Re: [Pan-devel] Re: Insane amount of memory required during compilation

[GISQUET Christophe]

Hello,

Duncan wrote:
> "Charles Kerr" <address@hidden> posted
> > > binary scorefile-test. For the record, this computer has 1GB RAM and 1GB
> > > swap.
> > Yep, this has been reported a couple of times on pan-users IIRC.

It really shows I should check pan-users before posting to pan-devel.

> x86_64 anyway, compiling pan with gcc 3.4.6 required very close to 1.3
> gigs of memory at once point in the process (with the scorefile stuff,with
> my CFLAGS).  Someone else reported this and I confirmed.

I think it indeed aborted at around 1.1GB. I had other memory intensive 
applications paused and flushed to swap, which explains it failed with my 
1+1GB setup.

> EULA), this now confirms the problem thru gcc 4.0.x and with at least
> 32-bit MSWormOS.  To my knowledge, we still haven't confirmed it on 32-bit
> x86 Linux, however.

Scratch my report, I thought CXX=g++-4.0.2 make would achieve what I expected 
but Makefile has g++ hardcoded. I launched again configure so as to use the 
proper version of g++ and it did compile without problem.

Once stripped the 4.0.2 binary size is 2196K compared to 2544 with 3.4.5. This 
shows a rather noticeable difference in inlining and probably template 
compilation management.

> FWIW, the only other package I can personally verify that  requires that
> sort of compile-time resources is KDE's kmail (also C++ based).  However,

I know a lib that was compilable under another compiler. While it would take 
like 30s with that compiler, g++ was taking around 8 minutes on the same 
computer and probably twice as more memory, although I never checked exactly 
the later figure.

> it still requires them with new gccs, while pan's requirement goes down
> substantially with gcc 4.1.x.

Too bad g++ often has regressions (and I'm not mixing this up with strictness 
increase).

Thanks for the heads up,
Christophe GISQUET