Make rare events count only after four occurrences and weigh them double, with options set where the coverage build honours them.
Hardware context: A buffer that reports normal operations and rare underflow and overflow events. Count rare events only after four occurrences and weigh them double.
A buffer reports an event code each cycle it does something. Codes 1 to 6 are normal operations, 0 an underflow and 7 an overflow. Underflow and overflow are rare, and the plan wants each seen four times before it counts, with the pair weighted double.
Write the covergroup ev_cov.
Signals in the Timeline
| Signal | Width | Role |
|---|---|---|
samples | 8 | input of each check |
Full specification
| Coverpoint | Bins | Options |
|---|---|---|
| common | normal[]: one per code 1 to 6 | defaults |
| rare | underflow = 0, overflow = 7 | weight 2; each bin needs 4 hits |
Each case samples a fresh instance and reads get_inst_coverage() times 100, rounded. The full check samples more lists than the examples show.
Examples
Example 1
One Overflow- Input
event codes: 1 2 3 7
- Output
coverage_percent_x100 = 1667
common 3 of 6 (50%); one overflow is not enough, so rare is 0%: (50 + 0 x 2) / 3 = 16.67%.
Example 2
Four Overflows- Input
event codes: 7 7 7 7 4
- Output
coverage_percent_x100 = 3889
overflow is covered: rare 50%. common 1 of 6: (16.67 + 50 x 2) / 3 = 38.89%.
Example 3
Two of Each- Input
event codes: 0 7 0 7 5 6
- Output
coverage_percent_x100 = 1111
Neither rare bin has four hits. common 2 of 6: 11.11%.
Constraints and Contract
- Keep the covergroup, coverpoint and bin names.
- The coverage build ignores
option.at_leastset in the covergroup (lesson C23): set it in the coverpoint.
Hints
Hint 1
Which coverpoint needs at_least, and which does not?
Hint 2
A weight of 2 on rare makes it count twice in the average.
Hint 3
Count the hits of 0 and of 7 separately.
What the grader checks: Your covergroup is sampled with 18 lists of event codes. Its instance coverage is compared with the value IEEE 1800-2023 19.7 and 19.11 give. Three lists are shown.
18 checks on Verilator coverage build · server. Run the examples first, then check the full regression. Full checks use the server judge. Downloads contain public samples only. A passing result establishes the tested cases, not a proof for all possible inputs.
See all 3 steps and the final capstone