Cross coverage valid bins with mutual exclusive coverpoints

Coverage
,
1

I have this example for 6 coverpoints that can take same 6 values, but never the same value between them at a certain point (mutually exclusive). These values are not necessarily consecutive.

For example (where the bins specifically does not have value 3):


bit [3:0] a0,a1,a2,a3,a4,a5;

covergroup cg;
  p0: coverpoint a0{
    bins v[] = {0,1,2,4,5,6};
  }
  p1: coverpoint a1
  {
    bins v[] = {0,1,2,4,5,6};
  }
  p2: coverpoint a2{
    bins v[] = {0,1,2,4,5,6};
  }
  p3: coverpoint a3{
    bins v[] = {0,1,2,4,5,6};
  }
  p4: coverpoint a4{
    bins v[] = {0,1,2,4,5,6};
  }
  p5: coverpoint a5{
    bins v[] = {0,1,2,4,5,6};
  }  
  CRS: cross p0,p1,p2,p3,p4,p5;

endgroup

With 2 values its simple but for more than 2 values it seems trickier. I suppose also there is a big penalty of performance of coverage between so many coverpoints.

What is the best way to make the cross coverage exclude the bins that are not mutual exclusive / or make only the valid bins that contain the mutual exclusive values ( for example valid crosses are {0,1,2,4,5,6},{6,1,5,4,2,0}…etc, where any value is not repeated 2 or more times.

2

In reply to ingamara:

Maybe you could specify the bins you want to ignore in the cross using the the with clause something like this

  CRS: cross p0, p1, p2, p3, p4, p5 {

    ignore_bins same      = CRS with ((p0 == p1) || (p0 == p2) || (p0 == p3) || (p0 == p4) || (p0 == p5) ||
                                      (p1 == p2) || (p1 == p3) || (p1 == p4) || (p1 == p5) ||
                                      (p2 == p3) || (p2 == p4) || (p2 == p5) ||
                                      (p3 == p4) || (p3 == p5) ||
                                      (p4 == p5)    
                                     );

  }

I think this basically leaves only the unique combinations in this case since v has 6 numbers 12345*6 = 6! = 720 (I could be wrong), I tried with 3 coverpoints with v = {0,1,2} and it seemed to work.
I know it does not look pretty, maybe there is a more efficient way to exclude them.

You can have a look to the 1800-2017 LRM section 19.6.1.2 Cross bin with covergroup expressions to get more details

HTH,
-R

3

In reply to rgarcia07:

Thanks, I did not find something better.