home *** CD-ROM | disk | FTP | other *** search
- > Hi there.
- > Passing integers to assembly procedures is easy.But passing
- >floating point numbers (eg A# etc) is somewhat complicated.We have to
- >use Fast Floating Point format(FFP). This means :
- >MMMMMMMMMMMMMMMMMMMMMMMMSEEEEEEE where M is the 24bit mantissa, S is
- >the sign and E is the exponent.Can anyone tell me how a number (let's
- >say 32) is represented in this format? Is it 32 1E0?
-
- The exponent is biased by 65; in other words, you subtract 65 from the
- exponent field to get the actual exponent. So 32 (1 * 2^5) would be:
-
- 10000000 00000000 00000000 01000110
-
- The mantissa is 1.000000... (there's an assumed binary point after the
- first digit), the sign bit is zero (positive), and the exponent is 5
- (70-65).
-
- --Andy Church (achurch@goober.mbhs.edu)
- WWW: http://www.mbhs.edu/~achurch/
- AMOS Web Site: http://www.mbhs.edu/~achurch/amos/
-
-