some clever bitmanipulations for VHDL please.

[completelyuseless]

vhdl invert vector

Hi all,

i have to synthesise some VHDL code that will manipulate bits in a certain way.

in a vector of say 'n' bits, i need to find the position of the most significant '1' in the vector and and invert the 0's that are to the left of it.

eg:
lets say a i have a vector that is 9 bits long, and arbitrarily it is say 000110100

there are 3 zeros before the most significant 1 in the obove vector, so i need a clever sort of masking system that will invert those zeros, making it 111110100.
Even if the position of that most significant '1' changes, the method must be general and work for any case, (provided the vector size does not change of course)

i know this can be done with a whole lot of if's and else's but id like to know if there is some clever quick masking system that i dont know about that exists to tackle this problem.

thanks in advance

 

[FvM]

vhdl invert

Similar operations can most easily be implemented by a FOR LOOP iterative scheme. Generally you should consider, that all equivalent design descriptions can be expected to end up in the same code due to the VHDL compiler's optimization action. Thus readability and compactness of code are the main criteria.

 

[lmtg]

vhdl inverting vector

I think you can either use a case statement or a loop..

 

[pev]

masking vhdl

This is how you do it.

input and output are signals of the same length.

process(clk)
  variable leading_zeros: boolean;
begin
  if rising_edge(clk) then
    for a in input'high dowoto input'low loop
      leading_zeros := true;
        if input(a) = '0' and leading_zeros then
           output(a) <= '1';
        else
           output(a) <= input(a);
           leading_zeros := false;
        end if;
     end loop;
   end if;
end process;

There is going to be syntax errors in this. But this is how you do it.