RE: [sv-ec] RE: [sv-bc] Can a function contain a fork/join/any/none?

To answer a question from Stu:

A check of a suite of 70 customer designs turned up no cases of
fork/join used in a function.

Even a check of a suite of thousands of internal regression tests
only turned up one case, and you would expect at least one case if
the suite provides full coverage.

And if someone did do this, it is easy enough to modify their code
not to do it.  Just replace the fork/join with begin/end.  Sometimes
users resist modifying any code, but this most commonly applies to
proven legacy designs.  Since fork/join is not generally synthesizable,
it would presumably only appear in the testbench, where there would
be less resistance to such a change.

So it can be argued that adding this restriction would not have any
significant negative impact on users.

Steven Sharp
sharp@cadence.com