Verilog Frequency Divider 〈PREMIUM | Anthology〉

always @(posedge clk) begin if (clk_en) begin // do something every 8 clocks end end If you must produce a real clock (e.g., for an external chip), route the divider output to a global clock buffer (BUFG in Xilinx, GCLK in Intel). This minimizes skew. 4. Advanced Dividers: Programmable and Wide-Range 4.1 Programmable Divider A programmable divider allows software to select ( N ) at runtime. This is a counter with a loadable terminal count .

module prog_divider #(parameter WIDTH=16) ( input clk, rst_n, input [WIDTH-1:0] divisor, // N value output reg clk_out ); reg [WIDTH-1:0] count; always @(posedge clk or negedge rst_n) begin if (!rst_n) begin count <= 0; clk_out <= 0; end else begin if (count == divisor - 1) begin count <= 0; clk_out <= ~clk_out; end else begin count <= count + 1; end end end endmodule verilog frequency divider

Output: 2 cycles high, 1 cycle low → 33% duty, frequency = clk/3. Fractional division (e.g., divide by 2.5) is essential for generating arbitrary frequencies from a fixed crystal. It is achieved by periodically swallowing clock edges using a phase accumulator. always @(posedge clk) begin if (clk_en) begin //

Caveat: The divisor value must be ≥ 2 and stable during operation. For very high input frequencies (e.g., 500 MHz in an ASIC), counter propagation delay may limit performance. Use synchronous prescalers with low-bit ripple counters or Johnson counters. Advanced Dividers: Programmable and Wide-Range 4

module div_by_3 ( input clk, rst_n, output reg clk_out ); reg [1:0] count; always @(posedge clk or negedge rst_n) begin if (!rst_n) begin count <= 0; clk_out <= 0; end else begin if (count == 2) begin // 0,1,2 -> 3 cycles count <= 0; clk_out <= ~clk_out; end else begin count <= count + 1; end end end endmodule

module clk_enable_div8 ( input clk, rst_n, output reg clk_en ); reg [2:0] count; always @(posedge clk or negedge rst_n) begin if (!rst_n) count <= 0; else count <= (count == 7) ? 0 : count + 1; end assign clk_en = (count == 7); // one cycle wide pulse endmodule Then downstream modules use: