Forum Discussion
2 Replies
- Altera_Forum
Honored Contributor
altfp_sqrt is using floating point numbers which have to be converted to integer before you send it to a DAC.
I think you are asking about a digital signal processing respectively number crunching problem which isn't specific to FPGA. You should get clear about the expected data number ranges. Pencil and paper method is probably the best way. - Altera_Forum
Honored Contributor
If you can afford to generate the square root 1 bit (of result) per clock (maybe 2 clocks at high clock rates), then the 'schoolboy' method works very well in binary and will probably use a lot less logic.
For 12 bit input I think you could just chain 6 subtract + conditional-mux blocks...