From: Kevin AT Quitt DOT net (Kevin D. Quitt)
Newsgroups: comp.os.msdos.djgpp
Subject: Re: Is DJGPP that efficient?
Date: Thu, 19 Dec 1996 17:15:23 GMT
Organization: Cruisin' from home
Lines: 27
Message-ID: <32c7778e.47755048@news.pacificnet.net>
References: <Chameleon DOT 850715753 DOT Kevin AT quitt> <32B50729 DOT 4D9A AT gbrmpa DOT gov DOT au> <850935362 DOT 27417 DOT 0 AT abwillms DOT demon DOT co DOT uk>
Reply-To: Kevin AT Quitt DOT net
NNTP-Posting-Host: 207.171.23.21
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
To: djgpp AT delorie DOT com
DJ-Gateway: from newsgroup comp.os.msdos.djgpp

On Wed, 18 Dec 1996 18:56:23 GMT, alaric AT abwillms DOT demon DOT co DOT uk (Alaric B.
Williams) wrote:


>>> it doesn't implement the basic trig function set (partial tangent), not
>>> to mention hyperbolics, or something as useful as simultaneous sin and cos.
>
>From my INTEL assembly format textbook:
>
>FPATAN: Partial Arctangent
>FPTAN: Partial Tangent
>FSINCOS: Sine and Cosine

And how much work do you have to do before the FSINCOS will actually work
on the data?  That's one of the Intel FPU's problems:  if the data aren't
within the correct range, you don't get a valid answer.  So what should be
a single instruction (and is on EVERY other machine) turns out to be an
entire subroutine.

Intel's problem all along:  the individual instructions are fast, but you
have to do a lot of work to use them.  Where's the gain?

--
#include <standard.disclaimer>
 _
Kevin D Quitt  USA 91351-4454           96.37% of all statistics are made up
Per the FCA, this email address may not be added to any commercial mail list