This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: A user's experiences with GCC-3.4 snapshots


On Tue, Sep 16, 2003 at 09:44:29PM +0200, Karel Gardas wrote:
> On Tue, 16 Sep 2003, Art Haas wrote:
> 
> > Parrot: GCC-3.4 has a _huge_ slowdown in compiling the 'core_ops_cg.c'
> > file - a C file with a large number of computed goto statements. In
> > GCC-3.2 this file compiled but required a huge amount of memory. GCC-3.3
> > fixed both of these shortcomings in 3.2 - the file compiled much more
> > quickly and with significantly less memory. GCC-3.4 has regressed
> > time-wise though, as the file now takes 5x or 6x as long to compile as
> > it does with GCC-3.3. There is also a compiler error generated when the
> > /usr/include/bits/mathinline.h file gets included, but I've lost my
> > build log with the error message. Grrr ...
> 
> Something like:
> 
> /usr/include/bits/mathinline.h: In member function `void FixedBase::adjust(Boolean)':
> /usr/include/bits/mathinline.h:429: error: can't find a register in class `FP_TOP_REG' while reloading `asm'
> 

I think that is it. The 'FP_TOP_REG' looks familiar ...

> If so, then I have also hit this issue with MICO (http://www.mico.org).
> http://gcc.gnu.org/ml/gcc/2003-09/msg00192.html
> 
> In addition, I think there is also memory consumption regression during
> compilation of some C++ code in gcc3.4. For example compiling file
> mico/orb/dii.cc from MICO, compiler consumes about 300MB of RAM while gcc
> 3.2-3.x was enough with 80-130 MB IIRC.
> 

I've seen more memory usage with both C and C++ code, though I don't
recall seeing anything as dramatic as that. The compiler is a little
slower that GCC-3.3 also, but that is an issue brought up by many people
over and over again, and I know the developers here enough about that.
The parrot example seems to me to be a good example of a C file that 
displays a dramatic slowdown for a particular code construction, and 
makes a good testcase for testing computed goto functionality in the
compiler.

Art

-- 
Man once surrendering his reason, has no remaining guard against absurdities
the most monstrous, and like a ship without rudder, is the sport of every wind.

-Thomas Jefferson to James Smith, 1822


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]