1 / 48

EEL4712 Digital Design

EEL4712 Digital Design. ROM. Instruction ROM example (1). LIBRARY ieee; USE ieee.std_logic_1164.all; USE ieee. numeric_std .all; ENTITY instruction_rom IS GENERIC ( w : INTEGER := 16; n : INTEGER := 8; m : INTEGER := 3); PORT (

nura
Télécharger la présentation

EEL4712 Digital Design

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. EEL4712 Digital Design

  2. ROM

  3. Instruction ROM example (1) LIBRARY ieee; USE ieee.std_logic_1164.all; USE ieee.numeric_std.all; ENTITY instruction_rom IS GENERIC ( w : INTEGER := 16; n : INTEGER := 8; m : INTEGER := 3); PORT ( Instr_addr : IN STD_LOGIC_VECTOR(m-1 DOWNTO 0); Instr : out STD_LOGIC_VECTOR(w-1 DOWNTO 0) ); END instruction_rom;

  4. Instruction ROM example (2) ARCHITECTURE ins_rom OF instruction_rom IS SIGNAL temp: INTEGER RANGE 0 TO n-1; TYPE vector_array IS ARRAY (0 to n-1) OF STD_LOGIC_VECTOR(w-1 DOWNTO 0); CONSTANT memory : vector_array := ( X"0000", X"D459", X"A870", X"7853", X"650D", X"642F", X"F742", X"F548"); BEGIN temp <= to_integer(unsigned(Instr_addr)); Instr <= memory(temp); END instruction_rom;

  5. Poor Design Practices & Their Remedies

  6. Initializations • Declarations of signals (and variables) with initialized values, such as • SIGNAL a : STD_LOGIC := ‘0’; cannot be synthesized, and thus should be avoided. • If present, they will be ignored by the synthesis tools. • Use set and reset signals instead.

  7. Dual-edge triggered register/counter (1) In FPGAs register/counter can change only at either rising (default) or falling edge of the clock. Dual-edge triggered clock is not synthesizable correctly, using either of the descriptions provided below.

  8. Dual-edge triggered register/counter (2) PROCESS (clk) BEGIN IF (clk’EVENT AND clk=‘1’ ) THEN counter <= counter + 1; ELSIF (clk’EVENT AND clk=‘0’ ) THEN counter <= counter + 1; END IF; END PROCESS;

  9. Dual-edge triggered register/counter (3) PROCESS (clk) BEGIN IF (clk’EVENT) THEN counter <= counter + 1; END IF; END PROCESS; PROCESS (clk) BEGIN counter <= counter + 1; END PROCESS;

  10. Poor Design Practices and Remedies • Synchronous design is the most important methodology • Poor practice in the past (to save chips) • Misuse of asynchronous reset • Misuse of gated clock • Disuse of derived clock • Dual-edge triggered register/counter

  11. Misuse of Asynchronous Reset • Problem • Glitches in transition 1001 (9) => 0000 (0) • Glitches in asyn_clr can reset the counter • How about timing analysis? (maximal clock rate) • Asynchronous reset should only be used for power-on initialization

  12. Misuse of Asynchronous Reset

  13. Decade (mod-10) Counter Architecturetwo_seg_archof mod10_counter is Constant TEN: integer : = 10; Signal r_reg : unsigned (3 downto 0); Signal r_next: unsigned (3 downto 0); Begin Process (clk, reset) Begin if (reset = ‘1’) then r_reg <= (others => ‘0’); elsif(clk’eventand clk=‘1’) then r_reg <= r-next; end if; End process; R_next<= (others => ‘0’) when r_reg = (TEN-1) else r_reg +1; -- output logic Q =< std_logic_vector (r_reg); Endtwo_seg_arch;

  14. Misuse of Gated Clock • Problem • Gated clock width can be narrow • Gated clock may pass glitches of en • Difficult to design clock distribution network

  15. Dedicated Clock Tree Network – H Tree

  16. Disuse of Derived Clock • Subsystems may run at different clock rate • Poor design: use a derived slow clock for slow subsystems • Multiple clock distribution network • How about timing analysis? (maximal clock rate)

  17. Disuse of Derived Clock - Solution • Better use a synchronous one-clock enable pulse

  18. Bad Design • E.g., second and minutes counter • Input: 1 MHz clock • Poor design

  19. Better Design

  20. Clock Divider – Same Frequency • Modifying clock phase to generate clock with different phases at the same frequency • Clock1 is the reference clock • Clock2 has  90° phase offset • Clock3 has  180°phase offset

