Hi Folks,

I suspect I know the answer but if anybody can chip in I'd be grateful.

I need to manipulate some binary numbers in an FPGA co-pro design I'm doing (div, square root, and possibly other roots and powers) and I got to thinking about converting the numbers to their log equivalent (base 2) doing the business and then convert back.

After spending all morning researching this at Queens university librray, and on line, I have come up with zillions of algorithms which use floating point, or require multiply or division or both, and therefore are useless to me.

I need the integer and fractional part of the result, not just integer.

If any of you have ideas about such algorithms, or where I can look for ideas (maybe obscure old textbooks etc), then let me know.

Many thanks,

Murdo.

PS: Application is for a MIMO RF decoder where I am comparing QR and Cholesky decompositions.