I'm facing this problem , i'm asked to implement a function in VHDL that takes an integer and returns a bit_vector , assumed that this integer is represented by 4 bits.
i don't want to use already built in function, i have to code the function.
I have made a function to convert from bit_vector to integer which was kinda of easy, but im stuck here :S
Any ideas how can i do it ?
Morten's is the correct answer but it's sometimes worth being open to alternative approaches...
As the question relates to a small (4-bit) range, a lookup table becomes attractive : I have assumed unsigned integers but it's easy to adapt.
subtype bv4 is bit_vector(3 downto 0);
constant LUT : array(0 to 15) of bv4 := (
"0000", "0001", "0010", "0011", "0100", "0101", "0110", "0111",
"1000", "1001", "1010", "1011", "1100", "1101", "1110", "1111");
function to_bv(n : natural) return bit_vector is
begin
return LUT(n);
end to_bv;
This will normally synthesise as you would hope rather than actually creating a ROM!
The VHDL standard packages is good inspiration for home brewed functions, and the numeric_bit package defines the to_unsigned function for conversion of natural type to unsigned type, which is the function VHDL actually uses for conversion to bit_vector. The function is implemented as:
function TO_UNSIGNED (ARG, SIZE: NATURAL) return UNSIGNED is
variable RESULT: UNSIGNED(SIZE-1 downto 0);
variable I_VAL: NATURAL := ARG;
begin
if (SIZE < 1) then return NAU;
end if;
for I in 0 to RESULT'LEFT loop
if (I_VAL mod 2) = 0 then
RESULT(I) := '0';
else
RESULT(I) := '1';
end if;
I_VAL := I_VAL/2;
end loop;
if not(I_VAL =0) then
assert NO_WARNING
report "NUMERIC_BIT.TO_UNSIGNED: vector truncated"
severity WARNING;
end if;
return RESULT;
end TO_UNSIGNED;
The initial if (SIZE < 1) and final if not(I_VAL =0) checks may be removed, if it is known that the function is never used with values that makes the checks relevant.
This leaves the for I in 0 to RESULT'LEFT loop that creates one result bit per iteration.
Based on Brian's answer, the constant LUT can be initialized using the TO_UNSIGNED function, to avoid the hand written literals:
function to_bv(n, size : natural) return bit_vector is
type bv_arr_t is array (0 to 2 ** size - 1) of bit_vector(size - 1 downto 0);
function bv_arr_init(size : natural) return bv_arr_t is
variable res_v : bv_arr_t;
begin
for i in 0 to 2 ** size - 1 loop
res_v(i) := bit_vector(TO_UNSIGNED(i, size));
end loop;
return res_v;
end function;
constant LUT : bv_arr_t := bv_arr_init(size);
begin
return LUT(n);
end to_bv;
