Andrew Coles wrote:
> 
> I've been looking at some code produced by egcs (with -S) and it doesn't
> contain FP opcodes, rather it calls functions to perform the FP maths
> operations. Is that because egcs uses it's own softfloat library or is it
> because the FP support code is not implemented by emulating the ARM FP
> opcodes?

The emulation is implemented for emulating FP opcodes, it's the the
performance is bad. Instead, it branches into functions stored in the C
library instead.

	Neil