How to shift a std_logic_vector by std_logic_vector using concatenation

Question

Say I have 2 std_logic_vectors:

inputA : std_logic_vector(31 downto 0)
inputB:  std_logic_vector(31 downto 0)

How do I shift inputA by inputB using concatenation?

I know how to shift left or right by 1 place but can't figure out how to shift N places to the right (or left). Note: this is a clockless circuit, and can't use standard vhdl shift operators.

Other techniques or ideas other than concatenation would be appreciated as well.

You probably can't do this in a clockless manner, since you'd need to have a programmable shift register or perform multiple rounds in a normal shift register. Doing this in a clockless manner wouldn't make sense since you would require 2^32 different shift registers for each possible path, and even then you would probably still require a clock. — slugonamission
– slugonamission, Commented Sep 17, 2012 at 14:01
@slugonamission: It's perfectly feasible without a clock - it's just a big bag of muxes... also known as a barrel-shifter — Martin Thompson
– Martin Thompson, Commented Sep 17, 2012 at 14:25
@MartinThompson - don't you ideally need to clock the value into the muxes to prevent hazards though? — slugonamission
– slugonamission, Commented Sep 17, 2012 at 14:29
@slugonamission: at some point, yes it'll need some registers on the extremities of the shifter. Once they are there, the timing tools will sort the rest out for you. — Martin Thompson
– Martin Thompson, Commented Sep 17, 2012 at 14:34

wjl · Accepted Answer · 2012-09-16 18:31:20Z

7

The simplest way to do this would be to something like this:

library ieee;
use ieee.numeric_std.all;
...
output <= std_logic_vector(unsigned(inputA) srl to_integer(unsigned(inputB)));

(BTW, being a clockless circuit has nothing to do with being able to use shift operators or not. What determines that is data types. This shift operation will be turned into the same logic by a synthesizer as you would get if you wrote something more complex with case statements all expanded out by hand.)

answered Sep 16, 2012 at 18:31

wjl

7,8752 gold badges38 silver badges44 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

Martin Thompson · Accepted Answer · 2012-09-17 14:24:41Z

3

I prefer wjl's approach, but given you asked specifically for a method using concatenation, try this:

function variable_shift(i : std_logic_vector, num_bits : integer) 
   return std_logic_vector is
   constant zeros : std_logic_vector(num_bits-1 downto 0) := (others => '0');
begin
   return i(i'high-num_bits downto i'low) & zeros;
end function;

(It could be written to take a second std_logic_vector for the num_bits parameter, but as it's fundamentally a number, I'd always use a number-based type for it)

edited Sep 17, 2012 at 14:24

answered Sep 17, 2012 at 10:15

Martin Thompson

16.9k1 gold badge40 silver badges58 bronze badges

Collectives™ on Stack Overflow

How to shift a std_logic_vector by std_logic_vector using concatenation

2 Answers 2

Comments

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related