From:	tyoung AT netaxis DOT com (tom young)
Subject:	Re: Math errors
Newsgroups:	comp.os.msdos.djgpp
References:	<6v30a8$1uc$1 AT news2 DOT saix DOT net> <36159F6E DOT 244FC966 AT cartsys DOT com>
X-Newsreader:	TIN [version 1.2 PL2]
Lines:	32
Message-ID:	<NGpR1.1674$O33.3450147@news6.ispnews.com>
NNTP-Posting-Host:	198.69.103.4
X-Trace:	news6.ispnews.com 907421485 198.69.103.4 (Sat, 03 Oct 1998 09:31:25 EDT)
NNTP-Posting-Date:	Sat, 03 Oct 1998 09:31:25 EDT
Organization:	ISPNews http://ispnews.com
Date:	Sat, 03 Oct 1998 13:31:25 GMT
To:	djgpp AT delorie DOT com
DJ-Gateway:	from newsgroup comp.os.msdos.djgpp

I get the following on both msdos/DJGPP and Linux/gcc ( with ric, which
uses math.h doubles: 

precision(7)
P2=4*asin(1);\n
P2;\n; sin(100000000*P2);\n;sin(10000000*P2);\n
6.2831853
-0.0000001
0.0000000

Not too shabby!  Your nephew must be using enormous numbers.  Probably
avoidable.

Nate Eldredge (nate AT cartsys DOT com) wrote:
: Rylan wrote:
: > 
: > Hi,
: > 
: > My nephew uses both DJGPP and Watcom, and recently complained that the DJGPP
: > C math COS/SIN functions start returning invalid values (larger than -1 and
: > 1) for large input values - he says that Watcom and real mode Pascal returns
: > correct values for inputs to their SIN/COS functions for which DJGPP returns
: > invalid values.
: > 
: > Any idea where the problem is? His email: grigi AT acenet DOT co DOT za

: Post an example?
: -- 

: Nate Eldredge
: nate AT cartsys DOT com

- Raw text -