Range must be bounded by constant expressions

SystemVerilog
1

I am trying to implement the following in systemverilog.

always_comb
    begin
	for(int i=0;i<=124;i=i+6)
	    begin
	        offset = DFF_data[ ((2*i)+1):(2*i) ]; // Error !	
		data_output[i+offset]= data_input[i];
	    end
   end

I get the following error in the line marked above: Range must be bounded by constant expressions. I can see the error ocuurs beacuse of the use of “i” in defining the range of array DFF_data. But writing this without “i” seems impractical. I am using QuestaSim 10.1d. Has anyone come across such a situation?

2

Verilog integral expressions must have a constant width determined at compile time.

You need to use a indexed part-select which allows you to specify a variable base expression with a constant width. (this comes from Verilog. system verilog - Indexing vectors and arrays with +: - Stack Overflow)

logic [15:0] down_vect;
logic [0:15] up_vect;
down_vect[lsb_base_expr +: width_expr]
up_vect[msb_base_expr +: width_expr]
down_vect[msb_base_expr -: width_expr]
up_vect[lsb_base_expr -: width_expr]

So you can write

offset = DFF_data[ (2*i) +: 2 ];
5

In reply to vishalkewlani:

Hello Guys,

I am facing the same problem with 2variables.

offset = DFF_data[ ((2j)+7):(2i) ];

Here i am having the 2 variables i&j.
So how to do part select.
Means to say that my width will vary.

Please answer the question.

6

In reply to raja.kuraku:

You can use a shift and mask operation

offset = (DFF_data >> 2*i) & 2**(width)-1

I’ll leave it to you or someone else to figure out the width expression and make sure I have the right operator precedence.

7

In reply to dave_59:

module wr_strb;
bit [15:0] Data = 'hFFFF;
bit [15:0] WDATA;

task strobe(int r);
WDATA[(4*r-1):0] = 'hx;
WDATA[15:(4*r)] =  Data[15:(4*r)];
$display("WDATA = %0h",WDATA);
endtask

initial
begin
strobe(2);
end
endmodule

Hi Dave,
i’m getting an error : “Range must be bounded by constant expression” for the above code.
If you have any solution for this please guide me.

thanks and regards,
Saroj Kumar Sahoo

8

In reply to saroj kumar sahoo:

function void strobe(int r);
  bit [15:0] mask;
  mask = (15'b1 << 4*r) - 1;
  WDATA =  Data ^ (mask & 15'hXXXX);
  $display("WDATA = %0h",WDATA);
endfunction
9

In reply to dave_59:

Thank you very much Dave.

10

data_in[7*(j+1)+j:8*j] // while i am doing this in sv this error comes----> Range must be bounded by constant expressions how to get the same functionality

11

In reply to mansi m joshi:

data_in[j+:8]
12

hello dave in this above example if only 8 bit slicing needs for ex : [7:0] [15:8] [23:16] [31:24] [39:32] [47:40] [55:48] [63:56] and so on

13

its done by using this : data_in[8*(j)+:8]

14

Hi Dave,

It looks ssingh’s original intention was to add a working implementation choosing 2-bit out of the data:
offset = DFF_data[ ((2*i)+1):(2*i) ];
So the equivalent should become:
offset = DFF_data[ (2*i) +: 2 ];
instead of choosing only width=1 with:
offset = DFF_data[ (2*i) +: 1 ];

1 Like