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: davem at redhat dot com, dewar at gnat dot com
- Cc: gcc at gcc dot gnu dot org, gdr at codesourcery dot com, scott at coyotegulch dot com
- Date: Sun, 19 May 2002 00:39:39 -0400 (EDT)
- Subject: Re: gcc compile-time performance
<<In fact, I found this bug so quickly (and it has been with us for
SO LONG, since February 2001 at the very least) that I really doubt
anyone has had any serious looks at profiles of GCC building something
interesting with -O2.
Well that's certainly a very encouraging result
<<I vow to bring that kind of compilation speed back to GCC,
irrespective of negative depressing comments like those of yours.
Well I am not sure you can fully meet your vow (to get to the speed of
the MIPS compilers) but if you can just get back to 2.8 or even 2.95
that would be pretty encouraging.
I am certainly happy to be proved wrong in my pessimism about improving
compiler performance. The pessimism had come from seeing a steady
decline in performance with not many silver linings, so your result
with CSE certainly qualifies as a silver lining.
perhaps what we need is more sohpisticated profiling. Mark had reported
that there were no hot spots, and you seem to have not only found one
but extinguished it.
But CSE is surely not run at -O0? so I assume this only helps in -O1/-O2
(that's still VERY useful of course).