Hi,
> However, the auxiliary math routins in libgcc.a (.so) will be the same, and > not optimized for the given CPU type. > Gcc compiler should be build with "multilib" support enabled, in this case > the build provides specialized libgcc.a and startup .o modules for each > possible "-mcpu", in separate subdirs in > /usr/lib/gcc//gcc-/. Under squeeze : $ apt-file search libgcc.a gcc-4.1: /usr/lib/gcc/x86_64-linux-gnu/4.1/libgcc.a gcc-4.1-multilib: /usr/lib/gcc/x86_64-linux-gnu/4.1.2/32/libgcc.a gcc-4.1-multilib: /usr/lib/gcc/x86_64-linux-gnu/4.1/32/libgcc.a gcc-4.3: /usr/lib/gcc/x86_64-linux-gnu/4.3/libgcc.a gcc-4.3-multilib: /usr/lib/gcc/x86_64-linux-gnu/4.3/32/libgcc.a gcc-4.4: /usr/lib/gcc/x86_64-linux-gnu/4.4/libgcc.a gcc-4.4-multilib: /usr/lib/gcc/x86_64-linux-gnu/4.4/32/libgcc.a So with apt-build you can rebuild libgcc.a with the options you want. > (I wonder - is it the big gain having them optimized?) Only the rigour of experimental tests will tell you :) (raw translation of "Seul de la rigueur expérimentale naîtra la vérité" :) ) > Thanks a lot for reply. You are welcome, xavier -- To UNSUBSCRIBE, email to debian-powerpc-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/ceb8df6f6b9ec8cb91590b19fdcbf...@ipnagenda.in2p3.fr
