as we all know, MS released the windows CE 3.0 source not too long ago (
http://www.microsoft.com/windows/embedd ... efault.asp). anyway, being a e125 user, I was pretty interested when I saw float.c. here is a quote in the comments:<br><br><br>This module implements a software emulation of the IEEE single and double floating operations. It is required on MIPS processors since the hardware does not fully support all of the operations required by the IEEE standard. In particular, infinitives and Nans are not handled by the hardware, but rather cause an exception. On receipt of the exception, a software emulation of the floating operation is performed to determine the real result of the operation and if an exception will actually be raised.<br> Since floating exceptions are rather rare events, this routine is written in C. Should a higher performance implementation be required, then the algorithms contained herein, can be used to guide a higher performance assembly language implementation.<br><br><br>Well, floating point ops aren't rare in PQ! Does anyone think that porting this lib to asm would be the answer to speeding up PQ for the MIPS? Just my .02