Xref: news-dnh.mv.net comp.os.msdos.djgpp:4354
From: Charles Sandmann <sandmann AT clio DOT rice DOT edu>
Newsgroups: comp.os.msdos.djgpp
Subject: Re: Speed of 387 FPU emulator!
Date: Sun, 14 Jan 1996 12:05:47 CST
Organization: Rice University, Houston, Texas
Lines: 17
Message-ID: <30f945fb.sandmann@clio.rice.edu>
References: <4d88lp$jrb AT bagan DOT srce DOT hr>
Reply-To: sandmann AT clio DOT rice DOT edu
NNTP-Posting-Host: clio.rice.edu
To: djgpp AT delorie DOT com
DJ-Gateway: from newsgroup comp.os.msdos.djgpp

> I was wondering is it normal for 387 FPU emulator ing DJGPP to be so slow?
> I made a program that calculates 5000x sinus function, and it took more that
> 30 seconds! I made the same program in Borland C and it took less than a
> second!

I would say this performance is abnormal, EMU387 is usually no more than a 
factor of 2X slower.

If performance and accuracy is important for FPU emulation, you probably 
should get the WMEMU package.  It is better than the Borland libraries, and is
the code used in Linux. It does have much more restrictive requirements
on re-distribution, however.

Improving any FPU emulation is a dead issue.  When 486 processors with FPUs
are $25 (and 387s are essentially free), and talented programmers going for
$50+ an hour, it just doesn't make any sense.  The source is available if
it does make sense to someone else.