Hello, I'm doing a lot of float calculation in my app and I'm a bit frightened by the poor float performance... I'm using the cygwin toolchain, are there any specific flags I should give to the compiler? I know the 2.0 sdk has an option to produce float friendly instructions (rather than thumb) but I have no clue how to do this on the cygwin toolchain. EDIT: I've seen there's a vectored float coprocessor in the iPhone... anyone knows if we can access it using the toolchain / official 2.0 sdk? Any hints would be appreciated.