Typo in constraint distribution ranges in cv32e40p_pulp_hwloop_instr_lib.sv

[Mike Thompson]

In a discussion of pull-request #2305 there was mention of a set of typos in cv32e40p_pulp_hwloop_instr_lib.sv. For example:

foreach(hwloop_count[i])
          hwloop_count[i] dist {[0:400] := 10, [401:1023] := 300, [1024:2047] := 150,
                                [2048:4094] := 50, 4095 := 300};

A comment from @dd-vaibhavjain indicated that he noticed a typo just now for hwloop_count range. As the typo was very minor, I approved and merged the change, so this typo is now at the head of the cv32e40p/dev branch. The purpose of this Issue is to ensure we review this and either fix or accept the range.

[Vaibhav Jain]

Hi @MikeOpenHWGroup , this was already fixed in that PR.

[Mike Thompson]

Really? Starting at cv32e40p_pulp_hwloop_instr_lib.sv line #1101 we have:

          hwloop_counti[1] dist {[0:400] := 10, [401:1023] := 300, [1024:2047] := 150,
                                 [2048:4094] := 50, 4095 := 300};

          hwloop_count[1] dist {[0:400] := 10, [401:1023] := 300, [1024:2047] := 150,
                                [2048:4094] := 50, 4095 := 300};

Is that what you want?

[Vaibhav Jain]

Yes, as far as typo is concerned it is updated for that. But if you are asking about the distribution itself, do you see any issue that we need to address?

[Mike Thompson]

if you are asking about the distribution itself, do you see any issue that we need to address?

Earlier I was concerned because it seemed that hits of hwloop_count[1] == 4095 would be hard to achieve. That seems to have been addressed. The probability of hwloop_count[1] == 0 seems low.

Anyway, you are happy with the distribution so I will close this issue. If functional coverage has holes we may need to tune the distribution.