GCC 3.3 compile speed regression - AN ANSWER

[Robert Dewar]

> All of my measurements suggest that improving the speed of optimization
> phases is irrelevant.  What's slow is the front end: parsing and 
> semantic analysis.  -O0 compilations are slow, not just -O3.

Well here is a data point that contradicts this conjecture. GNAT compilation
has also slowed down significantly. I will try to post some figures in the near
future. Since the GNAT front end is running on 2.8.1 and 3.2.x we have a direct
comparison which for sure does NOT include any influence from the C or C++
front end!