From:	Jason Dagit <thedagit AT mail DOT coos DOT or DOT us>
Newsgroups:	comp.os.msdos.djgpp
Subject:	Re: [Q]Computing speed in C++
Date:	Tue, 03 Mar 1998 23:09:02 -0800
Organization:	Dagit Enterprises
Lines:	34
Message-ID:	<34FCFE0E.FB6A8309@mail.coos.or.us>
References:	<34FCB769 DOT 42BEF1A8 AT gong DOT snu DOT ac DOT kr> <34FCCC11 DOT E5F43542 AT concentric DOT net>
Reply-To:	thedagit AT mail DOT coos DOT or DOT us
NNTP-Posting-Host:	coosbay2-35.transport.com
Mime-Version:	1.0
To:	djgpp AT delorie DOT com
DJ-Gateway:	from newsgroup comp.os.msdos.djgpp

D. Huizenga wrote:
> Bum-Seok Hyun wrote:
> > So I made a test for comparing computing time.
> > The result is
> > c code         : 21 sec
> > c++ code       : 57 sec
> > Fortran77 code : 115 sec
> >
<snip>
Well, compare your C to Fortran time.  If they were said to be close
then C/C++ difference in time would be very close.  Or perhaps, the just
plain C part of GCC is so much faster than the C++ part that it gives
this difference in time.  They are interesting numbers to look at
though.  Perhaps it shows why the newest verion of Netscape is so slow
(I think I heard it is 50% C++ now).

> > Does this mean that C++(or DJGPP) is not suitable
> > for numerical computation?
> >

I think DJGPP is suitiable for numerical computation, just use C++
sparingly if speed is an issue.

 
> I wouldn't know about C++, but DJGPP is certainly suitable for
> computation.  Have you seen Quake? It was written in DJGPP, although I
> am sure a great deal  of assembly was also used.

What did they use to do the windows versions of Quake?  Such as Quake 2?
> --
> Dan Huizenga
> Email: Skis AT Concentric DOT net
> Home: http://www.concentric.net/~skis
> (Not much there right now, except BGui and some [uuhhggg] VB stuff)

- Raw text -