This is the mail archive of the
mailing list for the GCC project.
Re: gcc compile-time performance
- From: dewar at gnat dot com (Robert Dewar)
- To: dewar at gnat dot com, gdr at codesourcery dot com
- Cc: gcc at gcc dot gnu dot org, scott at coyotegulch dot com
- Date: Sat, 18 May 2002 12:02:01 -0400 (EDT)
- Subject: Re: gcc compile-time performance
> The view that GCC is slow and should remain slow is ridiculous.
The view that GCC is slow is hardly ridiculous, it is (as you apparently
emphatically agree) the case!
The view that it will remain slow is merely realistic. I don't see any way
of significantly improving compile time performance that is realistic. For
me a significant improvement would be a factor of several.
If it keeps getting slower by 20-30% a year, that does not bother me, since
that is overwhelmed by an increase in hardware speed (this rate of increase
applies to all environments).
Now I would say that if it gets 20-30% slower and I see nothing in return
then that's worrisome, but if I see in return greater conformance to the
standard (it would have been nice if GCC had been ahead on all sets of
tests in Dr. Dobbs), or significantly better run time performance, then
I consider that fine. As a result of hardware improvements I can expect
to see an absolute improvement of perhaps a factor of 2 every year. And
again that applies to all environments, even in the incompetently run
french university, those old HP machines won't last for ever, and even
if they manage to replace them with some other obsolete junk, it won't
be *as* obsolete :-)