Reply to thread

Message: [QUOTE="Larry Evans, post: 4236472"] [snip] Thanks Robert. Now I'm wondering why, given a contiguous set of tag values, a vector of function pointers wouldn't work just as well as a switch. So, I changed the .mk file to include another macro, FUN_VEC, and added another macro with the same name to the .cpp file to chose between the switch or vector of function pointers implementation. The two .s output files are attached. It appears, in the case of: switch.U.10.s the the jump table stores a vector of labels, and the switch jumps to the label and then calls the function; whereas in: switch.D.10.s the function pointers are stored in fun_vec and then an index into fun_vec is made and then the call is made: movq fun_vec(,%rax,8), %rax call *%rax This seems to me to be about the same speed. I would even think that the fun_vec implementation would be a bit faster since indexing into an array, then jumping to an instruction address would seem to me faster than jumping to a instruction address and then calling a function at that instruction. Of course, I admit I know very little about how assembly works :( I'd appreciate any insights you might have. My system is a pc and the g++ -v shows: g++ -v Using built-in specs. Target: x86_64-linux-gnu Configured with: ../src/configure -v --with-pkgversion='Ubuntu 4.4.3-4ubuntu5' --with-bugurl=file:///usr/share/doc/gcc-4.4/README.Bugs --enable-languages=c,c++,fortran,objc,obj-c++ --prefix=/usr --enable-shared --enable-multiarch --enable-linker-build-id --with-system-zlib --libexecdir=/usr/lib --without-included-gettext --enable-threads=posix --with-gxx-include-dir=/usr/include/c++/4.4 --program-suffix=-4.4 --enable-nls --enable-clocale=gnu --enable-libstdcxx-debug --enable-plugin --enable-objc-gc --disable-werror --with-arch-32=i486 --with-tune=generic --enable-checking=release --build=x86_64-linux-gnu --host=x86_64-linux-gnu --target=x86_64-linux-gnu Thread model: posix gcc version 4.4.3 (Ubuntu 4.4.3-4ubuntu5) TIA. -Larry [/QUOTE]

Verification