Rand Versus Randc

Where the interviewer goes next. Open a node to practise that follow-up.

WhyBreak ItCompareScale It

1. Rand Versus Randc
2. Break ItA Randc Field That Kept RepeatingL1 Junior

   The opcode field is randc, yet the log shows the same opcode twice in a row. The generator does tr = new; tr.randomize(); for every transaction. Why?

   gen: op=2
   gen: op=2
   gen: op=0
   gen: op=3

   Reveal the Model Answer

   The cycle state lives in the object, and every new starts a fresh cycle with its first value chosen at random. Randomize one generator object repeatedly and copy its result into each transaction, or keep the randc field in a long-lived object.

   Self-Check

   Tick each key point you covered in your answer.

   Key point 1 The cycle state lives in the object, and every new starts a fresh cycle with its first value chosen at random. Key point 2 Randomize one generator object repeatedly and copy its result into each transaction, or keep the randc field in a long-lived object.

   Misconception

   "randc remembers its cycle for the whole simulation."

   It remembers per object. The twin constructs a new object for every call and repeats appear; one object for the whole test removes them.

   Run It

   A new randc object for every call, ten windows of four.

   RUNVerilator · simulator service

   Edit the code, then run it

   OutputWaves

   RunReset codeExpected: FAIL: each new object starts its own randc cycle, so values repeated within four calls

   See the Fix

   class opcode_gen;
     randc bit [1:0] op;
   endclass
   module tb;
     opcode_gen g;
     int repeats = 0;
     initial begin
       g = new;                          // the fix: one generator object for the whole test
       repeat (10) begin
         bit [3:0] seen;
         seen = 0;
         repeat (4) begin
           void'(g.randomize());
           seen[g.op] = 1;
         end
         if (seen != 4'hf) repeats++;
       end
       if (repeats > 0) $display("FAIL: each new object starts its own randc cycle, so values repeated within four calls");
       else $display("PASS: one object kept one cycle, so every window covered all four values");
     end
   endmodule
   

   Prints: PASS: one object kept one cycle, so every window covered all four values

   • Lesson 5.1Rand Fields and Checked Randomization
   • Lesson 5.6Random Stability and Callbacks
3. WhyWhy Randc Fields Stay SmallL1 Junior

   Why not declare a 32-bit address randc?

   Reveal the Model Answer

   The solver has to remember which values of the current cycle are used, which costs memory per value of the range. So the standard lets tools limit the width of a randc variable, but not below 8 bits. A 32-bit randc may be rejected and is not portable. For addresses, a constraint plus coverage does the job.

   Self-Check

   Tick each key point you covered in your answer.

   Key point 1 The solver has to remember which values of the current cycle are used, which costs memory per value of the range. Key point 2 So the standard lets tools limit the width of a randc variable, but not below 8 bits. Key point 3 A 32-bit randc may be rejected and is not portable. Key point 4 For addresses, a constraint plus coverage does the job.

   Spec Misquote

   "randc works on any width; it just takes longer."

   Tools are allowed to cap randc width, and only 8 bits is guaranteed.

   Source IEEE 1800-2017 §18.4.2

   • Lesson 5.1Rand Fields and Checked Randomization
4. WhyWhy Randc Variables Are Solved FirstL2 Mid-Level

   Why are randc variables solved before the rand variables, and what can that break?

   Reveal the Model Answer

   To keep its cycle, the solver picks the randc value first, from the values not yet used. Then it solves the rand fields around it. If a constraint ties the rand fields to that value and no solution exists, the call fails. So limiting a randc field through other fields is fragile.

   Self-Check

   Tick each key point you covered in your answer.

   Key point 1 To keep its cycle, the solver picks the randc value first, from the values not yet used. Key point 2 Then it solves the rand fields around it. Key point 3 If a constraint ties the rand fields to that value and no solution exists, the call fails. Key point 4 So limiting a randc field through other fields is fragile.

   Spec Misquote

   "Add solve len before addr to order a randc field."

   The standard does not allow randc variables in solve...before. They are always solved first.

   Source IEEE 1800-2017 §18.5.10

   • Lesson 5.1Rand Fields and Checked Randomization
5. CompareRandc Versus Shuffling a ListL1 JuniorL2 Mid-Level

   Asked if they mention covering a list of values

   You must visit seven specific register addresses, in random order, exactly once each. Compare randc with filling an array and calling shuffle().

   Reveal the Model Answer

   randc cycles over every value its type and constraints allow. Seven scattered addresses therefore need an index variable, a range constraint and a lookup table. An array of the seven addresses with shuffle() states the intent directly. It works for any element type and gives a new order on every call. Use shuffle() for an explicit list, and randc when the set is a small contiguous range.

   Self-Check

   Tick each key point you covered in your answer.

   Key point 1 randc cycles over every value its type and constraints allow. Key point 2 Seven scattered addresses therefore need an index variable, a range constraint and a lookup table. Key point 3 An array of the seven addresses with shuffle() states the intent directly. Key point 4 It works for any element type and gives a new order on every call. Key point 5 Use shuffle() for an explicit list, and randc when the set is a small contiguous range.

   Misconception

   "shuffle() takes constraints like randomize()."

   It is an array ordering method with no constraint support; it just permutes the elements in place.

   Source IEEE 1800-2017 §7.12.2

   • Lesson 2.4Array Methods and with Expressions
   • Lesson 5.1Rand Fields and Checked Randomization
6. Scale ItWhy Randc Is Not a Coverage StrategyL2 Mid-LevelL3 Senior

   A colleague proposes making every enumerated field randc "so coverage closes automatically". What do you tell them?

   Reveal the Model Answer

   randc guarantees one field's values cycle, but bugs live in combinations and sequences: opcode with length, back-to-back writes, a command during reset. Those are crosses and transitions that no per-field cycle controls. Close coverage by measuring the coverage model, finding holes and steering constraints or adding directed sequences at them.

   Self-Check

   Tick each key point you covered in your answer.

   Key point 1 randc guarantees one field's values cycle, but bugs live in combinations and sequences: opcode with length, back-to-back writes, a command during reset. Key point 2 Those are crosses and transitions that no per-field cycle controls. Key point 3 Close coverage by measuring the coverage model, finding holes and steering constraints or adding directed sequences at them.

   What Interviewers Listen For (Senior)

   A good senior answer moves the talk from stimulus tricks to the coverage model. It names crosses and transitions as the items that close last.

   Wrong Altitude

   Agrees, and adds randc to the remaining fields.

   It optimises the easy part, single-field values, and leaves the cross coverage that closes late untouched.

   • Lesson 5.1Rand Fields and Checked Randomization
   • Lesson 7.5Coverage and Tests for the Checker
   • Problem 7.5.8Two Coverpoints for the FIFO