  21. Clock Division by Two • If the clock we need is simply the system clock divided by two, we can implement a simple divider using a flip-flop and inverter: https://surf-vhdl.com/how-to-implement-clock-divider-vhdl/

  22. Clock division by power of two • This clock divider can be implemented using a running simple wrap around counter • The counter LSB flips at half the clock rate, this means that it can be used as clock divider by two. • The bit k rate is 1/(2^(k+1)) clock rate https://surf-vhdl.com/how-to-implement-clock-divider-vhdl/

  23. Clock Divider – Different Frequency • The frequency divider objective is to reduce the input frequency. It is implemented through the use of the scaling factor and a counter. • Scaling factor: the relation between the input frequency and the desired output frequency: • Assuming an input frequency of 50MHz and provided we need an output frequency of 200Hz, we yield: • Therefore, the counter of the frequency divider generates the output signal of 200Hz each 250000 cycles.

  24. Clock Divider (3) architecture Behavioral of clk200Hz is signal temporal: STD_LOGIC; signal counter : integer range 0 to 124999 := 0; begin frequency_divider: process (reset, clk_in) begin if (reset = '1') then temporal <= '0'; counter <= 0; elsifrising_edge(clk_in) then if (counter = 124999) then temporal <= NOT(temporal); counter <= 0; else counter <= counter + 1; end if; end if; end process; clk_out <= temporal; end Behavioral; Library IEEE; use IEEE.STD_LOGIC_1164.ALL; entity clk200Hz is Port ( clk_in : in STD_LOGIC; reset : in STD_LOGIC; clk_out: out STD_LOGIC ); end clk200Hz; A clock signal is a square wave with a 50% of duty cycle(same time active and inactive); for this case, 125000 cycles active and 125000 cycles inactive. Since the counter begins at zero, the superior limit is (250000/2)-1=125000 - 1. Why 124999 not 250000?

  25. Packages

  26. Packages • A package in VHDL is a collection of functions, procedures, shared variables, constants, files, aliases, types, subtypes, attributes, and components. • A package file is often (but not always) used in conjunction with a unique VHDL library.

  27. Example: XOR Tree A 2-d array (P) is needed to capture intermediate wires P30 shows the value of 0th wire in stage 3 WIDTH #Stage = log2 (WIDTH)

  28. XOR Tree library ieee; use ieee.std_logic_1164.all; use work.util_pkg.all; entity reduced_xor is generic(WIDTH: natural:=8); port( a: in std_logic_vector(WIDTH-1 downto 0); y: out std_logic );end reduced_xor;

  29. XOR Tree (2) architecture gen_tree_arch of reduced_xor is constant STAGE: natural:= log2c(WIDTH); signal p: std_logic_2d(STAGE downto 0, WIDTH-1 downto 0); Begin -- rename input signal in_gen: for i in 0 to (WIDTH-1) generate p(STAGE, i) <= a(i); end generate; -- replicated structure stage_gen: for s in (STAGE-1) downto 0 generate row_gen: for r in 0 to (2**s-1) generate p(s, r) <= p(s+1, 2*r) xor p(s+1, 2*r+1); end generate; end generate; -- rename output signal y <= p(0, 0); end gen_tree_arch;

  30. Util_pkg (1) library ieee; use ieee.std_logic_1164.all; package util_pkg is type std_logic_2d is array(integer range <>, integer range <>) of std_logic; function log2c (n: integer) return integer; end util_pkg ;

  31. Util_pkg (2) --package body package body util_pkg is function log2c(n: integer) return integer is variable m, p: integer; begin m := 0; p := 1; while p < n loop m := m + 1; p := p * 2; end loop; return m; end log2c; end util_pkg;

  32. Array Data Type • Array Type

  33. XOR Tree with Arbitrary Input Size (1) begin -- rename input signal in_gen: for i in 0 to (WIDTH-1) generate p(STAGE,i) <= a(i); end generate; -- padding 0’s pad0_gen: if WIDTH < (2**STAGE) generate zero_gen: for i in WIDTH to (2**STAGE-1) generate p(STAGE,i) <= '0’; end generate; end generate;

  34. XOR Tree with Arbitrary Input Size (2) -- replicated structure stage_gen: for s in (STAGE-1) downto 0 generate row_gen: for r in 0 to (2**s-1) generate p(s,r) <= p(s+1,2*r) xor p(s+1,2*r+1); end generate; end generate; -- rename output signal y <= p(0,0); end gen_tree2_arch;

  35. Unconstrained Array Types

  36. Predefined Unconstrained Array Types • Predefined • bit_vector array of bits • string array of characters • Defined in the std_logic_1164 package: • type std_logic_vector is array (natural range <>) of std_logic;

