i need to do some optimization in gcc for a machine that has memory to memory instructions . these are faster than loading to a register and then storing. ÃÂ is such an optimization already in any of the ports that anyone is aware of. ÃÂ thanks. ÃÂ reed