C++ compile time (again)

[Steven Bosscher]

Wolfgang Bangerth wrote:

>It's C++-compile-time-regression-rant time again. Compile times for a
>complete build of the deal.II library on a PIII 800 MHz:
>
>  gcc2.95 :  46 min
>  mainline:  89 min
>  tree-ssa: 118 min
>
>The last two numbers are generated with --enable-checking, but apart
>

--enable-checking makes a very significant difference, so you're not 
making fair comparisons here.  Do you have the numbers with checking 
disabled?

tree-ssa, well, by now we all know it needs a few nitro boosts before 
it's allowed to go on the freeway.  People are working on that.

>from that I don't think further comments are necessary. Switching to a
>new compiler can be frustrating if you have to wait 2 hours when you
>know it could be less than one...
>
>Now, what troubles me (and Gerald, who has been pointing this out for
>ages now) is that it seems as if these numbers don't change
>anything. Everytime Gerald brings this up, answers are "Oh, yes, good 
>you remind us that there is C++ out there, we seem to have forgotten".
>
That is not fair to, say, the CodeSourcery people.

>Why is it that everyone is always looking at C code for compile time?
>
If C speeds up, it usually helps _all_ languages, so speedin gup C is a 
good thing :-)

>It just doesn't make sense if we constantly surprise ourselves by
>pointing out the existence of real-life C++ codes. Testing compile
>time speed for C++ needs to be done as a normal part of gcc
>development. It's just impossible to figure out what made gcc slow if
>
Yes but how?   Is anyone testing/publishing C++ compile time performance 
on a regular basis?  There are three SPEC testers and an autocrasher, 
but there's not enough C++ in there, as we've seen many times now...

Gr.
Steven