Dmitry wrote:
Concerning FP benchmarking. All FP code is originally written by DEC and not optimized for small cpus. I just took that code (from newlib) and made it 'runnable'. If you want to improve it -- be my guest.I think you mean the library functions are from DEC. The actual floating point arithmetic routines are the software FP support built into GCC, aren't they? I don't think that code came from DEC.
Regards, Steve
