| ??? 07/26/02 15:44 Read: times |
#26363 - RE: floating point code size- Luca |
The Infineon 80C517/537 contain a "MDU" (Multiplication/Division Unit). This is indeed intended for integer arithmetic; 32/16, 16/16, 16*16. But it also has a 32-bit normalise which can be used to speed up floating point conversions. I vaguely remember using this feature a long, long time ago ...
Rob. |



