Path: utzoo!attcan!uunet!timbuk!sumac02!tbb
From: tbb@sumac02.cray.com (Todd Blachowiak)
Newsgroups: comp.lang.fortran
Subject: Re: Random Number Functions--Generate versus Input
Message-ID: <140719.13816@timbuk.cray.com>
Date: 12 Nov 90 20:51:50 GMT
References: <1990Nov8.170334.24579@athena.mit.edu> <1990Nov8.215752.7075@athena.mit.edu>
Reply-To: tbb@sumac02.cray.com (Todd Blachowiak)
Organization: Cray Research, Inc.
Lines: 35

In article <1990Nov8.215752.7075@athena.mit.edu>, oliver@athena.mit.edu (James D. Oliver III) writes:
|> I guess I need to make this clear: the random number genertor itself is
|> vectorizable, however, the values of interest to me are *functions* of
|> these randome variables.  To be more specific, I'm generating a series of
|> randomly oriented 3-D unit vectors.  The fastest way to do this is to
|> generate two random numbers for x and y and calculate z as long as x**2 +
|> y**2 <= 1.  This process, however, involves a conditional loop back to the
|> ranf() function and thus is not vectorizable.  However, the fact that this
|> method is not vectorizable is just of incidental interest, since it remains
|> the fastest way of doing it.

would something like this work for you.

     x(i) = ranf()
     y(i) = (1-x(i)*x(i))*ranf()
     z(i) = 1-x(i)*x(i)-y(i)*y(i)

I believe this would vectorize for you and it gets rid of your conditional.
Also, you do not have to make any wasted calls to the random number generator.

|> 
|> What I'd like to know is, *given* this way of generating my values, should
|> I expect it to be faster than reading the values from a file?  Like I said,
|> for this particular case I can run a benchmark, but I was wondering if
|> there was a general rule of thumb for the speed of I/O operations.
|> 
|> 
|> --
|> ____________________________
|> 	Jim Oliver   
|> 	oliver@athena.mit.edu /	joliver@hstbme.mit.edu
|> 	oliver%mitwccf.BITNET@MITVMA.MIT.EDU

-- 
Todd Blachowiak    tbb@cray.com