Tremendous increase in compile times for 3.4.1 with -mcpu=G5

[Bradley Lucier <lucier at math dot purdue dot edu>]

It seems that specifying -mcpu=G5 instead of -mcpu=G4 leads to a 
tremendous increase in compile time for 3.4.1.

Here's the compiler I'm using:

[descartes:~/programs/gambc40b4-devel] lucier% gcc -v
Reading specs from 
/pkgs/gcc-3.4.1/lib/gcc/powerpc-apple-darwin7.5.0/3.4.1/specs
Configured with: ../configure --prefix=/pkgs/gcc-3.4.1
Thread model: posix
gcc version 3.4.1

Here's a typical compile command in building Gambit-C:

gcc -mcpu=G4 -I../include -no-cpp-precomp -Wall -W -Wno-unused -O1 
-fno-math-errno -fschedule-insns2 -fno-trapping-math 
-fno-strict-aliasing -fomit-frame-pointer -fPIC -fno-common 
-DHAVE_CONFIG_H -c _gsc_.c

The time for make -j with -mcpu=G5:

633.250u 15.960s 9:57.25 108.6% 0+0k 1+101io 0pf+0w

The time for make -j with -mcpu=G4:

107.190u 13.420s 1:23.98 143.6% 0+0k 0+63io 0pf+0w

A typicial bad file can be found at

http://www.math.purdue.edu/~lucier/GNATS/GNATS-12/_num.i.gz

I filed a PR: 16975.

Brad

PS:  I'd report it against 3.5 on powerpc, but I haven't been able to 
get 3.5 to bootstrap for a while, when I can get it to bootstrap I 
often can't get it to build my applications, and the few times it has 
built my applications it generates incorrect code.  It's somewhat 
frustrating.