[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Pgcc in Deb

On Mon, Apr 03, 2000 at 04:17:51AM -0700, Joseph Carter wrote:

> We all know how screwed up the former is.  The latter has (and has had for
> some time) several very obnoxious bugs which result in bad code on certain

Definately - pgcc should be approached with some caution.  It's also
been known to play up with XFree86 and Mesa.

> non-trivial applications.  Those patches and improvements found in pgcc
> get added to egcs as soon as they are known not to do stupid things anyway

That's not entirely true - some of the stuff in pgcc isn't in mainline
gcc because there is no copyright assignment on file and no likelyhood
of getting one.  Other bits are probably stable enough on Intel but need 
work to avoid pessimising or bugs on other architectures.

> so it's worthwhile simply to stick with egcs and use the optimizations it
> provides.

> In almost all cases you will not be seeing any noticable difference in
> execution speed.

Depends on the program, of course.  For all the effort it takes, it's
often worth trying a pgcc build of compute-intensive programs, although
you should be able to check that it is correctly optimized - not to
mention do some benchmarking to verify that you're actually getting
something worthwhile from the exercise.

Most programs aren't that compute-intensive, and gcc is more reliable -
just blindly using pgcc is probably not a good idea.

Mark Brown  mailto:broonie@tardis.ed.ac.uk   (Trying to avoid grumpiness)
EUFS        http://www.eusa.ed.ac.uk/societies/filmsoc/

Reply to: