ant-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Steve Loughran" <stev...@iseran.com>
Subject Re: Problem compiling Fortran files using gcc compiler and cpptasks
Date Fri, 30 Aug 2002 16:52:05 GMT

----- Original Message -----
From: "Martin Gainty" <mgainty@hotmail.com>
To: <ant-dev@jakarta.apache.org>
Cc: <steve_l@iseran.com>
Sent: Friday, August 30, 2002 9:23 AM
Subject: Re: Problem compiling Fortran files using gcc compiler and cpptasks


> Steve:
>
> Isnt it better for her to bite the OOA bullet and start deisgning classes

depends on the quantity of code *and* the skillset of the developers. The
reason physisicsts use Fortran is that is what they have always used. Though
there is a niche of java for physics.

> (I think Vector would make a great abstract class..)

performance. Imagine you have an 'event' that takes place thousands of times
a sec, you need to analyse it enough to decide if it is 'interesting' enough
to be worth recording. Kind of equivalent of what Echelon does with voice
calls; the more real time prefiltering you can do the better. Then once you
have the data you need to analyse it as much as possible. Fortran has led
the way in vectorizing compilers, though the codeplay C/C++ compiler for
win32 looks like it can do a good job of vectorizing stuff for the MMX/SSE
registersets.


> Is there enough support for Fortran to justify the effort to incorporate
support for Ant??

There is some fortran support in <cc> already; someone wanted it.



--
To unsubscribe, e-mail:   <mailto:ant-dev-unsubscribe@jakarta.apache.org>
For additional commands, e-mail: <mailto:ant-dev-help@jakarta.apache.org>


Mime
View raw message