Dear Members, I am writing a verilog module that can count the number of zeros in my input data flowing in. Each input sample is 10 bit. And I only want to count zeros if 4 consecutive samples are zero. For example, for the input samples: [74 0 0 0 0 0 54 23 0 0 0 84 0 0 0 0 0 0 12] I would count the string of first 5 zeros. Then ignore the next 3 zeros because I only want to consider consecutive 4 zeros only. Then the next 6 zeros are considered. And the total number of zeros become, 5 + 6 = 11. I have been to count the number of zeros in any string but not able to accumulate them to get the final answer. I would be much thankful if someone can guide me. I am attaching my module as well as testbench. Notice that the test bench reads 10 bit samples from a txt file, namely, 'P281000BSR.txt', the test bench works fine, only the module is to be rectified. According to my matlab code, the answer for this txt file should be 127 in decimal.

I'm doing VHDL and so I can't serve a solution, but I would do it this fairly easy way: I would initially set asumto 0 and I would set acounterto 0. And then I would walk through the data and increment thecounterevery step I get a 0. When the input data is not 0 then I would compare thecounterwith 4 and if its more than 4 then I would add thecounterto thesumand reset thecounterto 0.

:
Edited by Moderator

Lothar M. wrote: > I'm doing VHDL and so I can't serve a solution I must correct myself... ;-) In Verilog this looks like that:

module BSR( input wire clk, //input wire clk_enable, input wire reset, input signed [9:0] data_in, output reg [9:0] counter, output reg [9:0] acc ); always @ (posedge clk) begin if(reset) begin counter = 0; acc <= 0; end else begin if (data_in == 0) begin counter = counter + 1'b1; end else begin if (counter > 2'b11) begin acc <= counter + acc; end counter = 0; end end end endmodule |

:
Edited by Moderator

Dear Miller, I must say, you are genius. I was so confused about the implementation in Verilog. Thanks so much. I would be great if you would clarify using the blocking assignment for counter. Is this a good practice? How does it affect our design? I was trying to implement the algorithm with non-blocking (<=) statements only. Thanks again.

Usman A. wrote: > I would be great if you would clarify using the blocking assignment for > counter. Is this a good practice? As I said: usually I do VHDL and so I simply don't know, but simulation looks fine... ;-)