On x86 and GCC 3.4.2, the resulting asm code when leaving a stackframe is (in intel_syntax) mov esp,ebp pop ebp .(results in 0x89EC5D or 0x8BE55D) The AMD optimisation guide for AMD64 tells you not to use mov esp, ebp - pop ebp but to use "leave"(just 0xC9)! Perhaps using "enter x(16bit), imm8" is better than "push ebp - mov ebp, esp - sub esp, x"
I hope you can improve performance with these optimisations. ___________________________________________________________ Gesendet von Yahoo! Mail - Jetzt mit 1GB Speicher kostenlos - Hier anmelden: http://mail.yahoo.de