«

[harish.avenue]

Pls help me in declaring and initializing an array of 10 elements & each of 32 bits in vhdl...

[jeppe]

try this

http://jjmk.dk/MMMI/Lessons/07_Memory/No1_ROM/index.htm

not the Answer but some inspiration.

[harish.avenue]

can we initialize the array with the elements declared in the entity like, in my case i hv declared variables (arr0,arr1,arr2,arr3,arr4,arr5,arr6,arr7,arr8,arr9 ) with bit_vector(31 downto 0) in entity..
ex:-

type type_name is array(9 downto 0) of std_logic_vector(31 downto 0);
signal sig_name : type_name:=(arr0,arr1,arr2,arr3,arr4,arr5,arr6,arr 7,arr8,arr9);

[jeppe]

Well .... Bit_vector must then be converted to Std_Logic_vector.

Or even better must Std_logic_vector be used at both declarations.

[harish.avenue]

in the following code, does the array declaration is correct & does the function

vect2int

help in converting the vector to integer, and will it assign the integer the values to k, maxvalue & minvalue...

-------------------------------
library ieee;
use ieee.std_logic_1164.all;
use ieee.std_logic_arith.all;
use ieee.std_logic_unsigned.all;-- Uncomment the following library declaration if using
-- arithmetic functions with Signed or Unsigned values
use IEEE.NUMERIC_STD.ALL;

-- Uncomment the following library declaration if instantiating
-- any Xilinx primitives in this code.
--library UNISIM;
--use UNISIM.VComponents.all;

entity newcode is
generic(N: integer:=9; M: integer:=31);
Port ( clk : in STD_LOGIC;
rst : in STD_LOGIC;
start : in STD_LOGIC;
pw0 : in bit_vector(M downto 0);
pw1 : in bit_vector(M downto 0);
pw2 : in bit_vector(M downto 0);
pw3 : in bit_vector(M downto 0);
pw4 : in bit_vector(M downto 0);
pw5 : in bit_vector(M downto 0);
pw6 : in bit_vector(M downto 0);
pw7 : in bit_vector(M downto 0);
pw8 : in bit_vector(M downto 0);
pw9 : in bit_vector(M downto 0);

ppr0 : in bit_vector(M downto 0);
ppr1 : in bit_vector(M downto 0);
ppr2 : in bit_vector(M downto 0);
ppr3 : in bit_vector(M downto 0);
ppr4 : in bit_vector(M downto 0);
ppr5 : in bit_vector(M downto 0);
ppr6 : in bit_vector(M downto 0);
ppr7 : in bit_vector(M downto 0);
ppr8 : in bit_vector(M downto 0);
ppr9 : in bit_vector(M downto 0);

cnt0: in bit_vector(M downto 0);
cnt1: in bit_vector(M downto 0);
cnt2: in bit_vector(M downto 0);
cnt3: in bit_vector(M downto 0);
cnt4: in bit_vector(M downto 0);
cnt5: in bit_vector(M downto 0);
cnt6: in bit_vector(M downto 0);
cnt7: in bit_vector(M downto 0);
cnt8: in bit_vector(M downto 0);
cnt9: in bit_vector(M downto 0);

opt_sgn : out bit);
end newcode;

architecture Behavioral of newcode is

type ppr_arr is array(0 to N) of bit_vector(M downto 0);
signal ppr : ppr_arr:=(ppr0,ppr1,ppr2,ppr3,ppr4,ppr5,ppr6,ppr7, ppr8,ppr9);

type pw_arr is array(0 to N) of bit_vector(M downto 0);
signal pw : pw_arr:=(pw0,pw1,pw2,pw3,pw4,pw5,pw6,pw7,pw8,pw9);

type cnt_arr is array(0 to N) of bit_vector(M downto 0);
signal cnt : cnt_arr:=(cnt0,cnt1,cnt2,cnt3,cnt4,cnt5,cnt6,cnt7, cnt8,cnt9);


function vect2int(v: in bit_vector(M downto 0)) return integer is
variable result:integer:=0;
begin
for i in 0 to M loop
if v(i)='1' then
result:=result+2**i;
end if;
end loop;
return result;
end vect2int;

begin
process(clk,start,rst)
variable v1:bit:='0';
variable k:natural:=0;
variable count1:integer:=0;
variable maxvalue1:integer:=0;
variable maxvalue2:integer:=0;

begin

if(start='1' and rst='0') then
if clk'event and clk='1' then
for i in 0 to N loop

k:=(vect2int(cnt(i)));
maxvalue1:=(vect2int(pw(i)));
maxvalue2:=(vect2int(pw(i)));

for j in 0 to(vect2int(cnt(i))) loop

v1:='1';
if (count1<=maxvalue1) then
count1:=count1 + 1;
else
count1:=0;
end if;

v1:='0';

if (count1<=maxvalue2) then
count1:=count1 + 1;
else
count1:=0;
end if;
end loop;
end loop;
end if;
end if;
opt_sgn <= v1;
end process;

end Behavioral;

[jeppe]

It's seems your about to implement a C-program in VHDL and doing this must you keep in mind that C will translated into machine-code and executed by a Programmable Final State Machine = CPU.

The VHDL program intended to be Hardware Description which translated into digital logic (perhaps a Non Programmable Final State Machine) which solves your task.

Ok - it seems your only planing to read from the arrays in you program and hence will be allowed to initialize the array with concurrent code like.

Arr(0) <= ppr0;
Arr(1) <= ppr1;

or perhaps even better, could you put this in a clock driven process in order to use F/F for storage.

[harish.avenue]

ya i'm reading the array in order to compute the output...
is that function correct?.... cant it be synthesisable?....will that convert the bitvector of 32bit in to an integer???

[jeppe]

I would say yes ( this not your problem)

The real problem will be if your algoritme is to complex.
Your decided to use clock-driven process and you use variables as well - good
But if the tasks in the loop actually require more then one clk-rising-edge will the
synthesize fail and you will properly not be able to simulate eigter.

Solution: Expand the loop into several states.