Re: C++ optimization: compile time + memory consumption regression on gcc3.3 branch

[Mike Stump <mstump at apple dot com>]

On Friday, February 28, 2003, at 05:59 AM, Karel Gardas wrote:
> two months or so ago, I've noticed that one of MICO core file is not
> compilable by gcc3.3 - i.e. I've thought that gcc goes into infinite 
> loop
> while compiling it. Today I'm trying to find where is the problem to 
> fill
> proper bugreport. I've found that gcc is not in infinite loop, but just
> that it spend too long time in ``expand'' and consumes too much memory 
> in
> comparison with gcc3.2.2
>
> Rought comparison:
>
> gcc3.2.2: time 20s memory 58MB (reported by top)
> gcc3.3:   time 900s memory about 250MB (reported by top)

If would be helpful to us if you could track this down with something 
like gperf and collect numbers before and after and tell us what 
routines are dying.  If not, just include the .ii in the bug report, 
and hope and pray someone else will do the work.