arrays - More resource efficient way to get the maximum of the last 512 values -
i have written vhdl code stores last 512 values of input signal , calculates largest of stored values. code works uses lot of lut resources of fpga. purpose of code calculate largest value of last 512 samples, there more resource efficient way of achieving this? (it's important calculates largest of last 512 values , not largest value observed input, latter achieved storing single number).
alternatively there someway code vhdl such synthesiser implement array block ram (bram) instead of luts?
the synthesiser using labview fpga (which believe uses xilinx ise internally compile/synthesise vhdl).
my current code shown below:
library ieee; use ieee.std_logic_1164.all; use ieee.numeric_std.all; entity recentmax port ( clk : in std_logic; reset : in std_logic; inputsignal : in std_logic_vector(15 downto 0); max : out std_logic_vector(15 downto 0) ); end recentmax; architecture rtl of recentmax -- declarations type array512 array(0 511) of signed(15 downto 0); signal pastvals : array512; type array256 array(0 255) of signed(15 downto 0); signal result : array256; signal calculationstate : unsigned(1 downto 0); signal nlefttocompute : unsigned(8 downto 0); begin -- behaviour process(clk) begin if(rising_edge(clk)) if(reset = '1') -- reset values in pastvals'low pastvals'high loop pastvals(i) <= (others => '0'); end loop; in result'low result'high loop result(i) <= (others => '0'); end loop; calculationstate <= to_unsigned(0, 2); max <= std_logic_vector(to_signed(0, 16)); nlefttocompute <= to_unsigned(256, 9); else -- stuff case to_integer(calculationstate) when 0 => in pastvals'low pastvals'high-1 loop pastvals(i+1) <= pastvals(i); end loop; pastvals(0) <= signed(inputsignal); max <= std_logic_vector(result(0)); nlefttocompute <= to_unsigned(256, 9); calculationstate <= to_unsigned(1, 2); when 1 => in 0 255 loop if (i <= to_integer(nlefttocompute)-1) if pastvals(i*2) > pastvals(i*2+1) result(i) <= pastvals(i*2); else result(i) <= pastvals(i*2+1); end if; end if; end loop; nlefttocompute <= shift_right(nlefttocompute, 1); calculationstate <= to_unsigned(2, 2); when 2 =>; in 0 127 loop if (i <= to_integer(nlefttocompute)-1) if result(i*2) > result(i*2+1) result(i) <= result(i*2); else result(i) <= result(i*2+1); end if; end if; end loop; if nlefttocompute > 2 nlefttocompute <= shift_right(nlefttocompute, 1); else calculationstate <= to_unsigned(0, 2); end if; when others => --- nothing - shouldn't here end case; end if; end if; end process; end rtl;
there 2 possibilities want.
first can choose instantiate bram creating dedicated process , array, synthesizer choose use block ram instead of 512 luts.
constant data_width : integer := 16; constant add_width : integer := 9; -- 512 addresses constant dpram_depth : integer := 2**add_width; -- depth of memory type dpram array (0 dpram_depth - 1) of std_logic_vector(data_width - 1 downto 0); signal mem : dpram; and process :
dpram_gen_p : process (clk_i) begin if (rising_edge(clk_i)) if (wr_req_i = '1') mem(wr_add_s) <= wr_data_i; end if; rd_data_s <= mem(rd_add_s); end if; end process; for main synthesizer, syntax implemented block ram. then, instead of pastvals signal need use data port of ram. careful reading cycles since need 1 clock cycle change address of read interface (rd_add_s) , 1 read data (rd_data_s).
the second option (according me easiest , fastest) implement fifo memory (you can use xilinx ip core generator) of size 512 words.
then need write in fifo until it's full , read data word word until it's empty , register highest value did in design.
Comments
Post a Comment