Hi, Le Mon, 5 Nov 2007 11:31:13 +0000, ninjaboy <[EMAIL PROTECTED]> a écrit :
> Yes, that makes sense, but why on AMD64 X2 is better rep insn than > unrolled? In my userspace test case rep is slower than unrolled. My knowledge of microprocessor internals is unfortunately too limited to give a precise answer to your question. However, you can see the code that sets X86_FEATURE_REP_GOOD at http://lxr.free-electrons.com/source/arch/x86_64/kernel/setup.c#574. The above comment says « On C+ stepping K8 rep microcode works well for copy/memset » You can read section 3.7.6 of the « Intel® 64 and IA-32 Architectures Optimization Reference Manual » for more informations about using "rep" for optimization. This manual is available at http://www.intel.com/products/processor/manuals/index.htm. Sincerly, Thomas -- Thomas Petazzoni - [EMAIL PROTECTED] http://{thomas,sos,kos}.enix.org - http://www.toulibre.org http://www.{livret,agenda}dulibre.org -- To unsubscribe from this list: send an email with "unsubscribe kernelnewbies" to [EMAIL PROTECTED] Please read the FAQ at http://kernelnewbies.org/FAQ
