions are helpful, > and there are some times when it could be actively harmful. Consider, also, that PRO isn't a technology; it's just the cutesy name for the 80686, since Intel can't patent numbers. And MMX is a set of instructions, not something Intel owns; Cyrix and AMD use it too. A PPro has 5 (I believe) instruction pipelines, as opposed to the Pentium's 2, so it runs a fair bit faster. Pentium II has faster clock speeds and a slot interface, so I believe it has a wider CPU bus, but it also has half-speed L2 cache, so it doesn't always beat either; I've seen benchmarks where a P-200MMX beat a PII-233 on MMX-related benchmarks, and a PPro200 (overclocked to 233) beat a PII-266 on mathematical calcs. HOWEVER: MMX chips also have improved speculative execution, and a few other fuzzy logic functions, so a P200MMX will run a little fast on a few things than one without MMX; but that's tech progression independant of MMX AFAIK. > MMX basically refers to the ability to use the floating-point registers > to do simple adds and bit-masks to 8 bytes at a time instead of one > 80-bit floating point number. 99.4% of all code doesn't fit this > application. > -- > oooOOO O O O o * * * * * * > o ___ _________ _________ ________ _________ _________ ___==_ > V_=_=_DW ===--- Don Wilde [don@PartsNow.com] [http://www.PartsNow.com ] > /oo0000oo-oo--oo-ooo---ooo-ooo---ooo-ooo--ooo-ooo---ooo-ooo---ooo-oo--oo *-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-* | FreeBSD; the way computers were meant to be | * "The only reason I'm burning my candle at both ends, is * | that I haven't figured out how to light the middle yet."| * fullermd@futuresouth.com :-} MAtthew Fuller * | http://keystone.westminster.edu/~fullermd | *-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*