  37. User-defined Unconstrained Array Types type std_logic_2d is array(integer range <>, integer range <>) of std_logic; • Examples: • signal s1: std_logic_2d(3 downto 0, 5 downto 0); • signal s2: std_logic_2d(15 downto 0, 31 downto 0); • signal s3: std_logic_2d(7 downto 0, 1 downto 0);

  38. User-defined Unconstrained Array Type defined in a package use work.util_pkg.all; entity …. generic ( ROW: natural; COL: natural) ); port ( p1: in std_logic_2d(ROW-1 downto 0, COL-1 downto 0); …….. ); architecture signal s1: std_logic_2d(ROW-1 downto 0, COL-1 downto 0);

  39. Array-of-Arrays Data Type constant ROW : natural := 4; constant COL : natural := 6; type sram_row_by_col is array (ROW -1 downto 0) of std_logic_vector(COL – 1 downto 0); …… signal t1: sram_row_by_col; signal v1: std_logic_vector(COL-1 downto 0); signal b1: std_logic; …… t1 <= (“000101”, “101001”, others => (others => ‘0’)); b1 <= t1(3)(0); v1 <= t1(2);

  40. Attributes of Arrays and Array Types

  41. Array Attributes A’left(N) left bound of index range of dimension N of A A’right(N) right bound of index range of dimension N of A A’low(N) lower bound of index range of dimension N of A A’high(N) upper bound of index range of dimension N of A A’range(N) index range of dimension N of A A’reverse_range(N) reversed index range of dimension N of A A’length(N) length of index range of dimension N of A A’ascending(N) true if index range of dimension N of A is an ascending range, false otherwise

  42. Array Attributes - Examples type A is array (1 to 4, 31 downto0); A’left(1) = 1 A’right(2) = 0 A’low(1) = 1 A’high(2) = 31 A’range(1) = 1 to 4 A’length(2) = 32 A’ascending(2) = false

  43. Unconstrained PARITY Generator (1) LIBRARY ieee; USE ieee.std_logic_1164.all; ENTITY parity IS PORT( parity_in : IN STD_LOGIC_VECTOR; parity_out : OUT STD_LOGIC ); END parity;

  44. Unconstrained PARITY Generator (2) ARCHITECTURE parity_dataflow OF parity IS CONSTANT width: natural := parity_in’length; SIGNAL xor_out: STD_LOGIC_VECTOR (width-1 DOWNTO 0); BEGIN xor_out(0) <= parity_in(0); G2: FOR i IN 1 TO width-1 GENERATE xor_out(i) <= xor_out(i-1) XOR parity_in(i); END GENERATE G2; parity_out <= xor_out(width-1); END parity_dataflow;

  45. Unconstrained PARITY Generator (3) • Will the previous code work for the following types of signal parity_in? • std_logic_vector(7 downto 0); • std_logic_vector(0 to 7); • std_logic_vector(15 downto 8); • std_logic_vector(8 to 15);

  46. Unconstrained PARITY Generator (4) ARCHITECTURE parity_dataflow OF parity IS CONSTANT width: natural := parity_in’length; SIGNAL xor_out: STD_LOGIC_VECTOR (width-1 DOWNTO 0); SIGNAL pp: STD_LOGIC_VECTOR (width-1 DOWNTO 0); BEGIN pp <= parity_in; xor_out(0) <= pp(0); G2: FOR i IN 1 TO width-1 GENERATE xor_out(i) <= xor_out(i-1) XOR pp(i); END GENERATE G2; parity_out <= xor_out(width-1); END parity_dataflow;

  47. References • Stephen Brown and ZvonkoVranesic. 2004. Fundamentals of Digital Logic with VHDL Design. McGraw-Hill, Inc., New York, NY, USA. • https://ece.gmu.edu/coursewebpages/ECE/ECE545/F18/viewgraphs/ECE545_lecture_8_regular.pdf • https://ece.gmu.edu/coursewebpages/ECE/ECE545/F18/viewgraphs/ECE545_lecture_9_FSM_refresher.pdf • P. Chu, RTL Hardware Design using VHDL • Chapter 14, Parameterized Design: Principle • Chapter 15, (except 15.4) Parameterized Design: Practice • https://surf-vhdl.com/how-to-implement-clock-divider-vhdl/ • https://forums.xilinx.com/t5/Xilinx-Evaluation-Boards/Generation-of-4-Phase-Shifted-clock-Signals-from-DCM/td-p/330395

  48. Questions?

More Related