Is there anyway to use -fwhole-program optimization with gcc4ti? I tried concatenating all of my source files together and using it, but I was told there was an "undefined reference to __main". I then tried adding the externally_visible attribute to my main function, but I still got the same error.

To see if it would have any benefit, I manually made all of the functions in the concatenated source file static, and got a binary that was 11% smaller.