Hello I was hoping you might be able to answer a question for me please. I am wondering how to disable the long-long, as --enable-long-long seems to be the default, and -ansi does not seem to disable it. I am using gcc as a cross compiler, --version says: powerpc-603e-linux-gcc (GCC) 3.4.3 My target is an embedded 32bit PowerPC. What I am trying to do is to disable the use of floating point instructions that seem to be used to move long long variables (or other 8 octet structs) without using the -msoft-float option. You might be trying to figure out why I'd want to. I am finding that the penalty of doing a fpu context switch is way way higher than what is saved by using a 64bit fp load/store instruction to do a move versus using a pair of 32bit int load/store instructions. However, there are places where we do indeed want to do floating point operations, and here I do want hardware floating point enabled. Here the overhead of the context switch is justified for doing real floating point. Any way I can achieve this? Thanks! Regards, Robert Austen --------------------------------------------------------------------- Robert Austen B.Sc.,P.Eng. Willowglen Systems Inc. 8522 Davies Road Total SCADA Solutions Edmonton Alberta Tel: (780) 465 1530 CANADA T6E 4Y5 austen@xxxxxxxxxxxxx Fax: (780) 465 0130