I'm a beginner at vhdl and I'm trying to make a 8 bit divider using shift and subtract method. The code for the divider is okay, I guess, but I'm having problems in simulating the inputs. When I use the Verilog test fixture, my inputs which should be (dividend=51 [00110011b], divisor=19 [00010011b]) always end up becoming (dividend=49 [00110001b], divisor=49 [00110001b])) in iSim, and then the outputs (Quotient and remainder = xxxxxxx (unknown)) instead of (quotient=2[00000010b] and remainder=13 [00001101b])... As I mentioned before, I'm a beginner at vhdl and I've been at this for five hours, I've tried my best to search for a solution online, but failed.... so any help or alternative solution would be greatly appreciated. I'm using Xilinx ISE 14.3 and iSim.
Here is the code for the 8 bit binary divider.
library IEEE;
use IEEE.STD_LOGIC_1164.all;
use IEEE.STD_LOGIC_UNSIGNED.all;
use IEEE.STD_LOGIC_ARITH.all;
entity division is
generic(SIZE: INTEGER := 8);
port(reset: in STD_LOGIC; --reset
en: in STD_LOGIC; --enable
clk: in STD_LOGIC; --clock
num: in STD_LOGIC_VECTOR((SIZE - 1) downto 0); --dividend
den: in STD_LOGIC_VECTOR((SIZE - 1) downto 0); --divisor
res: out STD_LOGIC_VECTOR((SIZE - 1) downto 0); --result/quotient
rm: out STD_LOGIC_VECTOR((SIZE - 1) downto 0) --remainder
);
end division;
architecture behav of division is
signal bufreg: STD_LOGIC_VECTOR((2 * SIZE - 1) downto 0); --signal array to hold both accumulator and dividend registers as one i.e bufreg(18 bits)
signal dbuf: STD_LOGIC_VECTOR((SIZE - 1) downto 0); --signal array to hold the divisor
signal count: INTEGER range 0 to SIZE; --count to determine when to stop
alias ADreg is bufreg((2 * SIZE - 1) downto SIZE); --ADreg is is alias for top half of bufreg register(17th to 9th bit)
alias DVNDreg is bufreg((SIZE - 1) downto 0); --DVNDreg is is alias for bottom half of bufreg register(8th to 0th bit)
begin
--our process begins here
p_001: process(reset, en, clk)
begin
if reset = '1' then
res <= (others => '0');
rm <= (others => '0');
count <= 0;
elsif rising_edge(clk) then
if en = '1' then
case count is
when 0 =>
ADreg <= (others => '0');
DVNDreg <= num;
dbuf <= den;
res <= DVNDreg;
rm <= ADreg;
count <= count + 1;
when others =>
if bufreg((2 * SIZE - 2) downto (SIZE - 1)) >= dbuf then
ADreg <= '0' & (bufreg((2 * SIZE - 3) downto (SIZE - 1)) - dbuf((SIZE - 2) downto 0));
DVNDreg <= DVNDreg ((SIZE - 2) downto 0) & '1';
else
bufreg <= bufreg((2 * SIZE - 2) downto 0) & '0';
end if;
if count /= SIZE then
count <= count + 1;
else
count <= 0;
end if;
end case;
end if;
end if;
end process;
end behav;
And here is the code for the Verilog test fixture (.v file):
module BINdivisionTEST;
// Inputs
reg reset;
reg en;
reg clk;
reg [7:0] num ;
reg [7:0] den ;
// Outputs
wire [7:0] res;
wire [7:0] rm;
// Instantiate the Unit Under Test (UUT)
division uut (
.reset(reset),
.en(en),
.clk(clk),
.num(num),
.den(den),
.res(res),
.rm(rm)
);
initial begin
// Initialize Inputs
reset = 0;
en = 0;
clk = 0;
//num = 0;
//den = 0;
// Wait 100 ns for global reset to finish
#100;
en = 1;
#100;
clk=1;
num <= "00110011" ;
den <= "00010011" ;
// Add stimulus here
end
endmodule
My biggest problems seem to be initialization of the inputs and probably some mistaken assignments, misuse of the test fixture file and some bad coding. I should also add that the divider code was compiled in vhdl, ISE 14.3. I'm extremely sorry if this has been answered before and I don't mean to piss anyone off by uploading bad and amateur code. If this kind of problem has been addressed before could you please provide me with a link. Again, any help is greatly appreciated.
